x>EPA
          Una.
          Environm
          A n'ncy
           Offic-
           Ecological Eft;-:
EPA-600/5-79-001h
 i.irv 1979

Methods Development
for Assessing Air
Pollution Control
Benefits
Volume II,
Experiments in Valuing
Non-market Goods:
A Case Study of Alternative
Benefit Measures of
Air  Pollution Control in the
South  Coast Air Basin of
Southern California

-------
                         OTHER VOLUMES OF THIS STUDY

Volume I, Experiments in the Economics of Air Pollution Epidemiology,
  EPA-600/5-79-001a.

     This volume employs the analytical and empirical methods of economics to
develop hypotheses on disease etiologies and to value labor productivity and
consumer losses due to air pollution-induced mortality and morbidity.

Volume III, A Preliminary Assessment of Air Pollution Damages for
  Selected Crops within Southern California, EPA-600/5-79-001c.

     This volume Investigates the economic benefits that would accrue from
reductions in oxidant/ozone air pollution-induced damages to 14 annual
vegetable and field crops in southern California.

Volume IV, Studies on Partial Equilibrium Approaches to Valuation of
  Environmental Amenities, EPA-600/5-79-001d.

     The research detailed in this volume explores various facets of the two
central project objectives that have not been given adequate attention in the
previous volumes.

Volume V, Executive Summary, EPA-600/5-79-001e.

     This volume provides a 23 page summary of the findings of the first four
volumes of the study.
This document is available to the public through the National Technical
Information Service, Springfield, Virginia 22161.

-------
                                                      EPA-600/6-79-001b
                                                      February 1979
                   METHODS DEVELOPMENT FOR ASSESSING
                      TRADEOFFS IN ENVIRONMENTAL
                              MANAGEMENT

                               Volume II

                Experiments in Valuing Non-Market Goods:
                   A Case Study of Alternative Benefit
                    Measures of Air Pollution Control
                       in the South Coast Air Basin
                        of Southern California

                                   by

        David S. Brookshire, Ralph C. d'Arge and William D. Schulze*
                         University of Wyoming
                        Laramie, Wyoming  82071

                             Mark A. Thayer
                        University of New Mexico

                         USEPA Grant //R805059010

                             Project Officer:
                             Dr. Alan Carlin
                 Office of Health and Ecological Effects
                   Office of Research and Development
                  U.S. Environmental Protection Agency
                          Washington, D.C.  20460
                 OFFICE OF HEALTH AND ECOLOGICAL EFFECTS
                   OFFICE OF RESEARCH AND DEVELOPMENT
                  U.S. ENVIRONMENTAL PROTECTION AGENCY
                          WASHINGTON, D.C.  20460
      *William  D.  Schulze's  contribution  to  this  volume was performed while he
was an Associate Professor of Economics at the University or Southern
California.

-------
                                DISCLAIMER


     This report has been reviewed by the Office of Health and Ecological
Effects, Office of Research and Development, U.S. Environmental Protection
Agency, and approved for publication.  Approval does not signify that the
contents necessarily reflect the views and policies of the U.S. Environmental
Protection Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
                                      ii

-------
                                  PREFACE


     The motivation for this volume can be traced to the authors' convictions
that the valuation of non-market goods through the application of economic
analysis could be accomplished.  The various contributions to the literature
that formed the underpinnings of this effort are many and diverse.  Yet, the
work by Alan Randall in setting forth the framework and first empirical ap-
plication of the iterative bidding technique for valuing non-market goods
must be noted.  Many individuals including Drs. Fred Blank, Robert Rowe,
Robert Horst, Jr., Alan Randall, Mr. Larry Eubanks, Mr. Berry Ives, Mr. Rex
Adam have provided worthwhile comments and criticisms.  None of these indiv-
iduals are responsible, however, for the results.
                                    iii

-------
                                 ABSTRACT


     In this study, the empirical results obtained from two experiments
to measure the health and aesthetic benefits of air pollution control in
the South Coast Air Basin of southern California are reported.  Each
experiment involved the same six neighborhood pairs, where the pairings
were made on the basis of similarities in housing characteristics, socio-
economic factors, distances to beaches and services, average temperatures,
and subjective indicators of housing quality.  The elements of each pair
differed substantially only in terms of air quality.  Data on actual market
transactions, as registered in single-family residential property transactions,
and on stated preferences for air quality, as revealed in neighborhood surveys,
were collected.  It was expected that a relation would exist between what
people do pay for air quality as reflected in property value differences, and
what they say they will pay, provided there are no incentives for them to dis-
tort their bids.

     Given various assumptions on income, location, aggregation by areas,
specific housing characteristics, and knowledge of the health effects of air
pollution, both the survey and the property value experiments yielded estimates
of willingness-to-pay in early 1978 dollars for an improvement from "poor" to
"fair" air quality of from $20 to $150 per month per household.  The results,
therefore, indicate that air quality deterioration in the Los Angeles area has
had substantial negative effects on housing prices and that these effects are
comparable in magnitude to what people say they are willing to pay for improved
air quality.
                                     iv

-------
                                 CONTENTS

Abstract      	iv

Figures       	vii

Tables        	viii

Chapter I     Introduction to Volume II	1

Chapter II    Theoretical Foundations	2
     2.1      Valuing Non-Market Goods:  A Common Theoretical Basis .  . .2
     2.2      The Substitution Framework	8
     2.3      Iterative Bidding Techniques  	 16
     2.4      Biases and Limitations of the Contingent
                Valuation Approaches  	 19
     2.5      Advantages in Using Survey Instruments  	 29
     2.6      A Summary of Recent Case Studies	40

Chapter III   Paired Sample Methodology:  The South Coast Air Basin .  . 50
     3.1      Rationale for Paired Sampling	50
     3.2      Socioeconomic Control Considerations  	 51
     3.3      Census Tract Pairings 	 52
     3.4      Description of Paired Areas 	 53
     3.5      Ambient Concentrations for the Paired Areas 	 58

Chapter IV    The South Coast Survey Questionnaire Study  	 65
     4.1      Survey Instrument Design  	 65
     4.2      The Photographs Accompanying the Survey 	 71
     4.3      The Surveying Procedures	77
     4.4      Preliminary Empirical Results for the Iterative
                Bidding Portion of The Survey Instrument Study  .... 81

Chapter V     The South Coast Property Value Study	106
     5.1      Overview	306
     5.2      Data Characteristics	10.8
     5.3      Empirical Analysis	114
     5.4      Summary	130

Chapter VI    Preliminary Comparisons Between Property Values
                and Iterative Bidding Results	133

-------
Appendices
                                                                     Page
    A.     The Survey Instrument for the South Coast Air Basin .  .  .  137
    B.     Health Pamphlet Employed in the Learning Experiment .  .  .  157
    C.     Mean Bids by Area by Type for Alternative Clean-up Dates   172
    D.     Preliminary Regression Relationships for Selected Variables
            in the South Coast Experiment  	  179
    E.     Variable List for the South Coast Air Basin Experiment   .  190
    F.     Paired Sample Area Maps	203

References	216
                                    vi

-------
                                 FIGURES

Number                                                                Page

2.1     Willingness to pay Comparison for Property Values versus
         Iterative Bidding Results 	
                                                                   •
2.2     Effect of An Improvement in Information on Consumer's
         Surplus   	31

2.3     Effects of Costly Exchange	34

3.1     Isopleths for Nitrogen Dioxide Levels in the South Coast Air
         Basin   	60

3.2     Isopleths for Total Oxidant Levels in the South Coast Air
         Basin   	61

3.3     Isopleths for Total Suspended Particulates in the South Coast
         Air Basin	62

3.4     Isopleths for Nitrogen Dioxide, Total Oxidant and Total
         Suspended Particulates in the South Coast Air Basin 	 63

4.1     Information Collective Flow for Survey Instrument  	 66

4.2     Information Sequence in Survey Instruments 	 69

4.3     Observation Paths in the South Coast Air Basin   	  .  . 76

4.4a    Photographs Depicting Observation Paths for "Good" Visibility  78

4.4b    Photographs Depicting Observation Paths for "Fair" Visibility  79

4.4c    Photographs Depicting Observation Paths for "Poor" Visibility  80
                                    vii

-------
                                  TABLES

Number                                                                Page
 2.1  Summary of Information Requirements and Results for Each
       Approach	   17

 3.1  U.S..Census Information for the Paired Areas  	   54

 3.2  South Coast Air Basin Pollutant Information .,.,.,.,...   59

 3.3  Daily Maximum Hourly Average Concentrations of Various Pollutants
       in the South Coast Air Basin	   64

 4.1  Outdoor Activity and Cost List  ................   67

 4.2  Indoor Activity and Cost List	   68

 4.3  Survey Instrument Type Breakdown	   82

 4.4a Mean Bids by Area by Type	,	   84

 4.4b Mean Bids by Area by Type .	   85

 4.5  Results of the t-tests Regarding the Equality of Area Mean
       Bids to Zero	   86

 4.6  Results of the Bid Equality Tests of the Paired Communities .  .   88

 4.7  Results of the t-tests for the Equality of the Mean Bids by
       Sample Area by Bidding Vehicle	„	   90

 4.8  Test for Means for Starting Point Bias	,	   92

 4.9  Results of the t-tests for the Equality of the Mean Bids for
       Observed versus Derived Bids by Sample ARea  .....'....   95

 4.10 Results of the t-tests for Comparing the Sequencing Effects in
       Each Step of the Bidding Process	,	   97

 4.11 Results of the t-tests for Comparing the Effects of Different
       Completion Dates of Cleanup in Each Step of  the Bidding Process  99
                                    viii

-------
Number                                                               Page

A.12   Results of the t-tests for Comparing the Effects of
        Different Completion Dates of Cleanup in Each Step of
        the Bidding Processa,b,c	100

A. 13   Mean Bids by Area by Type for the "New" Data Set	102

4.14   Results of the t-tests for Comparing the Equality of the
        Mean Bids Obtained from the "Old" and the "New" Data
        Sets in Each Step of the Bidding Process3'^	103

5.1    Variables Used in Analysis of Housing Market  ........  110

5.2    Average Housing Characteristics ...............  113
5.3    Number of Homes in the South Coast Air Basin by Air Quality
       Categories	,	,  ,  115
5.4    Air Quality Definition	* *  .  *  ,  .  115
5.5    Sale Price Differentials Attributed to Air Quality  ......  117

5.6    Benefits - Paired Sample Methodology	119

5.7    Estimates Econometric Equations (Linear)  	  122

5.8    Benefits - Linear Econometric Methodology N0? (TSP) 	  123

5.9    Estimated Econometric Equations 	  126

5.10   Estimated Willingness to Pay Equations (N02)  ........  128

5.11   Benefits - Multi-Step Econometric Methodology 	  129

6.1    Alternative Estimates of Monthly Bids by Household, Total
        Benefits for Air Quality Improvement in the South Coast Air
        Basin	134

A.I    Indoor Activity and Cost List .......... 	  138

A. 2    Outdoor Activity and Cost List	139

B.I    Pollution Levels and Standards (Parts Per Million)  	  161

B.2    Carbon Monoxide Effects3(National Standard:  40 parts  per
        million/one hour exposure)  .	162

B.3    Sulphur Dioxide Effects  (National Standard:  .5 parts  per
        million/one hour exposure)  	  163

B.4    Summary of Experimental  Data on Ozone Effects3  (National Stan-
        dard:  .08 parts per million/one hour exposure)   	165

B.5    Nitrogen Dioxide Effects3  (National Standard:   .05 parts per
        million/one hour exposure)  	  167

                                    ix

-------
number                                                               Page

C.I    Mean Bids by Area by Type (Completion Date of Cleanup:  2 yrs) 173

C.2    Mean Bids by Area by Type (Completion Date of Cleanup: 10 yrs) 176

D.I    Preliminary Regression Equations for Bids a' >C'               180

D.2    Preliminary Regression Equations for the Aggregated "A"
        and "B" Areas  . . .....................  181

D.3    Preliminary Regression Equations for the Paired Areas  ' ' '   182

D.4    Preliminary Regression Equations for the Paired Areas a' )C)   183

D.5    Preliminary Regression Equations for the Paired Areas a't)»c'
D.6    Preliminary Regression Equations for the Paired Areas  ' ' '   185

D.7    Preliminary Regression Equations for the Paried Areas ' ' '    186

D.8    Preliminary Regression Equations for the Paired Areas ' ' '    J87

-------
                                 CHAPTER I

                         INTRODUCTION TO VOLUME II

     Determining the benefits of non—market, public, or collective goods
such as environmental quality has become more important as increased regul-
ation by government agencies has imposed heavy costs on the private and
public sector.  The question "Do benefits of environmental programs exceed
costs?"  is being asked with increasing frequency.  This study attempts both
to compare methodologies for estimating the benefits of environmental control
and to provide specific estimates of the benefits of air pollution control
for selected areas of the South Coast Air Basin of Southern California.

     Although a number of methodologies are available, here we focus on
two; the use of survey instruments and on the use of property value differ-
entials to measure the value of air quality.  The overall focus of Chapter
II is primarily to set the existing work on non-market valuation in per-
spective.  We initially present a theoretical basis for the variety of
valuation approaches in Section 2.1 where the theoretical linkages between
techniques such as the property value and survey approaches are shown.  The
structure of what is termed the survey instrument substitution approach will
be discussed in detail in Section 2.2.  The iterative bidding technique also
using the survey instrument will be presented in Section 2.3.  Endemic to
both the substitution and iterative bidding approach are potential biases
and other limitations on what we term the contingent valuation approach in
which hypothetical situations are utilized.  In Section 2.5, arguments are
presented which suggest some advantages to the contingent valuation approach.
These arguments have been typically ignored while the bias issues in Section
2.4 have tended to dominate the propositions in existing literature.  Section
2.6 presents a brief review of the existing work using contingent valuation
methods.  The remainder of the volume is devoted first to reporting in
Chapter III on the paired sample methodology used in the South Coast Air
Basin study.  Then, the design and results of the South Coast survey instru-
ment are presented in Chapter IV including summary estimates of total air
pollution control benefits in the South Coast Air Basin.

-------
                               CHAPTER II
                         THEORETICAL FOUNDATIONS

2.1  Valuing Non-Market Goods:  A Common Theoretical Basis

     The provision or control of collective goods and bads has fallen
implicitly to the public sector.  Given the economists goal of efficiency,
implying a price oriented framework, an immediate problem arises.  That is,
the price of a public good or intangible cannot readily be observed in
a market setting yet the employment of benefit-cost analysis requires some
form of price information for such goods as aesthetics.  How then is
the requisite information obtained?

     Several approaches have been recently subjected to theoretical
development and empirical scrutiny for ascertaining the value of non-
market goods.  One such approach is a direct valuation method which,
simply stated, asks the consumer to bid in a highly structered situation
the dollar values for alternative levels of provision of the public
commodity in question [Davis, 1963; Kurz, 1974; Bohm, 1972; Bohm, 1971;
Randall, et.al.,  1974;  Brookshire, et.al., 1976; Blank, et.al., 1977;
Randall, et.al., 1978;  Thayer and Schulze, 1977; Brookshire and Randall,
1978].  Another approach is the travel cost method which has had many
empirical applications [Knetsch, 1963; Clawson and Knetsch, 1966;
Davis and Knetsch, 1966; Pearse, 1968].  In the case of air pollution
there have been more than several property value studies:  [Ridker and
Henning, 1967; Anderson and Crocker, 1971; Freeman, 1974].  These
property value studies are based on the hedonic market approach which is
theoretically founded on the household production model of the consumer,
[Lancaster, 1966; Rosen, 1974; Muellbauer, 1974; Hori, 1975].  Consumers
are conceptualized as combining private and public commodities through a
household production function to produce characteristics that the
consumer values.  In these studies, observed variations in market prices
are associated with characteristics such as environmental quality.
Finally, questionnaire approaches have been employed to gather data on
household technology and preferences.  A primary motivation in the
development of this indirect valuation approach was to provide a cross-
check for the empirical estimates derived from direct valuation methods,
[Blank, et.al.,  1977].   Actual empirical efforts to apply this indirect
valuation method have concentrated on producing data useful in estimating
the household technology, [Blank, et.al., 1977; Brookshire, Randall, et.al.,
1977].

     The "direct and indirect valuation methods rely on what are essentially
a set of hypothetical situations, both in terms of the level of provision
                                      2

-------
of the public commodity, as well as in  terms of  the property rights
structure associated with use of the public commodity.  Therefore, the
empirical values which are produced are contingent upon the hypothetical
structures which are presented to the consumer.—  It is for this reason
that these two methods have come to be  known as  "contingent valuation"
methods.^./

     In the contingent valuation format consumers are queried as to
willingness to pay, willingness to accept compensation, past and current
experiences, potential expenditures adjustments, and so forth in estimating
compensated demand functions for public commodities.  Thus the motivation
for contingent valuation approaches is  to produce valuation measures that
can be used in benefit-cost analysis under the Pareto improvement
criterion.

     The emphasis to date in developing non-market valuation techniques
has been toward choosing a well-defined public good and designing a survey
instrument for the valuation procedure.  This process involves issues of
replication,A'  bias testing,A'  and methodological cross-checks .A'

     These efforts have fallen short in several ways.   First, the variety
of approaches have no common theoretical framework.  Second, no acknowledg-
ment of the various characteristics of  the good have been set forth.
That is, the good "air quality" can have an aesthetic characteristic, a
health characteristic, plus others possibly.  Thus a bid or value placed
on changing levels of air quality,  where only a single characteristic such
as aesthetics is bid upon, will possibly produce a lower bound estimate.
Finally, the lack of a common theoretical framework has precluded designing
a survey instrument which obtains information for every individual,
enabling a cross testing of various methodological approaches for valua-
tion.-^'

     The variety of approaches used to value public goods lack a common
theoretical basis.  Whether the analysist employs contingent, actual
observed behavior or market prices, the results have been based on narrow
theoretical structures which have little relationship to others.  That is,
the initial assumption sets are not identical and differing modeling
structures further aggrevate the problem.

     Certain characteristics must exist in a common modeling structure,
such as the possibility of consumer substitutions across activities and
sites,  and must include site or activity specific levels of environmental
quality.  Individual utility can then be specified as a function of levels
of activities,  A,, .  .  -^A^, .  .  . .A^ (where the subscripts denote either
sites or different activities for a given site) as a function of environ-
mental quality for each environmentally related activity or site,
Q]_, . . . ,Qi,  .  .  . ,Qn (where we take increases in Q.j_ as increasing
environmental quality), and as a function of a composite commodity X.
Utility is then a quasi-concave function.

               U(A, ,  .  .  .,A ;  Q   .  .   . ,Q ; X),                      (.2.1)
                  i         n   1         n

-------
where 3U/3A. = U^ >_ 0, 3U/3Q. = U^ > 0, and 3U/3X = U  > 0 so utility is
           1    A           i    4 —                 x ~
increasing in A., Q., and X.  Of course, a number of assumptions on the

separability of U are obvious given environmental quality is related to
specific activities in the model.  However, we do not pursue that issue
here.  Rather, we focus on the form of an economic unit's marginal
willingness to pay for environmental quality.

     The budget constraint necessary to specify the individuals optimization
problem is given as:

                              n
                          Y - Z   PA-X>0                       (2.2)
                             1=1   i i

or income Y minus the sum of expenditures on environmentally related
activities   n
             £  P.A. (P.  is taken as the price of activity i which may, in
            i=l  1 1   x
fact, represent joint consumption of several market commodities) minus
expenditures for the composite consumption commodity X (price is taken as
unity to simplify the analysis).

     For a given vector of environmental quality, an economic actor will
then choose to allocate his activities such that (2.1) is maximized
subject to (2.2) which in turn implies that:

     U*          UA
     Y~  <  P., (rf  - P.) A. = 0, A. >_  0     i = 1,2, . .  .,n,     (2.3)
     ux  -   i   ux     11       i

or the marginal rate of substitution between activity i and the composite
commodity X equals the price of activity i - if that activity is chosen
(A  > 0).   We, of course, assume X > 0.
  i
     To determine the marginal willingness to pay for environmental
quality at a particular site, for example i = 1, we set utility as given in
equation (2.1) equal to a constant and totally differentiate the resulting
expression.   By then taking the total differential of equation (2.2),
setting dQ.  = 0 for i ^ 1 and by using (2.3) we obtain:

                          dY      n       dPi     Un
                          —   =  E   A   	-  -  —^-                (2.A)
                          dQl    i=1  Ai  dQl     ux                    ^

                                          (a)     (b)

as the change in income necessary to offset a change in environmental
quality at site 1.—   Another expression for dY/dQ can be obtained simply
by taking the total differential of the budget constraint, equation (2.2),
(again setting dQ  = 0 for i ^ 1) :
                 i

-------
               ^  -   "l   A    i  +   ?  P   ^  +  «
               d"l    1=1   * <%     1-1 ?1  %     %'            *  '5)

                            (c)           (d)          (e)

presuming that the dA./dQ. are consistent with constant utility.  Comparing
the two expressions for marginal willingness to pay implies that since  the
terms  (a) and  (c) in equations (2.4) and  (2.5) respectively are identical,
that :

               n     dA.            U'

                          +             <0'
so the sum of the terms (d) and (e) in equation  (2.5) are negative.

     The interpretation of (2.5) provides the basis for comparing various
methodologies.  If the objective is to determine the marginal willingness to
pay for environmental quality dY/dQ , one obvious approach is to simply
postulate in a survey instrument that Q  changes by a small amount,
and request information on the contingent willingness of the individual to
accept compensation for a decrease in quality or payment to prevent a
decrease in quality.  This direct approach, however, is open to questions
of bias, a topic we take up in Section 2.4.

     A second approach, which we term the substitution approach, is to
assume that prices of activities do not change in response to change in
environmental quality.  For many situations this may well be a reasonable
approximation.  For example, if an energy development such as a powerplant
disrupts a recreation site, recreationists may respond by driving further
to other alternate sites.  If no entrance fees are employed or if such
fees are institutionally fixed, if driving costs, the price of gasoline,
etc., and prices of recreation equipment don't change, then the
assumption that dP. = o appears to be a good one.  In that case the
marginal willingness to pay becomes identical to (2.6) above or:

               jv      n      dA.     ,v
               £l   =  r  p   _ i  +  dX  <  n                       (2 71
                                             °'                       (2'7)
Where prices are known, estimates of the value of environmental quality can
be obtained empirically by collecting data on dA./dQ  (the change in the
pattern of recreation activities in response to a change in quality), and on
dX/dQ  (the change in expenditures not related to recreation activities).
Of course, the change in environmental quality can be contingent, resulting
in changes in activities, or actual cross-sectional or time series data
can be employed where environmental quality varies over space or time.  In
any case, all studies to date focusing on substitution of activities or
commodities in response to changes in environmental quality that we are
aware of  have assumed prices to be fixed.

-------
     In contrast  to the actual or contingent  substitution approaches,  the
hedonic approach, focusing on  price effects of  changes  in environmental
quality effectively assumes both the allocation of  some activities and
other expenditures is  invariant to changes in quality  (dA./dQ  = 0,  for
some i, and dX/dQ = 0), but also assumes  that  all  prices other than P  ,
the price associated with A and in turn  Q1? remain  fixed  (dPi/dQi =  0,
V. <£ 1).  Thus,  from equation  (2.5):

                                    dP
                          -^   -  A    l                               a K\
                          dQl  -  Ai dQ^   •                            (2'8)
As an example  of  this  appraoch, consider a study which  uses changes  in
property values  of homes near  streams and  lakes in  response to changes  in
water quality  as  a measure of  the value  of water quality improvements.
Serious questions must be raised however,  concerning the reality of  the
assumptions that  other prices  and levels of other activities are fixed.
For example, if  water  quality  deteriorates in a small lake, local
residents may  well substitute  other recreation  alternatives so property
values will not  fully  capture  the willingness to pay for water quality.

     In summary,  the marginal willingness to pay of individuals for
environmental  quality  can be determined  as shown in our theoretical
context by three  approaches.   First, individuals can be directly asked  to
provide their  marginal willingness to pay:  dY/dQ-,.  Second, assuming no
price changes  occur, information can be  collected on dA^/dQ-i and dX/dQj,
the substitution  of activities and expenditures which occurs in response
to a change in environmental quality.  From this data one can impute a
marginal willingness to pay.   Third, assuming the allocation of activities
and expenditures  is invariant  to a quality change and assuming all prices
but one are also  invariant, the change in  the single remaining price, dP-j_,
can be used to impute  environmental benefits.   Of the three approaches,
'the one which  requires the fewest a priori assumptions  and minimal data
collection is  the first, contingent valuations  derived  from survey instru-
ments.  However,  this  direct approach remains to be systematically compared
to other methods.  The empirical portions  of  this research attempt to
address this situation.

     One final point needs to  be made with respect  to non-marginal changes
in environmental  quality which require in  turn  that proper measures  of
willingness to pay as  opposed  to marginal willingness to pay be utilized
for comparing  alternative methodologies.   In  the empirical studies presented
below, individuals were asked  to bid on  non-marginal changes in air
quality.  These  direct non-marginal bids are  then compared to the changes
in property value which are associated with a similar shift in environmental
quality.  What then is the theoretical relationship between the property
value measure  of  willingness to pay as compared to  the  direct asking
approach?  If  we  assume that property values  capture the entire willingness
to pay for clean  air,  then Figure 2.1 provides  an answer.

     In Figure 2.1 monthly rent or equivalent monthly payments for owner
owned homes is plotted on the  vertical axis.  On the horizontal axis we
plot air quality.  Now, hedonic price theory  implies that if people
prefer clean air, rents should rise across a  region (everything else

-------
                                 Figure 2.1
         Willingness  to  pay  Comparison  for  Property Values versus
                          Iterative  Bidding  Results
Rent/Month
     AR
                                                      Air Quality

-------
held constant) as the air quality improves.  This results in the rent
gradient denoted by R in Figure 2.1.  Individuals with different preferences,
tastes, over different levels of air quality will locate at different
points along R.  Thus, an individual A with indifference curve 1^ chooses
to live in an area with poor air quality, Ai , and pay lower rent, while
an individual B who prefers clean air with an indifference curve Ig will
chose to live in an area with better air quality, A.2, but must give up
income to pay the higher associated rent.  How, if we ask individual B
located at air quality level A2 how much at the most he or she is willing
to pay to improve air quality at that location to A-j, the individual should
be willing to pay $B per month as shown in Figure 2.1.  Note, however,
that if we compare the rents of air quality at location A2 to those at a
location with improved air quality, A3, equivalent to that specified in
the hypothetical question above, the rent difference is AR, as shown in
Figure 2.1, which exceeds the bid B.  This occurs because although the
rent gradient gives the same valuation at the margin as the bid, (the
indifference curve Ig and the rent gradient R have the same slope at A2),
when non-marginal changes in air quality are employed, the rent gradient
moves across individuals of differing tastes with respect to air quality.
In other words, the rent gradient may overestimate willingness to pay
because higher rents in clean air air areas are associated with especially
sensitive individuals and not with the general population..§/  Thus,
although as we have shown in preceding arguments, property value studies
may underestimate marginal willingness to pay, they may also overestimate
non-marginal or total willingness to pay.

2.2  The Substitution Framework

     The analytical framework of the substitution approach is that of the
household production function approach to consumer behavior.  Three
inter-related substitution approaches are possible to implement.  Although
each approach contains empirical characteristics which distinguish them,
they are all the result of a single analytical structure.  It is believed
to be very important to provide consistency checks within the overall
research effort and within, where possible, a single approach.   As such,
each approach, although requiring a separate set of assumptions and
empirical structures, should be able to generate substantially identical
outputs in analytical terms which can be compared.  The analytical structure
along with a description of the three sub-approaches will be presented
in general terms.

     Specifically, this section discusses in detail the hedonic and
substitution approaches set forth in the contingent valuation framework
in the previous section (see equations 2.7 and 2.8).   Clearly,  considera-
tion of the approaches principally involves issues of replication.   While
the overall substitution framework is discussed, the reader should be
forewarned that no empirical results are presented in this report,
However, the necessary data was collected as described below.  Thus,  a
comparison between alternative approaches is forthcoming.^./

     The essential elements in the substitution framework are:   (1) a
utility function (set of preferences); (2) a household technology;  (3)
budget and time constraints;JJi'  and (4) the prices of marketed  goods.

-------
 The  economic  agent  is  presumed  to  derive utility or  satisfaction  from a
 set  of  characteristics which  are produced within the user  unit  itself.
 Inputs  into the  individual's  production  process  are  the  market  goods  which
 the  individual purchases  given  the individual's  income and market  prices.
 Analytically, the individual's  problem is that of maximizing  a  utility
 function, U(Z),  which  is  a  function of the vector of characteristics,  Z,
 subject  to the user's  technology,  Z = F(X),  or means of  transforming  a
 vector  of market goods, X,  into  characteristics,  and subject  to the
 individual's  income  constraint,  P  X = I,  where P is the transpose of  the
 vector  of market prices and I is the individual's fixed  income.

     That is:

                          Max    U(Z)
                            Z

                          s.t.   Z  = F(X)

                                 I  = PTX                             (2,9)

     The dimensions  of  this problem are  perhaps  best illustrated by way of
 a two-stage optimization  procedure.   The  first  stage of the  problem would
 be:

                          Min    PTX
                           X

                          s.t.   Z  = F(X).                             (2.1Q).

 Here, of course, the problem is  to  minimize  the  expenditures  the individual
 makes in producing a given vector of characteristics, Z. The  Lagrangrian for
 this problem  is:

                          L = PTX + A [I  -  F(X)]                       (2.11)

 where A is a multiplier.  First  order interior conditions  are:

                          3L/3X  = P  - F'(X)A = 0
                                                                     (2.12)
                          3L/9A  = Z  - F(X) = Q

 The solution  to  this system of necessary conditions  gives,  first,  a
 system of input  demand functions:

                          X = X(P,Z),                                (2,13)

which are functions of prices and the fixed vector of produced character-
 istics, and second,  a vector of  shadow costs, or hedonic prices for the
 produced characteristics:

                          A = aPX°/3~Z = P/F' (X)                      (2.14)

where X  denotes the optimal value of X.—

-------
     In addition, the minimum value for expenditures is obtained by
multiplying the optimal demands for the inputs by their presumed fixed
input prices.  The resulting function is known as a cost function, .12/
and it gives the maximum expenditure to achieve a given output vector.
It is, of course, a function of input prices out, and the output vector.
chosen:

                          C(P,Z) = PTX(P,Z)                           (2.15)

Note an interesting property of this function.  The partial derivative of
the cost function taken with respect to the characteristic vector is  the
marginal cost of producing the characteristic, which is the hedonic
price.
                          9C(P,Z)/3Z = 3P9Z = A                   (2.16)

Here, is the motivation for the empirical studies utilizing what has come
to be known as the hedonic price technique.  The approach appears quite
simple in that simply relating observed expenditures on measures of
characteristics produces hedonic price equations when the estimated      ,
relation is partially differentiated with respect to a characteristic. —
     The second stage in solving the individual's problem is written:

                          Max     U(Z)
                           Z

                          s.t.    I = C(P,Z)                         (2.17)

This problem chooses that combination of characteristics that the
individual will produce in order to maximize his utility subject to the
condition that his budget be exhausted.

     The Lagrangian for this problem is:

                          L = U(Z) + ccff - C(P,Z)]                   (2.18)

where = is a multiplier.  First order interior conditions are:

                          3L/8Z = U'(Z) - 3C(P,Z)/3Z <* = o

                                = I - C(P,Z) = 0                     (2.19)
This solution to this system of necessary conditions gives, first, a
system of characteristic demand functions:

                          A = Z(A,T)                                 (2.20)

which are functions of hedonic prices, X, since A = 9C(')/3Z, and income,
and second the marginal utility of income:

                          « = 3U(Z)/3I                               (2.21)
                                    10

-------
     The dual to Che problem solved in stage 2 is written:

                          Min     C(P,Z)                              (2.22)
                           Z

                          s.t.    U = U(Z)

The Lagrangian for this problem is written:

                          L = C(.P,Z) + YfU - U(Z)]                    (2.23)

and the necessary conditions for an interior optimum are:

                          3L/3Z = 3C(P,Z)/3Z - U'y = 0                (2.24)

                          3L/3y = "U - U(Z) = 0                        (2.24)

Solution to this system of equations yields income compensated demand
functions for the characteristics:

                          Z* = Z*(X,U)                                (2.25)

which are functions of the constant utility index and hedonic prices.  In
addition note that since hedonic prices are functions of the market
prices of inputs, the income compensated demand for the characteristics
can also be expressed as functions of market prices and a constant
utility level.  Furthermore, equations (2.25) could be substituted into
equations (2.13) to obtain income compensated demand functions for the
inputs as in'equation (2.26).

                          X* - X*(P,U)                                (2.26)

     It is also possible to define an expenditure function as the minimum
expenditure necessary to achieve a given utility level.  The expenditure
function can be obtained by multiplying the compensated characteristic
demand functions by the vector of hedonic prices as in equation (2.27).

                          M(A,U) = XTZ*(A,U)                          (2.27)

Similarly, the income compensated input demand functions might be
multiplied by the vector of market prices in order to obtain an expenditure
function relating directly to input markets.  However,  since income
compensated demands for the characteristics were utilized in obtaining
income compensated input demands, and since hedonic prices are functions
of input prices, deriving the expenditure function with respect to
characteristics would be equivalent to deriving an expenditure function
looking at the input side.

     Now that the basic analytical structure has been presented,  each
specific approach will be outlined.   Each approach is embodied in the
analytical structure and is distinguished by the point within that
structure at which empirical estimation begins and with respect to the

                                   11

-------
information required to implement each approach.  List 2.1 presents an
outline of each approach.  The following discussion will follow the
outline presented in this list.

     The expenditure equation approach is primarily involved in estimating
hedonic prices and ordinary characteristic demand functions.  The first
step in this procedure is to estimate a cost function, C(P,Z).  This is
done by relating estimated expenditures by individuals to individual
observations of characteristics, Z.  Note that since interest at this
point is only with obtaining estimates for A, the hedonic prices, it is
not necessary to actually obtain information on market prices, P, in
order to obtain estimates of A.  Partially differentiating the estimated
cost function with respect to each characteristic will provide A\(" denotes
estimates, and i indexes a characteristic).

     In simple cases, where the technology is nonjoint and linear
homogeneous, C(P,Z) will be linear in Z [Pollak and Wachter, 1975].
However, when the technology is joint, it will be the case the C(P,Z)
will be nonlinear in Z.  When this is true, differentiation of the
estimated cost function with respect to the characteristics yields a
set of hedonic prices which are functions of the characteristics them-
selves.  It is therefore possible at this point to estimate individual
hedonic prices, Z. , for each individual user k in the sample (see A.3 in
List 2.1).  Now  1  that individual prices have been obtained, characteris-
tic demand functions can be estimated by relating individual observations
for characteristics to estimated individual hedonic prices and incomes
(see A.4 in List 2.1).M/

     There would appear to be at least three problem areas in this
approach.  First, it must be noted that the procedure is estimating a
system of characteristic demand functions.  Standard econometric procedures
suggest that the two-stage least squares estimation technique be utilized.
This procedure will yield consistent but not unbiased estimates of
hedonic prices [Theil, 1971].

     Second, an identification problem exists in that the data utilized
in the estimation procedures is simultaneously determined by supply and
demand considerations.J-5/   Thus, even considering that there is a system
of demands to estimate, there is also a system of characteristic supply
curves for which, unfortunately, it appears there is not enough informa-
tion to allow a solution to the identification problem to be devised.  Work
will be undertaken on this problem in the hope of identifying an
additional set of information that could be obtained from the substitution
portion of the questionnaire which will allow solution of the identifica-
tion problem.

     Finally, there is some apprehension with respect to the step which
estimates individual hedonic prices on the basis of individual characteris-
tic demands, and then turns around to estimate demand curves using the
same information.  This procedure seems to be somewhat circular, and may
seem somewhat more questionable in the case when C(P,Z) is estimated to be
linear in Z.  In this case, how is the expenditure equation approach
able to estimate characteristic demand curves since there would be no
                                   12

-------
                                 List  2.1

                           PROCEDURE OUTLINE

A.   Expenditure Equation:

     1.   Estimate C(P,Z) by relating  expenditures  to  Z.

     2.   3C/3Z = A; C nonlinear in Z  ->•  X  = f (Z) .
                   .*.
     3.   Estimate A   = f(Z.,} where  k  represents  individual  observations
                    ik      ik
          and i represents characteristics.
                                                 .«,
     4.   Estimate Z = Z(A,I) by relating  Z  to  A   and  I  .
                                           i     i      i


B.   Expenditure Function:

     1.   Estimate Z = F(X) by relating  Z. to X. to obtain  technology
          for representative household.  1     1

     2.   Assume U(Z).

     3.   1 and 2 -> specific form for M(X,U) and Z* = Z*(A,tJ),  both of
          which are derived.


C.   Cost Function:

     1.   Estimate C(P,Z) by relating expenditures to Z  and P  .
                                                       i      i

     2.   Define marginal rate of substitution between Z  and Z , as
                                                        1      J
                             , ac(P.Z)
                             '  -  - and
                                                                    . . . , n .
           . .            j£    '    97
                         i         1
                                   J                    3R.      3R.
     3.    If  R.(Z) is continuously dif f erentiable, and — - R  - — 1  R  +
              1                                        8Z1  j   SZ-j^   i


          3R±      3R
          "J7~~  ~  "92  = 0, i,  j = 2, 3, .  .  .,n, then U(Z) is the solution
            J        i

          to dZ  + R (Z)dZi  + Ro(Z)dZ_ + .  .  .  + R (Z)dZ  = 0.
               1    ^     1     3     J            n     n


     4.    From 1,Z = F(X) can be infer£ed in form by picking a Z and
          graphing the isocost  line C(Y,P)  for  different prices P'.  This
          procedure will trace  out an isoquant.

     5.    From 3 and 4, M(A,U)  and Z* = Z*(A,U) can be derived.
                                   13

-------
variation in P?  The expenditure function approach is primarily involved
in deriving income compensated characteristic demand curves.  The first
step in this procedure  involves estimation of the household technology.
Information on the technology is derived from the substitution question-
naire.  The method by which the technology is to be estimated is to relate
individual observations on characteristics with individual input
observations.   This will result in a representative household technology
(B.I in List 2.1).

     The second step in this procedure is to assume various forms for
utility functions,  U(Z).  It is, of course, necessary to either obtain
information on reasonable utility functions or to assume forms for the
utility function in order to derive compensated demand functions.  Given
a utility function and an estimated technology, compensated characteristic
demand functions are derived by solving the problem represented by
equation set (2.22) above.

     There appears to be two major problem areas in this approach.  First,
the necessity of having to assume forms for U(Z) weakens the approach in
terms of believability and applicability.  This approach could be
strengthened by estimating marginal rates of substitution between the
characteristics in order to limit the possible class of utility functions
which is consistent with the empirical results.  The second problem is,
of course, an identification problem with regard to estimating the
household technology.  The data observations which will be obtained from
the household substitution questionnaire will embody both demand con-
siderations and technology since they are presumed to be equilibrium data.
Again, additional work now must be initiated, providing an information set
which will allow solution of the identification problem.

     The motivation for the cost function approach is increased generality.
By using duality theory [Shepherd, 1970; Uzawa, 1964; Hall, 1973;
Diewert, 1974] in combination with the theory of integrability of demand,
[Samuelson, 1950; Hurwicz, 1971] it is hoped that fewer a priori assump-
tions will have to be made in generating income compensated characteristic
demand functions.  The first step in this approach is to estimate C(P,Z)
by relating expenditures to characteristics and input market prices, P.
This is similar to step 1 in the expenditure equation approach, except
that in that approach it was- not necessary to include P in the estimation
procedure.  Utilizing P in estimating C(P,Z) is necessary in this approach
if it is to be later possible to identify the form of the household
technology (see C.4 in List 2.1).

     Given an estimated C(P,Z), it is possible to derive marginal rates of
substitution between the characteristics which can be defined as:

                                       3C(P,Z)
                                         j> y
                            R,, (Z) = -	i                         (2^28)
                             13        3C(P,Z)
                                          3Z.
                                            3
                                    14

-------
          3C(P Z)
Note that 	k^T—   is nothing more  than  the hedonic price  for  character-
            o & .
istic i.  From traditional consumer  theory it  is known  to  be the case
that the marginal rate of substitution between any two  characteristics
will be equal  to the ratio of the characteristics prices at an optimum.
Once R  .(Z) are calculated across all characteristics entering the
      ij
utility function, this approach  turns to  the  theory of  integrability of
demand in ord*er to infer a utility  function up to a monotonic  transforma-
tion from information about the  marginal  rates of substitution.

     Defining Ri;L(Z) = R. (Z) for i  = 2,3, . .  . ,n, if the  R. (Z) are
continuously differentiaole and  if:                        1

     (3R /3Z )R. - (3R./3ZJR. + (3R /3Z.) -  (3R /3Z )  = 0           (2.29)
        ilj      j   1  i      13       ji

               . . .,n,

then U(Z), up to a monotonic transformation, is the solution to the
following equation [Hori, 1975].

     dZ, + R (Z)dZ,, + R  (Z)dZ., + .   . .  +  R (Z)dZ  = 0                (2.30)
       1    2     ^    3     J            n     n

If one assumes input-output separability, the  utility-maximizing level of
Z does not change.  The problem  is  then to find the dollar expenditure
necessary to maintain the given, utility-maximizing level of Z, which is
simply 3C(-)/3 public good.

     The fourth step in this approach is  to derive the  form of the
technology from C(P,Z).  This is done by  picking a Z or output level, i.e.,
level of characteristics, and varying the vector of prices.  This
essentially causes variation in  an isocost line.   The envelope of such line
will trace out an isoquant and therefor provide information on marginal
rates of technical substitution  from which the structure of the household
technology can be inferred.

     Finally,  now that both the  structure of utility and technology have
been inferred from C(P,Z),  derivation of  income compensated characteristic
demand curves can be derived from the problem  represented by equations
(2.22) above.

     This  approach is quite general in that duality theory of cost and
production function have derived a series of propositions which hold
regardless of the particular form of the  technology [Hall,  1973;  Diewert,
1974].  This allows specification of a class of cost functions which are
reasonable forms for such functions to take.    This not only simplifies
empirical  estimation, since certain forms for  C(P,Z)  are not reasonable,
it also strengthens the empirical results such that the possiblity of
specification error may be lessened.

     Although this approach perhaps allows more generality by not
imposing a set of restrictive assumptions for  empirical application, the
approach concurrently requires considerably more  information which is
                                    15

-------
very difficult to obtain.  That is, it is necessary not only to know
expenditures and quantities of characteristics, but prices must also be
known.

     Table 2.1 contains a summary of each approach, its information
requirements, and derivable results.  The most costly approach in terms
of information requirements is the cost function approach, which must
obtain information on the prices of input goods.  However, the cost
function approach has the advantage of not requiring the assumption of
a specific utility function in order to obtain an estimated Z* = Z*(A,U).
However, its exact specification limits the domain of possible forms of
utility functions.  Such an assumption is crucial to the expenditure
function approach.  The expenditure equation approach, while not
necessarily having to assume. U(Z), is aided by such an assumption because
it facilitates derivatiori of estimable forms for Z = Z(P,I).  In addition,
derivation of Z* = Z*(A,U) cannot be obtained withou_t a form for U(Z).'
In each approach it is possible to derive Z* = Z*(A,U) which, of course,
will be where consistency of the approach is ultimately to be tested.

     Blank, et.al. , - (1977) set forth the basic methodology for
obtaining the necessary information for the substitution approaches.  Three
steps can be delineated in this process.  First, the respondent's
initial situation is established.  This is the current level of activities,
locations, and expenditures (fixed and variable).  Second, the respondent
is presented with the contingency such as either an increase or decrease
in environmental quality.  Third, the respondent is asked how, if any,
expenditures or activity patterns would change as a result of a change
in environmental quality.  Steps 2 and 3 are then repeated.  From this
information the analyist  is able to perform the necessary estimation
procedures.  The actual survey instrument employed for this study will
be described in Chapter IV.

2.3  Iterative Bidding Techniques

     The iterative bidding technique involves a direct determination of
economic values from data which represent responses of economic actors to
contingencies posited to them via a survey instrument.  Assuming that the
good under question is of the public good variety, the individual him-
self has no choice as  to  the amount he consumes.  Thus, of the four
Hicksian measures of consumers surplus, only the surpluses are relevant
in valuing changes in air quality.  The individual's problem is then one
of responding to proposed contingencies.  Two types of responses can be
delineated:  willingness to pay (¥TP) and willingness to accept
compensation (WTA).  Thus the bid offered to the individual and the
subsequent welfare position for the WTP can be represented as:

                          U(Q,Y - B) = U(Q',Y)                       (2.31)

and for the WTA as:

                          U(Q,Y) = U(Q',Y 4- C),                      (.2.32)

where U (•) is the individual's utility function, Q is air quality, Y is
                                   16

-------
                                  Table 2.1

                  Summary- of Information Requirements and
                          Results for Each Approach
Approach
Expenditure
Equation
Expenditure
Function
Cost
Function


Information Estimate
z c(p,z)b
z
X Z = F(X)
Y3 C(P,Z)d
P
*7 T? ( 7 ^
Z RljtZ)

Assumptions Result
Z = Z(X,I)
U(Z) Z* = Z*(A,U)C
U(Z) A*(X,U)
U(Z)
Z = F(X)
Z* = Z*(A,U)
      Y denotes expenditures on market input goods.
      Relate C to Z only.
      In order to derive Z* = Z*(X,I),  it is necessary that a form for U(Z)
be assumed.
      Relate C to Z and P.
                                         17

-------
concave and B or C are bids made or compensation received.  Inherent in
the consumer surplus measures represented in equations (2.31) and (2.32)
are the notions of initial position of the consumer (i.e. the current
situation) and the individual's rights structures in relation to the good
in question (i.e. the current level of air quality).  Depending upon the
relationship of the individual's initial endowment  (Y), position (Q),
the rights (either Q' or Q), the delineation between equivalent and
compensating surpluses for WTP and WTA can be set forth.   Thus, the
Hicksian compensating and equivalent measures are conceptually different
in that the reference welfare level is different.  The compensating
measure is defined as the amount of compensation, paid or received which
would keep the consumer at his initial welfare level assuming the change
takes place.  The equivalent measure is the amount of compensation, paid
or received, which would bring the consumer to his subsequent welfare
level in the absence of the change.  To this extent that the different
Hicksian measures are empirically different, except in the quite un-
likely circumstance that the two alternative quantities,  Q' and Q", of
the public good Q, where Q" is larger and ceteris paribus preferred.
The four relevant measures of value are the following:

     1.   Willingness to pay to avoid Q'

             E
             0' Y- 0" Y'  0"
             x » -1 > x j1 > x

     2.   Willingness to pay to obtain Q"

          WTPC
             Q'.Y; Q',Y; Q"

     3.   Willingness to accept compensation and take Q'

          UTA
          WiV,Y: Q",Y; Q'

     4.   Willingness to accept compensation and forego an offer of Q"

             E
          WTV,Y; Q',Y; Q'

where the superscript E indicates the equivalent measure, and C indicates
the compensating measure, the first subscript specifies the individual's
rights in terms of the bundle of goods (Q* or Q") and his endowment of
the numeraire, Y, the second subscript indicates the starting bundle of
goods and endowment, and the third subscript indicates his final bundle
of goods after he has paid his WTP or accepted his WTA.  His final
endowment will be Y plus or minus the amount he actually pays or accepts,
respectively.

     The four measures of value bear the following quantitative relation-
ship, in absolute value terms;16/

UTP               — T.7TP               s 17TA               = UTA
   n' Y- n" Y- n"      o' Y- n1 Y- n" — Wian" Y- n" Y- n1      n" Y- n' Y-n"
   X> 1 > X> -1 > X       X> X J V> I- J X^       X> X > X 5 I > <^       X> * > X> * > 4
                                                                . . .(2.33)
                                   18

-------
     How then are the relevant measures to be obtained?  The Randall, et.al,
 (1974 a and b) study introduced several features which have for the most
 part been retained in later iterative bidding studies.  "The hypothetical
 market is defined and described in substantial detail:

     a.   The alternative levels of provision of the qir quality are
 described in quantity, quality, location and time dimensions verbally and
 wherever  possible depicted in photograph sets to ensure uniform per-
 ception across the respondent population.

     b.   A hypothetical market is created in a substantial degree of
 institutional detail.  Exclusive mechanisms are often expressly introduced
 or alternatively the respondent is assured that all users of the good
 will pay equally (e.g., through tax increments, increments in the price
 of associated services, or charges collected in special funds).  The
 method of payment, called the payment vehicle, is specified and is chosen
 for its feasibility, its familiarity to respondents, preferably as a
 result of its customary use in similar contexts in the real world, and
 sometimes for its policy relevance.

     c.   The respondent reacts to prices posed by an enumerator,
 indicating whether he would, in a WTP case, pay the price or go without the
 good.   The price is varied iteratively, until the price at which the
 respondent is indifferent is identified.   The procedure simulates the
 respondent's typical market experience, where he is confronted with
 specified goods at stated prices and must decide to buy or not to buy."—
 The iterative bidding process represents an attempt to establish a
 hypothetical market having many of the features of existing markets.
 Chapter IV will discuss in detail the iterative bidding format employed
 in this study.

                                                                  18/
 2.4  Biases and Limitations of the Contingent Valuation Approaches-—

     Since the seminal article by Samuelson (1954), general agreement
 among economists suggests that any effort to value public goods will  be
 plagued by the incentive structure facing individual consumers thus
 encouraging them to misrepresent their true preferences.   That is, the
 consumer would believe himself to be better off by not paying for
 provision of a public good while at the same time enjoying consumption
 of the public good because others have paid for its provision.   This
 point of view represents the classic argument for why markets fail to
 provide public goods, and why valuation methods are expected to reveal
values that are biased.

     Certain specific concerns have been identified in pertinent litera-
 ture that all contingent valuation studies must address.   These
 concerns can be of the following types:  biased valuations resulting
 from incentives in the survey instrument producing biased responses
 and structural characteristics of the survey instrument inducing biased
 responses.   The former includes information bias, vehicle bias and
 starting point bias.  Let us consider the nature of each bias in turn
 and its empirical evidence to date.
                                    19

-------
Strategic Behavior

     The substitution approach is an indirect method of eliciting willing-
ness to pay for the public good.   It would not be expected that the
substitution game would be subject to problems of strategic behavior,
believing ,that the consumer has insufficient information on alternatives
to misrepresent his preferences.   Even the consumer who may be perceptive
enough to realize that the method is designed to infer values from his
pattern of activities would be unable to determine, with any accuracy,
the relative values that he is revealing by his pattern of activities.
The consumer could thus only confound, not bias the resulting valuation.

     The iterative bidding approach, however, is a direct valuation
method that would be expected to be characterized by significant incentives
to misrepresent true preferences.  Strategic bias exists in revealed
valuations when the consumer attempts to influence the outcome of the
valuation process by his announced valuation.  The particular case of the
free-rider problem occurs when the individual underrepresents his bid,
hoping to pay as little as possible and still have the desired level of
the public commodity provided.  Incentives for strategic behavior appear
to depend on the mechanism by which the public commodity is to be provided.
In the context of iterative bidding for environmental quality incentives
to misrepresent preferences should depend on how an individual's tax share
is hypothetically determined, on the individual's desire to have changes in
environmental quality, and on his belief as to the extent which others
desire to have changes in environmental quality.

     In order to examine influences on strategic behavior in the context
of iterative bidding formats, a typical game structure which has had
empirical use [Brookshire, Randall, et.al., 1977; Blank, et.al., 1977]
will be discussed.  The consumer is asked to reveal his willingness to pay
for changes in a public good given that, if changes in the provision
level are actually provided, the consumer will have to pay as his tax
share the mean of all bids, as will all other members of the community.
Kurz (1974) discussed this game structure as his "Experiment 2."  He
argues that if consumers act as though they are perfect competitors, in the
sense that they do not believe that their bid will influence the mean bid
and thus their payment, then they will reveal their true valuations.

     In order to examine the incentives that a consumer might have
under this game framework, let us look at any one individual i.A2.'
Assume that there are n individuals in a community which is considering
changing the level of provision of environmental quality, Q, to Q'.  The
cost of having all n individuals reveal their demands would be too
costly, so only k individuals are to be sampled on their preferences
for Q'.  Let £ denote the number of individuals sampled before individual
i, and m denote the number of individuals sampled after i.  The following
notation will be used:

     B. = bid revealed by individual j (J = 1,2, . . . ,£; £ <_ k - 1)

     B  = bid revealed by individual h (h = 1,2, . . .,m; m = k - £ - 1)
      h

                                   20

-------
      /s
      B. = bid revealed by  individual  i
      i          •        <             -

      B. = individual i's true bid

      The mean bid for the  total  sample  can  be  represented  as;
                           £         m    -^
                           E   B. +  Z    B  + B.
                          1=1   J   *_,   &     i
The mean bid for the sample prior  to  the bid by  individual  i  is:
                    B. =
                      .      —         -    .                         C2<35)


     Assume individual i is asked to reveal  his bid  for Q'.   He  is
informed that he must pay the eventual mean  bid as seen in.  (2.34).  A
reasonable assumption about i's motives would be  to  suppose  if he
desires Q' , and if individual i desires to play strategically for a level
of Q', he will attempt to influence 13 to be  equal to Bj_, his true bid.
Certainly no attempt would he made to influence the  outcome  such that
B > Bj occurs.  If individual i knew that he was  the final bidder and that
the mean for the sample before he was sampled was as in (2.35)_,  then  he
would determine his bid to be:

                                      a
                          B. = kB. -  £  B.  (£ =  k - 1),              (2.36)
However, only one individual i could ever be so fortunate to know  the
mean of the sample thus far and be the last bidder.

     If individual i is- not the last individual, he would have to  know £,
         JJ        m                                           _
m, and ( E  B  +  Z  B ) in order to bid strategically to set B =  B  -  In
        J-l  j   h=l  h"                                           i
case he had such information, he would determine his bid to be:
                                                                      (2.37)
This is a great deal of ^information to have available.  Perhaps individual
i could ask to be given B-^ before he would make his bid, but there would
                           m
be no way for him to know  Z  B, , even if he were also told m.  The
                          h=l


                                    21
^
.
1
a
£
Z
1=J
+ m +
B. +
1)B.
E Bh
h=l h

-------
point is, a great deal of information is needed for any individual i
to be effective in strategic bidding under this payment framework,
assuming of course that the individual would desire the sample mean
bid to represent his true bid.

     What if individual i wanted to be a free-rider and wanted to pay
nothing for Qr?  If he bids zero, what does he gain?  He risks pulling
B down somewhat, perhaps even jeopardizing acceptance of Q', and still if
Q' were accepted he will have to pay some positive amount.  Perhaps
individual i "succeeds" in paying less, but he cannot be a free-rider in the
true sense.

     What this discussion has attempted to point out is that this type of
iterative bidding structure implies information requirements far too
great for any individual to effectively bid strategically.  In the face
of such information requirements, it would appear reasonable for the
-individual to reveal his true bid.—Note that individuals can
misrepresent their bids, but they could not be certain as to the extent
they would be acting in their favor (whatever they perceive _that to be),
or whether their actions could really alter the outcome for B perceptably.
In other words, consumers would have the "incentives" to behave as perfect
competitors when confronted with this bidding framework.

     Another reason for discussing this particular bidding game experiment
is that Brookshire, et.al., (1976), and Blank, et.al., (1977), have used
this framework to test for strategic behavior.  Both studies were
attempting to place a monetary value on changes in environmental quality,
which was defined in terms of changes in visibility resulting from changes
in emissions from coal-fired electric generating plants.  Both studies
concluded that strategic bias was not evident in the sample data generated
where the consumer was told he would have to pay the mean of the sample.
Further, Blank, et.al., (1977) specifically developed a set of bidding
formats that attempted to provide the consumer with information that he
might use in bidding strategically.  Specifically, individuals were
allowed to reveal their bids, but were then told that the mean values
based on other studies in similar communities was $X.  The individuals
were then allowed to revise their bids.  Only one out of every 40
individuals revised their bids.   Even given additional information that
would be potentially useful in formulating strategic bids, the consumers
did not revise their bids.  This suggests the absence of strategic
behavior tendencies.

     A number of other studies provide empirical information on the
existence of strategic behavior in revealing the valuation for public
commodities.  Bohm (1972) utilized an experimental approach which forces
actual payment for the publicly provided commodity (public television).
Bohm's conclusion was that strategic behavior was an insignificant part
of this experiment.  However, in a hypothetical context he did discover
significant strategic bias in elicited bids.-^i/  An interesting attempt
at examining strategic behavior and mechanisms designed to dispel
motives toward such behavior was that of Babb and Scherr (1975).   Babb
and Scherr used an experimental setting and three alternative mechanisms
to reveal the valuation for two publicly provided commodities:   a concert
                                   22

-------
fund and a library fund.  They utilized the Clarke  tax mechanism, a leas
familiar valuation mechanism associated with Loehman and Whins ton. Q.971).,
and a "control" mechanism in which  the individual would pay his actual bid.
The third mechanism was called a "voluntary" mechanism and was the control
mechanism because strategic behavior was- expected to be present under
such a system.  Bahh and Scherr found little evidence supporting the
existence of strategic behavior.  In fact, the lowest valuations tended
to be generated under the Clarke tax mechanism rather than the voluntary
mechanism as would have been expected..^!/  In debriefing sessions, it
was discovered that very few individuals attempted  to free ride.  The
respondents indicated the following reasons for not being a free-rider:
(1) feeling of being cheap; (2) funds were worthwhile; and (3) altruistic
reasons.  This set of empirical results is consistent with recent
conjectures by Johans.en (19.77) and  Smith. (1977). as  to why the free-rider
problem may not be of the importance traditionally attached to it by
economi&ts.

     It is not being suggested that strategic behavior may not generally
be a problem in the valuation of public commodities.  The empirical
evidence to support such a conclusion is really not available as yet,
although it could certainly be considered suggestive.  However,  contingent
valuation methods, especially properly structured bidding formats,
seem to provide a reliable framework within which to reveal values for
public commodities.  These methods,  at least in the studies to date,  are
not plagued by the problem of strategic behavior,

Vehicle Bias

     Iterative bidding formats, unlike the substitution approach,  require
some form of payment mechanism by which the good in question is valued.
Early studies of the iterative bidding process, [Randall,  et.al.,  1974],
suggested the need for a realistic payment mechanism to mechanistically
create a market.  Devices employed to date have included entrance fees,
tax structures and utility bills as a, form of payment.

     Essentially,  vehicle bias, arises when the valuation results
demonstrate that either the mean bids or the number of protest votes
varies significantly across vehicles.  Reasons for this type of  result
lie potentially in the respondent interpreting the vehicle as anything
but a form of payment,  A manner of  misinterpretation is when the
vehicle itself represents a change in the rights structure facing the
respondent.  In this case the responses could be a confounding between
a dollar estimate of the public good in question and a "vote" via dollar
amounts on the proposed rights or institutional change.

     Additionally from economic theory,  an individual's substitution
possibilities associated with alternative payment mechanisms, are  different.
When a payment vehicle allows  the individual to substitute over  a wider
range of current commodities purchased,  then the bid should  be higher or
compensation should be related to adjustments in disposal  income  or
wealth,  where the individual has the greatest latitude for potential
substitution.   Practically,  however, a believable payment  mechanism

                                   23

-------
related to income adjustment cannot, in general, be applied.  For
example, surveys are often taken at recreational sites away from the
recreationists' locale or state.  In this case, a wage tax  (or income
compensation) may not be viewed as realistically payable by the
recreationist.  Thus-, there is a tradeoff between accuracy associated
with a less then ideal method of payment and the believability of the
vehicle for payment or compensation.  The reduction in substitution
possibilities for a more believable payment vehicle is likely to reduce
the contingent expenditure or increase the contingent compensation
estimate.

     Randall, et.al., (1978),  Brookshire, Randall, et.al.,  (1977) failed
to observe vehicle bias at statistically significant levels.  However,
Blank, et.al., (1977) did report the existence of vehicle bias.

Starting Point Bias

     The contingent valuation approach commences with questions on payment
(and/or compensation) for hypothetical changes in environmental attributes.
It has been found in most sample surveys that it is better  to ask the
recreationist (or any type of interviewee) a question with a "yes" or
"no" answer than a question requiring explicit calculations [see Randall,
1974; Brookshire, et.al., 1977].  It is presumed the recreationist can more
accurately respond to the yes/no question framework, although to our
knowledge this proposition has not been analytically tested for responses
to contingent valuation questions:.  Given the proposition that yes/no
responses are desirable, it is necessary to suggest a starting bid or
minimal level of compensation.  The potential bias arises with starting
points from at least two possible sources.  First, the bid itself may
suggest to the individual the approximate range of appropriate bids.
Thus, the individual may respond differently depending on the magnitude
of the starting bid.  Second, if the individual values time highly, he
may become "bored" or irritated with going through a lengthy bidding
process.  In consequence, if the suggested starting bid is substantially
different from his actual willingness to pay, the bidding process may
yield inaccurate or only roughly approximate results.  The effect of these
two types of starting point biases- may substantially influence the accuracy
of contingent valuation and therefore the usefulness of the approach for
assessment of preferences.

     Several studies have explored whether starting point bias exists
[Brookshire, Randall, et.al. , 1977; Thayer and Schulze, 19.76; Randall,
et.al.,  19.78; Blank, et.al., 19.77].  Only in Blank, et.al.  , (19.77)
has starting point bias been observed in the valuation results.

Limitations in the Structural Characteristics of the Contingent Framework

     Let us turn to the possible confounding of the iterative bidding
process stemming from the structural characteristics of the contingency
framework.±£/  These problems have in the past been termed hypothetical '
bias problems.  Essential to the contingent framework is a clear,
concise survey- instrument incorporating the points made earlier.  Not
only must it fulfill certain requirements from the economist's

                                    24

-------
perspective, but the public commodity must be defined in cognitive and
comprehensive terms.  If these requirements are fulfilled, there remains
a question of how contingent a contingency can he set forth to a
respondent and still receive a, valuation response that is dependable and
interpretahle.  This potential problem can be viewed in terms of
rights and initial endox^ents as presented in Section 2.3.

     Iterative bidding processes propose contingencies to individuals,
often in terms of proposed reallocation of rights or increasing the price
of maintaining an existing right.  However, survey instruments to date
have typically proposed small changes in the rights structure or in the
price of maintaining existing situations.

     Psychologists, Ajzen and Fis.hbein (1977), have specified the
conditions under which behavioral intentions should predict behavior.
The behavioral intention  and the actual behavior should correspond, in
terms of the action, its context, its target and its time frame.  The
iterative bidding format meets these conditions remarkably well.  The
only major problems that may be expected to arise relate to context;  if
the context in the bidding format departs from the policy context in
the real world,  one may expect some difference betx^een stated behavioral
intention and actual behavior.  This provides a warning for researchers
who want their data to be predictive.  On the other hand,  it introduces
a major difficulty in evaluation and validation of the results of
contingent valuation efforts.  Lack of correspondence between stated
behavioral intention (e.g.  "I would sacrifice for clean air") and actual
behavior (e.g.,  I drive an old clunker without emissions controls).
is often explained by differences in context.  That is,  in the real world,
there is no effective market in which one can directly obtain cleaner air
without the cost of some increase in motoring expenses.   Opportunities to
treat contingent valuation data as testable and thus refutable hypotheses
are hard to find.

     While one of the important advantages of contingent valuation
techniques is that they permit exploration of new and different situations,
there are some limits to their value for this purpose..  Extremely large
departures from known and familiar contexts may impede cognition and
comprehension,  reduce the credibility or plausibility of the hypothetical
market,  and in extreme cases, introduce an element of confusion in the
interpretation of responses.

     For instance,  if a large change in endowments or rights was proposed
yet the contingency is still  anchored in the existing rights and endowment
structure,  a change in the individual's production relationship would most
likely require readjustment.   This question would then arise as to the
production relationship or the subsequent one served as  the basis of
valuation.   Further, is there a relationship between the two values or
are the two valuations from the contingencies non-comparable?

     The problem that arises  in choosing which value or  discerning the
linkage, given the contingencies were anchored in the existing rights a,nd
endowment structure, is that, in the large contingency case,  a reallocation

                                   25

-------
of fixed equipment expenditures would seem to be necessary to validate
the response to the proposed contingency.  That is, by posing a large
contingency change, the individual is forced relative to the small change
into possibly a costly exchange.  It is reasonable to assume that, faced
with a charge of $20-$50 in a utility bill, the respondent may be able
to make adjustments in disposal income easily, but it is difficult to
envision such a process taking place when faced with a $1,000 adjustment.

     Setting aside issues of partial versus general equilibrium adjustments
and exchange costs, one must ask the reliability of a proposed contingency
which is not readily linkable to the respondent's existing endowments and/
or rights structure.  Under these circumstances, an extreme contingency
anchored yet far removed from the initial endowments and rights may induce
disorientation in the bidding format and thus in the bids.  The bid could
be "noise," or a "vote" on the contingencies themselves much like certain
vehicles elicit a "vote."

     In fact, iterative bidding frameworks to date have assumed implicitly
zero exchange costs.  Current iterative bidding processes have relied
on a partial equilibrium framework and assumed that the proposed con-
tingencies had an effective zero exchange cost.  Leaving aside how a
general equilibrium iterative bidding format might be designed, the current
processes arrive at a value that is not necessarily "true,"—  However,
the current iterative bidding practitioner merely takes refuge in the
standard assumptions of benefit-cost analysis and assumes a partial
equilibrium framework.

     Setting aside the general equilibrium bidding problem and examining
briefly, as simply as possible, the effect of exchange costs on bids, the
Bradford bid curve framework implicitly assumes a world of frictionless
contingent markets, and thus a bid is- void of any exchange costs,
However, when large contingencies are propos-ed, the assumption that a bid
is void of the potential exchange costs becomes unrealistic.  To examine
the effect of exchange costs on a Bradford bid curve, let us assume the
following simple utility relationship:

                          U(E,Y) = U(E,Y~B-Z),

where E is the public good, Y is income, B is the bid, and Z is exchange
cost associated with a contingency.  Assuming UF >_ 0; Uy >_ 0, the derived
bid curve is:

                            _ a + by(E' - E) - bE'Z
                          B            ' + c)
if Z = f(AE), where f > 0. and f" > Q, then 3B/9Z < 0 where b > d and
c > Q.-iL-i/  Thus, as the proposed contingency is further from the initial
endowment or the existing rights, B decreases as Z Increases.  If B
represents the respondent's bids in a frictionless world of zero exchange
costs, the greater the underestimate of the aggregate bid will be, the
greater is Z.


                                    26

-------
     Randall, et.al.,  (1974b), Blank, et.al.,  (1977), and Brookshire,
Randall, et,al.,  (1977), using iterative bidding  techniques, and Hammack
and Brown  (1974), using open-ended questions,  found differences between WTP
and WTA many times greater than those predicted by equation  (2.33).  In
addition, many respondents were simply unwilling  to respond to WTA
questions,  preferring  instead to make a statement to the effect that
"there is no amount of compensation  large enough  ..."  This seems to
be another  case of the context correspondence  problem whereby the initial
rights and  endowments  as well as- the terminal  rights and endowments are
far removed from  the existing situation.

     Since  consumer surplus measures are tied  to  initial rights and
endowments, it is plausible to argue proposed  contingencies and thus
bids must be anchored  in the existing rights and  endowment structure to
be reliable.  In  this  case, a person's responses  to a contingency are
path dependent in that previous experience and preferences "direct" the
response.   Thus,  if the individual is in a state  of the world, SOTW,
whereby the preference set is formed by t-1 experiences, responses to
contingency will be forthcoming from the perspective of tastes, production,
and exchange costs of  SOTW .  For small contingency changes, the individual
may view the adjustment as costless and employ "familiar" preferences
in answers.

     However, some bidding formats to achieve  certain surplus measures
place the respondent in contingencies that are not a small deviation
from some SOTW  to some reasonable close contingency but which in fact
represents a discrete movement to a SOTW .  In this situation the
individual has no realistic t-1 experiences to draw upon.   Furthermore,
a confounding might arise in that responses that will be forthcoming
might rely upon the SOTW  information.   Now this potentially will
present a confounding because there exists no a priori reason x^hy tastes
and exchange and production costs are necessarily identical or even map
in a systematic manner into a SOTW  relative to a SOTW .  In this case
there is no basis to assume that preferences are also identical.   The
contingency posed in terms of SOTW  is possibly being answered in terms
of information from SOTW  with no reason to assume the information is
relevant in terms of posited contingency based on the SOTW .  Thus,
where the payment of compensation is not customary' in the real world and
the rights in the real world are opposite to those posited in the
hypothetical market, answers to WTA  (or SOTW-B) questions seem highly
unreliable.
                 •p*
     Finally, WTP  questions also may not be immune from context
correspondence difficulties.   Questions asking willingness to pay to
avoid a threatened welfare loss may generate some responses protesting
the imposition of the welfare loss, if that imposition is  seen as
violating either the existing structure of rights or the respondent's
perception of what is right in the sense of being "ethically proper."
Again,  a subset of respondents may interpret the question  as an opportunity
to vote "no" to a referendum on the threatened imposition  of the  welfare
loss, rather than a command to indicate what adjustments would be made
to the threatened narrowing of the opportunity set.

                                   27

-------
     Iterative bidding studies have recognized the possibility of protest
votes arising from objections to context, whether it be a change in rights
or a welfare loss.  However, identified protest votes may be only the
beginning.  It is possible that a bidding format which generates a high
proportion of identifiable protest votes may also elicit responses which
are biased downward.  Bidding formats which elicit a high proportion of
protest votes should be screened out at the pretest stage.
                                                         r
     The above considerations suggest that, a priori, WTP  questions,
which introduce a hypothetical market in which the respondent can buy
improved situations not currently provided and not expected to be
provided free of charge, may be considered the most promising within
limits.  Pretesting of WTA  and WTP  questions is highly important, with
high incidences of protest votes and unexpected differences between the
means obtained and the mean responses to WTP  questions being indicators
of context correspondence problems.  In this respect, the existence of
a rigorous method of deriving the expected differences between alternative
value measures, as presented in equation (2.33), is of value.  It provides
a test of the null hypothesis that actual differences do not differ from
expected differences, and a method of deriving, where necessary, the
policy relevant value measure from the measure which provides the most
accurate empirical data.  How is the latter measure to be identified?
It is the measure which is derived from the hypothetical market which
exhibits the highest degree of context correspondence.  This answer is
not entirely satisfactory, since it is based on the notion that the best
method guarantees the best results, rather than a rigorous test of the
results as refutable hypotheses.

     Let us reiterate that contingent valuation techniques have several
distinct advantages over the alternative methods which are available for
the valuation of non-market goods.  Contingent markets minimize transactions
costs, permit "trade" in non-exclusive and public goods, and generate data
in a form totally consistent with theoretical models of valuation for
public goods.  The use of contingent markets introduces the possibility
of a variety of influences which may bias or otherwise distort the
results obtained.

     We are of the opinion, supported by considerable but admittedly
inconclusive evidence, that these distorting influences are not endemic
to well designed contingent markets, and that careful pretesting will
expose poorly designed contingent markets.  Nevertheless, we recognize
that, for the very same reason that contingent valuation techniques are
used (i.e., the absence of observable markets in the good under study),
testing of contingent valuation data as refutable hypotheses is usually
not possible.'  Replication, however, is possible using the same methods
with different samples or several different conceptually sound methods to
value the same good.  Replication, while unable to provide conclusive
evidence of validity, is to be encouraged and the results of replication
attempts thus far are encouraging, c.f. Randall, et.al., (1974a and b),
and Brookshire, et.al., (1976).
                                    28

-------
2.5  Advantages in Using Survey  Instrument

Introduction

     Contingent valuation approaches are soundly grounded in economic
theory, however, a major point of contention among economists has been the
use of survey instruments for gathering data.22.'  An additional area of
concern has been whether individuals exhibit strategic behavior when
responding to survey instruments.   In spite of practically no empirical
evidence to support the existence of strategic behavior by individuals
when responding to a survey instrument about environmental and aesthetic
phenomena, economists have expended enormous intellectual energies in
devising ways to cause individuals  to reveal their behavior and their
preferences truthfully when responding to questions about these and other
non-marketed goods.££/  Perhaps because of their complexity few of these
devices have found their way into actual survey instrument construction.
Nevertheless, the sheer volume of papers devoted to the issue of obtaining
accurate revelations of preferences for non-marketed goods gives weight
to any assertion that economists distrust empirical results based on data
generated by survey instruments.

     The purpose of this section is not to debate the reality of strategic
behavior or other biases.  Instead, the intent is to raise the possibility
that economists, by their near-exclusive devotion to the strategic
behavior problem, may, at their own apparently unrecognized cost,  have
neglected many of the analytical and empirical advantages to be reaped
through the use of survey instruments.   That is, they may have concentrated
on the costs while disregarding the benefits.

Need Survey Instruments be Hypothetical?

     Fromm (1968) and many other economists strongly believe that
hypothetical questions generate fictional and therefore inaccurate
answers.   These inaccuracies,  if one judges by the relative literature
emphasis, are caused by incentives the individual has to give untruthful
answers.   The incentives stem from the perceived advantages which xrould
be accrued to the individual if he behaves strategically.   One knows,
presumedly,  that the answers are untruthful because the individual's
observed behavior and the preferences this behavior reveals are often not
consistent with the individual's statements about his preferences.   If one
believes that hypothetical statements are imaginary (fictional),  then he
would hardly be surprised by these discrepancies.   Another interpretation
is, however,  possible.

     The dictionary defines a hypothetical proposition as  a conditional
proposition,  i.e.,  an "if X,  then Y" statement.   A hypothetical question
would then be a conditional statement in the subjective mood,  an "if X
were .  .  .,  then .  .  .?" statement.   In a survey setting,  the hypothetical
question is posed by the interviewer to the respondent;  the respondent
then states how he intends to behave in the posited situation.   Thus,  for
example,  as is frequently done in surveys,  the respondent  might be
shown a number of pictures of different landscapes and be  asked his
expectations about his budget and/or time allocations for  each of  the
                                   29

-------
depicted landscapes.

     Formally, the problem set before the respondent seems no different
than the problem he faces when he plans on the basis of a weather
forecast 60 spend tomorrow afternoon at a picnic.  The respondent's
ultimately realized activities and his planned activities are neither
instantaneous nor coincidental.  If an updated forecast is received that
alters the expected weather,  he may change his plans so that he spends
only enough time at the picnic to each lunch.  Realizing that meteorology
is an inexact and conditional science, he will be prepared to change his
plans again on receipt of new information.

     It would indeed be surprising then if frequent discrepancies did not
occur between responses to hypothetical questions and eventually observable
behavior.  The answer provided to a hypothetical question is tentative and
contingent, just as the potential picnicker's plans are tentative and
contingent.  Both the picnicker and the respondent will adapt their
plans according to the information they receive and the changes in their
curcumstances.  The key point is that the contingent answer is still
acceptable given the well defined circumstances that were presented to the
respondent.  The question of inaccuracy is not whether given a change in
circumstances the observable behavior pattern changes but whether the
contingent answer can be observed when the defined circumstances have not
changed.  Only if the answers relate to the past rather than intended
behavior will a simple comparison of answers with actual behavior suffice
to ascertain the accuracy of the answers.  Otherwise, one must explain how
the individual responds to new information and circumstances in order to
perform the compaTison.

     Even if the previous argument is accepted, the question remains as to
how contingent answers fit into the consumer's surplus framework.  This
framework provides the analytical engine by which economists attach
values to non-marketed goods.

     Assuming for simplicity that the respondent's demand for an activity
is weakly complementary in the non-marketed good of interest, it is easy to
illustrate the relation between a hypothetical environmental or aesthetic
state and consumer's surplus._ri'   In Figure 2.2, participation in the
activity with which the non-marketed good is associated is assumed to have
an invariant opportunity cost of p.  This opportunity cost is independent of
the level of availability of the non-marketed good.  The D curve in
Figure 2.2 gives the individual's income-compensated demand function for
an activity, A, averaged over all possible levels of the non-marketed good.
For example, A might be a fishing activity and the non-marketed good
might be atmospheric visibility.

     The ability to see distant mountains from the fishing location is
assumed to enhance the utility obtained from the fishing activity.  As
shown in the figure, the efficient plan for the individual with no
forecast of the availability of the non-marketed good is to look forward
to undertaking the activity at level 3Q.-12'  At this level, the marginal
value he attaches to an additional planned unit of the activity just

                                    30

-------
9
P
                      Figure  2.2

        Effect  of  An  Improvement  in  Information
                on Consumer's  Surplus
                                                    (D|O
      M
                          31

-------
equals his opportunity cost.  The consumer surplus he expects to obtain
from the activity, once he actually participated in it, is the area
above the opportunity cost line and beneath the demand function.  In short,
the area under the "average demand" function, D, is the individual's
mathematical expectation of the valuation he will attach to his planned
activity levels, once realized.

     Now suppose the individual receives additional information about the
availability of the non-marketed good.  Again for simplicity, assume that
the additional information will indicate whether the atmosphere will be
clear, C, or murky, M, on the day:he plans to undertake his fishing
activity.  The manner in which the fisherman will revise his estimates
about the probability of clear or murky conditions can be described by
Bayes' (1764) rule.jii'  For instance, if the improved information predicts
clear atmospheric conditions, the fisherman's subjective evaluation of his
average compensated demand function will be (D|c).  The level of the
activity he will then plan to undertake will increase to aQ.  Moreover,
the area (b-d-e-f) gives the increase in expected utility if "clear"
is the forecast of atmospheric visibility.  Similarly, if the forecast
is "murky," the fisherman's expected utility level will be reduced to a~,
and the area (b-d-h-g) gives the loss in expected utility due to the
forecast.

     In essence, the consumer surplus an individual expects to obtain from
the availability of a non-marketed good can be extremely sensitive to the
state of his information about this availability.  It is this expectation
that determines his commitment of resources and time — his observable
behavior..I?/  Customary treatments of consumer surplus refer to the
surplus an individual obtained from actually participating in an activity,
given (implicitly) his state of information at the instant of the actual
participation decision.  His information at this instant need not be
complete.  When dealing with a hypothetical situation involving a public
good, the consumer surplus measure refers to the value the individual
expects to obtain.  This decision is dependent on the state of information
about the availability of the public good at the time he is deciding
whether to participate in the activity.  The former situation refers
to the surplus associated with D; the latter situation refers to surpluses
associated with demand functions similar to (DJC) and (DJM).  Expectations
can, in principle, be equally disappointed or fulfilled with D as with
(DJC) or (D|M).  The substance of consumer surplus is not at all altered
by increasing the possibility of information acquisition.  This dismissal
of the use of survey instruments because of their hypothetical nature
seems little more than an insistence that reality conform to analytical
habit and convenience of the economist.

Survey Instruments and Benefit-Cost Analysis

     By attributing discrepancies in stated and realized choices solely
to strategic behavior, economists,  as the preceding discussion argues, may
have often misconstrued the meaning of data acquired by survey techniques.
In addition to strategic behavior and the acquisition of information,
there exists another and potentially more important reason for these


                                   32

-------
discrepancies:  for non-marketed goods,  the hypothetical world circum-
stances posited in instruments differ  from the circumstances  in  the
world of observable behavior.  In this subsection, it will be argued  that
the circumstances in the world of the  instruments correspond  more
closely to the analytical foundations  of benefit-cost analysis.  That is,
data gathered by survey instruments may  often, for non-marketed  goods,
be more consistent with economic theory  than is data generated by
observable, realized behavior.

     Benefit-cost analysis is an attempt to ascertain the quantity of
some numeraire (i.e., current dollars) that the gainers and losers from
some proposed public investment will consider equivalent in value to  their
respective gains and losses.  The price  structure, where price is a
sufficient measure of social as well as  private value, represents the
only terms with which the world with or without a public investment is
evaluated.  Prices, as generated by market exchange and adjusted in
proportion to excess demand, embody all  relevant information  about
relative economic scarcities and are a sufficient means of allocating
resources of their socially most highly valued uses._'   The  benefit-
cost analyst is trying to ascertain what individuals are willing to pay
and/or would have to be paid for the public investment in a world
where markets are pervasive.

     If realized market behavior is used as the data base to  establish
these valuations,  the analyst uses propositions from economic theory  for
two purposes:  (1) to infer what the price structure would be in a world
of pervasive markets; and (2) to reason  from the pervasive market price
structure -to the implied consumer valuations.   When survey instrument
responses are employed for the data base, the first step can be avoided
if the conditions posited in the instrument correspond to a world of
pervasive markets.  One might reasonably question whether the conditions
corresponding to a world of pervasive markets are sufficiently close  to
a respondent's experiences to be meaningful to him.  This justifiable
doubt must be weighed,  however, against  the difficulties of carrying
through the analytical exercises necessary to construct a pervasive
market price structure from initial knowledge of the price structures of
a world where markets for many goods are not pervasive.   The way in which
this difficulty is customarily avoided when using observable, realized
prices is to assume (for simplicity) that the observed prices correspond
to those in a world of pervasive markets.

     It is a relatively easy task to construct examples that make apparent
the difficulties of reasoning to pervasive markets from observations on
non-pervasive markets.   Consider costs of exchange, a phenomenon present
whenever valuable resources (e.g.,  time,  information, legal and police
services,  etc.) must be expended to perform the exchange process.

     In Figure 2.3 the individual's initial endowment of Y^ and ^2 is at Q*
When exchange processes become costly,  the individual's  budget constraint
will vary according to his initial endowment.   This is because the costs of
the act of exchanging Y-,  for Y2 differ from the costs of exchanging Y2 for
Y^.  For example,  from the perspective of a single individual, the cost of

                                   33

-------
                    Figure 2.3

             Effects of Costly Exchange
Z
H
                            34

-------
engaging  in a  transaction in which he  is  to  exchange automobiles  that he
owns  for  clean air may differ  from these  same costs in a  transaction where
he is exchanging clean air  for automobiles.  If  the exchange act  is
costly, an initial endowment of Q implies a  budget constraint of  VQV,
whereas if the exchange act is costless,  the budget constraint is MM,
the customary  form which is an integral part of  derivations of demand
functions and  their associated consumer surpluses.  When  the individual
completes "his  exchanges during the period, he will select Y^ and  Y^ as an
optimum if MM  is operative.  If VQV is the operative budget constraint,
he will select Y£ and Y£.   If some point on  MM other than Q constitutes
the intiial endowment, costly acts of  exchange will mean  that a budget
constraint different from either VQV or MM may be operative because the
costs of  exchange acts may differ by the relative quantities of the goods
in the initial endowment as well as. by types of  goods.  Thus, the
individual's budget constraint may vaxy according to the  form in which his.
initial endowment was accumulated, although  the market value of this
endowment may be identical for many combinations of Y^ and Yj'  Since cos,ts
of the exchange act differ according to the  original (-YpY2). combination,
each combination will result in a different  and  generally nonlinear budget
constraint.  It follows that, from the individual's perspective, a dollar
is not an invariant pecuniary measure.  Instead, the subjective value of
an additional dollar depends on the form of  the  income change, i.e., on
the good  in which the increment is embodied.  Moreover, it appears that
realized market behavior is dependent not only on money incomes and
relative  prices of goods, but also upon the  combination of goods  the
individual starts with and the relative and  absolute costs of exchange
associated with those goods.  These costs of exchange acts are probably
neither trivial nor similar across individuals.

     The huge sums spent on industries (law, middlemen, etc.) whose major
or sole purpose is to facilitate exchanges attests to the non-trivialness.
In addition,  if exchange act inputs,  including native intelligence and
training, are not distributed equally across the population, and if these
inputs contribute positively to the effectiveness of an individual in
producing exchanges,  then costs of exchange  acts will not be similar
across individuals.

     If realized market behavior depends on  the costs of the exchange act
for the bundle of goods an individual holds, if, for the same bundle of
goods, these costs differ across individuals, and if individuals do not
hold similar goods bundles,  then the analytical  effort required to infer
what the price structure would be in a world of  pervasive markets must
clearly be greater (probably much greater), than when all individuals
have no exchange act  costs and when budge constraints are therefore
invariant with respect to the bundle of goods held.   Rather than facing
these and similar analytical complexities directly in order to construct
the price structure  of a world of pervasive markets,  or rather than
simply dismissing the problem as an offensive bother,  it may often be more
effective to  question the individual  about his  responses where he is to
assume that markets  are pervasive.   That is, the individual is allowed to
respond directly to  a perturbation in a hypothetical world of pervasive
markets rather than  having the investigator try  to infer what the

                                    35

-------
individual would do in a world of this sort from information about a
world where markets are not pervasive.

     An individual may be able to state his preferences for a particular
state-of-the-world quite clearly.  However, if markets are nonexistent or
incomplete (as they in fact are for a great many aesthetic and
environmental goods), the individual may have no means to communicate
these preferences.  The very lack of markets is due to the costs of
forming and maintaining them and the costs- of the act of exchange.  In
& survey instrument, a hypothetical (contingent), world can be constructed
in which costless means of communication are available.  On occasion,
therefore, the individual's preferences are perhaps more readily inferred
from his statements rather than from his behavior.  The individual who
drives a 1965 Plymouth Valiant and states that he is "for" clean air
has no market in which he can directly exchange his old heap for some
clean air.  The survey instrument provides this market.

Can Survey Instruments Reduce A Priori Assumptions?

     The ability of the human mind to cope with complex reality is
limited.  Successful grappling requires that the dimensionality of
reality be reduced.  When trying to establish the collection of values
individuals place upon non-marketed goods, there are at least two
general ways to reduce drastically- a number of parameters that must be
estimated.  First, one can draw upon a priori restrictions from the economic
theory of the consumer.  Second, an experimental approach to the question
of data can be adopted.

     Economists who have ever seriously worked with problems of consumer
analysis are thoroughly familiar with three fruitful a priori restrictions
(additivity, homogeneity,  and symmetry) that come from the neo-classical
demand theory of Slutsky(1915) and Hicks (1934).  Further reductions in
dimensionality of the parameter space in. which estimation is to be
carried out can be achieved by judicious invocation of various separability
conditions.^-'   Finally, some recent developments in the application of
mathematical duality principles to consumer theory sometimes allow one to
reduce the number of parameters to be estimated without having to impose
particular monotonicity .and curvature properties upon the consumer's
maximization problem.—

     The second general class of means for reducing the parameter space
includes experimental as well as survey techniques.  These techniques
are advantageous, even though widely neglected in economics, because they
permit the investigator to control the number and levels of different
physical contexts and adaptation opportunities to which the individual
must respond.22J  Disturbances imposed by confounding variables upon the
responses of interest are therefore at least partially controlled for in
the data generating exercise.  This contrasts with the standard practice
of placing sole reliance in an ex post fashion upon the application of
multi-variate parametric estimation techniques.   For a given number of
observations,  survey instruments increase degrees of freedom and the
efficiency of estimators.

                                    36

-------
     The use of experimental and survey techniques to reduce the parameter
space may be advantageous in addition to statistical considerations.  Often,
as noted above, the investigator imposes, ex post, various separability
conditions upon market-generated data in order to make it more tractable.
These separability conditions may imply, for example, that beer drinking
at the local tavern is not a substitute for cross-country skiing.  The
conditions are imposed without consulting the individuals whose responses
are registered in the market data.   They are instead generated by what the
investigator intuitively feels to be "reasonable," and required for
analytical convenience.  It is not obvious that the investigator's
"feelings" and the framework he uses in accounting for what is and what
is not important is to be preferred to actually providing the respondent
with the opportunity to state how he would respond to alternative
contingencies.  The details to be abstracted from are presented to the
respondent rather than being left to the fertile and usually clever mind
of the investigator.  In both situations, simplifications are made that
will permit the investigator to work with the data.  In the survey
instrument case, however, the respondent gets the opportunity to weigh
the importance of these confounding variables in making his choices.  In
the observed behavior case, the investigator is presuming he knows as
well as the respondent, from the respondent's perspective, what is and is
not an irrelevant alternative.  Survey instruments allow the domain in
which the response data is generated to conform to the structures of
the underlying analytical model rather than forcing,  via a set of
possibly tenuous assumptions (e.g., the absence of jointness, the
presence of perfect competition,  etc.),  the real world generated data
to conform to the preconceptions of the model.

     A slightly different facet of  the above point arises with the
recognition that much market data used by economists for empirical
analysis is collected by possibly untrained agents many times removed
from the economist-user.  Often,  this data is collected as by-products
of the activities of organizations  whose interests are far removed from
and possibly much less disinterested than the research economist.^Z'
The old saw about lying with statistics can just as readily refer to the
manner in which data are organized  for presentation as to the manner in
which already organized data are employed for estimation purposes.   Except
possibly in the case of direct investigator observation of market responses,
the generation of response data via survey instruments or experimental
means can make the specific connection between the reporting of data and
its uses for testing hypotheses more strong and certain.   The investigator
then has no choice but to accept the responsibility for the survey
data generated under his direction.  He must accept ultimate responsibility
for the origin of the data, as well as the analytical model and the
estimation procedures used to test hypotheses.

Survey Instruments and Property Right Structures

     Market prices acting as devices to signal and coordinate the
purchases and activities of disparate individuals work well where
resource contributions are easily ascertained and reciprocated by
rewards.  For example, the spot exchange of two currencies requires
no  statement of the terms other than the exchange ratio.  When
                                    17

-------
cardinally measurable and perfectly homogeneous commodities such as
currencies are exchanged, the parties to the enterprise need only count the
quantities exchanged to establish, what they have obtained.

     In valuing environmental goods, there are two issues at hand.  First,
given an existing property right structure and assignment, what is the
value of the good?  Second, what would be the value of the environmental
good if the property rights to the good were to be reassigned or restructur-
ed?  The first issue, while important, can be assumed accomplished if the
second issue can be answered.

     The problem in answering the second issue is that for most environ-
mental and aesthetic goods, the costs of exchange cannot be assumed to be
as trivial as in the currency exchange ratio example.  If one adopts an
economic efficiency perspective, there are harsh impediments to tracing the
parties initially responsible for the environmental or aesthetic effect,
detailing the actual levels of the effect, and finally ascertaining the
contributions of each perpetrator of the effect.

     When these costs of the act of exchange exist, the economic structure
itself becomes a variable of the decision problem.  The problem can be
viewed as one of finding a set of obligations for each individual's
behavior pattern so that his costs and rewards are made less dependent on
his joint relations with other individuals using the same non-marketed
good.  Rules of evidence and procedure are established for all users.
Likely and important contingencies will be specified and appropriate
responses will be stipulated.  The objective being that easily measured
performance standards will be formulated.  In short, the assignment of
property rights as well as the property rights structure itself is changed.
These reassignments and restructurings of property rights have been a
major means by which environmental and aesthetic insults have been controlled,
It is likely they will continue to be so.

     There exist analytical devices in economics that allow one to ascertain
the effect of property rights reassignments of an environmental or aesthetic
good upon consumer valuations..387  These valuations can be established with
time and budget allocation data obtained by everyday behavioral observations
or by survey instruments.  However, where the conditions of use, exclusion,
or alienation are altered (i.e., property rights are restructured), there is
no everyday behavior to observe, except insofar as one is willing to draw
analogies from observed behavioral responses to changes in the property
rights structures of other goods.  If one knew what the availability of
the environmental good would be under the property rights restructuring, it
might seem possible, if one had everyday behavioral observations on consumer
time and budget allocations at the same level of availability, to determine
the change in consumer valuation due to the property right restructuring.
However, the purpose of the restructuring is to reduce the costs of the
act of exchange and, as we argued this reduction can alter
the value the consumer attaches to a given level of availability.  Further-
more, since consumer valuations will, through either the market or the
political process, influence the level of availability, how is one going
to reason from the level of availability to consumer valuations for the


                                   38

-------
restructured property right?  Economic analysis does not yet have a
sufficient understanding of the reciprocal relations between costs of the
act of exchange and property rights structures, nor between these costs and
various demand phenomena, to permit the ready testing of detailed empirical
generalizations in a wide variety of settings.  Thus the only really sound
way of obtaining an estimate of whether the net benefits of restructuring a
particular property right is positive, if one insists upon employing
observed everyday behavior, would be to perform the restructuring and observe
the results.  In some circles, this is simply known as trial and error.
To measure is not necessarily to understand.  Trial and error can be an
extremely costly way to perform research because the errors are real rather
than hypothetical.  In contrast, survey instruments allox^ one to investigate
the behavioral responses to a wide variety of property rights structures
without involving the citizenry in the traumas of what often is euphemistic-
ally termed social experimentation.

     One obviously directly cannot observe everyday behavioral responses to
property rights structures that have never existed.  Similarly, one
cannot directly observe the everyday behavioral responses of individuals
who have never participated in activities involving the environmental or
aesthetic good at the levels- at which the good has been historically
available.  If some of the proposed levels of availability have not been
historically available, and if some former non-participants would become
participants- at these new levels,  the use of data on observed behavior to
ascertain valuations would mean that the valuations of the would-be
participants play no part in determining the valuation.  For each proposed
level of availability, the use of observed, realized behavior to establish
valuations will mean that only historical participants are to count.  Those
who have not participated historically have no opportunity to communicate
their preferences.  Survey instruments, because they allow the researcher
to introduce ranges of availability of the environmental or aesthetic
good that are broader than historical experience, allow the values of
historical non-participants to become relevant.

Conclusions

     The preceding is a taxonomic discussion of some reasons why survey
instruments may often be a superior means of generating data with which
to value environmental and aesthetic goods.  We have argued that economists
have erred in viewing the situations these instruments posit as necessarily
fictional; that the data generated by survey instruments may,  for non-
marketed goods and the activities with which they are associated, accord
more closely with the condition of received economic theory; that survey
instruments can make it easier to remove the difficulties of estimation
and interpretation introduced by confounding variables; and that survey
instruments often permit one to deal more readily with  phenomena that have
not been in the range of historical experience.   These are indeed sub-
stantial advantages that economists have not adequately recognized or
appreciated.  Nevertheless,  whatever the advantages, a major disadvantage
remains.  Until detailed analytical knowledge is acquired of the manner  in
which expectations are formed, there exists no way to refute empirical
propositions- established from survey instruments that inquire into
expected behavior.  In this sense, survey instruments- are non-scientific.

                                    39

-------
2.6  A Summary of Recent Case Studies

     This section reports, in chronological order, four studies which have
attempted to value environmental quality, and thus form the basis for much
of the current report.  The first two studies, the Lake Powell Experiment
and the Farmington Experiment attempted to value air quality near existing or
proposed coal-fired powerplants.  The third study, the Geothermal Experiment,
examined the impact of proposed geothermal powerplant development on an
existing recreation area, calculating possible damages.  The fourth and
most recent study, the Wildlife Experiment, examined the value of wild-
life to recreators in areas which may be impacted by strip mining of coal.

The Lake Powell Experiment

     Lake Powell, with an annual visitation now approaching two million
visitor days, is an excellent example of the tradeoff mentioned above.
The lake was formed by the filling of Glen Canyon and retains the steep
cliffs, rugged terrain features, and scenic vistas one associated with the
Grand Canyon, but are here available to pleasure boaters and other
recreators.  Construction of the Navajo generating station located at the
southern end of Lake Powell, was completed in 1976.  Another larger power-
plant, the Kaiparowits Project, was also proposed for construction near
Lake Powell and became an issue of substantial public concern if not the
primary issue for environmental groups in the Southwest.

     As part of the Lake Powell research project, during the summer of 1974
recreators at Lake Powell were interviewed in an attempt to determine the
aggregate willingness to pay to prevent construction of the proposed
Kaiporowitz plant  [see Brookshire, et.al., 1976].  Photographs of the
existing Navajo powerplant which all of the recreators had seen, (stacks
remain visible more than 20 miles up the lake), were shown to recreators
both with visible pollution emanating from the stacks and with the stacks
alone.  Recreators were then asked what entrance fee they would be willing
to pay to prevent construction of another similar plant; first, where only
pollution would be visible from the lake itself, and second, where both
stacks and pollution would be visible.

     The analysis of the data attempted primarily to deal with strategic
bias.  As noted above, if recreators believed that a uniform entrance fee
might actually be set on the basis of the average bid of the sample survey
to prevent construction, or believed that construction plans might
be affected by the research results, then "environmentalists" might well
bid very high, and "developers" might well bid zero dollars in an attempt
to bias the results.  A theoretical model of strategic bias was constructed
to explain the distribution of obs.erved bids: which would likely be bimodal
rather than normally distributed if strategic bias was present.  The fact
that the actual distribution of bids was noramlly distributed was thus
taken as evidence that strategic bias was not present.  It was conjectured
by Brookshire, et.al. , (19.76), that the absence of strategic bias was due
to the hypothetical nature of the experiment; few respondents felt that
their answers- would effect real world outcomes.  The remainder of the
research was devoted to specifying an econometric model of the bidding
game results to estimate income effects by group.  Recreators were

-------
divided into four categories, developed and remote campers, and visitors
and residents of the nearby  town of Page, Arizona.  Although  the effect of
individual income by group on bids was statistically significant at the
99% level, the income effects were all very small.  It was shown then that
both theoretically and empirically the small income effect impled:  (1)
that a compensated variation measure would not differ practically from the
equivalent variation measure used in the experiment; and (2)  that income
redistribution between groups would not significantly effect  the aggregate
bid.

     The average bid per family or recreator group was $2.77  in additional
entrance fees in 1974 dollars, and the total annual bid, which can be
interpreted as an aggregate marginal willingness to pay to prevent one
additional powerplant near Lake Powell, was over $700,000.  Two points
should be made about these results.  First, they show impressive
consistencies both with the one previous bidding game study [Randall,, et.al. „
1974] in the region as well as with the succeeding Farmington experiment
discussed below.  Second, if the results are accepted as indicative of
recreator preferences in general for the entire region,  the canyon lands
of southeastern Utah, and if the bids are extrapolated to all the effected
recreation areas as well as Lake Powell, the aggregate bid would approach
$20 million per year since there are some 15 national parks and recreation
areas within a 100-mile radius of the proposed Kaiporowitz site.

                         39/
The Farmington Experiment—

     This study attempted to establish the economic value of visibility
over long distances within the Four Corners Region.  The southwest is
characterized by vast spaces and open vistas unencumbered by  industrial-
commercial development,  urban development,  or airborne pollutants.   The
major focus of the study was to attempt to establish how recreationists
and residents value continuing to be able to see over long distances.
Clearly,  the ability to  observe long distances is an almost pure public
good.  The use by anyone does not interfere with use by anyone else.
In addition,  efforts were made to examine the extent of certain biases
including:  information,  strategic, starting point, and vehicle bias on
compensating and equivalent variation measures of consumer surplus.   The
Farmington Experiment also included a  (first) attempt to examine contingent
behavior  changes in response to visibility changes, i.e.,  how people
allocated time between indoor and outdoor activities.

     A survey questionnaire was given to recreationists and residents in
the Four  Corners Region  of New Mexico and Arizona.   The interviewee was
shown a set of pictures  depicting visible ranges from 25 to 75 miles and
asked to  bid across them.   The pictures were taken at the same location.

     Two  rather distinct methodologies were used to examine contingent
valuations for visibility.  The first assumed a utility function with
arguments of visibility  and income and asked the respondent a sequence of
questions on maximum willingness to pay and minimum compensation.   The
second utilized a  utility- function with time spent on indoor and  outdoor
recreation as- the relevant arguments.   With this function in mind,  a


                                    41

-------
sequence of questions were asked the respondent on adjustments in time
allocations by activity where cRanges occurred in visibility»  Thus, the
first approach is an attempt to measure the left-hand side of equation
('2.4) or (.2.5)., while the second, based on contingent behavioral changes,
attempts to measure components of th.e right-hand side of these equations.

     As part of the contingent expenditures approach, direct tests were
made for strategic bias, information bias, vehicle bias, and starting point
bias.  Strategic bias was evaluated by two means.  First, the "game" was
structured so the individual presumed that he would have to pay the
"average" bid, not his own.  The presumption was that if his bid were below
the mean bid and he desired to increase the magnitude of the aggregate bid,
he would bid higher.  Alternatively, if his goal were to reduce the mean
bid, he would revise his bid downward.  Only in the extreme case when the
individual's maximum bid is identical to the mean bid would there be no
incentive for the individual to change.  In addition to this process, the
individual was questioned about his bid being too low.  It was suggested
that his bid was not sufficient to keep powerplant emissions at present
levels for sustained high, quality ambient air, and was then asked if he
would revise his bid.  In only one case did we observe an individual
acting strategically and it turned out to be an Economics Professor from
the local Junior College!  However, fully one-third revised their bid when
confronted with the possibility that their bid was insufficient.  Whether
this latter result is indicative of the presence of strategic downward
bias in initial bids or the effect of new information cannot be ascertained.
Individuals may be acting strategically by subjectively forming their
preferences as to the effect of their maximum bid, selecting the bid
appropriately, and then not revising it.  However, it appears to be an
additional indication along with the results of Brookshire, et.al.,
(1976) that individuals generally do not act strategically, at least
in a meaningful manner to bias the outcome of the result.

     In addition to the tests on strategic bias, analysis was made of
various forms of information bias, essentially trying to establish influenc-
es of various aspects of the game.  It was observed that the higher the
starting bid suggested by the interviewer, the higher the maximum
willingness to pay (equivalent variation) estimate derived from the study.
Thus, if the interviewer suggested a bid of $1.QO higher, on the average
individuals would bid about $.6Q more at a maximum.  Also, the choice
of the method of payment influenced the magnitude of the bid significantly,
as would be anticipated from economic theory.  The bid should increase,
the greater the number of substitutions there are in the form of the vehicle
used to make payment; and this was observed in the results, i.e., individ-
uals were willing to bid higher when confronted with a "payroll tax" than
with an increase in entrance fees*  Finally, it was observed that whether
the individual was given previous information on average bids or not, had
a substantial impact on the maximum bid.  We do not wish to suggest these
results indicate any final conclusions with regard to the information
bias problem with, contingent valuations approach, but they are suggestive
that for these approaches to be accurate, one must be very careful with
the vehicle used for payment and the amount and quality of information
given to the interviewee upon initiation of the questionnaire,

                                    42

-------
     The contingent behavior component of  the questionnaire attempted
through contingent changes in  time allocation to  infer an expenditure
function and compensated demand relation for visibility.  Various procedures
were utilized to approximate the compensated demand curve, primarily by
postulating an exact form of a utility function and estimating a time
related household technology.

     The mean bid per recreationist family per month was $4.06 while their
minimum compensation per month was $17.40..  The compensated substitutions
approach led to estimates ranging from approximately $5.00 per month for the
case where the receptor had no entitlement to clean air to approximately
$280.00 per month with complete entitlement.  However, these estimates are
not directly comparable because the contingent behavior estimates include
residents in addition to recreationists which should increase the magnitude
of the estimate.

     Both the Randall, et.al., (1974) and Brookshire, et.al., (1976)
studies only obtained equivalent variation bids.  The following comparisons
are therefore limited to the EV bids.  Using the  sales tax as a vehicle,
Randall, et.al., (1974) reported yearly mean bids of $85.00 (A to C) and
$50.0.0 (B to C)  per household.  Our yearly mean bids for the most
comparable situations were $82.20 and $57.00.  If one considers that the
Randall, et.al., (1974) figures should be increased by 37 percent to
account for inflation between 1972 and 1976 and that, on the other hand,
the Randall, et.al.,  (1974) figures should be higher as respondents are
also bidding on soil banks and transmission lines, these figures are
very comparable.

     The overall mean for situation A to C in the Brookshire, et.al., (1976)
study was $2.77  per month with standard error of  the mean ($.19).   Adjusted
for the 6.6 percent inflation between the time periods of the studies,
the comparison values are $2.95 and ($.20).  The overall mean for
recreationists for the comparable situation was $4.56 (_$1.11), which is
considerably different.  However, the mean bid ($2.44 and $.23) for
$1.00 starting bids in the Farmington Experiment,  while still statistically
different,  is much closer.

     The Farmington Experiment demonstrated reasonable replicative
consistency with other studies.  It also demonstrated that questionnaire
biases may be serious in attempting to utilize contingent valuation
methodologies.   Extrapolating the equivalent variation measures to all
recreationists using the Navajo reservoir,  an annual estimate of $916,000
is obtained which is an estimate roughly consistent with that in the Lake
Powell Experiment.

                     40/
Geothermal Experiment—

     The Jemez Mountains of New Mexico are both scenic - characterized  by
brightly colored rock outcroppings and forest areas --• and a major  recreation
resource with, fishing, camp grounds,  hiking trails, and hot springs all
located on U.S.  Forest Service lands.   However,  the Jemez Mountains also
contain one of the major geothermal resources in the southwest.   Geothermal

                                   43

-------
leases have been let by the U.S. Forest Service on land which is now used
.solely by recreators.

     Both a bidding game and a contingent site substitution approach were
used to estimate environmental damages to recreators from possible
geothermal development [see Thayer and Schulze, 1977].  Recreators were
shown both photographs of geothermal development in similar mountainous
terrain and a map of the location of possible development relative to
recreation areas.  Noise levels and emissions characteristics were
described in detail.  A bidding game was then conducted with a vehicle
which was a uniform entrance fee to prevent development.  Alternatively,
respondents were asked to indicate what their contingent recreation plan
would be (what sites would they visit including new substitute sites and
how often) if development x^ere to occur.  The subsample which responded to
the site substitution question, was then also asked what they would bid
in the form of a uniform entrance fee to prevent development.  Finally,
starting point for the bidding game was varied from $1.00 to $1Q.OQ in
various subsamples.  Thus, the study was structured to test:  (1) if the
bidding game and site substitution results were consistent; (2) if informa-
tion on alternative new substitute sites would effect bidding game results;
and (3) for starting point bias.

     A set of theoretical models was constructed to estimate a consistent
measure of willingness to pay to prevent development from both measures,
the bidding game and additional travel costs associated with alternative
recreation plans.  Additionally, the model was modified to explain
information bias; how changes in perceived costs of alternative (driving
costs) should affect bids, and to explain starting point bias, individuals
either tradeoff their honest bid against the length of the bidding
process or wish to "please" the interviewer by trading off their honest
bid against what they perceive as the "desired" response.

     The results of the experiment were as follows:  Thirty-five percent
of the respondents indicated they would no longer visit the Jemez area
if development occurred.   This resulted in about a 50% contingent decrease
in visitation.  About 65% of the respondents indicated they would visit
alternative sites more frequently, usually the Pecos Forest area.  Bids
averaged $2.35 per visitor party day while the site substitution measure
yielded a. range of $2.03-$2.S4 depending on the assumed driving cost per
mile.  The results appear to be consistent for the two approaches and
imply an annualized aggregate bid to prevent construction of about
$30.0,0.00. for a 50 megawatt plant.

     More surprising, however, were the results for information and
starting point bias experiments.  Neither bias was statistically significant.
The obvious question is:   Why are these results different from those of the
Farmington experiment? -  which indicated that both information and
starting point would likely be serious problems.  The best explanation
that can be given at this point is that the value of the change -in an
environmental quality proposed in the two studies was more precisely
perceived by respondents  in the geothermal experiment than in the
Farmington experiment.  In other words, respondents would more easily
relate the costs to themselves of "losing" in part a recreation area than

                                    44

-------
they could determine the costs of a change in visibility.
                       A 1 /
The Wildlife Experiment—'

     Through contingent expenditure and behavior approaches, this study
attempted to develop a methodology for valuing wildlife experiences.
The valuations were developed to enable policymakers to judge which sites
may be reserved from energy developments that would seriously impinge on
wildlife.  Hunters and wildlife observers were queried as to their
willingness to pay for "encounters" with various types of wildlife.  The
species examined, all within Wyoming, were elk (Cerrus Canadensis),
cottontail (Sylvilagus Spp.),  coyote (Lanis Latrans),  grizzly bear
(Urrus Horribilis), bighorn sheep (Ovis Canadensis),  trout  (Salmo Spp.),
dipper (Circulus Mexicanus)  and brown creeper (Certhia Familians).
The assumed utility function had as arguments the number of encounters
and length of activity.  Thus, the study attempted to measure both the left
and (components of) the right-hand side of equations (2.4) and (2.5).
Prices for purchase of private goods for the hunting, fishing, or observa-
tion experience were presumed to be constant, which appears, except for
inflationary factors, to be a reasonable assumption.

     A type of vehicle bias was observed as bids were recorded on license
or access fees and also utility bill adjustments.  Difficulties were
encountered in convincing some respondents that competition between
energy development and wildlife herds would be sufficient reason for
utility bill adjustments to be a plausible payment mechanism.   Starting
point bias was tested for,  but was not found to substantially affect the
bids on species commonly hunted.  Thus,  this additional evidence appears
to substantiate the comparison between the Lake Powell  and Farmington
Experiments which led us to propose that the more clearly identified the
change in environmental attribute is, the lower the probability of
encountering starting point bias.

     This experiment also examined contingent valuation approaches applied
to the concept of option demand for grizzly and bighorn sheep hunting *
Preliminary evaluation of the responses  indicated,  this may be an effective
approach for obtaining option value and  existence value extimates.
Valuation analyses have not been fully exploited in this study as yet.
But, preliminary results indicate that for elk the average compensating
surplus measure is $72.00 per year to increase expected encounters from 0
to 5 per day of elk hunting.  Some private clubs which  specialize in elk
hunting in Wyoming charge entrance fees  ranging from  $85.00 to $150.00 per
year or roughly equivalent  to the compensating surplus  measure for elk
obtained through contingent valuation approaches.

     The four case studies  discussed above have shown an impressive
consistency both in results and in the evolution of techniques to deal with
the bias problem.  Bias is,  of course, inherent in using contingent
responses to value environmental quality.   The view of  these researchers
is that problems of strategic, information, vehicles, and starting point
bias are all surmountable with proper questionnaire design,  modeling,  and
econometric analysis.

                                    45

-------
                           FOOTNOTES - .CHAPTER II

     — The dictionary defines a hypothetical proposition as a conditional
proposition, i.e., an "if X, then Y" statement.  A hypothetical question
would then be a conditional statement in the subjective mood, an "if X,
then .  .  .?" statement.  In a survey setting, the hypothetical question is
posed by the interviewer to the respondent; the interviewee then states how
he would alter his activities in response to the posited situation.  Formally,
the problem set before the respondent seems no different than the problem
he faces when he plans on the basis of a weather forecast to spend tomorrow
afternoon at a picnic  [Blank, et. al. , 1977].

     2/
     - See Schulze and d'Arge (19.78).

     3/
     — See Brookshire, et. al. (1976) which produced similar values to these
presented in Randall, et. al. (1974a).

     LI
     - See Brookshire, et. al. (1976); Blank, et. al.  (1977); Thayer and
Schulze (1977); Randall, et. al. (1977); and Brookshire, Randall, et. al.,
(1977).


     — See Blank, et. al. (1977) and Brookshire, Randall, et. al. (1977).


     — An exception to this is the study by Brookshire, Randall, et. al. (1977)
which employed an iterative bidding procedure and also obtained data necessary
for a travel cost comparison.  Additionally, Blank, et. al.  (1977) and
Brookshire, Randall, et. al. (1977) employed the substitution approach and
an iterative bidding approach in separate questionnaires.

     — This is equivalent to the compensating variation measure of consumer
surplus where the initial level of utility is maintained.  See, for example,
Mishan (1971).

     a I
     — Distributional effects are ignored at this point.

     9/
     — Preliminary analysis of the substitution data set does suggest that
the number of substitutions of activities was not great when faced with a
contingent change in air quality.

    —-In what follows, the time constraint is omitted but adds little
difficulty except in the dimensions of characteristics and user production
functions.


    —Although implicit here, technology also influences the form of the
input demand functions.  Changes in technology will change these functions.
                                    46

-------
    12/
    — A  large body of  literature  exists  which  explores  the  dual  relationship
between cost  functions  and  production  functions  or  technologies.   See,  for
example,  Shephard  (1970), Uzawa  (1964), Hall  (1973),  and  Diewert  (1974).

    13/
    — It  should be noted that if  C(P,Z)  is nonlinear in  Z,  then  A will be
a function of both P and Z, i.e.,  A =  A(P,Z).  For  discussions of  the hedonic
price approach see Rosen (1974), Muellbauer  (1974), Lucas  (1975),  and
Crocker (1975).  Empirical  application of  this approach  is not as  simple as
it would appear primarily because  of an identification problem.   On  this
point see  Rosen (1974)  and  Crocker  (1975).

    14 /
    — It must be pointed out that income  compensated characteristic demand
functions are not estimated in this approach.  The  reason is that  there is
no reason  to believe that the estimated expenditures x^hich represent the
dependent variable in estimating C(P,Z) are generated with utility held
constant.


    — See Crocker (1975) for a more detailed discussion of  the nature  of
the identification problem.


    — A complete discussion of the surplus measures in relation to rights
structure and starting points is in Brookshire and Randall (1978).  Surplus
measures and their relationship to Bradford bid curves which have been a
focal point in non-market valuation is thoroughly discussed.   Randall and
Stoll (1978), extended the analysis of Willig (1976) to permit its application
to the valuation of changes  in commodity space.


    — This description is quoted from Brookshire and Randall (19.78).

    18/
    — Certain arguments presented in this section draw heavily upon Brookshire
and Randall (19.78).  These are designated  throughout the section.

    19/
    — This discussion is taken from Brookshire and Eubanks  (1978).

    20/
    ~~ If this is  not the case then we are at a loss as to the incentive
structure the rational individual is operating under.

    21/
    — Individuals were not  confronted with the prospect of paying the
mean bid,  thus a definitive  statement in the context of our discussion is
impossible in terms of strategic bias.

    22/
    — A number of respondents indicated an incentive to behave strategically
under the Clarke tax mechanism that may have been overlooked.  Several of
the respondents stated "that they attempted to reveal demands slightly lower
than the other participants" to achieve a negative variable charge under the
Clarke system.  [Babb and Scherr,  1972, p. 46].

-------
    23/
    — This subsection explicitly reproduces pages 19-26 from Brookshire
and Randall (1978).

    24/
    — An effort reported here from the South Coast Air Basin survey
indicates a more generalized framework is possible.

    25/
    — The assumption of b > 0 represents the case of convex indifference
curves.

    f\ s i
    — This section is a paper presented at the EDRA 9 Environmental
Aesthetics Symposium, the University of Arizona, Tucson entitled "The
Use of Survey Instruments in Economic Valuations of Environmental Goods"
by David S. Brookshire and Thomas D. Crocker.

    27/
    — The following statement by Fromm (1968) exemplifies the attitude:
"Furthermore, it is well known that surveys that ask hypothetical questions
rarely enjoy accurate responses." (p.  174)  A lengthy discussion of the use
uf questionnaires in the paper on which the Fromm.(1968) effort is a commentary
is summarily dismissed with this single unsupported statement,

    o o /
    — Originally set forth by Wicksell un 1869,  the  public  goods
preference revelation problem was rediscovered by Samuelson (1955).  The
first reasonably complete preference revelation device is in Clarke (1971).
Smith  (1977) provides an up-to-date review of the problem and its suggested
solutions.

    297
    — According to MMler (1974, PP- 183-189.), weak complementarity exists if
the quantity demanded of a private good or activity is zero when the marginal
utility of the public good is zero.  The condition permits one to avoid
having to solve for utility and expenditure functions when trying to establish
the demand for a public good by exploiting its connections with private,
marketed goods.

    30/
    — A good elementary presentation of Bayes   (1764) rule is available
in Raiffa  (1970, pp. 17-21).


    — Adaptive behavior, once having committed one's self and experiencing
unanticipated regret or satisfaction thereby, can be treated as the
acquisition of further information.

    32/
    — As used here, "social" refers solely to a world in which all voluntary
gains  from exchange, given the initial distribution of income, are exhausted.
Only under classical conditions  (an absence of nonconvexities, irreducible
uncertainty, coordination costs leading to externalities, and less than
complete contingent claims markets), does current economic knowledge demonstrate
that market prices alone would be sufficient  to make efficient (Pareto-optimal)
allocations attainable.
                                     48

-------
    347
    — Perhaps the best overall review of the state of demand theory from the
perspective of the development of a_ priori restrictions to assist in
estimation problems is Goldberger (1967),


    — Diewert (1974) reviews the applications of duality theory to economic
problems.


    — An illustration of this technique can be found in Blank, et_. al.   (1977)
and Brookshire and Randall, et, al. (1977).  In the former, picture sets
presented to individuals represented pre-determined levels of visibility
(defined in terms of visible range).  This allows the linkage of physical
parameters to valuation estimates.  In the latter case, landscape types were
classified for an elk hunting experience.

    37/
    — Even with Census Bureau data, the economist does not know all the
"adjustments" that have been undertaken to make the data presentable.

    no/
    — If there is an increase in pollution,  the amount the sufferer would have
to be paid in order to be willing to accept the increase is consistent with
the polluter being liable for the damages he causes.  The amount the consumer
would be willing to pay to prevent the increase implies that the polluter
has zero liability for any harm he imposes upon the sufferer.

    39/
    — This study was supported by the Electric Power Research Institute,
Palo Alto,  California, to the University of Wyoming.  EPRI does not assume
any liability for the completeness of  research,  or usefulness  of the
results.

    40/
    — The. research reported here was  supported by a NSF grant entitled,
"An Economic and  Environmental Analysis of Solar and Geothermal Energy
Sources."

    41/
    — Portions of this study were funded by  the U.S.  Fish and Wildlife
Service contract  numbers 14-16-0009-77-002 and 14-16-0009-77-003 with
the University of Wyoming.
                                     49

-------
                                CHAPTER III

           PAIRED SAMPLE METHODOLOGY:  THE SOUTH COAST AIR BASIN

3.1  Rationale for Paired Sampling

     The previous chapter presented an overview of the theoretical and
conceptual structure of various non-market valuation techniques.  In order
to enable a cross-check between the iterative bidding technique and the
substitution approach involving primary data collection and a secondary data
property value study, a common sampling methodology is needed.  Given the
variable of perturbation is air quality, an ideal sample methodology would
control for all the factors influencing the valuation.  This, of course, is
impossible.  The approach settled upon was to form pairs of census tracts
in the South Coast Air Basin (SCAB) holding socioeconomic type character-
istics constant yet allowing a variation in air quality across pairs.
Throughout the SCAB are located air monitoring stations providing readings
on Ozone (0_), Nitrogen Dioxide (N0?), Nitric Oxide (NO ), Carbon Monoxide
           ,3                       £.                   X
(CO), Hydrocarbons (HC), Sulfur Dioxide (S0~), particulate matter, wind and
in some cases lead (Pb), and oxidant levels.  Our aim in this sampling pro-
cedure was to relate as closely as possible the readings of these consti-
tuents of air pollution to the surrounding census tract populations.

     Given the locations of the air monitoring stations in the SCAB, we were
able to identify surrounding census tracts.  For these census tracts the
Department of Commerce provides excellent demographic information.  This
information is used for three specific purposes:  (1) define the census
tract parameters and characteristics; (2) designate census tracts represen-
tative of the SMSA as a whole; and (3) provide the means for matching census
tracts in the test areas to similar census tracts in the control area.  The
goal was to control, by careful•choice of the study areas, for as many
potential influencing factors that might explain differences in preferences
toward environmental health and amenity levels.

     Thus the aim of the sampling procedure was to determine paired areas
in the SCAB that are similar in all relevant characteristics except air
quality.  If the mean values of the relevant characteristics are not signi-
ficantly different across areas,  the difference in valuation of amenities
and environmental health effects given by an individual household in an
area characterized by clean air,  versus the valuation given by an individ-
ual in an area characterized by diminished air quality, should only be
due to the existence of pollution in their environment.


                                     50

-------
3.2 Socioeconomic Control Considerations

     Certainly many variables affect an individual's valuation of a public
good such as air quality.  Variables of influence that should be considered
are:   (1) median income; (2) mean income;  (3) percent high school graduates;
(4) total population;  (5) percent non-white;  (6) percent 0-19 years old;
(7) percent 20-34 years old; (8) percent 35-64 years old; (9) percent 65
and older; (10) percent male; (11) percent in construction industry;  (12)
percent in manufacturing industry;  (13) percent in other jobs; (14) median
school years completed; (15) number of persons per household; (16) median
housing value; (17) median age of structures; (18) structural density, i.e.
percent private residences;  (19) average temperature; (20) miles to beach;
(21) miles to Los Angeles International Airport; and (22) miles to major
interchange.  Each variable  represents a characteristic of the census tract.
The characteristics provide  information as to the demographic profile of a
census tract both in a qualitative  (pertaining to the population) and quan-
tative (measures of a physical or structural nature) sense.

     Consider education as a necessary control variable.  Education is a
valuable and expensive commodity.  It allows people potentially a greater
appreciation and awareness of life, its alternatives, and its shortcomings,
among other things.  Furthermore, education possibly could make one aware
of the effects of air pollution on one's health and enjoyment of life, make
one more aware of the interdependencies and externalities of the problem,
and may make one more demanding that something be done to alleviate the
problem.   To control for these and other possible effects from educational
differences  among  households,  we used the variable of percent high school
graduates over the age of twenty-five years.  This measure gives the
general level of education of the inhabitants of a census tract.

     Certain people are physically affected by air pollution more than are
others.  The older one is the less able are one's physical defenses to
neutralize the effects of diminished air quality.  Also, the younger the
child, there are more years he must live in a polluted environment and there-
fore he is more likely to develop, for example,  asthma or other bronchial
complications.  Thus we controlled for the age distributions across census
tracts where possible.

     The following age groups were used:    (1) 0-19 years:  the age which
children are most likely still at home and still dependent upon the head
of the household; (2) 20-34 years:  newly established households;  child
bearing age;  (3)  35-65 years:  this age group is representative of more
established households, usually couples whose children are growing up or
may have already left home; and  (4) 65 years and older.

     General social and cultural factors were controlled for by using
census data on percentage white, percentage black, and percentage other.
Such influences may enter through risk preferences and time horizons, and
attitude toward one's health.

     Lave and Nagin (1974)  in their study of various influences on mortality
discovered that certain occupations have a higher mortality rate than


                                     51

-------
others.  Since job environments expose individuals to varying levels of
diminished air quality, three categories were considered:  (1) exposure
to air pollution — percent employed in the construction industry;
(2) occupational hazards unrelated to ambient air pollution —
percent employed in the manufacturing industry; and (3) absence of air
pollution — all other jobs not considered above with the exception
of farming.  The percentages of each were found to be heterogeneous in
the census tracts.  Therefore, occupational exposure is a question that
was explored with the survey instrument rather than from the census
tract data.

     The most complicated of categories has to do with residential location.
There are many factors that enter into a locational decision, some being
more specific than others in determining the exact location.  The value an
individual places on the health and amenity effects from air pollution gives
an indication of his incentive to change his place of residence.  Those with
high preference for avoiding the problems and dangers of polluted air will
have expressed this preference by moving to a census tract with clean air.

     There are many reasons why people choose to locate in one area as
opposed to another and air quality is just one of a whole myriad of consid-
erations.  One of the most important considerations is the job location.
Another consideration is the type of community.  The study focuses upon
households so we therefore desire to sample in areas with a high concen-
tration of private residences, or the so-called "bedroom communities."  A
private residence is defined as limited to one family homes on less than
10 acres of land and with no business on the property.  A third consideration
is a set of convenience factors that might determine location as the dis-
tance from physical points of importance.  Among these are miles to the
beach, miles to an airport and miles to a major interchange.  Proximity to
a recreational center is usually an important consideration.  The main
recreational activities in Southern California center around the beaches.
Miles to the airport is important because of its transportation and also
to avoid the noise that affects a very wide portion of Los Angeles.  Miles
to a major interchange is important for getting anywhere in the .Los Angeles
area, such as schools, jobs, shopping centers, and recreational areas.

     Income allows for a greater variety of lifestyles and expressions.
Differences in income also result in many of the behavioral differences
attributed to the other variables such as education, race, and age.  For
our income variable we used the mean income for each census tract.  We also
separated the responses into income classes to see if the marginal valu-
ations with respect to income are constant.across income classes.

3.3  Census Tract Pairings

     In the preceding section our aim was to suggest those variables that
may affect the value an individual gives in response to the air quality
survey instrument.  We desired to control for as many of these variables
as possible in advance of the actual survey.  These variables were con-
sidered when pairing census tracts.  Each variable is a characteristic of
the census tract and of its inhabitants, characteristics that in some way
are expected to influence people's valuations.  By looking at the
                                     52

-------
differences in these characteristics, homogeneous-pairs of census tracts.
were identified to the extent possible.

     From 1970 census data  (U.S. Bureau of Census 1970) we have values for 22
variables for each of the census tracts..?/  The total number of census tracts
in the South Coast Air Basin number in the thousands.  The principle criteria
in the area selection process for our sample plan was to include some areas
with clean air and other areas with various levels of diminished air quality.

     In choosing the preliminary sample pairs of areas, an attempt was made  to
to include some areas that  not only met the above criterion but that also had a
view of some natural or man-made phenomena that is in some way unique and
outstanding.  The existence of this"view" is not subject to quality measure
but is treated as a binary variable; either it exists or it doesn't.  To
determine areas with a view and those without, the researchers did an
onsight inspection of the South Coast Air Basin.

     There was a preliminary choice of 77 census tracts.  The differences
between the 77 census tracts for each variable in the data set were cal-
culated.  From an original  set of 22 variables, nine x^ere considered of
major importance in determining the similarity of census tracts:  (1) mean
income; (2) percent high school graduates; (3) percent non-white; (4) median
housing value; (5) number of persons per household; (6) percent private
residences; (7) median year structure built;  (8) total census tract popula-
tion; (9)  percent over 60 years old.  Table 3.1 presents the values by
paired area for the chosen variable.

     Much of the preliminary sample pairing was done by comparing the rel-
evant variables across census tracts in the manner indicated above.  How-
ever, it became apparent that the most efficient method of obtaining final
sample areas would be to conduct field observations with whatever inform-
ation was currently available and pick several potential sample areas during
these field examinations.  This set of sample areas was then subjected to
the same test procedures as had been conducted in the previous sample area
selection efforts to determine final pairs.

3.4 Description of Paired Areas

     The results of this pairing effort are summarized in the following
pages.  Each of the sample areas is described with the area with which it
was matched.  The match was made on the basis of differing air quality and
constant control variables.

1.   Canoga Park and El Monte

     Canoga Park:   Northern half of census tract //1345

     Boundaries:    North:  Saticoy Street
                   East:    Variel Avenue
                   South:  Sherman Way
                   West:    Topanga Canyon Boulevard

     Air Quality:   Fair
                                   53

-------
               Table 3.1
U.S. Census Information  for the Paired Areas


City
I.
Canoga Park.
El Konce
II.
Culver City
Montebello

III.
Newport
Beach
Pacific
Palisades
IV.
Irvine
Palos Verdea
V.
Enclno
La Canada
VI.
Hunclngton
Beach

Redundo Beach


Tract
Number

1345
4334

7026
5301.02
53QQ.Q2


630.01

2627.02

525
6704.02

1396
4607


993.03

6205.01
6205.02

Total
Population

5Q12
7516

7372
3858
3478


7421

3915

9337
8088

3593
507Q


4091

6608
7179

Mean
Income

8821
8211

15,750
13,808
18,858


25,592

35,419

14,059
26,118

36,242
30,647


9,859

11,815
10,501
Median
Housing
Value

20,800
17,900

30,800
27,500
38,400


50,000+

50,000+

33,100
50,000+

50,000+
50,0001-


18,800

23,600
23,000
Nunber of
Persons Per
Household

2.44
2.82

3.50
3.1Q
3.10


2.83

3.10

2.63
3.86

3.32
3.29


2.22

3.28
3.13
Percent
High School
Graduates

46
38

76
60
74


90

89

86
93

83
87


60

63
59
Percent
Private
Residences

15
25

69
56
54


49

76

76
82

55
89


42

56
32
Median Tear
Structure
Built

59
59

59
59
64


64

49

70
68

59
59


59

59
59
Percent
Over Sixty
Years Old

15
13

7
19
10


10

18

28
3

9
14


20

6
6

Percent
Xon-Vhite

10
3

6
4
24


0

Q

4
3

2
1


2

2
2


Air Quality

fair
poor

good
poor
poor


fair

good

fair
good

fair
poor


poor (exotics and
aulfates)
good
good

-------
     El Monte:
              Census tract #4334
     Boundaries:   North:

                   East:

                   South:
                   West:

     Air Quality:  Poor
                      Garvey Avenue to Peck Road and north on Peck to
                      Valley Boulevard.
                      southeast on Valley to Mountain View, southern
                      boundary.
                      Schmidt Road
                      Edwards Avenue
     This is a pairing of the lowest income communities in the sample plan.
Both are inland and subject to high summer temperatures.  The census data
show that the pairing is a good match, with the exception that Canoga Park
has ten percent fewer private residences.  Field observations indicated that
the areas are very similar in appearance.
2.
Culver City and Montebello

Culver City:   Census tract #2026

Bounaries:  North:  Cota Street to Jefferson Boulevard, Jefferson to
                    Overland Avenue, Overland to Northgate Street.
            East:   City Line
            South:  San Diego Freeway, Jefferson and Boulevard and Playa
                    Street.
            West:   Ballana Creek
     Air Wuality:

     Montebello:


     Boundaries:
              Fair

              Census tract #5301.02 part of #5300.02 in the northeast
              part of the area.
              North:
              East:
              South:
              West:
Lincoln Avenue
Montebello Boulevard
Whittier Boulevard
Wilcox Avenue
     Air Quality:  Poor
     This grouping involves areas of upper low income population.  Montebello
is farther inland, hence subject to higher summer temperatures.  Census data
show that Montebello has slightly fewer private residences, somewhat newer
homes and more people over age 60.  The greater percentage of non-white
population in tract 5300.02 is concentrated outside the area  chosen for
sampling.  In general, the data show a fairly good match and the field
check confirmed this.  Nevertheless, the difference in summer temperatures
is a source of potential difficulty for the empirical analysis.

3.   Newport Beach and Pacific Palisaides

     Newport Beach:  central portion of census tract #630.01 (West Cliff
                     section)
                                     55

-------
     Boundaries:   Northwest:   Irvine Avenue
                   Northeast:   Nottingham Drive
                   Southeast and Southwest:  Westcliff Drive and Santiago
                   Lane

     Air Quality:  fair

     Pacific Palisaides:  northeast portion of census tract #2627.02, area
                          southwest of intersection Sunset Boulevard and
                          Chautanqua Boulevard bordering both sides of
                          Pampas Ricas.

     Boundaries:   North:  Toyopa Drive
                   East:   Toyopa Drive
                   South:  Corona del Mar
                   West:   Alma Real Drive

     Air Quality:  good

     The Pacific Palisades neighborhood appears to be solidly middle class
with homes somewhat on the large side, with a mixture of styles, including
some that are two and three stories.  These homes are generally 20 to 25
years old and trees and shrubbery are well developed, affording some degree
of seclusion.  The lots are not very large relative to the size of the homes.
There are no ocean views, but  the area is fairly close to the bluffs.  The
neighborhood in Newport Beach is on top of a hill and the backs of at least
some of the homes on Nottingham Street overlook upper Newport Bay.  The area
also appears to be solidly middle class.  The homes are somewhat newer but
compare favorably in terras of their size and lot size.  Census data indi-
cate substantial differences in mean income, percent private residences,
median age of structures and percent of population over 60 years old in the
whole tracts.  Field observations show that these differences are not as
strong in the actual neighborhoods.chosen, except in the age of the homes.
Both areas have easy access to beaches, are somewhat removed from com-
mercial/industrial areas and are comparable in terms of income levels and
lifestyle.

4.  Irvine and Palos Verdes

     Irvine:  Greentree homes, small portion of census tract #525.

     Boundaries:  Northwest:  Culver Road
                  Northeast:  Walnut Avenue
                  Southeast:  Yale Avenue
                  Southwest:  boundary of the development

     Air Quality:  fair

     Palos Verdes:  Beechgate Drive area, portion of census tract #6704.02.

     Boundaries:  North:  Silver Spur
                  South:  Crest Road

     Air Quality:  Good
                                    56

-------
     The census data on this pair does not match well in terms of income,
persons per household or percentage of people over 60 years old.  However,
the Irvine tract, #525, is a very large tract encompassing much agricultural
land and several retirement communities.  Therefore, census data do not ac-
curately reflect the situation in the chosen area.  The Greentree Homes de-
velopment is two to three years newer than in Palos Verdes; the houses may
be slightly smaller but the lots are of similar size.  Inspection showed
that the two areas are comparable in terms of lifestyle and age structure.
Both are upper middle class areas, have about equal accessibility to beaches
and are in areas of similar temperatures.  Also, both are located very close
to "classy" shopping centers and main arteries.  They differ in that Palos
Verdes is hilly, while Irvine is flat.  Greentree Homes is an enclosed
development, where the Palos Verdes area is not.  The field observations
indicated that this was one of the better pairings.

5.  Encino and La Canada
     Encino:

     Boundaries:
portion of census tract #1396.

North:  Ventura Freeway
East:   Balboa Pvoad
South:  Rancho Street
West:   White Oak Avenue
     Air Quality:  fair

     La Canada:
     Boundaries:
south-central portion of census tract #4607, vicinity of
Chevy Chase Drive and Berkshire Drive.

North:  Foothill Boulevard
East:   Foothill Freeway
South:  Highland Drive
West:   hills west of Chevy Chase Drive and south of
        Descanso Drive
     Air Quality:  poor

     Encino has a commercial strip (Ventura Boulevard) through the center,
while the La Canada area has a similar development around the fringe.  The
Encino area consists mainly of ranch style houses with fenced yards and
gates across the driveways.  It is hilly, the homes are secluded because of
trees and bushes, and there are several private roads scattered throughout.
La Canada has a very similar appearance.  Both are inland with similar sum-
mer temperatures and incomes are comparably high.  The census data show
significant differences only in the percentage of private residences.  How-
ever, the areas chosen are almost entirely private residences; this should
not be a problem.  This is a very good match as seen in both the field
experience and in the census data.

6.  Huntington Beach and Redondo Beach

     Htmtington Beach:  central portion of census tract #993.03.
                                    57

-------
     Boundaries:   North:  Adams Avenue
                   East:   Beach Boulevard
                   South:  Frankfort Avenue
                   West:   Alabama Street

     Air Quality:  poor

     Redondo Beach:  eastern portions of census tracts #6205.01 and 6205.02.

     Boundaries:   North:  Manhattan Beach Boulevard
                   East:   Inglewood Avenue
                   South:  Artesia Boulevard
                   West:   Rindge Lane

     Air Quality:  good

     Both areas are in beach communities with similar income levels and
temperatures and about equal accessibility to beaches.  Both border closely
to commercial strips.  The Redondo Beach area is very homogeneous in the
type and quality of houses.  Most are small, stucco block houses with 800
to 1000 square feet of floor space.  They have small yards and are moder-
ately well kept.  There are a few duplexes and apartments mixed in x^ith the
single family dwelling units.  The Huntington Beach area is not as homo-
geneous as Redondo Beach.  The average house and lot size is about the same
but the variance is greater.  Railroad tracks run close to the western
boundary (Alabama Street) and that vicinity was avoided in the sampling.  The
census data match well except that Huntington Beach area has more older
people and fewer persons per household.  These two measures are probably
related and are reflected in the areas chosen.  The field observations indi-
cated that this is a fairly good match and probably the best available in
the two communities.

3.5 Ambient Concentrations for the Paired Areas_

     Employing the data from monitoring stations in the South Coast Air
Basin, Table 3.2 was constructed.  Consideration was given to the basic
wind patterns in the area.

     Focusing on total oxidants, nitrogen dioxide and total suspended par-
ticulates, isopleth maps were constructed for each  pollutants.   Finally,  an
average isopleth was constructed.  Figures 3.1-3.4 represent these maps.
Finally, Table 3.3 presents the arithmetic average for 1975 for the daily
maximum hourly average concentrations for the sample areas employed in
Figures 3. 1 - 3.4.
                                    58

-------
                                                                   Table  3.2

                                             South Coast Air Basin Pollutant Information
                        Daily Maximum Hourly  Average Concentrations
                      of Various Pollutants in  South Coast Air  Basin)
                               (Arithmetic Average - -1975)
Hourly  Average Concentration of Various     Total Suspended
Pollutants in South Coast Air Basin      Particulaces by Hi-Volu=e
       (Arithmetic Average - 1975)      Method (Fiberglass  Filter)
                                        (Arithmetic Moan  -  1975)
^^•^Pollutant
Station ^~\^^
Anahein
Azusa
Burbank
Coata Mesa
El Toro
La Habra
L^guna Beach
Lennox
Long Beach
Los Angeles
Lynvood
Neuhall
Pasadena
Pomona
Reseda
West Los Angeles'
Whictier
Redondo Beach
Santa Ana Canyon
Los Alam'itos
Total
Oxidanc
(pphm)
3.7
10. 7
8.6
4.3
1.9*
6.4
3.3
3.8
3.3
8.0
4.2
9.0
10.5
9.8
9.9
5.9
6.1
-
-
~
Carbon
Monoxide
(ppct)
6.5
5.6
10.7
11.4
3.4*
8.1
4.5
10.4
7.2
10.0
11.1
5.0
8.9
6.1
7.6
7.9
6.7
-
-
—
Sulfur
Dioxide
(pphm)
2.2
2.6
2.4
2.6
1.1*
2.9
1.8
5.1
5.2
3.3
3.9
1.8
2.5
2.8
1.6
2.6
5.8
3.2
.4
5.3
Nitrogen
Dioxide
(pphm)
9.2
10.6
12.7
6.5
6.7*
10.9
9.8
10.1
11.0
12.9
9.2
5.6
14.1
11.9
11.8
13.4
12.5
-
-
~
Total
Hydrocarbons
(PP»)
4'. 6
5.0
6.2
-
1.7
3.8
-
5.3
-
4.4
-
3.9
4.2
4.0
3.7
5.3
5.1
-
-'
3.4
Total
Oxidnnt
Cpphra)
1.2
3.6
3.0
1.8
.9*
1.8
.9
1.7
1.5
3.0
1.9
3.5
3.6
3.3
4.2
2.5
2.3
-
-
~
Carbon
Monoxide
(ppn)
2.9
3.7
5.8
4.8
1.7*
3.3
2.3
4.2
4.2
4.7
5.9
2.7
3.9
3.3
3.7
2.9
3.0
-
-
~
Sulfur
Dioxide
(pphm)
.6
1.5
1.5
.9
.2*
.8
.5
2.0
2.1
2.0
1.9
1.2
1.6
1.4
1.1
1.5
2.5
1.4
.1
1.3
Nitrogen
Dioxide
(pphn)
5.4
6.0
7.4
3.0
3.8*
6.4
3.9
5.6
6.2
6.7
5.2
3.2
8.2
7.2
6.4
6.8
7.2
-
-
—
Total
Hyd rocarbons
(ppn)
2.9
3.7
3.8
-
1.3
2.2
-
2.9.
-
2.8
-
2.6
2.9
3.3
2.1
2.9
3.2
-
-
1.9
Total Suspended
Particulates (yg/
-------
                          Isopleths for Nitrogen  Dioxide Levels in  the South Coast Air Basin
ON
o
                                                        Figure 3.1
                                                             La  Canada
                                                              II.i
               Pacific
               Palish
                                                             Montebello
                                                                 If 12-1
                                   if
                                 Culver City
               Nitrogen Oloxidt
               (PPHM)
               Poor:  >il  pphm
               Pair:  9-11  ppNn
               Good:  <9 pphm
                •  Indicates air monitoring station

-------
           Tsopleths for  Total  Qxidant Levels  in thg_Sou.t:h  Coast  Air  Basin
                               £3 a a a
                         Verdes
                           JL 3.0
Poor:  >8 pph.-n
Fair;  3.5-8 pphm
Gox»d:  <3.5 pphm
 •  indicates air sionttoring station

-------
o\
                 Isopleths for  Total  Suspended Particulates  in the South Coast  Air Basin

                                                    Jigure  3.3


                                                            La Canada
                                                             130
                Total. Suspended Partfculates
                Poor:  >110 ug/m .,
                Fair:  90-110 ug/w
                Good:  <90 ug/m-*

                •  Indicates air monitoring station.

-------
              Isopleths for  Nitrogen Dioxide, Total Oxidant and  Total
             Total  Suspended Particulates  in the  South Coast Air Basin
                                      Figure  3.4
Canoga Park
  Atr Quality  In the South Coast Air Baain'*'
   Based upon an ln
-------
                                  Table 3.3

           Daily Maximum Hourly Average Concentrations of Various
                   Pollutants in the South Coast Air Basin

                         (Arithmetic Average - 1975)
Montebello

Culver City

'Canoga Park

El Monte

Encino

Pacific Palisades

Newport Beach

Irvine

Palos Verdes

Redondo Beach

Huntington Beach

La Canada
Oxidants
pphm)
7.0
5.8
7.5
10.1
7.0
3.0
4.0
4.0
2.0
3.5
3.6
10.2
Nitrogen
Dioxide
(pphm)
12.7
13.0
11.0
12.5
9.1
6.2
6.7
6.8
6.3
8.5
9.0
11.2
Total Sus]
Particul;
(yg/m'
115
88
110
116
105
78
80
75
67
85
115
130
                                     64

-------
                                 CHAPTER IV

                  THE SOUTH COAST SURVEY QUESTIONNAIRE STUDY

4.1 Survey Instrument Design

     Chapter II reviewed the theoretical and conceptual state of the art in
employing the contingent claims mechanisms.  The essential questions addres-
sed implicitly in the discussion with regard to aesthetics and health ef-
fects in the South Coast Air Basin are:  whether a valuation for an enviro-
mental good can be disaggregated into characteristic parts, the relative
efficacy and consistency of bidding and substitution formats in accomplishing
this task, and whether a survey instrument can be properly designed enabling
the estimation of the overall contingent valuation equation.  This chapter
will present the structural design of the survey instrument, the method of
choosing the accompanying photographs, the survey implementation procedure,
and preliminary statistical results from the iterative bidding component of
the survey instrument.

     The structural components and the directional flow of the survey instru-
ment are presented in Figure 4.1.  Many types of information are sought by
the survey instrument.  The first component can be viewed as establishing
baseline information about the respondent.  The respondent's current indoor
and outdoor recreational activities, costs of both types of activities,
location of the activities, the frequency and duration of activities, and
the importance of the activities are established.  The respondent is held
to a "typical week" time budget for indoor and outdoor activities that was
initially established in the questioning process.—  This information was
then entered on the indoor/outdoor activity and cost lists in Tables 4.1 and
4.2.

     At this point the interviewer presented information relating to either
aesthetic effects of visibility or health effects in the South Coast Air
Basin.  Recalling the earlier discussion about information bias, the alter-
native initiation points for beginning the valuation process were a poten-
tial factor in the final results.  That is, in disaggregating an environ-
mental good down into characteristic components, does the order in which
the characteristics are presented affect not only the final summed valuation
of the good but the characteristic parts valuation?  In order to test this
hypothesis, information was obtained as presented in Figure 4.2.  First, the
sample population was broken into two groups:  those mailed a health bro-
chure (as in Appendix B) and those provided no additional aesthetic or
health information other than that presented in the survey instrument. Sec-

                                    65

-------
            Information  Collective Flow for Survey Instrument
                               Figure A.I

Indoor
Costs

<
j uaseiine iniormation f
l/
Indoor
Activities

\
Outdoor
Activities



Outdoor
Costs



Expenditures
and Income

Indoor
Activities
1


\
Outdoor
Activities
._. I
J,
In f nr™n f \ nn \f- 	 . — 	 —
I No Substitutions


Bid on Aesthetics
Plus Acute  Health
                    Substitute
Indoor
Activities



' \


Outdoor
Activities

                   Information
Bid on Aesthetics
Plus Acute Health
Plus Chronic
Health
                    Substitute
     Respondent's
     Evaluation
                                            No Substitutions
                                                                    (Step 1)


                                                                    (Step 2)
                                                                    (Step 3)
                                                                    (Step
(Step  5)
                                                                    (Step 6)
                                            No Substitutions |
                                                                    (Step 7)
Indoor
Activities




f
Outdoor
Activities

1 Hornr* I.ivinp I/


                                                                    (Step 8)
(Step  9)
                                    66

-------
           Table 4.1




Outdoor Activity and Cost List
Activity
Outdoor Spectator
Sports
Tennis
BikinR
Beach Activities
General Exercise
Fishing
Swindling
Sailing
Jogging/Walking
Hobbies, Arts 4 Crafts
Outdoor Gardening or
Fixing up House
Golf
Hiking
Camping
v7














Organized Sports Events 1
Individual Sports
Events

Hours
Per Week
A
















Other (specify) ; 1
B

















C




D




I
























Times
Per Week
A

















B

















C

















D

















Location
(Map Grid)
.A

















B

















C

















D

















Miles
Traveled
A

















B

















C

















D

















Direct
Costs
A
















B
















1
C

















D

















% Day

















Equipment
Replacement
Costs

















loportance


















-------
          Table 4-2




Indoor Activity and  Cost  List
Activity
Indoor Spectator Events
Indoor Tennis
Raquetball, Handball
Table Tennis
Bowling
Indoor Gardening or
rr, Fixing up House
General Exercise
Organized Sports Eventa
Reading
Television
Movies
Club Activities,
Organlzat Ions
Individual Sports
Swiraning
Visiting Neighbors or
Friends
Other (specify)
v7
















Hours
Per Week
A
















B
















C
















D
















Times
Per Week
A
















B
















C
















D
















Location
(Map Grid)
A
















B
















C
















D
















Miles
Traveled
A P
































C
















D
















Direct
Costs
A
















B
















C

-














D
















Z Day
















Equipment
Replacement
Costs
















Importance

















-------
                                                 Figure  4.2

                                 Information  Sequence  in Survey  Instruments
                1. No Heath Information
                   Presented
Baseline
Info Sec-
II.   Health Brochure
     Previously Mailed
                                                     A.  Aesthetic    (l)Aesthetic   (2)Aesthetic     (3)Total
                                                                      and  Acute      Acute,  Chronic   Check
                                                     B.  Acute
                                                       (l)Acute
                                                      Chronic
(2)Acute
Chronic
Aesthetic
(3)Total
Check
                                                     C.
                                                       (1)
                                                     D.
                                                      (1)
(2)
(3)
(2)
(3)

-------
ond, these two groups were further  subdivided into two additional categories
according to the sequence of information presented in the survey instrument.
Either a single individual was asked valuation questions about air
quality characteristics in an aesthetic affects, aesthetic plus acute
health affects and aesthetic plus acute plus chronic heath affects
sequence or acute health affects, acute plus chronic health affects,
and acute plus chronic health plus  aesthetic affects.  Data collected in
this manner would allow a statistical  test of ordering and initiation
point effects in the overall valuation effort."L'


     An iterative bidding format was administered based on a contingency
perturbation from the existing conditions presented to the respondent.  This
represented an improvement from the original condition in the resident's
air (i.e., poor to fair, etc.).  The bid was established using either a util-
ity bill or a lump sum monthly payment as the vehicle.  Further, in order to
be able to observe any individual time discounting, the clean-up period was
set forth as either 2 or 10 years.  Additionally, three alternative starting
points of $1, $10, $50 were employed to initiate the actual bidding process.
Finally, some respondents were handed  a "life table" that would show the
total amount they would pay as long as they lived in Los Angeles depending
upon their bidu  Thus, the iterative bidding format within the survey instru-
ment employed structural characteristics that allowed for eventual testing
of all the potential bias discussed earlier.

     After the recording of the maximum bid, the interviewer moved to step
3.  This step initially established the following:   (1) the respondent had
stated a willingness to pay for improvements in air quality (even if zero);
(2) the respondent had less money overall as a result; and (3) thus (1) and
(2) indicate they value clean air and  thus they have traded income for clean
airu  Then the respondent was queried  as to whether  the improved air quality
conditions would alter their current activity patterns in any or all cate-
gories (i.e., time, duration, place and/or type).  Thus column B of Tables
A.I and 4.2 were filled out with  the time constraint being checked .A'


      The beginning of step 4 essentially repeated steps 2 and 3 in procedure,
 however, the information content was different.  Consider previous bidding
 games as a focal point.  Typically, the process would involve yet another
 perturbation of the environmental  good in question.  However, we are inter-
 ested in attempting to disaggregate the characteristic parts of the enviro-
 mental good air quality into aesthetic affects, and acute and chronic health
 affectSo  Thus step 4, depending on whether step 2 begins with information
 on aesthetic or acute affects, presented either acute health affects infor-
 mation for the former point or chronic health affects information for the
 latter initiation point.  The same initial vehicle was employed.  If the
 "life table" was used earlier it again was made available.  The bidding
 began with the last maximum stated bid of the previous step.   Step 5 again
 repeated earlier conditions (i.e., the trade of income for less health
 effect) and the outdoor and indoor activity/cost lists were filled in for
 column C.

                                     70

-------
     At the initiation of step 6,  information regarding the last remaining
characteristic of the "good" air quality was provided which was either health
affects for the aesthetic initiation point or aesthetic effects for the acute
initiation point.  Then the procedure for the iterative bidding was repeated.
Similarly, siren 7 renres^nted a renetition nf the substitution sections.

     Finally, at the termination of step 7,  a final review of bidding struc-
ture and the substitution answers  x^ere reviewed.  The respondent was
allowed any adjustments that were  deemed necessary.—

     Upon completing the iterative bidding and substitution sections, a
series of general information questions covering socioeconomic information,
property value information of the  residence, type of residence (i.e., number
of stories, pool, rooms, etc.), reasons for  current locational choice, health
related questions (i.e., heart trouble, medication, etc.), and attitudinal
questions relating to air quality  were administered.  This is step 8 in
Figure 4.1.

     Finally, step 9 involved a respondent's evaluation of the survey (i.e.,
relevant, policy oriented, etc.) and an enumerator evaluation.

4.2 The Photographs Accompanying the Survey

     The survey instrument in depicting air  quality in the South Coast Air
Basin employed picture sets.  This section will discuss the underlying
considerations in constructing the picture set employed in the South Coast
Air Basin.

     Visibility is dependent on light.  Light is a form of energy, made up
of electromagnetic energy, and is  really a form of matter made up of indi-
vidual particles (photons).  Light travels in streams and is subject to any
interference in its path.  Light waves can bend, spread,  interfere with one
another, and react with obstacles.  Visibility is the state or quality of
being perceivable to the eye.  It  is a subjective term in its common usage,
referring to the general clarity of the air.  In its more strict use, visi-
bility is defined as the farthest  distance that any object of suitable size
can be visually identified without the aid of magnifying  instruments.  Both
the common and strict definitions  of visibility suffer from lack of precise
meaning because of the many variables which  are difficult to control.  There-
fore, it is important that more precise definitions of visibility be ex-
plored in order to use the concept accurately..

     There are three characteristics of a light wave that are of concern:
(1) its intensity, which is related to the height of the  wave crests and
indirectly determines brightness of the light; (2) the wave length, which
depends on the distance between crests and largely determines color;  and
(3) its polorization, the angular  orientation of the crests.  These three
characteristics are influenced by  what happens when the light waves come in
contact with other matter.  In particular, we are concerned with how these
characteristics affect changes in  visibility as light waves interact with
particulate matter in the atmosphere.
                                    71

-------
      There are two issues in the way  light  affects visibility.  First is the
 ability of an object to reflect light  in such patterns as define the visual
 characteristics of the object.  Second is the ability of that reflected
 light to reach the observer in such a  way as to differentiate the charac-
 teristics of an object from the background.  First,., let us assume that
 every object, except a perfectly black object, reflects light some distance.
 Further, if the light reflected from  an object reaches an observer and that
 object is distinguished from the background, it is said that the object is
 visible to that observer.  Visibility  is not only dependent on light but
 upon the distance between the object  and the observer.  As the distance
 increases, less and less light reflected from the object reaches the ob-
 server until the object is no longer  distinguished from the background.
 When the observer can no longer distinguish the object from the background,
 the object is said to be beyond the visible range.  In summary, the visi-
 bility of an object illuminated by light depends upon the apparent contrast
 between the object and its background,  the  ability of the observer to dis-
 tinguish the object from its background, the size of the object and the
 angle of reflection, and the condition and  technique of observing.

      Three definitions of visibility are commonly found in the literature.

      Visual Range:  A dark object is moved  through the atmosphere toward the
      horizon sky.   As the distance between  the object and observer increases,
      contrast between the object and horizon sky decreases.   At some dis-
      tance the contrast between object and  horizon sky becomes too small
      to be distinguished, and the object "vanishes."  The distance between
      the observer and the object at the "vanishing point" is the visual
      range.

      Prevailing Visibility:  The greatest visibility which is obtained or
      surpassed around at least half of the  horizon circle, but not neces-
      sarily in continuous sectors.

      Meteorological Range:  The distance at which the contrast of an object
      is reduced to the point where the human eye can no longer distinguish
      it from the background, or that distance for which the contrast
      transmittance of the atmosphere  is two percent.

      It is possible under a certain set of  circumstances to measure visi-
bility by using photography.  Stephens  (1949) developed a method for mea-
suring photographically the "extent to  which visual range has been reduced
by haze."

      Briefly, the technique involves photographing (on black and white film)
black objects that are far enough away  to be obscured.  Then the photo-
graphic densities of the objects and the adjacent sky are measured on the
negative.  Calculated from these relative densities are the visual range,
distance of the object, and contrast of the  film.

      The theory of photographic photometry, used to calculate long-range
visibility, as in Roberts', et. al.  (1974) study of visibility measurements in
the Painted Desert, states that if a "black  object of sufficient size is
moved through the atmosphere away from  the observer, the object will appear
to become brighter as the distance from the  observer increases, even though
                                     72

-------
 the level of  illumination remains constant."  The apparent increase in
 brightness is  the result of light being scattered as it moves toward the
 observer by suspended aerosol particles in the air between the object and the
 observer.  The effect of an apparent increase in brightness illustrates the
 reduction of  light as it moves through air that contains particulates.  It
 is this effect that  is detected by the technique of photographic photometry.

      The apparatus  utilized in photometry is very simple.  All that is needed
 are a camera,  a positive gray scale, some means of measuring the distance
 from the observation point to the object photographed, and a densitometer.
 A densitometer is the devise by which the relative densities on the negative
 are measured.

      Unfortunately, photographic photometry has various problems which may
 cause problems in insuring reliable results.   First, for the purposes of this
 study, it is  crucial to differentiate between visibility reductions due to
 natural haze  and polluted haze.  We are attempting to measure the increase
 in haziness (the decrease in visibility) made by pollution.  Photometry
 simply measures the  visible range, without regard to the differentiation of
 natural and polluted haze.

      There are interrelations among the specifications for the object, the
densitometer,  and the camera.  The size of the image on the negative whose
 density is to be measured depends on the size and distance of the object
 and on the focal length of the lens.   The minimum size of the image that can
 be used depends on the characteristics of the densitometer utilized.   To
 further complicate the interrelation, since with any ordinary lens the
 illumination at the  focal plane rapidly decreases toward the edges of the
 frame, it is necessary to find what area of the negative is satisfactorily
 uniform in relation  to the particular camera utilized.   This illumination
 function is found by trial and error and is beyond the scope of our present
 efforts.

      Ideally, data  should be obtained by photographing distinct objects in
 each possible direction once each hour from each sample area and from one
 general area.^-'  At  least two distinct objects should be included in each
 observation path to  insure that,  regardless of the atmospheric conditions,
 data can be obtained from the photographs.   It is desirable to include views
 in all quadrants because of visibility and meteorological differentials
 across areas.   Observation points were chosen with these criteria.

      We define "object" as some unique natural or man-made phenomenon in the
 landscape that is distinct from its immediate surroundings.   "Observation
 path" is the line of sight.   It is important  to properly identify and locate
 objects in the observation path,  certainly if accurate measurements are to be
 made.   It further helps the respondent gain perspective when viewing the
 picture set.   Proper identification and measurement of  objects in each
 observation path chosen was accomplished using city and topological maps as
well as visual inspection.

      "Observation point" is simply the place in each area from which photo-
 graphs were taken.   Earlier in this report  we noted that one criteria in the

                                     73

-------
selection of some  sample  areas  in  the SCAB was  a view.  Therefore, obser-
vation points in the sample  that are with  a view were cnosen with'the aim of
representing to respondents within each sample  area a scene that they typ-
ically observe.   In this way, it was intended that the photographs would
merely serve as a reminder to respondents  of the changes in visibility due to
air quality.

      Certainly the most  important consideration is what is contained in each
observation path and therefore  in each picture.  Ideally, each observation
path should have at least two readily recognizable objects with which the
majority of respondents are  familiar, allowing  them to estimate easily visibility
by the contrast of those  objects.  The observation paths and the objects
therein should be concerned  foremost with  the portrayal of a visibility
gradient (in our study, "good," "fair," and"poor"), and should be very care-
ful to exclude objects that may trigger bias in the respondent in responding
to something besides visibility (or health affects).  For example, a free-
way interchange in the picture may stir up negative feelings in the respondent
even before the respondent considers the impact of changes in visibility.
Such undesirable characteristics in the observation paths may increase the
chance for bias in the valuation procedure.

      Another very important consideration was  to  insure consistency in field
operations.  There was a  standard operating procedure at. each observation
point.  Each photograph was  taken with identical equipment.  In order to as
closely as possible duplicate the quality of photographs from each location,
each photograph was taken with the same model Minolta SLR camera,  135 mm
lens (used for the photographs the respondents saw), 55 mm lens (to record
on film the local weather conditions for future reference), and the same high
quality professional color film.jj/

      Crucial to the photographic data collection  effort and the standard-
ization of field operations was for each photograph to be accurately logged.
Thus, each frame of exposed  film was logged and each step in the procedure
was carefully recorded so as to minimize discrepancies between observation
points.

      For each exposed frame, the researcher kept  a record of various char-
acteristics.  The time and date of exposure is  important in order to co-
ordinate the data the research  team collects in the field with the data
collected by the local airports and local  air monitoring stations.  The F-
Stop (aperature opening)  and shutter speed were recorded so as to further
estimate changes in visibility.  Since the photographs were going to be shown
to respondents as well as analyzed, it was important to insure the quality
of the photographs.  By quality, we mean that each photograph must be an
accurate rendition of the air quality as prevailed during each exposure.  In
order to insure a proper  exposure, each photograph was "bracketed."  That is,
for each photograph one frame was taken with a  normal meter reading, then
one frame of the same observation path was "underexposed" (meaning one F-
Stop above normal keeping the same shutter speed as for the normal photo-
graph) , then one frame "overexposed."  In this way we were assured that the
best possible representation on film of each observation path was .produced.

                                     74

-------
     It should be emphasized that the objective of these photographs  is  to
portray to respondents changes in visibility due to changes in air quality.
Observation points and observation paths were chosen primarily on the  cri-
teria  already listed, but once these sites were chosen, pictures could
only be used if the changes in air quality were such as to represent  the
range from clarity to visual obscurity that is typical for each area during
the year.   Of course, changes in visibility due to changes in air quality is
quite out of our control.  Therefore, we could only use those photographs
in the survey instrument portion of this study that were indeed representa-
tive of the typical range of visibility for any one area.   Such photographs
could only be obtained if the air quality was "right."  The results of this
effort are summarized below.

     Photographs were taken from seven sample area observation points and
from one general site observation point.   By "general site" we mean some
area or view that would likely be familiar to most of the  respondents no
matter where they lived in the South Coast Air Basin.

     Figure 4.3 entitled "Los Angeles Observation Paths" depicts the seven
sample areas and the Griffith Park Observatory.   The map is scried as shown
and each vector eminating from specific observation points represents fif-
teen miles.

     The Griffith site afforded three excellent observation paths:  (1) to-
ward downtown Los Angeles, with large buildings approximately five miles
from the observation point;  (2)  down Western Avenue toward large buildings
approximately four miles away and toward  two sets of hills in the back-
ground;  and (3)  southwest toward large buildings near Beverly Hills.

     Recall that we have six pairs of sample areas:

     1)   Canoga Park*            El Monte

     2)   Culver City*            Montebello*

     3)   Newport Beach*          Pacific  Palisades

     4)   Irvine                  Palos Verdes*

     5)   Encino*                 La Canada*

     6)   Huntington Beach        Redondo  Beach

     Those areas marked with an  asterik (*)  were chosen as  tentative  sites
for observation points.   A brief description of  each observation path from
each site  is as  follows:

     La Canada:   (4)  northeast  across the basin  toward  mountains;  and (5)
northwest  through the basin toward the mountains.
     Encino:  (6)  northeast  toward large  buildings  with mountains  in  the
background; (7)  north toward two sets of  large buildings at different
distances  with mountains in the  background;  (8)  north-northwest toward
large buildings with mountains  in background;  and (9) west down Ventura

                                    75

-------
                               Figure  4.3
          Observation  Paths  in the  South  Coast  Air Basin
           Los Kngeles
        Observation Paths
                                               Canada
                                     Griffith Park
                                       Observatory
                                                  A     Montebello
 ,  , Miles
 012345

Each vector «  15 miles
                                      76

-------
Boulevard toward the mountains.

     Canoga Park:   (10) north-northeast toward large buildings with moun-
tains in the background;  (11) north toward sets of large buildings with
mountains in the background; and  (12) west toward a set of large buildings
with mountains in the background.

     Culver Cit.y:   (13) northwest toward a set of large buildings with
mountains in the background; (14) west toward buildings in Santa Monica; and
(15) southwest toward two large buildings in Marina Del Ray.

     Palos Verdes:   (16) north toward buildings in Beverly Hills; and (17)
north-northeastern  toward large buildings in downtown Los Angeles with moun-
tains (Griffith Park) in  the background.

     Montebello:  (18) south-southwest toward buildings; and  (19) southeast
toward Whittier with hills on one side of the observation path.

     Newport Beach:  (20) northeast toward buildings with two sets of moun-
tains in the background; and (21) east across the Bay toward hills with moun-
tains in the background.

     On numerous occasions, photographs were taken from the eight obser-
vation points and the twenty-one observation paths.  For each observation
path, of course, the attempt was to photograph "good," "fair," and "poor"
days of visibility.  This was successfully accomplished for the Griffith
Park site, but, was  unsuccessful for all specif ic. samnl R areas p.xcent Encino.
For the other areas, we were unable to obtain the necessary gradients in the
photographs that would represent the typical range of visibility for each
area.

     The photographs used in the asking games for each sample area were the
observation paths from Griffith Park toward downtown Los Angeles and down
Western Avenue.   Figures 4.4a-c present the actual photographs in a black
and white version.  The visibility for picture set A (poor) was estimated at
2 miles, for picture set B (fair) at 12 miles, and for picture set C (good)
at 28 miles.

     The researchers were unable to obtain a poor air quality picture set
for the Griffith Park area with the same light and color characteristics
as the good and fair picture sets, although pictures x^ere obtained for this
location of approximately 2 miles.  In consequence, the researchers sub-
stituted a picture  set with similar foreground and light and color charac-
teristics taken at  approximately the same time in Orange County, California.

4.3  The Surveying  Procedures

     This section will detail the actual sampling procedures given the
sample plan discussed earlier.   The first task was to identify a group
in each paired area to receive a health brochure.   The second task was the
actual administering of the survey instrument.
                                    77

-------

                        Figure 4.4a
                           (Good)
Photograph Depicting Observation Paths for "Good" Visibility
                             78

-------
                        Figure 4.4b
                           (Fair)
Photograph Depicting Observation Paths for "Fair" Visibility
                             79

-------
                         Figure 4,4c
                            (Poor)
Photographs Depicting Observation Paths for "Poor" Visibility
                               80

-------
     Late in September 1977, a team of University of Wyoming and University of
Southern California personnel contacted by telephone a random sample of the
population in each of the twelve final sample areas.  The team was equipped
with reverse telephone directories- (i.e., phone directories listing people
by address instead of by name) which enabled accurate isolation of names and
addresses of potential respondents within the boundaries of each sample area.
Once streets and addresses were located within each sample area, a random
number generating table was utilized to pick the names of potential respond-
ents from each street.—'   This was done to insure that no bias would be
introduced into the telephone sampling process.

     People were then randomly contacted by telephone until at least thirty
people per area had agreed to cooperate wjLth the research team in the air
quality study.  Thus a minimum of 360 people, distributed over twelve sample
areas in the Los Angeles Basin, were to be the respondents in the asking game
portion of the study.

     Then in Spring 1978, half of the potential respondents were sent a
health pamphlet entitled "Air Pollution and'"Health,"  The half of the poten-
tial respondents receiving this pamphlet was to be the group upon which we
would test a learning hypothesis of the asking games.  Approximately 180
potential respondents comprised this group.

     In early March 1978, a research team comprised of staff and graduate stu-
dents from the University of Wyoming and a similar team from the University
of New Mexico went to Los Angeles to begin the survey instrument portion of
the study.  The two teams were divided into four groups in order to sample
each of the twelve sample areas most efficiently.

     The first order of business was to contact each of the potential re-
spondents by telephone and set up appointments with them.  Although most of
the potential respondents could be reached by telephone,  an unexpectedly
high percentage of persons who had previously agreed to cooperate with us
in the study declined the interview.  This drastically cut the potential
respondent list and forced alternative methods of sampling.

     Because of this setback, the sampling process was broken into three
parts.   First, of course, we arranged  interviews with those respondents
from our original lists who had said they were willing to cooperate and who
were still interested.  Second, we once again utilized the reverse telephone
directories to set up new appointments with people in each sample area.
Third,  each sample area was canvassed by members of the research team by a
door-to-door method.  By this procedure the sample size was approximately
345 interviews..successfully completed by the research team.  Table 4.3
presents the breakdown of the resulting sample by type of survey instrument.

4.4 Preliminary Empirical Results for the  Iterative Bidding Portion of
    The Survey Instrument Study

     This section will present preliminary results of the iterative bidding
format  portion of the contingent valuation study.  Initial bias tests will
be presented including vehicle bias,  starting point bias, and the potential

                                    81

-------
           Table 4.3




Survey Instrument Type Breakdown
Questionnaire
Type
Location
La Canada (A ->• B)
La Canada (A -* C)
El Monte (A •* B)
El Monte (A -» C)
Nontebello (A ->• B)
Kontebello (A .+ C)
JCanoga Park (B -» C)
"Encino (B ->- C)
Culver City (B -» C)
Pacific
Palisades (C -f C*)
Redondo Beach (C •* C*)
Palos Verdes (C -* C*)
Huntington
Beach (B -» Cl
Newport Bench (B f C)
Irvine (B -+ c)
TOTALS
Health Pamphlet
Aesthetic
Kale
1




2


1
3
3

1

1
12
Female
1
4


1
3

1
1

1
3
2
4
3
24
Acute
Male
1



1
1
It
1
1
1
3


3

16
Female
1






2
2
2
2
2
1


12
No Health Pamphlet
Aesthetic
Male
2
2
2
5
3
3
6
8
4
5
5
1
5
4
7
62
Female
7
5
5
7
5
5
it
6
8
3
5
7
8
7
7
89
Acute
Male
4
3
6
2
4
2
3
' 5
4
1
4

7
2
4
51
Female
1
4
3
6
4
4
3
5
9
6
4
7
16
2
5
79
Total by Area
by Type
1-°.
14
16
2C
18
20
20
28
30
2]
27
20
40
22
27
345

-------
for sequencing effects.  Additionally, some preliminary regression results
will be discussed in Appendix D.

     Table 4.4a,b presents the mean total bids by area partitioned by proposed
clean-up date.  The results in Table 4.4a range from $47.75 per month for the
Pacific Palisades area to $4.50Per month in the Newport Beach area.I/  This
difference in mean bids between Pacific Palisades and Newport Beach is not
fully understood at this time.  As was set forth in the theoretical Chapter
II, these results are not commensurate with economic expectations.  This
problem can be investigated when the substitution results are integrated into
the analysis.

     Appendix C presents tables for cumulative mean bids by sequence where ei-
ther aesthetic or acute information is presented first by area and differentiat-
ing between a 2 or 10 year clean up time horizon.  All other potential effects
such as biases are assumed zero.  The results presented in the appendices form
the basis of some simple statistical tests.  The tests to be considered are
whether:
     1.  the area mean bids are significantly different from zero;

     2.  the aesthetic, acute, chronic, and total bids for the paired
         areas are significantly different;

     3.  the results indicate the existence of starting point, vehicle
         or sequencing bias; and

     4.  the results indicate different bidding behavior when individuals
         were offered different completion dates for cleanup.

     The results of the t-tests regarding the equality of area mean bids
being statistically different from zero are presented in Table 4.5.  Of
interest is whether the results of the test allow the null hypothesis to
be rejected.  In all but the one case of Montebello area for the chronic
bid, the null hypothesis is rejected with 90% confidence.  Some cases sug-
gest higher levels of confidence.  Thus, we can initially infer that in
all areas, the values individuals place on the. three characteristics of air
quality under consideration tend to be non-zero.

     In Table 4.6, the equality between bids between the paired areas is
tested for the three characteristics and the total bid.  Only two pairs
reject the null hypothesis that the two areas' mean bids are equal:  Paci-
fic Palisades/Newport Beach and Culver City/Montebello.  The former was for
aesthetic, acute, and the total bid while the latter was for the acute health
bid only.  The purpose of this test can be seen in reference to the discus-
sion in the contingent valuation theory section.  At issue was the difference
between a bid from the property value study in comparison to the iterative
bidding study.  Recall that a contingency proposed to an individual was
moving him along an indifference curve.  Assuming that each area represents
a homogeneous set of preferences which differ across areas, the test in
Table 4.6 asks whether the movement in dollar amount is the same across the
paired areas.


                                    83

-------
                                 Table  4.4a

                            Mi-.HI  UUit by Ari-.n by  Type*

                      (C<>M|ilt:Lliiii Date  t>C CltMiiupi   2  Yr«.)

Arc.l
El Kin to (A - »)
Ei Horn c (A - C)
la C.in.1il--i U •• H)
U Con.ida (\ • C)
Honccbello (A •, B)
Hontcbcllo (A - C)
ConOfa Pnrk (b -» C)
Culver City (B . C)
Enclno IB • C)
Huntlncton Bcacli (B - C)
Irvine (B •• C)
Mcvporc Bc.iclv '  Hie  rL'S|H)'Kj*-iU  In
h.is !.<•»•" f-iik- vitli  rc»pcc
colU-clJon of bit!-*; .-uirt  {
life  c.Mc II.-I.H nr li.vi uni
A  1 Iff l:il»Io  U(.-|'U-i:. tin-
(or viirltiu!!  t-xi'f(-n.-d II (e
Lion  In tMA t.iMc  h.is been Hint  of  btrlct  aadlttvi:
Ity  cffrct.   In obc.ttnlnf,  the mo .in  bidw:   (1) no
I».TI(C  with  re;: eel  lo Che  bitldinp.  SCIIUIMICC ;  (2) rxo
i.i.itic  vlmthkT   lit-.jl Ch p;impliU't I'.lt*  or !us  not bt-ctl
 juiv.nscv ol"  i c Interview;  (3) n*> <:l f fcr^nt iatlon
i  to  itie 4lJff rtiit  pro|io:icd vrhlcli-« for  the
•'•)  no lUrfcn- iJatlou h.in  lin.n n-nlc ulivthirr a
      *St:nul.iril error oE

      *S:im|»lv MJru  of cue
":.UKk" cotuu ri-.irii* of  llu- »'Uciii-rt muntlily bJds
iq..ui;i.

hu  mr>.in tiid In all  cnsus.

 C.IHC Jn  All  <-.»Hon.

-------
                                Table   4.4b

                            HL.III IMUti by An-u by Type*

                           iU-llon D.uc of  Clc-.inun:   10 yra.)

Ar,-a
El ('.jut.- (A - B)
Kl tloiuc (A - C)
U Cnii.ida (A - t)
U Canada (A - C)
Hontcbcllo (h -. B)
Honlubcllo (A - C)
Canoc'i I'nrk (» - C)
Culver city (8 - C)
Enci:io (1) - C)
Huiuinstor. Bi-ach (8 - C)
Uvtnc (B -, C)
Hcvport Bench (3 - C)
Pacific fjlluadcs (C - C«)
Palos Vi.ciics (C - C*)
Kedondo tloicli (C •• C«)

At-sttu-tlc
Bin
1.67
(!.f)7)«
(,)._..
1.17
.(1.54)
3.60
(1.79)
14.43
(7..S5)
2.70
(1.32)
(101
4.38
(1.52)
(8l_
3.48
(1.19)
(HI
U.08
(4.41)
(12).
3.27
(1.93)
(l.D...
10.22
(3 . 30)
(I?)
10.90
(4.00)
liil_
1.15
(0.61)
5.58
(2.14)
'(12)
5.36
(1.24)
(IV)
12.46
(4.68)
(12)
•Van Rica
Aruie !k-.-llt!i
U 1 tl
11.89
(5.28)
L'i)
5.33
(3.07)
ty
15.60
(9.'»0)
	 (.IPJ
10.7!
(6.85)
_. O)
8.80
(4.73)
02L
1.38
(0.78)
	 (6)
2.07
(0.94)
OJJ
8.54
(?.54)
— —iLU.
5.36
(3.19)
(U)
10.79
(2.83)
tZ£I_
10.54
(3.92)
(12)
2.00
(1.09)
.___ (1C)
14.83
(3.45)
(12)
13.09
(5.12)
(")
6.96
(4.05)
(12)
(S./ao,,a)
Chronic ik'.ilth
BU
1.11
(1.11)
1.17
(O.S3)
(6)
9.50
(8.96)
(10)
0.43
(0.43)
(7)
1.70
(0.67)
(10)
0.75
(0.53)
(8)
0.48
(0.33)
ill)
4.08
(2.10)
(12)
1.45
(0.65)
(11)
6.84
(2.3S)
(19)
1.94
(0.92)
(12)
3.45
(2.06)
(10)
59.67
(39.58)
(1?)
2.73
(1.80)
(11) ._
4.42
(1.73)
.(12) ..

Total
Bid
14.67
(5.55)
(9)
10.67
(2.97)
(6)
28.90
(52.73)
(10)
25.57
(7.86)
(7)
13.20
(5.69)
(10)
6.50
(2.28)
(S)
6.03
(1.26)
(11)
23.71
(8.35)
(12)
10.09
(3.61)
28.42
(5.91)
_yjy
23.38
(5.96)
_oy_
6.60
(2.89)
(10J
80.08
(41.34)
(12)
21.18
(5.40)
. X'U .
23.83
(9.05)
(12).__
      M'lic implicit .issii;a;it ion  ^  this table  In?; bi-cn  tli.n
of h!rt:=  (or r.icli air  qti.iliiy cld'CU  1«  ol»t:\liil-.ij*,  il\c T.
olU-cUoa of  hiiiy; ;i i»il  ('•) no <\ 1 ( fcn-ut J-iL l<
Jfc  CtbK- l-:is or h.Tb  m»L  It•.•*•» »;liuvn Lo  the  r<-M-
 li!f  ciMc  iN-plrta  ilit- "stoc'h"  countcrpartti oC
or v.ir louy L-x]K-ccod  1 If c;"i>.Tnn,
                                                               of  otrict adriUivtty
                                                              .in bic-j:   (1) no
                                                                        :; (2)  no
                                                                        ;  nut  tictn
                                                                        •iul.lt Ion
                                                                         the
                                                                     cioulUly bida
      *SiniuUu-d  t-rror  of  the mean bid in  all caeca.

-------
                                                      Table 4.5


                       Results of the t-tests Regarding the Equality of Area Mean Bids to Zero*



                                        H :   The mean bid is equal to zero**
                                         o

                                        H :   The mean bid is greater than zero
Name of Area
El Monte
El Monte
La Canada
La Canada
Montebello
Mon tebello
Canoga Park
Culver City
Encino
Type of
Contingency
(A - B)
(A + C)
(A -+ B)
(A -* C)
(A •+ B)
(A -* C)
(B •* C)
(B + C)
(B -<• C)
n
20
13
17
17
19
19
19
28
28
Aesthetic Bid
Reject H at 95%
0
Reject H at 95%
o
Reject H at 95%
0
'
Reject H at 95%
o
Peject H at 99%
0
Reject H at 90%
0
Reject H at 99%
Reject H at 99%
°
Reject H at 99%
o
Acute Health Bid
Reject H at 95%
o
Reject H at 95%
o
Reject H at 90%
o
Reject H at 95%
o
Reject H at 99%
o
Reject H at 99%
o
Reject H at 99%
o
Reject H at 99%
o
Reject H at 99%
Chronic Health Bid
Reject H at 95%
0
Reject H at 95%
o
Accept H
o
Reject H at 95%
o
Reject H at 95%
o
Accept H
o
Reject H at 95%
0
Reject H at 99%
o
Reject H at 93%
(continued)

-------
                                                      Table 4.5
                                                     (continued)
Name of Area
Huntington Beach
Irvine
Newport Beach
Pacific Palisades
Palos Verdes
Redondo Beach
Type of
Contingency
(B ->• C)
(B •* C)
(B + C)
(C ->- C*)
(C -*• C*)
(C ->• C*)
n
38
27
20
20
19
26
Aesthetic Bid
Reject H at 99%
o
Reject H at 99%
0
Reject H at 99%
Reject H at 95%
o
Reject H at 99%
o
Reject H at 99%
0
Acute Health Bid
Reject H at 99%
o
Reject H 'at 99%
o
Reject H at 95%
o
Reject H at 95%
o
Reject H at 99%
o
Reject H at 99%
o
Chronic Health Bid
Reject H at 99%
o
Reject H at 99%
o
Reject H at 95%
0
Reject H at 90%
o
Reject H at 95%
0
Reject H at 99%
o
     *The bids for each air quality effect are assumed to be strictly separable.   In obtaining the mean bids,
no differentiation is made with respect to:  (a) different bidding sequences;  (b) vehicle used;  (c) starting
bid; (d) health pamphlet versus no health pamphlet; or (e) life table versus no life table (f) completion date of
cleanup.
    **0ne-tail test t     T~-—   where p = area mean bid for a certain air quality effect
                           //—
                                        s = sample standard deviation
                                        n = sample size

-------
                                                           Table 4.6

                                 Results of the Bid Equality Tests of the Paired Communities*

                                              H :   The two mean bids are equal
                                               o
                                              H :   The two mean bids are unequal
Paired Areas
Pacific Palisades
Newport Beach
Canoga Park
El Monte
Irvine
Palos Verdes
Encino
La Canada
Huntington Beach
Redondo Beach
Culver City
Montebello
N
20
20
19
33
27
19
28
34
40
26
28
38
Aesthetic Bid
Reject H at 99%
Accept H
Accept H
o
Accept H
Accept H
o
Accept H
Acute Health Bid
Reject H at 99%
o
Accept H
o
Accept H
o
Accept H
0
Accept H
o
Reject H at 95%
0
Chronic Health Bid
Accept H
o
Accept H
o
Accept H
o
Accept H
o
Accept H
o
Accept H
o
Total Bid
Reject H at 95%
o
Accept H
o
Accept H
o
Accept H
o
Accept H
o
Accept H
o
CO
CO
          *The statistical test employed was a two-tailed t-test of the null hypothesis (H )  that the mean bids for
     each of the paired communities were equal.   The test also used a pooled variance estimate in the celculation of
     the test-statistic.   The information in the major cells  of the table reports the level of significance for the
     statistical tests.  Rejection of H  at the reported significance level means that the test failed to reject H  at
     a higher level of significance.   Only three significance levels were tested:  90%, 95%,  and 99%.  Failure to feject
     H  means that H  could not be rejected at the 90% level  or greater.   The purpose of the  above tests is to check if
     tnWe is a stat2stically significant difference between  the mean bids of the areas within the same pair.   Throughout
     this analysis, the bids for each air quality effect are  assumed to be strictly separable.  In obtaining the mean
     bids no differentiation is made with respect to:   Ca)  bidding sequence; (b)  vehicle used; (c) starting bid;
     (d) health pamphlet versus no health pamphlet;  or (_&}  life table versus no life table (f) completion date of cleanup.

-------
     In an iterative bidding format, various types of biases might be
introduced by the structure of the survey instrument.  In this s.tudy, the
types of biases selected for examination were vehicle bias, starting point
bias, and information sequence Bias.

     A test of means was conducted between the monthly utility bill and the
lump sum payment mechanism for the areas- by characteristic bid and for the
total bid.  Table 4.7 presents the results.  The null hypothesis set forth
was that the mean bids were equal irrespective of the Bidding vehicle.  For
Montebello, Canoga Park, Encino, Huntington Beach, Newport Beach, Pacific
Palisades, Palos Verdes, and Redondo Beach, the null hypothesis is accepted
for the total bid.  However, for Irvine, Culver City, La Canada, and El Monte,
we reject the null hypothesis, at least at the 90% confidence level, for the
total bid.  No obvious reason exists at this point in time for this result.
The principle problem area then appears to be in the aesthetic bids.z.'

     A second bias of concern is that of starting point bias.  Recall from
previous discussions that starting point bias results from the final bid
being definitely related to the starting bid, i.e.,  the higher the starting
point, the higher will be the final bid, thus suggesting a type of inform-
ation bias.  Table 4.8 presents the results of a test for starting point
bias.  The structure of the test was as follows.  Three starting points of
$1, $10, and $50 were employed in the survey instrument.  This results in
three potential comparisons of starting points for the resulting mean bids:
(1) $1 to $10; (2) $1 to $50; and (3) $10 to $50.  The null hypothesis was
whether the total mean bids were equal within the three combinations of mean
bids ignoring all other potential effects.  For the $1 to $10 pair, the null
hypothesis of no effect was rejected in La Canada and Encino.  The. $1 to
$50 pair was rejected for La Canada and Montebello.   Finally, the $10 to $50
pair was rejected only for Redondo Beach.

     To fully understand why the isolated cases indicate starting point bias,
a greater understanding would require consideration of other systematic
effects in the data set.  However, preliminary evidence based on Table 4.8
suggests that starting point bias is not a major problem for all of the
iterative bidding results.

     Another area of consideration is the question of sequencing of inform-
ation affecting the bid structure not only for the air quality characteris-
tic bids, but also the final bid.  Recall that bids  were collected according
to the following sequences:

     1.  aesthetic, aesthetic plus acute, and aesthetic plus acute plus
         chronic, or,

     2.  acute,  acute plus chronic,  and acute plus chronic plus aesthetic.

The question of sequencing is whether the ordering of the bidding process
effects the size of the bid.  For instance,  will individuals bid a different
amount for aesthetic effects if it is first,  as in (1) above, compared to
being last as in (2) above.  Similarly, will the acute bids vary?  Addition-
ally, we are interested in whether the orderings presented in (1) and (2)

                                    39

-------
                                                      Table 4.7


                              Results of Che t-tests for che Equality of the Mean Bids

                                          by Sample Area by Bidding Vehicle5'1



                                         H :  The two .mean bids are equal
                                          o


                                         H :  The two mean bids are unequal
Name of Area
El Monte
La Canada
Montebello
Canoga Park
Culver City
Encino
Huntington Beach
Irvine
n^*
20
22
21
1
17
U
18
9
n2**
13
12
17
12
11
14
22
18
Aesthetic Bid
Accept H ***
0
Accept H
Accept H
o
Accept H
o
Reject H at 95%
Accept H
o
Accept H
o
Accept H
o
Acute Health Bid
Accept H
0
Accept H
o
Accept H
o
Accept H
o
Accept H
Accept H
o
Accept H
o
Reject H at: 90%
o
Chronic Health Bid
Accept H
o
Accept H
o
Accept H
o
Accept H
0
Accept H
0
Accept H
o
Accept H
o
Accept H
o
Total Bid
Reject H at 90%*>'<**
o
Reject H at 90%
o
Accept H
o
Accept H
o
Reject H at 90%
o
Accept H
o
Accept H
o
Reject H at 95%
o
(continued)

-------
                                                      Table 4.7
                                                     (conciiuied)
Name of Area
Newport Beach
Pacific Palisades
Palos Verdes
.
Redondo Beach
"1**
13
11
12
'
17
nz**
7
9 '
7
g
Aesthetic Bid
Accept H
o
Accept H
0
Reject H at 95%
Accept H
o
Acute Health Bid
Accept H
o
Accept H
Accept H
o
Accept H
o
Chronic Health Bid
Accept 1!
0
Accept H
0
Accept H
o
Accept H
o
Total Bid
Accept H
0
Accept H
0
Accept H
o
Accept H
. o
     "Throughout the questionnaires, rvo different vehicles of payment are employed.  Some Respondents are
proposer! to pay their bids in separate monthly payments, and some others in additions to their utility bills.
This tabJe is prepared to check if the choice of the payment vehicle has any statistically significant effect
on the mean of the bids.

     The bids for each bidding stage are assumed to be strictly separable.  In obtaining the mean bids no
differentiation is made with respect to:  (1) starting bid; (2) bidding sequence; (3) health pamphlet versus
no health pamphlet; and (4) life table versus no life table-(5) completion date of cleanup.

    **n^:  Sample size of the interviews in which the respondent was proposed to pay his bids in separate
           monthly payments.

    **n2:  Sample size of the interviews in which the respondent was proposed to pay his bids as additions to
           his utility bills.

   ***The tests are done for <* •= O.Q1, * = 0.05, and =; = 0,10^  "Accept HO" means HQ is accepted for 1-* ~ Q.90
and higher,  i.e.,  for « <_ 0.10.   "Reject Ho at X%" means Ho is rejected at the given X% but is accepted at the
next higher 1-= value, i.e., if x% - 90%, then Hn is accepted at l-« - 95%.

-------
                                  Table A.8



                   Test of Means  for  Starting  Point  Bias*
                         H :   Mean  bids  are  equal
                          o


                         H :   Mean  bids  are  unequal

Name of Area

El Monte



La Canada



Montebello



Canoga Park



Culver City



Encino



Huntington Beach



Irvine


Sample
n
8
8
15

13
13
12

16
16
12

7
7
6

9
9
11

9
9
11

8
8
15

7
7
12

Sizes
"2
15
9
9

12
9
9

12
9
9

6
6
6

11
8
8

11
8
8

15
15
15

12
8
8


Starting Point Pairs
1-10
1-50
10-50

1-10
1-50
10-50

1-10
1-50
10-50

1-10
1-50
10-50

1-10
1-50
10-50

1-10
1-50
10-50

1-10
1-50
10-50

1-10
1-50
10-50

Totals
Accept H **
Accept H°
Accept H°
o
Reject H at
Reject 11° at
Accept H°
o
Accept H
Reject H° at
Accept H°
o
Accept H
Accept H°
Accept H°
o
Accept H
Accept H°
Accept H°
o
Reject H at
Accept H°
Accept H
o
Accept H
Accept H°
Accept H°
o
Accept H
Accept H°
Accept H°
o




95%
90%



90%










90%











(continued)
                                       92

-------
                                  Table  4.8
                                 (continued)

Name of Area


Newport Beach



Pacific Palisades



Palos Verdes



Redondo Beach



Sample
n.
1
9
9
6

7
7
6

2
2
9

10
10
10


Sizes
n
2
6
5
5

6
7
7

9
8
8

10
5
5



Starting Point Pairs

1-10
1-50
10-50

1-10
1-50
10-50

1-10
1-50
10-50

1-10
1-50
10-50

Totals
Accept H
Accept H°
Accept H°
o
Accept H
Accept H°
Accept H°
o
Accept H
Accept H°
Accept H°
o
Accept H
Reject H° at
Reject H° at
o
i













95%
90%


     *The purpose of  this  table  is  to check  if  there  is  any  significant
influence of the starting  bid  offered by  the  interviewer on  the total bid
of the respondent.   In calculating  the mean  total  bids:   (1)  no differentiation
has been made with respect to  the sequence  that  the air  quality effects are
presented; (2) no differentiation has been made  whether  a health pamphlet has
or has not been sent  to the  respondent in advance  of  the interview;  (3) no
differentiation has been made  with  respect  to  the  different  proposed vehicles
for the collection of bids;  and  (4) no differentiation has been made whether
a life table has or has not  been shown to the  respondent during the  interview.
A life table depicts  the "stock" counterparts  of the  elicited monthly bids for
various expected lifespans.(S) no differentiation has been made  with respect  to
the different dates of cleanup.
    **Accept H  •+ H  is accepted for 1 - = =  0.90  and higher; i.e.,  for
= < 0.10.     °    °
                                       93

-------
will give different total bids.  Ideally, the sequencing or ordering of bid-
ding information will not affect the results.  In an attempt to test for
sequencing affects, two separate tests of -means were"conducted•  The first
test involved a comparison by area By bid type of the mean values of the
observed bids against the derived Bids.  If an assumption of additivity is
made in the bids, then we can obtain an aesthetic observed bid (from 1 above)
and a derived aesthetic bid (from 2 above).  The question is then whether
the two bids differ.—'  That is, does- the order in which we obtain bids affect
the magnitude for the bid.   Table 4.9 presents the results of this test,  ^or
aesthetic bids, El Monte (A -> B) ,±±>  La Canada (A -*• C), Canoga Park, Encino,
Huntington Beach, Irvine Palos Verdes, and Redondo Beach the null hypothesis
was rejected.  The null hypothesis was rejected for the acute bids in La
Canada (A -*• C), Culver City, Encino, Huntington Beach, Newport Beach, and
Palos Verdes.

     The -null hypothesis was rejected for chronic bids for La Canada (A •> C)
and Newport Beach.  Finally, the null hypothesis was rejected for the total
mean bids only in Newport Beach and Pacific Palisades.  What can be con-
cluded from this set of results?  First, the test does not completely resolve
the issue of sequencing.  In some cases, the mean bids that were observed
are statistically different under the assumption of linear additivity.   Sec-
ond, keeping the first point in mind, we note that the total bid does appear
to be insensitive to the bidding across different orderings of character-
istics of the environmental good air quality.

     A second test to further investigate the extent of sequencing effects
is to compare each step of the bidding process irrespective of the subject
(i.e., acute or aesthetic information) of the bid.  The null hypothesis is
then to compare the mean values of step 1, the mean differences in values
of step 2 from step 1, the mean difference in values from step 2 to step 3,
and the total bid.  Table A.10 presents these results.  For the first bid-
ding step, only Palos Verdes had the null hypothesis rejected.  The null
hypothesis for the second bidding step was rejected for Pacific Palisades,
Newport Beach and Irvine.  For the third bidding step only El Monte was
rejected.  Finally, the null hypothesis was rejected for Pacific Palisades
and Newport Beach.—'   What can we conclude about sequencing from this test?
First, again no definitive statement can be made regarding the existence or
non-existence of sequencing.  The results suggest that regardless of the
information being bid upon, the step size (i.e., bid difference from the last
step) is independent of the information underlying the bid.  Second, ir-
respective of the bidding order, the total Md is insensitive to orHer effects.—''

     The results of the t-tests comparing the effects of different completion
dates of cleanup for each area are presented in Table 4.11.  Additionally,
Table 4.12 presents similar results for each of the paired areas.  The null
hypothesis of this test was that the bids are equal no matter the completion
date for the cleanup.  The null hypothesis was rejected only in isolated
cases such as Canoga Park in Table 4.11.  The implication of this result is
that individuals appear not to view the magnitude of their bid being signi-
ficantly determined by the proposed cleanup date.
                                    94

-------
                                                                Table <..9


                                        Results of the t-tests for che Equality of the Mean Bids

                                             for Observed versus Derived Bids by Sample Area*



                                                (Two-tail test; Pooled Variance Estimate)


                                                   H :  The two mean bids are equal
                                                    o

                                                   H :  The two mean bids are unequal
Name of Area
El Monte
El Monte
La Canada
La Canada
Montebello
MontebeJ Id
Canoga Park
Culver City
Enclno
Type of
Contingency
' (A - B)
(A -* C)
(A -» B)
(A - C)
(A -, B)
(A -» C)
(8 - C)
(B •* C)
(8 -» C)
-
No. of
"l
7
10
11
11
10
12
9
13 1
15
Oba.
nz
13
3
6
6
9
6
10
15
13
Aesthetic Bid
Reject H at 90X
o
Accept H
o
Accept: H
o
Reject H at 90%
o
Accept H
o
Accept H
o
Reject H at 952
o
Accept H
0
Reject H at 99%
o
Acute Health Bid
Accept H
Accept H
0
Accept H
o
Reject H at 902
0
Accept H
o
Accept H
0
Accept: H
Reject H at 90% i
o .
Reject H at 99X
0
Chronic Health Did
Accept H
o
Accept H
o
Accept H
0
Reject H at 95%
o
Accept H
0
Accept H
o
Accept ii
o
Accept H
o
Accept H
Total Bid
Accept H
o
Accept H
o
Accept H
0
Accept H
r 0
Accept H
o
Accept H
o
Accept H
o
Accept H
o
Accept H
o
(continued)

-------
                                                                Table 4.9
                                                               (continued)
Name of Area
Huntington Beach
Irvine
Newport Beach
Pacific Palisades
Palos Verdes
Redondo Beach
. Type of
Contingency
(B •» C)
(B -» C)
(B -* C)
(C - C*)
(C * C*)
(C •* C*)
No. of
"l
16
18
14
10
11
14
Obs.
"2
22
9
6
10
8
12
Aesthetic Bid
Reject H at 90%
o
Reject H at 95%
o
Accept H
Accept il
o
Reject H at 99%
o
Acute Health Bid
Reject H at 90Z
0
Accept H
Reject H at 95*
o
Accept H
0
Reject H at 90%
o
Reject H at 95% j Accept HQ
'
Chronic Health Bid
Accept 11
o
Accept H
o
Reject H at 952
o
Accept H
o
Accept H
o
Accept II
o
Total Bid
Accept H
o
Accept 11
o
Reject H at 95%
O
Reject H at 90%
o
Accept H
o
Accept H
o
     *Across the questionnaires,  the effects of air quality are introduced  in  two  different  sequences:   (1)  Aesthetic  •+  Acute  Health -«•
Chronic Health; (2) Acute Health -1- Chronic Health •* Aesthetic.   The bids  for each  effect  are assumed  to  be separable.  The  purpose of
the above tests is to check if there is any significant  influence of the  sequence  of  presentation  of  the air quality effects on  the
mean bids for each effect.   For example,  the mean aesthetic bid obtained  by the  first sequence  for some  area is  compared with  the mean
aesthetic bid obtained by the second bidding sequence for the same area.  These  tests of  significance are repeated  for each mean bid
and for each area to find out the "sequencing effect" on bids.  ' In obtaining the mean bids,  no  differentiation is made with respect to:
(a) vehicle used; (b) starting bid; (c) health pamphlet  versus  no health  pamphlet;  and (d) life table versus no  life table., (f)
pletion  date of clean-up.

-------
                                                      Table 4.10


                            Results of the t-tests for Comparing the Sequencing Effects

                                         in Each Step of the Bidding Process


                                     (Two-tailed test; Pooled Variance Estimates)


                                              H :   The bids are equal
                                               O

                                              H :   The bids are unequal
Name of Area
El Monte
La Canada
Montebello
Canoga Park
Culver City
Encino
Huntington Beach
Irvine
"1*
17
22
22
9
13
15
16
18
V
16
12
16
10
15
13
24
9
First Bid
Accept H **
0
Accept H
o
Accept H
o
Accept H
o
Accept H
o
Accept H
0
Accept H
o
Accept H
o
Second Bid
Accept H
0
Accept H
o
Accept H
o
Accept H
0
Accept H
o
Accept H
o
Accept H
o
Reject H at 90%
o
Third Bid
Reject H at 90%
o
Accept H
O
Accept H
o
Accept H
o
Accept H
o
Accept H
0
Accept H
0
Accept H
o
Total Bid
Accept H
0
Accept H
o
Accept H
o
Accept H
0
Accept H
o
Accept H
o
Accept H
o
Accept H
0
(continued)

-------
                                                                   Table 4.10
                                                                  (continued)
oo
Name of Area
Newport Beach
Pacific Palisades
Palos Verdes
Redondo Beach
V
14
10
11
14
V
6
10
8
12
First Bid
Accept H
o
Accept H
o
Reject H at 95%
Accept H
o
Second Bid
Reject H at 95%
o
Reject H at 90%
0
Accept H
o
Accept H
0
Third Bid
Accept H
o
Accept H
o
Accept H
Accept H
o
i 	 • • 	
Ttotal Bid
Reject H at 95%
o
Reject H at 90%
o
Accept H
o
Accept H
o
     *This table presents the results of the t-tests done to determine whether or not there is a significant
difference between the means of the first bid,  the difference between the first and second bids, and the
second and third bids irrespective of bidding sequence (i.e., whether the Aesthetic or Acute bid was asked
first in the questionnaire).
                 **n  = chose questionnaires  which  ask Aesthetic  question first,   n  = those questionnaires which ask. .Acute
             question first.

-------
                                                               7-ihle A.11
                                            ReaulCs of th« t-testa for Conparing the Effecta
                                                of Different Conplet ion Dates of Cleanup
                                                 in Each Step of the Bidding Process a«b

                                              (Two-tailed test; Pooled Variance Estimates)

                                                       H :  The bida nr« of]ual.
                                                       H :   The bids arc unequal.
Area
Zl Konte
El Monte
La Canada
La Canada
Hontcbello
Hontcbcl io
Canoga Park
Culver City
EDC tOO
Hontington Beach
Irvine
Nevporc Be-ich
Pacific PjJlsades
Pa Los Verdes
Redondo Beach
. Type of
Contingency
A •' B
A * C
A •* B
A -> C
A * B
A ' C
B -* C
B - C
Nunber of
Observations
V
10
7
7
10
9
11
8
16
B - C 17
B » C
B -» C
B •• C
C - C*
C -• C«
C - C*
19
15
10
8
8
14
V
9
6
10
7
10
s
11
12
11
20
12
10
12
11
12
Mean Bids
Aesthetic Bid
Accept H
Accept H
o
Accept H
o
Accept K
Accept H
o
Accept H
o
Accept H
o
Acct-'pt H
o
Accept H
Accept H
o
Accept H
0
Accu-pt H
o
Accept H
0
Accept H
0
Accept H
0
Acute Health Bid
Accept H
o
Accept H
0
Accept H
o
Accept H
o
Accept H
Q
Accept H
o
Reject H at 99Z
0
Accept H
o
Accept H
0
Accept H
o
Accept H
0
Accent H
0
Accept H
o
Accept I!
o
Accept H
o
Chronic iie.il th Bid
Accept H
o
Accept H
o
Accept. 15
o
Accept H
o
Accept H
o
Accept t!
O
Accept 11
o
Accept H
o
Ac cup I H
o
Accept H
Accept IJ
0
Accep: H
o
Accept K
O
Accept H
o
Accept H
o
Tot.il Bid
Accept H
o
Accept H
o
Accept H
o
Accept* H
o
Accept H
o
Accept H
o
Reject H at. 951
o
Accept K
0
Reluct H at 90*
o
Accept H
o . ,
Accept H
Accept H
o
Accept H
o
Accept H
o
Accept H
0
      The bids for each bidding stage are assumed to be strictly separable.  In obtaining the juean bids, no differentiation la Bade
with respect to:  (1) bidding sequence; (2) starting bid; (3) bidding vehicle; (4) health pamphlet veraus no health pamphlet; and
(5) life table versus no life table.

     b
      The tests are done for « - 0.01. - - Q.05, and - " 0.10.  "Accept H " Hi-ana H  la accepted for « - 0.01, « - 0.05, and - - 0.10.
"fc«J«ct H  at XI" aeans that H  is rejected at XZ, but is accepted at the°nei!t higher I - * value; e.g., If XI • 901. then H  Is
redacted  for « • 0.10 but Is Accepted for « • 0.05 ond « - 0.01.                                     •                     °

     •o^ . aaaple tiz, Of interviews with propoted completion date of cleanup of 2 years.

     *»2 " »Hple a-iM of interviews with propose* coapletiea dace et eleoaup of 10 yeare.

-------
                                                         Table  4.11


                                             Results  of  the  t-cesis  for'Copipnring  the  Effects
                                                 of  Different  Completion  D.ito.s  of  Cleanup
                                                 in  Each rtcp  of  the iUJciin;;  Process ai°.c

                                               (Two-tailed test;  Pooled Variance Estimates)

                                                        H :  The  bids  are equal.
                                                        H  ;   The  bids  are  unequal.
Paired Areas
El Monte-
Canoga Park
Montebelio-
Culver Citv
La Canada-
Enclno
Huntlngton Beach-
Redondo Beach
Newport Beach-
Pacific Palisades
Number of
Observations
V
25
35
34
33
18
Irvine- j 23
Palos Verdes i
V
26
30
28
32
22
23
Mean Bids
Aesthetic Bid
Accept H
o
Accept H
Accept H
r 0
Accept H
o
Accept H
o
Accept H
0
Acute Health Bid
Accept I!
0
Accept HQ
Accept H
o
Accept H
o
Accept H
o
Accept H
o
Chronic Heal th Bid
Accept H
o
Accept 11
0
Accept H
o
Accept H
o
Accept K
o
Accept H
o
Total Bid
Accept H
o
Accept H
o
Accept H
o
Accept H
o
Accept H
o
Accept H
0
Aggregate Data Set j 169
161
Accept H
o
Accept H
0
Accept H
0
Accept H
0
      In obtaining the results* the total number of Interviews in each paired area is divided into two parts with respect to
their proposed completion dates of cleanup (i.e., 2 years versus 10 years) and the t-testp are done to test whether this
difference has any significant influence on the mean bids.


      The tests are done for « - 0.01, " - 0.05, and « - 0.10.  "Accept H " means U  is accepted for all three of the
« levels.                                                                °         °
      The bids for each bidding stage are assu.-r.ed to be strictly separable.  In obtaining the Bean bids no differentiation la
made with respect to:  (1) bidding sequence; '(2) starting bid; (3) bidding vehicle; (4) health pamphlet versus no health
pamphlet; and (5) life table versus no life table.

     *n  - sample size of interviews with proposed completion date of cleanup of 2 years.

     *n  - eample alze of interviews with proposed completion data of cleanup of 10 y&ara.

-------
     The previous results presented rely in many cases on small sample sizes
for the statistical tests due to the partitioning required.  Recall that
several types of bias as well as game structure questions had to be examined.
In view of the small sample sizes for a few areas, additional questionnaires
were administered.  Table 4.13 presents the mean Bid results of these ad-
ditional interviews.  Before integrating into the basic data set, it was
decided to test whether the "new" data was significantly different from the
"old" data in terms of mean values.  Results of the tests are presented in
Table 4.14.  Culver City for the total bid category is the only significantly
different result from the "old" data set.  This is due to one of the indiv-
iduals bidding an exceptionally larger sum than others as noted in the foot-
note in Table 4.13.
                                     101

-------
                                Table 4.13

             Mean Bids by Area by Type  for the "New" Data  Set*
Area
Canoga Park (B->-C)
Canoga Park (B-vC)
Culver City (E->C)
Encino (B->C)
Encino (B->£)
Pacific Palisades
(C->C*)
Pacific Palisades
(C-vC*)
Completion
Date of
Cleanup
2 years
10 years
10 years
2 years
10 years
2 years
10 years

Aesthetic B
3.20
(1.29)**
(10)***
3.00
(2.00)
(5)
6.25
(6.25)
(4)
6.67
(6.67)
(3)
4.17
(1.54)
(6)
6.00
(3.70)
(6)
16.67
(23.58)
(9)
Mean-
Acute
H.B.
9.18
(2.12)
(10)
9.00
(2.92)
(5)
27.50
(8.29)
(4)
13.33
(7.26)
(3)
6.67
(3.07)
(6)
9.17
(2.71)
(6)
12.89
(5.75)
(9)
Bid ($ /Month)
Chronic
H.B.
8.00
(3.27)-
(10)
10.00
(5.70)
(5)
28,75
(16.25)
(4)
1.67
(1.67)
(3)
8.33
(6.41)
(6)
6.67
(2.98)
(6)
6.67
(2.89)
(9)

Total
Bid
20.30
(3.41)
(10)
22.00
(8.15)
(5)
62.50****
(21.07)
(4)
21.67
(12.02)
(3)
19.17
(6.76)
(6)
22.67
(6.31)
(6)
35.11
(13.35)
(9)
   *The bids for each  bidding  stage  are  assumed  to  be  strictly  separable.   In
    obtaining the mean bids  no differentiation has  been  made  with  respect  to  1)
    bidding sequence,  2)  starting  bid, 3)  bidding vehicle,  4) health  pamphlet
    versus no health pamphlet,  and 5)  life table versus  no  life  table.   A life
    table depicts the  "stock"  counterparts of  the elicited  monthly bids  for
    various expected life spans.

  **Standard error of  the mean bid in  all  cases.

 ***Sample size of each case in all  cases.

****Individual total bids for  Culver City  were as follows:
       I:  $100
      II:  $ 75
     III:  $ 25
      IV:  $ 50
Aes.
 0
 0
 0
25
Ac.
25
50
10
25
Ch.
75
25
15
 0
                                           102

-------
                                                     Table A.IA

                  Results of the t-tests for Comparing the Equality of the Mean Bids Obtained from
                     the "Old" and the "New" Data Sets in Each Step of the Bidding Process3'

                                     (Two-tailed test; Pooled Variance Estimate)
                                               H :  The bids are equal.
                                               H-' :  The bids are unequal.
Completion
Type of Date of
Area Contingency Cleanup
Canoga Park
Canoga Park
Encino
Encino
Pacific Palisades
Pacific Palisades
Culver City
B-K:
B-K:
B-K:
B-K:
C-K:*
C-K:*
B-vC
2
10
2
10
2
10
10
years
years
years
years
years
years
years
Number of
Observations
Nl N2
8
11
17
11
8
12
12
10
5
3
6
6
9
A
Mean Bid
Aesthetic
Bid
Accept
Accept
Accept
Accept
Accept
Accept
Accept
Ho
Ho
Ho
Ho
Ho
Ho
Ho
Acute Health Bid
Accept
Reject
Accept
Accept
Reject
Accept
Reject
"o
HQ at 95%
Ho
Ho
H at 95%
H0
H at 95%
Chronic
Health Bid
Accept HQ
Accept H
Accept H
Accept H
Accept H
Accept 1!
Accept H
Total Bid
Accept
Reject
Accept
Accept
Accept
Accept
Accept
Ho
HQ at 90%
Ho
Ho
Ho
Ho
Ho
 The bids for each bidding stage, are assumed to be strictly separable.  In obtaining the mean bids, no differentiation
 is made with respect to: 1) bidding sequence, 2) starting bid, 3) bidding vehicle, A) health pamphlet versus no health
 pamphlet, and 5) life table versus no life table.  The "old" and "new" data sets are a result of additional inter-
 viewing that was carried out to supplement the sample size in a few areas.
O.OS but is accepted for <* = 0.01.
 The tests are done for = = 0.01, <* = 0.05, and « = 0.10. "Accept H " means H  is accepted for <* = 0.01, « = 0.05,
 and = = 0.10.  "Reject H  at X%" means that H  is rejected at X2 , but is accepted at the next higher 1-= value;
 e.g.., if X% = 95%, then H  is rejected for « = 0.10, and

N-: Sample size of interviews from the "old" data set.
N^: Sample size of interviews from the "new" data set.

-------
                           FOOTNOTES;  CHAPTER_IV


     —  Some questions have been raised about employing a "typical week." as
the time period upon which to base, the analysis.  However, in a previous
study valuing wildlife, various time units- were employed (last trip, typical,
dairy formats) for hunting experience.  No statistical difference was found
in the activity responses.  See BrooRsfiire, Randall, et. al. (1977).
     21
     —  Note when the acute initiation point was employed the picture sets
were not made available until the bidding on the aesthetic characteristic
was begun.

     3/
     —  Note holding the respondent to the original time constraint for their
current situation implies no leisure-work tradeoff possibilities.  Arguments
can be'presented for or against this assumption.  However, we note that in
Blank, et. al.  (1977) that when this tradeoff was offered as part of the
substitution format, few respondents did make the trade.
     A/
     —  In a few cases the respondents bid their maximum willingness to pay
initially rather than on the characteristic points.

     —  The equipment used was a 135 mm Minolta single lens reflex camera
(SLR), 135 mm telephoto lens, 55 mm lens, tripod, Kodak Vericolor II profes-
sional color film, log sheets and a small ice chest.

     —  It should Be noted that film is affected a great deal by even mild
variations in temperature, especially heat.  Thus, it is critical for the
film to be protected before, during, and after use.  The film used for our
study was kept in a small portable ice chest until used; once the roll or
film was used, it was put back into its air tight container and then back
into the cooler until ready for processing.  As a further aid in protecting
the film, the researchers used rolls of twenty rather than 36 frames in an
effort to minimize the time the film was exposed to heat.

     —  Appendix E details the exact streets in the sample plan.
     o /
     —  Note the Pacific Palisades bid is for a C -> C* contingency, thus
implicitly employing a bid for a basin-wide improvement, not involving this
location directly.
     9/
     ~~  Further examination of vehicle bias will require breaking out
aesthetic observed versus aesthetic derived bids and conducting a t-test.

     —-  Note this applies to acute and aesthetic bids,  however, the chronic
bids are entirely derived.

     —  The (A -»- B) and (A -> C)  notation refers to type of contingency moves
for residents of the A (poor) air quality area.   The principle reason for
administering two types was- to avoid an overly long survey instrument.
                                     104

-------
     12/
     —  This is the identical result noted in the first sequencing test
which.by the structure of the tests must be the case.



     —  A follow-up on this thesis would be a te

against total for step 2 against total for step 3.
—  A follow-up on this thesis would be a test of the total step size 1
                                     105

-------
                                 Chapter V

                   THE SOUTH COAST PROPERTY VALUE STUDY

5.1  Overview

     Many different methodologies for valuing non-market or environmental
goods and services have been proposed.  However, none of these approaches
is universally accepted and debate remains over which methodology is most
appropriate.  New valuatioq methods such as the contingent valuation approach,
are marked by uncertainty and criticism from both professional and r.on-
professional audiences, and thus require replication and evidence of
internal consistency in order tp demonstrate validity

     The purpose of the research on property values presented here is to
provide the necessary comparison for the contingent valuation approach
which is described in detail above.  This is accomplished through an
analysis of the housing market within the sample plan communities of the
South Coast Air Basin located in Los Angeles and Orange Counties.  Specifi-
cally, this research asks if households will actually pay for cleaner air
in the form of higher property values for homes in clean air communities
and if this willingness to pay is comparable to the hypothetical willingness
to pay expressed in the survey instrument.

     Valuation of reductions in urban air pollution concentrations based
upon housing value differentials is the most common form of the hedonic
price procedure as developed by Rosen (1974), the basis of which is Lan-
caster's (1966) consumption theory.  This procedure assumes that access
to environmental (dis)amenities is capitalized in property values.  This
assumption is based on the premise that households are willing to pay a
premium for an otherwise identical home located in a clean air area versus
that located in a polluted area.  The capitalization can be. discovered by
isolating the impact of air quality in two alternative ways:  (1) by
developing a sample pairing system which minimizes the variation in housing
and community variables other than air quality and comparing housing values;
or (2) by regressing housing value data on air quality and other variables.
In the latter method,  the resulting empirical relationship is the basis for
a determination of the value of the environmental good.

     Previous housing value studies have concentrated on the regression
procedure.   The first significant empirical study of air quality and pro-
perty values was done by Ridker and Henning (1967),   The authors applied a
least squares regression model to cross-sectional data (compiled by census
tract) for the St.  Louis area.  In order to fully specify their model of
property values, variables corresponding to housing, location, neighborhood,
                                   106

-------
political jurisdiction and individual characteristics were included with air
pollution measures as independent variables.  A significant negative rela-
tionship was found between the. sulfation level (annual geometric mean) and
median property values.  Further, a property value increase of between $83
and $245 was associated x^ith each .25/mg./100cm2/day reduction in the
sulfation level.  This translates into total benefits of approximately $83
million if sulfation levels are reduced by .25 nig., or to an ambient
concentration level of .49 mg.

     This research was followed by a similar study by Anderson and Crocker
(1971) who analyzed the impact of air pollution on both renter and owner
occupied properties for St. Louis, Kansas City and Washington, B.C.  As in
the Ridker-Henning work, the basic unit of observation was the census tract and
cross-sectional data which was employed in a non-linear regression model.  The
Anderson-Crocker results confirmed the Ridker-Henning finding of a negative
and statistically significant relationship between air pollution (annual
arithmetic mean concentrations of sulfur oxide and suspended particulates)
and property values.  The same result was also found for rental property.

     Deyak and Smith (1978), in an effort to generalize these empirical
conclusions utilized an updated data base (1970 census) gathered from
representative SMSA's.   Their results provided added support for the findings
of Ridker-Henning and Anderson-Crocker.  However,  in andtaer paper, Smith-
Keyak (1975), using data on owner and renter occupied central city housing
in eighty-five cities,  which also formulated a residential location model
that included location pubiic services and tax effects, found that air
quality did not significantly affect property values.  This conclusion was
in accordance with the results found by Steele (1972) and later" Wieand
(1978).  Both authors found no statistically significant relationship to
exist between air pollution and property values.   The Wieand findings are
especially surprising since he employed essentially the same data base as
Ridker-Henning.  The major change was substituting monthly rent per acre
•in place of median property value as the dependent variable.

     These results indicate that  an analysis of housing markets can yield
information on the value of non-market goods.  However, they also demon-
strate the fragility of the methodology.  That is, all assumptions outlined
in Chapter II must be met and extreme care is required in model specification
and interpretation of the results.

     The analysis undertaken here encompasses three separate but related
approaches, with benefits from reduced air pollution in bid equivalent terms
(e.g., terms comparable to the contingent valuation results)  specified at
each level.  The first  approach involves a straightforward comparison of
average housing values  in the sample pair communities, standardizing only
for house size (square  feet of living area).   The  resulting differential
in sale price between paired communities,  which are theoretically identical
except with respect to   ambient air pollution concentrations,  is then
attributed to the disparity in air quality.   It should be noted that this
methodology relies quite heavily  on the successful operation of the sample
plan.   That is, the variation across pairs in all  other housing and neigh-
borhood characteristics (excepting air quality) must be minimal if the sale
price differential assigned to air pollution is to reflect accurately
                                     107

-------
individuals  true willingness to pay for clean air.

     In the second procedure we utilize an econometric estimate of the
impact of air quality on housing values to determine benefits of reduced
air pollution.  This portion of the study corresponds to the traditional
econometric analysis of the housing market and is an attempt to estimate
a linear relationship between a home's sale price and its supply of housing
and community attributes.  The value of an improvement in air quality is
then deduced from the resulting hedonic housing value equation.

     The final approach is a further refinement of the above methods and
consists of a multi-step procedure which makes allowance for air pollution
abatement to be valued differently by households with varying income and
initial pollutant concentrations.  This methodology was developed recently
in a paper by Harrison and Rubinfeld (1978).  The first step is to estimate
a hedonic housing value equation, similar to the second approach, but
allowing 'for non-linearities where appropriate in the functional form.  The
second step is to calculate the marginal willingness to pay for individuals
in each of the sample communities for a small change in air quality.  The
third step is to estimate a marginal willingness to pay equation as a
function of income and other houshold variables.  By integrating individual
marginal willingness to pay estimates, we at least partially overcome the
problems pointed out in Section 2.1.  Finally, we employ this latter rela-
tionship to determine benefits of air quality improvements.

     Each of the three appro-aches as described above can be viewed as a
part of a systematic analysis of housing market data in the communities
which comprise the sample plan.  Further, each procedure yields pollution
abatement benefit estimates which can be used to compare to the contingent
valuation experiment.  In addition to its usefulness as a comparability
exercise, this housing value analysis has advantages over previous studies
in that data is drawn as part of the sample plan which by its nature con-
trols for many exogenous factors not wholly explained in the standard treatment.
This, for example, tends to explain why our statistical "fit" is superior
to previous studies.  However, it should be kept in mind that sampling
is therefore appropriate for comparison to the contingent validation
experiment but is non-random and may lead to biased estimates of basin-wide
damages.

     The remainder of this chapter is organized as follows.  Section 5.2
describes the data base and sources utilized in the study.  In Section 5.3,
the three appraoches and their associated results are presented.  Section
5.A concludes'the analysis.

5.2  Data Characteristics

     The area under investigation is defined as the South Coast Air Basin.
However, this study utilizes data for the sample plan communities only (see
Chapter IV).  In this regard, the sample chosen for study is not entirely
random but rather a function of a pre-testing scheme.  This may not Be a
major restriction on either the methodology or results since the paired
communities are representative of the entire spectrum of living conditions
in Los Angeles and Orange Counties^  It sfiould also be noted that this study
                                    103

-------
is restricted to single family residences and the results are therefore
only possibly applicable  to other housing types (niohile homes s apartments,
condominiums).  Further, we concentrate on the owner market to the exclusion
of other markets (rental, leasing, etc.).

     Focusing upon the paired communities then, the data base was constructed
to enable the impact of air quality differentials on housing sale price to
be isolated.  Thus, the dependent variable in the analysis is the sale price
of owner occupied single family residences.  The independent variable set
consists of variables which correspond to three levels of aggregation:
house, neighborhood, and community.  The data base contains 719 independent
observations.  Table 5,1 describes further the data employed in the study.

     The housing characteristic data, obtained from the Market Data Center
(a computerized appraisal service centered in Los Angeles), pertains to
homes sold in the January, 1977 - March, 1978 time period and contains
information on nearly every important structural and/or quality attribute.
Table 5.2 provides summary statistics for many of the housing characteristic
variables for each of the sample communities.  It should be emphasized that
housing data of such quality (e.g., micro level of detail) is rarely
available for studies of this nature.  Usually outdated data which is overly
aggregate (for instance census tract averages) is employed.  These data
yield functions are relevant for the "census tract" household  and  are only
marginally relevant at the micro level.  However, in this study it was
imperative that data comparable to that used in the contingent valuation
experiement be utilized.  That is, since pollution abatement benefit estimates
were calculated at the household level in the contingent study, it was
necessary to generate similar estimates based on comparable data in this
validation exercise.

     In addition to the immediate characteristics of a home, other variables
which significantly affect its sale price are those that reflect the
condition of the neighborhood and community in which it is located.  That
is, the local tax and public goods expenditure rates, school quality, ethnic
composition, crime rates, proximity to employment centers (and in the South
Coast Air Basin, distance to the beach), and measures of the ambient air
quality which have a substantial impact on sale price.  Therefore, in order
to capture these impacts and to isolate the independent influence of air
quality, these variables are included in the econometric modeling.

     The measures of air quality used in the empirical analysis were obtained
from California Air Resources Board publications (1977).  This agency is
responsible for monitoring pollution levels in the Basin.  The South Coast
Air Shed, because of the existence of a large number of both mobile and
stationary sources combined with meteorological and topographical conditions
which limit the area's ability to disperse pollutants, has a long history of
pollution problems.  A relatively complete regional network of monitoring
stations has been developed.  This allows the measurement of ambient air
pollution levels rather than concentrations on isolated hotspots.   A detailed
exposition of air pollutants by area was given in Chapter III.

     In conclusion, we view the data b'ase ass-embled for the housing value
study as appropriate for comparability testing of the contingent valuation
                                     109

-------
                                                   Table 5.1

                                Variables Used in Analysis of Housing Market
Variable
Definition (assumed effect on
housing sale price)
Units
                     Source
Dependent
  Sale Price
Independent-Housing
  Sale Date
  Age



  Bathrooms

  Living Area


  Pool


  Fireplaces
Independent-Neighborhood
  Distance to Beach
Sale price of owner occupied single
family residences.
Month in which the home was sold
(positive indicator of inflation)

Age of home (negative indicator
of obsolence and quality of
structure)

Number of bathrooms (positive
indicator of quality)
Living .area (positive indicator
of the quantity of home)
Zero-one variable which indicates
the presence of a pool (positive
indicator of quality)

Number of fireplaces (positive
indicator of quality)
Distance to the nearest beach
(negative indicator of relative
proximity to main recreational
activity)
                                                                    ($1,000)
January 1977=1
March 1978-15

Years
Number

Square feec
Zero=no pool
One=pool
Number
Miles
                     Market Data
                     Center
Market Data
Center

Market Data
Center


Market Data
Center
Market Data
Center
Market Data
Center
Market Data
Center
Calculated
(continued)

-------
                                                   Table 5.1

                                                  (continued)
Variable
Definition  (assumed effect on
housing sale price)
                                                                    Units
                                                                                          Source
  School Quality
  Ethnic
  Population Density
  Housing Density
  Distance to
    Employment
  TSP
School quality as measured by student
percentile scores on the California
Assessment Test-12th grade math

Ethnic composition-percent white in
census tract(s) which contain sample
community (positive).

Population density in surrounding
census tract (negative indicator
of crowding)
Housing density in surrounding
census tract (negative indicator
of crowding)
Weighted distances to eight
employment centers in the South
Coast Air Basin (negative indicator
of proximity to employment)
Nitrogen dioxide concentrations
Concentrations of total suspended
particulates
                                           Percentile  *100
Percent  *100
                      Local School files
                      in  sample communities
1970 Cfnsus
People per square     1970  Cenfus
mile
Houses per square    1970 Census
mile
Miles/Employment     Calculated
Density
Parts per hundred    California Air
million (pphm)       Resources Board

Micrograms per   ^   California Air
cubic inecer (iig/m )   Resources Board
(continued)

-------
                                                        Table 5.1

                                                      (continued)
     Variable
Definition (assumed effect on
housing sale price)
Units
Source
     Independent-Community
       Public  Safety
         Expenditures
       Crime
      Tax
H-1
vo
Expenditures on public safety per         $/People
capita (positive indicator of
attempt to stop criminal activity)
Local crime rates (negative indicator     Crime/People
of peoples' perception of danger)

Community tax rate (negative measures     $/$l,000 of
cost of local public services)            home value
                     1976-77 Annual
                     Report Financial
                     Transactions
                     Concerning Cities
                     of California

                     FBI (1976)
                     1976-77 Annual
                     Report Financial
                     Transactions
                     Concerning Cities
                     of California

-------
                                                   Table  5.2

                                       Average  Housing  Characteristics
City
Canoga Park
El Monte
Culver City
Montebello
Orange
Whittier
Redondo Beach
Huntington Beach
Pacific Palisades
Newport Beach
Palos Verdes
Irvine
Encino
La Canada
Population
Sale
Price ($)
43,914
34,273
82,916
63,957
70,368
67,647
64,817
77,214
257,383
141,473
165,016
83,054
209,158
153,804
99,719
Sale Price/
Sq. Ft.
40.299
32.1
58.03
43.48
46.89
41.64
58.6
53.239
91.05
68.5
64.98
50.97
70.95
59.91
58.07
Living Area
(Sq. Ft.)
1089.68
1067.68
1428.73
1470.95
1500.58
1624.67
1104.18
1450.32
2826.67
2065.41
2539.44
1629.49
2947.84
2567.17
1717.1
Number of
Bathrooms
1.16
1.18
1.54
1.67
1.98
1.64
1.28
1.95
3.14
2.43
2.72
2.13
3.04
2.6
1.99
Number of
Fireplaces
.227
.18
.56
.67
.73
.95
.30
. 71
1.78
1.20
1.24
.95
1.44
1.45
.86
Age of Home
(Years)
32.9
35.6
26.1
28.1
16.4
32.3
27.0
13.3
28.1
15.5
13.5
4.4
16.2
33.3
19.2
     The property value study includes two more communities in the data base then did  the contingent
valuation study:  Orange and Whittier.

-------
experiments.  The reasons are threefold.  First, the housing characteristic
data is extremely detailed at the household level of aggregation, and
extensive in that a relatively large number of observations are considered.
Second, we have assembled a variety of neighborhoods and community variables
which enable the isolation of the air quality influence on housing values.
Third, the air pollution data is comprehensivei

5.3  Empirical Analysis

     As outlined in the introduction, each of the three stages of empirical
analysis undertaken in this study constitutes a separate attempt to capture.
the monetary impact which air quality differentials have on housing values.
Once discovered, these monetary estimates of the air quality effects are
translated into the value of improving air quality in the South Coast Air
Basin.  These calculations are later utilized to test the validity of the
contingent valuation experiment.

     The following household benefit equation is used and shows the inter-
relationships or common characteristics of the three approaches;

                           N                 N
                  AHB = [( E  AQI.*NH *AD )/ £  NH ]*CRF              (5.1)
                          1=1    X          1=1

where

     AHB = average annual benefits per household for a reduction in air
           pollution concentration.

     AQI = air quality improvement in area i (poor-fair, fair-good).

     NH  = number of homes in area i affected by the air quality change.

     AD  = average home sale price differential attributed to a one unit
           improvement in air quality.

     CRF = capital recovery factor.  This is the rate necessary to transform
           an initial capital investment into a series of equivalent annual
           charges including payment of both capital and interest.  In this
           study the CRF is assumed to be .0995 which corresponds to a
           .0925 interest rate and a payback period of 30 years.

     N   = number of specific areas affected by air quality improvement.  In
           this study N is restricted to two as benefits are calculated for
           upgrading the air quality in the poor areas to fair and in the
           fair areas to good (see Chapter III).

     The number of homes in the affected areas (see Table 5.3) and the air
quality improvement are common to each of the three methodologies.  Table
5.4 illustrates both the present air quality classifications-, the reductions
in nitrogen dioxide (NO ), and total suspended particulatea (TSP) which are
required to achieve significant improvement as measured by the relative
quality indicators.  This analysis was not able to effectively separate out
                                     114

-------
                                  Table 5.3        , .    _
              Number of  Homes  in South Coast  Air  Basin
             	bV Air  Quality  Categories	
                                      Number of Homes  in South  Coast Air Basin
Air Quality                               (Los Angeles  and Orange  Counties)


Poor                                                 1,056,325
Fair                                                   804,823
Good •                                                 • 228,772


Total                                                2,089,920
                                  Table 5.4

                       Air Quality Definition - NO
Air Quality
(Arithmetic  Average  1975 - pphm)                                 Classification

           > 11                                                       Poor
           9-11                                                       Fair
           < 9                                                        Good
           12.38   ,,„.                                           Average Pocr
              9CC^-^™                                            1       r-  •
             .55                                                 Average Fair
            6.9                                                  Average Good
                       Air Quality Definition - TSP
Air Quality
(Arithmetic  Average 1975 - yg/m )                               Classification

            > HO                                                    Poor
          90-100                                                    Fair
            < 90                                                     Good
           118.4  > ^--                                           Average Poor
           100.0   ^.                                            Average Fair
            78.8                                                 Average Good
                                         115

-------
the independent effects of each, of the pollutants due to collinearity in the
data set.  Therefore, all calculations employ only one. pollutant measure as a
proxy for the general air quality situation.  Nitrogen dioxide is usually
that variable, however, resulta are also presented for TSP for purposes of
comparison.

     The final element of equation (5.1), the sale price differential
attributable to differences in ambient air quality (which can be inter-
preted in this context as the willingness to pay for improved air quality)
is determined by the particular approach.  The first method determines this
parameter in the simplest manner, through comparison of sale prices in the
paired communities, and relies almost exclusively on the sample plan.  The
second is intermediate in complexity and employs traditional property value
analysis to determine the monetary effect of air quality differentials.  The
third approach is the most involved, attempting to account for variation in
preferences in the determination of willingness to pay through statistical
means-.  In this manner the air quality impact on individuals is explicitly
specified.  These methods are addressed in detail in the following three
subsections.

     Before discussing in detail the empirical investigation and correspond-
ing results, a few notes pertaining to the theoretical underpinnings of the
analysis are in order.  First, the capitalization of environmental goods
into housing values can be captured through such empirical work only if
certain assumptions concerning the economic behavior of individuals and
the functioning of the housing market are accepted.  These are:  (1)  con-
sumers must perceive differences in housing and neighborhood characteristics,
expect them to remain unchanged and act on these perceptions; (2) housing
markets should function reasonably well and be in short run equilibrium;
(3) environmental quality must be exogenously determined and differences in
environmental quality must be capitalized only in housing prices; and  (4)
all relevant hedonic price functions should be continuous with first and
second derivatives that exist (e.g., there must be sufficient variation in.
both housing and neighborhood characteristics, including air quality, to.
permit, continuity).  Second, it should be noted that this housing market
analysis is consistent with and indeed a substudy within the general
theoretical treatment developed in detail above.

Paired Sample Methodology

     The paired sample approach is limited by the ability of the sample plan
to pair communities which are virtually identical in every respect including
air quality.  No explicit effort is made to account for differences in sale
price induced by other housing or neighborhood characteristics.  This is
admittedly a naive approach, yet it produces two positive outputs.   First,
because this methodology only implicitly compensates for many home and
community variables, the resulting benefit estimates can be considered an
upper bound on the population's willingness to pay for reduced air pollution.
Second, if these benefit numbers closely parallel those of the other, more
refined econometric methods can be considered successful.

     Table 5.5 presents average sale prices in each of the sample communities
and air quality regions„  These figures are standardized for house size with
                                    116

-------
                     Table  5.5




Sale Price Differentials Attributed  to Air Quality
             Paired  Sample Methodology
Community
Canoga Park
El Monte
Culver City
Montebello
Orange
Whittier
Enclno
La Canada
Redondo Beach
Huntington Beach
Pacific Palisades
Newport Beach
Palos Verdes
Irvine
Air
Quality
Fair
Poor
Fair
Poor
Fair
Poor
Fair
Poor
Good
Fair
Good
Fair
Good
Fair
Avg. Home
Sale Price -
1717 Sq. Ft.
Home
69,193
55,116
99,645
74,655
80,516
71,491
121,826
102,868
100,790
91,412
156,342
117,608
111,573
87,622
Sale Price
Differential
«)
25.5
33.5
12.6
18.4
10.3
32.9
27.3
Avg. Sale Price
Differential
Poor-Fair, Fair-
Good ($) (%)


19,371 (25%)


27,498 .(28%)

                            117

-------
each assigned the population average 1717 square  foot home.  The illustrated
sale  price differentials are calculated utilizing the community with the
poorer air quality as- the base.  Further, the figures indicate  that the value
associated with an improvement in air quality from poor to fair is approxi-
mately 25% of the average poor community home price  (for the same sized
home).  A similar upgrade in classification from  fair to good is valued at
about 28% of the value of the average fair region home.  Translated into
monetary terms these figures represent approximately $19,000 and $27,000
per home, respectively.

     As all other components were specified previously, these price figures
complete the information required to compute annual household benefits using
the paired sample approach.  These benefits are presented in Table 5.6.  As
illustrated, the capitalized benefits are in excess of $39 billion for an
approximate 25-30 percent reduction in urban air  pollution concentrations
(poor-fair, fair-good).  This translates into nearly $4 billion in annual
terms' evaluated at 9.95 percent capital recovery  factor, which  is equivalent
to how much individuals would be willing to pay for cleaner air in the form
of higher house payments.  In order to transform  these figures  into bid
equivalent terms, they are weighted by the total  number of affected homes
and the days in a year.  Thus, based on the paired sample methodology, each
household is willing to pay $4.50/day or $135 per month for the stated air
quality improvement.  Although it is expected that this method produces
high benefit estimates, the above figure seems a  reasonable amount when one
considers the variety of impacts (health, aesthetic, etc.) associated with
deteriorated air quality.

     Although these willingness to pay figures seem interesting and
reasonable, this methodology possesses a number of obvious shortcomings
which may negate their significance.  These can be classified as follows,
First,  this 'methodology attributes the entire differential in average sale
price to the variation in air quality.  This explicitly neglects a variety
of other possible differences which could account for the disparity in sale
price (although at least partial compensation for these factors is incor-
porated in the sample plan).   That  is,  this approachs at least to some
extent,  is using air quality to proxy for many relevant neighborhood and
community variables.   Isolation of  the independent influence of air quality
may not be complete.

     The second problem with this methodology is that each household,
regardless of its characteristics,  is assumed to place an identical value
on the  reduction in air pollution.   Thus no allowance is made for household
differences which would imply varying valuations.   In the next subsection
we employ traditional property value analysis and attempt to solve the first
category of shortcomings.   However,  the latter problem is not effectively
addressed until the following subsection.

Econometric Approach - Linear Equation

     The underlying structure of the econometric approach, is a single
equation empirical model which purports to  explain the variation in sale
prices  of homes located in the South Coast  Air Basin.   The basic operational
                                     118

-------
                                   Table  5. 6

                    Benefits - Paired Sample Methodology
Change In Air Quality
Capitalized Benefits
  (Billion Dollars)
Annualized Benefits
 (Billion Dollars)
R = .0925, CRF =  .0995
Poor to Fair
Fair to Good
       20.46
       19.34
       2.04
       1.92
Total
                                     39.8
                                    3.96
                              Capitalized Benefits
                                      ($)
                             Annualized Benefits ($)
                             R = .0925, CFJ- = .0995
Per Home
Per Home Per Day
Per Home Per Month
      21,385
       2.30
       4.50
     135.00
                                         119

-------
tool is ordinary least squares (OLS).  The procedure is to regress the full
independent variable set (see Table 5.1 for a complete description of these
variables and their expected relationship to housing price) on the vector of
home s-ale prices.  The result of this econometric analysis of housing
market data is a hedonic housing value eq-uation.  The estimated coefficients
of such an equation provide information on the relative significance and
value of each of the independent variables.   That is, the coefficients
specify the effect that a unit change in a particular independent variable
has on sale price.

     In reference to the air quality variable, this procedure allows one to
focus on its impact while separating out the influence of other extraneous
variables.  Therefore, this analysis- yields two outputs concerning the
relationship of air quality differentials to housing price.  First, the
relative significance of air quality variations is determined and second,
the estimated coefficient pertaining to air quality implicitly measures its
monetary value.

     The initial objective then is to estimate a linear hedonic housing
equation which best fits the data.   However, there exist a number of
empirical problems which could prevent efficient estimation of the desired
relationship.  For instance, two problems which generally arise in property
value studies are misspecification bias (.the independent variable set is
incorrectly specified) and multicollinearity (members of the independent
variable set are highly correlated).  Either of these may produce biased
estimates.  Furthermore, it is essential that these biases be avoided since
the estimated coefficients become the basis for the benefit calculations.

     Misspecification can be adequately countered by including in the equa-
tion all relevant independent variables without including variables which
have no a_ priori (on theoretical grounds) relationship with the dependent
variable.  The data set used in this study is relatively complete, in that
it contains a large number of housing and neighborhood characteristics.
Further discussion of this subject, however, is postponed until the next
section where experiements which demonstrate the effect of specification
error are performed.

     Multicollinearity is a common problem in studies of this nature.
This occurs since many of the independent variables are integrally linked.
and therefore possess extremely high correlation coefficients.  For instance,
with respect to housing characteristics, living area, number of rooms,
number of bedrooms, etc.; they are so interconnected (each representing size
of home) that least squares estimation techniques cannot determine the
independent impacts that these variables have on housing values.   Therefore,
living area was chosen as the proxy variable for house size.  Note that
home quality is measured by the inclusion of fireplaces, pools, and number
of bathrooms.

     Similarly, the air pollution variables showed a high degree of correla-
tion.  Again the estimation procedures were unsuccessful in separating out
the independent influence of each of the pollutants.  Thus only one pollution
variable, usually NO  was utilized as a proxy for the general state of air

                                     120

-------
variables and accessibility to beaches.  However, this collinear relation-
ship was effectively broken fay the structure of the sample pairings.  Thus,
the simple correlations between NO,,, TSP, and distance to beach do not ex-
ceed .37.

     Our concern about multicollinearity among the neighborhood and com-
munity variables was also warranted when it was found that housing density
and population density were so highly correlated (.96 simple correlation)
that they were essentially measuring the same phenomena.  The solution was
to allow only one of these variables in any equation.  These empirical
problems aside, we next proceed to discuss the estimated hedonic housing
value equations.  However, it should be re-emphasized that the estimation
was accomplished within the bounds of these empirical difficulties.

     The equations which provided the best statistical fit of the data are
presented in Table 5.7.  The relatively high R^'s (: .83) indicate that a
large proportion of the variation in housing price is explained by the
variation in the independent variables.  Except for two aberrations all
coefficients possess the expected sign and are statistically significant.
The exceptions are age, which is positive related to house value and
significant, and local tax rates which have the anticipated relationship
but are statistically insignificant.  The former may occur since age may
not be an adequate measure of housing condition since many older homes in
the Los Angeles are of high quality.  The insignificance of local tax rates
seems puzzling.  However, this is probably a result of the linear functional
form since in the next subsections we find that taxes become significant
when non-linearities are introduced.  Furthermore,  the age variable assumes
the proper relationship in the non-linear equations.

     Further examination of Table 5.7 gives added insight into the deter-
minants of house value.  The air pollution variables both perform as expected
and are highly significant.  In addition, the coefficients on the pollution
variables are quite similar (-316.89 for TSP and -260.4 for N02 when N02 is
converted to ug/nP units) signifying that each, as  a proxy for pollution,
has a similar impact on house price.  The stability of the coefficients on
the non-pollution variables (they are virtually identical) is also striking.
This result suggests that households are averse to  pollution generally
rather than to any single pollutant.

     The quantitative significance of a unit change in any of the independent
variables is obtained by examining the coefficient  values.  For instance,
an increase of 100 square feet of living area would cause a $2866.8 increase
in the house price.   Likewise, the coefficient on sale date shows that sale
prices were'increasing by nearly $l,900/month over  the study period.
Employing this same type of analysis, benefits from a reduction in either
N02 or TSP can be calculated.   Therefore, using N02 as the proxy, an im-
provement in air quality from poor to fair infers capitalized benefits of
$14,445/home while a change from fair to good is valued at $13,526/home.
As in the previous methodology,  these figures together with data previously
generated (number of affected homes, etc.)  become the basis for calculating
average annual benefits [see equation (5.1)].   These benefit computations
are completed in Table 5.8 for both N02 and TSP (in parentheses).

                                     121

-------
                                Table 5.7



                  Estimated Econometric Equations (Linear)*




              Dependent Variable = Home Sale Price in Dollars
Independent Variable
Sale Date
Age
Living Area
Bathrooms
Pool
Fireplaces
Distance to Beach
Distance to Employment
Crime
School Quality
Ethnic Composition
Housing Density
Tax
Public Safety Expenditure
TSP
N°2
Constant
R2
Sum of Squared Residuals
Degrees of Freedom
NO Equation
1897.8
(7.0041)
313.3
(2.8236)
28.665
(13.516)
21.856
(9.2552)
10213
(3.216)
14107
(7.1613)
-436.55
(-.19769)
-22597.
(-9.635)
-564090.
(-2.7727)
208.91
(2.7353)
4178.3
(2.7697)
-5.5248
(-1.9503)
-8.7207
(-.68288)
59.189
(6.7578)

-5104.3
(-4.8851)
-324820
(-2.2395)
.832
496900
703
TSP Equation
1944.6
(7.0946)
192.41
(1.7823)
28.558
(13.272)
22.378
(9.3117)
11375
(3.5566)
13187
(6.6648)
761.27
(-3.4148)
-18370
(-6.3776)
-674680
(-3.0476)
171.94
(2.1777)
7442.9
(5.5327)
-7.9192
(-2.6061)
-3.0441
(-.22507)
56.278
(5.6769)
-316.89
(-2.7845)
-
-652150
(-5.0917)
.828
508200
703
*t-statistics are in  parentheses.
                                         122

-------
                                  Table  5.8

                  Benefits r Linear Econometric Methodology
                                  NO   (TSP)
Change- in Air Quality
Capitalized Benefits
 (Billion Dollars)
  Annualized Benefits
   (Billion Dollars)
R = .0925, CRF = .0995
Poor to Fair
Fair to Good
  15.3 (6.2)
  10.9 (5.4)
       1.52 (.61)
       1.08 (.54)
Total
  26.2 (11.6)
       2.6  (1.15)
                            Capitalized Benefits
                                    ($)
                            Annualized Benefits  ($)
                           R = .0925, CRF = .0995
Per Home
Per Home Per Day
Per Home Per Month
  14077.  (6233)
      1401.  (620)
       3.84  (1.70)
     115.20.  (51-.0)
                                     123

-------
     The benefit figures dictate discussion on two counts.  First, the large
discrepancy between the N02 based and TSP based benefits is a result not of
the respective regression coefficients but rather from the fact that present
NO- concentrations are much higher and demonstrate greater variability than
those for TSP.  For instance, the average poor community has an ambient NC^
level of 242.7 yg/nr whereas the TSP concentration for a similar community
is 118.4 pg/m-\  Also, the gap between the quality indicators (poor-fair,
fair-poor) and therefore the required improvement is much greater for NC^
than TSP in percentage terms.  Thus, since TSP concentrations are both
lower and more ubiquitous than N0~ concentrations, the benefits on reducing
the latter are correspondingly higher.

     Second, the linear econometric methodology yields benefit estimates
which are somewhat lower (~ 15% for NO^ calculations, ~ 62% for TSP cal-
culations) than those presented for the paired sample approach.   This is an
expected occurrence since the linear econometric study explicitly accounts
for the variation in non-pollution variables through statistical means.
Therefore, this method can be considered an improvement over the previous
examination of mean housing values.

     However, this approach is not without its associated problems.
Specifically, there has been much discussion in the property value liter-
ature that benefits based on a linear equation coefficient tend to overstate
the true willingness to pay for air pollution reductions (see Section 2.1
and references 1, 4, 5, 10, 11).  That is, it is generally accepted that
the air pollution coefficient may be employed to value marginal change but
its applicability for total benefit calculations (non-marginal changes)
requires that further assumptions be made.  For example, the linear equation
method contains the implicit assumption that every reduction in air pollu-
tion is valued identically by all households.  This neglects variations in
average benefits which may accrue to particular population groups identified
by income or susceptibility to present pollution concentrations.  In effect,
household preferences are assumed to be identical.  This limits the
acceptability of the linear econometric approach.  In the next subsection we
further refine this approach and address remaining issues.
                                          I/
Econometric Approach - Non-linear Equation

     The non-linear methodology is a multi-stage procedure, the objective of
which is to determine the benefits of air pollution abatement while allowing
different values for various individuals.  In essence, this method addresses
the major criticisms of the previous approaches but must effectively assume
the mathematical form of individual preferences.  The first step involves
the estimation of a hedonic housing value equation.  This is similar to the
previous analysis except that in this case we do not arbitrarily restrict
the functional form to be linear.  Non-linearities are to be expected in
an analysis of housing market data because:  (1) the market may not be in
long-run equilibrium; (2) there may exist disequilibrium supply conditions;
or (3) there are indivisibilities among housing and neighborhood character-
istics.  Therefore, in this step an attempt is made to find the functional
form which provides the best statistical fit for the data.

                                     124

-------
     The second step is to determine the marginal willingness to pay for
small changes in the air pollution data.  This is done by taking the
derivative of the hedonic housing value equation (obtained from the initial
step) with respect to air pollution and evaluating for each of the four-
teen sample communities.  Calculated for each community, this derivation
yields the within community average household willingness to pay for marginal
improvements in air quality.  Determination of the marginal willingness to
pay is accomplished at the community level of aggregation based on the
assumption that the individual households within the community are com-
pletely homogeneous.

     In the third step the marginal willingness to pay  figures (obtained
in previous step) are regressed on a set of community characteristics
(income and present pollution level) in order to estimate a marginal
willingness to pay schedule.  The resulting estimated equation provides
information on how various communities identified by these characteristics
value reductions in air quality.  Thus, differentiation along community
preference can be accounted for.  For instance, it is a widely held belief
that marginal willingness to pay increases with income.  This hypothesis
is tested in this step.

     The final step employs this latter estimated relationship to determine
the home sale price differential attributable to the previously specified
air quality improvements.   Mathematical integration of the relevant mar-
ginal willingness to pay equation (a function of the stated community
characteristics) accomplishes this task.  This final information component
is then inserted into equation (5.1) to derive average household benefits
in bid comparable terms.

     The results of the hedonic housing value equation estimation are pre-
sented in Table 5.9.  As measured by R^, the non-linear functional form
performs somewhat better than the linear equation.   In the N02 equation all
independent variables conform to our a_ priori expectations concerning the
relationship to sale price and all except ethnic composition are statisti-
cally significant at the 5% level ( 11  _>_ 1.645).   A similar statement holds
for the TSP equation except that crime replaces ethnic composition as the
only insignificant variable.  In their respective equations, the air pol-
lution variables are highly significant.  Note also that squared pollution
terms were utilized in the estimation.   It was found that these performed
better than either the first-order or cubic terms.   However, the performance
difference was not significant.   Therefore, further analysis (benefit cal-
culations, etc.) based on the equations containing the first or third order
terms was completed and is discussed below.

     The non-linear specification prevents straightforward analysis of the
quantitative impact of a unit change in an independent variable since the
effect depends upon the level of all other variables.  However, if NC>2 and
the other variables are assigned these mean values  than a unit improvement
in N02 (one pphm) is valued at $2,010.

     Before proceeding to the next procedural step,  a few comments concerning
the effect of misspecification bias are in order.   That is,  we conducted
experiments to see what would happen to the coefficient on air pollution if

                                     125

-------
                      Table 5.9




          Estimated Econometric Equations*




Dependent Variable - Log (Home Sale Price in $1,000)
Independent Variable
Sale Date
Age
Living Area
Bathrooms
Pool
Fireplaces
Distance to Beach
Distance to Employment
Crime
School Quality
Ethnic Composition
Population Density
Log (Tax)
Public Safety Expenditures
(TSP)2
(NO )
Constant
R2
Sum of Squared Residuals
Degrees of Freedom
NO Equation
.018439
(10.108)
-.0027044
(-3.5185)
.00019976
(14.024)
.14777
(9.2661)
.089959
(4.2096)
.10355
(7.8325)
-.014037
(-9.1443)
-.26979
(-11.663)
-2.2798
(-2.3574)
.00099327
(2.0286)
.0081532
(1.2523)
-.000067145
(-7.8422)
-.030991
(-1.8253)
.00032792
(5.1487)
-.0010374
(-2.6935)
4.2297
(6.2304)
.877
22.62
703
TSP Equation
.018924
(10.427)
-.0031401
(-4.1178)
.00019688
(13.896)
.15285
(9.6443)
.092764
(4.389)
.099225"
(7.5833)
-.013132
(-9.1824)
-.23201
(-9.1314)
-1.5245
(-1.5444)
.0010087
(2.0792)
.027307
(4.5564)
-.000061627
(-7.2705)
-.046438
(-2.7565)
.00028288
(4.8582)
-.000015702
(-4.1798)
—
2.3602
(3.8836)
.878
22.29
703
                            126

-------
certain neighborhood variables were omitted from 'the equation.  For example,
if distance to beach is excluded then  the air pollution coefficient
increases from .0010374 to .0034176.   Similarly, if population density is
omitted then the pollution coefficient increases to .0024284.  In each of
these cases the air pollution term serves as a measure of pollution and
other neighborhood disamenities as well.  These specification errors would
eventually result in biased benefit estimates.  Therefore, a fully specified
equation is crucial.

     The estimated equations shown in  Table 5.9 yield the marginal
willingness to pay for improvements in air quality by taking the derivative
with respect to the relevant air pollution variable.  This procedure
supplies information on the amount of money the average household in each
community would be willing to pay for  small changes in pollution levels.
This information, in conjunction with  community average income and pollution
levels, are the basic inputs to the third methodological step - estimation
of the willingness to pay equation.  Table 5.10 presents two formulations
of this equation for N02-  The first assumes a linear relationship while
the second postulates a log-log form.  As is indicated by the coefficients
both income and pollution are positively related to marginal willingness to
pay.  Thus, higher income communities  in poor air quality regions have the
greatest willingness to pay.   Similar results were discovered for the TSP
based equations but they are not presented.

     Given this analysis it then becomes possible to complete the multi-
step procedure and calculate:  (1)  the average sale price differential
attributable to changes in air quality; and (2) benefits derivable from
these changes in per home, per day units, through use of equation (5.1).
The first calculation is accomplished by integrating the xjillingness to
pay equations (assigning the income variable its mean value) over the
range of air quality improvement._/  In this manner, the reduction in
pollution consistent with the poor to fair improvement is valued at $5,793/
home for the linear N(>2 willingness to pay equation and $6,134/home for the
log-log NC>2 equation.  The values which correspond to the fair-good change
are $4,244/home and $4,468/home, respectively.  If TSP is used as the mea-
surement criteria then poor-fair is valued at $6,053/home (linear) and
$6,033/home (log-log) while fair-good is valued at $5,677/home (linear)
and $5,964/home (log-log).

     The above figures are translated into average benefits illustrated in
Table 5.11 through application of equation (5.1).  As can be seen from
examination of Table 5.11, daily household benefits calculated using the
multi-step procedure range from $1.40/day/home to $1.48/day/home or $42.00
and $44.40 per month, respectively for N0?.  These are considered our
"best" estimates since the technique used as their specification at least
addresses known methodological problems.   We correspondingly place the
most faith in them.   Further, the TSP based calculations remain fairly
constant at about $1.60/day/home,  so the daily household willingness £0 pay
to achieve the specified air quality improvements are relatively insensitive
to the pollutant used in the willingness to pay equation.   The TSP results
are also insensitive to the specification of the hedonic housing equation,
the first link in this methodology.  That is,  whether the first or third
order TSP term was used in this equation (rather than the squared term)
                                     127

-------
                                  Table 5.10

                Estimated Willingness to Pay Equations (NO )*

         Dependent Variable = Marginal Willingness to Pay in Dollars
Independent Variable                  Coefficient                  t-statistic

Constant                               -1601.3                      -2.7622
Income**                                .050051                      8.2662
NO  level                               162.67                       3.7832


 2
R  - .864
Degrees of Freedom = 11
       Dependent Variable = Log (Marginal Willingness to Pay in Dollars)
Independent Variable                  Coefficient                  t-statistic


Constant                               -6.4845                      -5.7025
Log (Income**)                          1.1473                       13.092
Log (N00)                               .87283                       6.1051


R  » .942
Degrees of Freedom = 11
     *These equations are based on the hedonic houring value equation which
utilizes (NO )2 as the air pollution measure.

    **The income variable is defined as average community income and in
dollars.
                                         128

-------
                                  Table  5.11

               Benefits - tyulti-Step Econometric Methodology*

             (A) NO  (TSP) - Linear Willingness to Pay Equation
Change in Air Quality
Poor to Fair
Fair to Poor
Total
Capitalized Benefits
(Billion Dollars)
6.12 (6. A)
3.42 (4.6)
9.56 (11.0)
Annualized Benefits
(Billion Dollars)
R = .0925, CRF = .0995
.61 (.637)
.34 (.458)
.95 (1.095)


Per Home
Per Home Per Day
Per Home Per Month
Capitalized Benefits
(?)
5136 (5910)
Annualized Benefits (?)
R - .0925, CRF = .0995
511 (588)
1.40 (1.61)
42.00 (48.30)


(B) N02 (TSP)
Change in Air Quality
Poor to Fair
Fair to Poor
Total
- Log-Log Willingness
Capitalized Benefits
(Billion Dollars)
6.5 (6.4)
3.6 (4.7)
10.1 (11.1)
to Pay Equation
Annualized Benefits
(Billion Dollars)
R = .0925, CRF = .0995
.645 (.64)
.355 (.47)
1.0 (1.1)


Per Home
Per Home Per Day
Per Home Per Month
Capitalized Benefits
($)
5427 (5964)
Annualized Benefits ($)
R = .0925, CRF = .0995
540 (593)
1.48 (1.63)
44.40 (48.90)
     *Note that in the estimated hedonic housing equation (step 1)  the
second order pollution terms were used.
                                          129

-------
had little effect on the eventual benefit calculations.  However, this was
not the case for NC^-  In this instance, daily household benefits fluctuated
from a low of $. 87/day/home or $26.10 per-month I(NC>2)  used in hedonic
housing equation and linear willingness to pay equation] to a high of $2.09/
day/home or $62.70 per month (first order NO,., term used in housing equation
and linear willingness to pay equation).

     In comparing these figures to the simpler property value approaches,
we again find an adjustment downward as the methodology becomes more re-
fined.  This is consistent with our conjecture that the paired sample
approach would yield upper bound benefits.  This result also provides
further support for the hypothesis that the linear econometric approach
overestimates the total willingness to pay for pollution reductions.  This
overestimation can be partially corrected by employing the final procedure
posited here.

     In conclusion, we have attempted to describe and utilize a multi-step
approach to the determination of air pollution abatement benefits.  Each
of the steps is linked to those that preceed it.  Therefore, benefit
calculations are a function of a hedonic housing value equation, the
resulting marginal willingness to pay data, and an estimated willingness
to pay schedule which yields the sale price differential attributable to
air quality.  Finally, our "best" estimates of daily household benefits
was $1.40/day/home calculated using the second order NO  term in the hedonic
housing equation and a linear willingness to pay equation.  However,
benefits could easily range from $.87/day/home to $2.09/day/home.

5.4  Summary

     This paper began with the premise that valuation of non-market com-
modities constitutes a socially desirable objective on efficiency and
equity grounds.  However, no methodology, which is generally accepted,
exists to accomplish this goal.  Therefore, any new experimental valuation
technique requires validation.   The analysis undertaken here is an attempt
to satisfy this requirement for the contingent valuation approach.

     This study can be viewed as a systematic investigation of housing
market data within the communities which comprise the sample plan.  It
consists of three separate approaches.  The first, the paired sample
methodology, is primarily based on the sample plan.  In this procedure we
attempted to determine the benefits derivable from air quality improvements
through a comparison of sale price averages in the paired communities.
This approach was found to be beset with a number of problems, yet the
upper bound of $4.50/home/day for a poor-fair, fair-good improvement was
determined.

     The second approach,, a linear econometric methodology, was an attempt
to utilize traditional property value analysis to develop benefit estimates.
The ordinary least squares regression technique was the basic tool used to
estimate a linear hedonic housing value equation.   The benefit calculations
derived from this equation were considered an improvement over the paired
sample approach since explicit account was made for a number of housing

                                     130

-------
and neighborhood variables.  Thus, this analysis provided more refined
benefit figures but they were still considered an overestimate since no
allowance was made for varying valuations of air quality changes dependent
on houshold characteristics.

     In the final approach, multi-step econometric methodology, we addressed
the criticisms which plagued the earlier approaches and developed more
refined benefit estimates.  Our best estimate of willingness to pay for the
specified air quality change (about a 30% reduction in average ambient levels)
was approximately $1.40/day/home ($42.00 per month).  This amount is based
on a hedonic housing equation which allows non-linearities [including using
(NOo)  as a proxy for air pollution] and either a linear or a log-log
willingness to pay equation.  However, this figure is not precise and
therefore we put the possible range of benefits at between $.87/day/home
($26.10 per month) and $2.09/day/home ($62.70 month).
                                    131

-------
                            FOOTNOTES:  CHAPTER V

     ~~ This analysis follow., closely the procedure developed by Harrison
and Rubinfeld (1978).

     2/
     — The formula used in these calculations is:

J            Pollution before
                             (WTP.)d   Pollution
            Pollution after

          where WTP  = f(income, pollution).
                   i
                                     132

-------
                                 Chapter VI

                  PRELIMINARY COMPARISONS BETWEEN PROPERTY
                    VALUES AND ITERATIVE BIDDING RESULTS

     The South Coast Air Basin experiment consisted of an attempt to value
air quality through examination of differences in property values and
through an interview survey instrument to measure willingness to pay.  Six
pairs of neighborhoods were selected for comparative purposes.  The pairings
were made on the basis of similarities of housing characteristics, socio-
economic factors, distance to beach and services, average temperature, and
subjective indicators of the "quality" of housing.  Thus, for each of the
pairs, an attempt was made to exclude effects on property values other than
differences in air quality.

     While the sample paired methodology was an attempt to establish com-
parability betxv>een results of the research designs, certain cautions should
be kept in mind.  These additional assumptions are that:

     1.   an implicit hypothesis exists such that there is a directional
         consistency between the types of biases of the two research
         designs;

     2.   in a theoretical sense, each research design is measuring the
         same "good;"

     3.   the groups being sampled are identical within the paired areas;

     4.   the time frames from which the valuation estimates are derived
         are assumed constant (i.e., equilibrium versus non-equilibrium
         contexts for individuals and markets); and

     5.   a problem exists in assigning proper weighting for a set of
         diverse samples.

     With these difficult qualifications in mind, let us turn to a pre-
liminary comparison of results obtained from the property value and sample
survey results.  Table 6.1 provides some extremely preliminary results on
monthly valuations by households of an arbitrary improvement in air quality
in the Los Angeles Basin of approximately 30%.  For the paired comparisons
property value study, the estimate per household with no adjustments for
household differences except in an areal and subjective sense (see Chapter
III), is approximately $135 per month.   Extrapolated^the results to the basin
as a whole yields an annual benefit from an improved  air quality improvement
of 30%,  a value of approximately $4 billion dollars.
                                    133

-------
                                   Table  6.1

             Alternative  Estimates of Monthly  Bids  by  Household,
                 Total Benefits  for Air  Quality  Improvement
                        in  the South Coast Air Basin

             (Approximate 30% Improvement in Ambient Air Quality)




Average ($) bid per house-
hold per month
Annual benefits (selected
areas and groups of the
South Coast Air Basin)
in billions of $'s)
Property Value Study

Paired
Communities

$135
$3.96




Linear
Regression

$51-115
$1.25-2.6



Non-
Linear
3-Step

$42*
$.95



Survey Study

Mean.
Bid

$29**
$.65



Preliminary
Regression
Results

$26***
$.58



     *Best estimate, possible range, $26-63 per month.

    **Based on maximum total bid with an adjustment for years to achieve
improvements in air quality.

   ***Based on maximum total bid equation wich an adjustment for the amount
of air pollution information available to the household.
                                        134

-------
     The other extreme is an estimate of, -its value of improved air quality
per month by household utilizing the preliminary results in Appendix D
from the survey.  This value is approximately $26 per month per household.
This yields a rough estimate of annual benefits from an approximate 30%
improvement in air quality of slightly more than $.5 billion dollars.

     Further, intermediate estimates are calculated on the basis of various
economic assumptions delineated in Chapters II and V.  By making various
assumptions with regard to the change of air quality in the Los Angeles
Basin, other estimates of improvements can be derived.  For example, if it
can be presumed that the various areal groups, when bidding from a
reasonably poor air quality to a reasonably good air quality, were bidding
on the basis that their area would be totally cleaned up, an alternative
estimate of the mean bid on an annual benefits basis is $1.07 billion
dollars.  This is comparable with the linear estimate derived from the pro-
perty value study.  For the reasons given earlier, these researchers believe
that, depending on assumptions, a range of willingness to pay for both
studies anywhere from a low of approximately $20-30 to a high of approxi-
mately $140-150 per month per household is obtained.

     It appears from these preliminary results and comparisons that con-
tingent valuation studies will tend to give a lower valuation of air quality
improvement than observing at the margin what happens in an extremely
volatile property market.  However, only after substantial in-depth
statistical examination and comparability checks between the two studies
will the researchers be able to state unequivocably how these valuations
may turn out.  The results compiled in this study suggest that survey
instruments, when compared to property value techniques, provide a rea-
sonable mechanism to obtain environmental quality benefit estimates.  The
survey approach has the advantages that:  (1) data can be collected at
low cost on specific environmental problems (the investigator is not tied
to the availability of existing data sets); (2) benefit measures can be
disaggregated across individuals and sources of benefits from various
characteristics such as aesthetic experiences and perceived health can be
obtained; and (3) a voluntary consumer statement of willingness to pay
gives some justification in and of itself for expenditures on air quality
and perhaps more generally on environmental quality programs.

     As a final caution,  it should be kept in mind that the South Coast Air
Basin studies were conducted in an area where individuals have both an
exceptionally well-defined pollution situation that they have encountered
and a well-developed hedonic price-property value market for clean air.
The effect of clean air on property values, and in turn, on the degree to
which people are aware of increased housing prices in high air quality
areas appears to be exceptionally well specified at this time in the
South Coast Air Basin.   Note further that 1970 property values on the basis
of several studies have shown a much weaker association with air quality
than those that were obtained utilizing the 1977-78 air quality data set
applied here.  We feel that this change reflects a substantial shift in
tastes and concern over air quality for this regional population.   Therefore,
it should be recognized that the results of this experiment may well not
be generalizable to other situations where the environmental commodity,

                                    135

-------
i.e., air quality, is not so well specified, either through actual market
prices or human perception.
                                     136

-------
                                 APPENDIX A
     This appendix presents the bas-ic survey instrument employed in the
South Coast Air Basin study.  As discussed, the initiation point for the
survey instrument was either acute health effects or aesthetic effects.
Three basic areas existed (good, fair, Bad),  Thus the following combinations
existed for survey instrument types.

     1.
     2.
     3.
     4.
     5.
     6.
     7.
Format
Format
Format
Format
Format
Format
Format
Format
for
for
for
for
for
for
for
for
A
A
A
A
B
B
C
C
area
area
area
area
area
area
area
area
moving
moving
moving
moving
moving
moving
to C*
to C*
to B" (Aesthetic)
to B (Acute)
to C (Aesthetic)
to C (Acute)
to C (Aesthetic)
to C (Acute)
(Aesthetic)
(Acute)
     The structure of the different combinations was identical.
1 is presented for illustrative purposes.
Combination
                                     137

-------
       Table  A.I
INDOOR ACTIVITY AND COST LIST
Activity
Indoor Spectator Events
Indoor Tennis
Raquetball, Handball
Table Te-nnis
Bowling
Indoor Gardening or
Fixing up House
General Exercise
Organized Sports Events
Reading
Television
Movies
Club Activities,
Organizations
Individual Sports
Swinnint?
Visiting Neighbors or
Friends
Other (specify)
V
















Hours
Per Week
A i B
































C
















D
















Tines
Per Week
A
















Bi C
1






























D









Location
(Hap Grid)
A ' B
1

|_






)


















\









C I D




|



















Miles
Traveled
A P , C i D
i








1






I
' 1 !

































1














Direct
Coses
A

i






B








1





i




C













D












I Day
Equipment
Replacement
Coses
\











I
1
1 :









i











Importance

















-------
      Table  A.2
OUTDOOR ACTIVITY AND COST LIST
Activity
Outdoor Spectator
Snorts
Tenni s
Biking
Be.sch Activities
General Exercise
Fishing
Sw i m.T. i n £
Sail in"
Jocelnr./'.Val'Kinc
Hobbies. Arts & Crjfts
Outdoor C.-rdening or
Fixing up House.'
Golf
Hiking
N/









Hours
Per Week
A













Canping
Organized Sports Events '
Individual Sports
Events

Other (spccliv)






I
B




C




D




l ;












'




















1
Times
Per Week
A

















B

















C


	

D





|
I




















Location
(Mao Grid)
A

















B

















C

















D









Miles
Traveled
t.
























B

















C












D












1








Direct
CostE
A

















B

















C
















D
















i
% Day

















Equipment
Replacement
Costs

















Importance


















-------
 GAME FOR A AREA MOVING TO R, AESTHETIC                   Questionnaire //	
                                                          Interviewer   //	

 INTRODUCTION

 1.  In a typical week how much day and night leisure time do you have available?
     This includes both weekdays and weekends.  By leisure, I mean the time you
     do not spend eating, sleeping, or working to earn a living.  	 hours

 2.  Has air pollution influenced where you have chosen to live?  Yes[ ]  No[ ]

 3.  Has air pollution influenced where you have chosen to work?  Yes[ ]  No[ ]

 4.  Would you classify the air quality in the area where you live as:
     Good[ ] Fair[ ]  Poor[ ]

 5.  Would you classify the air quality in the area where you work as:
     Good[ ] Fair[ ]  PoorI ]

5a.  What is your occupation?	
 6.  Are you aware of any health hazards due to air pollution?  Yes[ ] No[ ]

 7.  How long have you lived in the Los Angeles area? 	___	

 8.  How much longer do you plan to live in the los Angeles area?	
 9.  Do you think automobile emission standards should be:  Increased[ ]
     Decreased[ ] Kept the Same[ ]

 ADMINISTER INDOOR AND OUTDOOR ACTIVITY AND COST LISTS (TYPICAL WEEK) CHECK
 THE TOTAL TIME CONSTRAINT

 Bidding Game for Residents of Area A

     Here are three photographs representing average levels of visibility
     for the three different regions of the Los Angeles Area shown on this
     map.  Picture A represents poor visibility; Picture B represents fair
     visibility; and Picture C represents good visibility.

     Public officials are strongly considering the possibility of trying to
     reduce the levels of emissions throughout the Los Angeles Area.  Such
     action could require additional funds which might be generated by (a
     monthly charge, an extra charge in your utility bill) for as long as
     you live in the Los Angeles area.  These funds will be used to help
     finance air quality improvements in the Los Angeles area.
                                      140

-------
Aesthetic

     The Los Angeles area has some very "Beautiful background scenery.
     Because of automobile and industrial emissions', there is a haze which
     reduces and distorts the ability to see this scenery.  This means that
     many people have to leave Los Angeles and travel long distances to be
     able to enjoy views which could be visible from their homes if these
     emissions were reduced.

     As indicated by the map, you live in an area which has been classified
     as having poor air quality relative to the rest of the Los Angeles
     area.  Picture A represents the visibility level which typically occurs
     in your area.  I am only interested in how you value being able to see
     long distances.

     If the level of emissions could be reduced in the Los Angeles area so
     that visibility conditions would be represented by Picture B instead of
     A, not only in the B area but also in your area, and if the air would
     be cleaned up to this level in (2, 10) years, would you pay (a monthly
     charge, an extra charge in your utility bill) of ($1, $10, $50) for as
     long as you live in the Los Angeles area?

[ ] DO FOLLOWING ONLY IF CHECKED

LIFE TABLE:  Here is a table that might help you.  It shows the total amount
you would pay for as long as you live in Los Angeles for various amounts of
monthly payments.

RECORD MAXIMUM BID
     Would you consider moving to a new location in the los Angeles area
     if air quality were like Picture C everywhere?  Yes[ ] No[ ]

IF YES:

     Where would you move?  (GRID LOCATION ON MAP)	
                                    141

-------
Bidding - Aesthetics and Acute Health

     The questions I asked earlier were only concerned with, your perception
     of visibility.  This section deals with, not only visibility hut with
     short term health effects which may be aggravated by air pollution.
     Some pollutants in high concentrations cause eye irritation for many
     individuals.  Studies have shown that about half of the population
     experiences eye irritation under conditions represented by Picture A;
     about one-fourth experiences- eve irritation under conditions represented
     by Picture B'; while no one experiences- these irritating effects when
     conditions are represented by Picture C.

     Since vou reside in area A, which has Keen classified as having poor
     air quality, there is reduced visibility as well as irritating health
     effects as compared with B'.  If the level of emissions could be reduced
    'in the Los Angeles area so that visibility and irritating health effects
     were represented By Picture B not only in the B area but also in your
     area, and if the air would be cleaned up to this level in (2, 10) years,
     would you pay a (a monthly charge, an extra charge in your utility bill)
     of (START BIDDING WITH PREVIOUS MAXIMUM BID) for as long as you live
     in the Los Angeles area?

[ ] DO FOLLOWING ONLY IF CHECKED

Life Table:  Here is a table that might help you.  It shows the total amount
you would pay for as long as you live in Los Angeles for various amounts of
monthly payments.

RECORD MAXIMUM BID
REVISIONS IF NECESSARY
     Would you consider moving to a new location in the Los Angeles area
     if air quality were like Picture C everywhere?  Yes[ ] No [' ]

IF YES:

     Where would you move?  (GRID LOCATION ON MAP)	
                                     142

-------
 Substitutions

 Aesthetic; Original Position A; Movement to B


 NON-ZERO BIDS:

      You stated that you would pay $	 per month for as long as you live
      in the Los Angeles area if visibility improved from A to a condition
      like that shown in B, if this could be accomplished in  ('2, 10) years.
      If you actually paid this amount in order to help finance air pollution
      control programs, you would have less money to spend overall.  However,
      because you said you would pay some amount of money, you are indicating
      that clearer air is something you value.  Consequently, when conditions
      like B are achieved, even though you have paid money to help improve
      the visibility, this does not mean that your standard of living is
      worse than before because now you will be living and recreating in a.
      less polluted area.

      If the visibility conditions were to improve from A to B in your area,
      would the improved conditions change the pattern of your leisure
      activities?  This could be changes in time per week, location, or
      frequency.  Yes[ ] No[ ]

 IF NO, GO TO NEXT BIDDING GAME
 IF YES, THEN:

 A.  Administer indoor and outdoor activity and cost list
 B.  Check time Constraint	

 ZERO BIDS:

      Although you told me that you would not pay anything to have visibility
      improve throughout the Los Angeles area to like that shown in Picture B,
      would the improved conditions change the pattern of your leisure
      activities?  This could be changes in time per week, location, or
      frequency.  Yes[ ] No[ ]

IF NO, GO TO NEXT BIDDING GAME
IF YES,.THEN:

A.  Administer indoor and outdoor activity and cost lists
B.  Check time constraint
                                      143

-------
Substitutions

Aesthetic + Acute; Original Position A; Movement to R

NON-ZERO BIDS:

     With the extra information on possible short term health, effects when
     conditions are like A, you said that you would pay $      per month
     for as long as you lived in tne Los Angeles area if conditions im-
     proved from those associated with Picture A to conditions shown in
     Picture B, and if this could be accomplished in (2, 10) years.  As
     before, Because you said you would pay some amount of money, you are
     indicating that clearer air is something you value.  Consequently, when
     conditions like A are achieved, even though you have paid mqnay to help
     improve the visibility and to lessen short term health effects, this
     does not mean that your standard of living is worse than before because
     now you will be living and recreating in a less polluted area.

     If conditions- improved so that Picture B were representative of the
     entire area, with no visibility problems or irritating effects, would
     the improved conditions change the pattern of your leisure activities?
     Yes[ ] No[ ]

IF NO, GO TO NEXT BIDDING GAME
IF YES, THEN:

A.  Administer indoor and outdoor activity and cost lists
B.  Check time constraint	

ZERO BIDS:

     Although you told me you would not pay anything to have visibility
     conditions and short term health effects improve throughout the area
     to like those shown in Picture B, would the improved conditions change
     the pattern of your leisure activities?  Yes[ ] No[ ]

IF NO, GO TO NEXT BIDDING GAME
IF YES, THEN:

A.  Administer indoor and outdoor activity and cost lists
B.  Check time constraint
                                     144

-------
Substitutions

Aesthetic + Acute + Chronic; Original Position A; Movement  to  K

NON-ZERO BIDS:

     Given the information that continued exposure  to levels of air pol-
     lution like those shown In Picture A could actually reduce your  life
     expectancy, you said you would pay $	 per  month for as long  as you
     lived in the Los- Angeles- area if conditions improved from those  in A  to
     those shown in IS, and if this- could lie accomplished in (2, 10) years.
     Once again, I would like you to think of this  expenditure as  leaving
     you as well off as before you paid the money,  since you are now  living
     and recreating in a less- polluted area.

     If conditions- improved so that Picture B were  representative  of  the
     entire area, with no visibility problems- or short and  long term  health
     effects-, would the improved conditions change  the pattern of  your lei-
     sure activities  Yes[ ] No[ ]

IF NO, PROCEED TO GENERAL INFORMATION SECTION
IF YES, THEN:

A.   Administer indoor and outdoor activity and cost lists
B.   Check time constraint	

ZERO BIDS:

     Although you told me that you would not pay anything to have  visibility
     or short and long term health effects improve  throughout  the  Los
     Angeles area to like those shown in Picture B, would the  improved con-
     ditions change the pattern of your leisure activities?  Yes[  ] No[ ]

IF NO, GO TO GENERAL INFORMATION SECTION
IF YES, THEN:

A.   Administer indoor and outdoor activity and cost lists
B.   Check time constraint

-------
Bidding - Aesthetic, Acute Health, and Chronic Health Effects

     The quality of the air may also affect your long term health.  There is
     evidence that high concentrations of emissions as- represented in
     Pictures A and B have lasting effects- upon the respiratory and circul-
     atory systems in addition to eye irritation and reduced visibility.
     Evidence shows- that, on the average, people who live in areas with con-
     centrations- like those in Picture A can expect a reduced lifespan of up
     to 2 years compared with people who-live in conditions represented by
     B, and up to 3 years when compared with people who live in conditions
     represented by Picture C.

     If the level of emissions- could Be reduced in the Los Angeles area so
     that visibility, short and long term health conditions would be repre-
     sented By Picture B instead of A, not only in the 5 area but also in
     your area, and if the air would be cleaned up to this level in...("2, 10.)
     years, would you pay (a monthly charge, an extra charge in your utility
     bill) of (START BIDDING WITH PREVIOUS BID) for as long as you live in
     the Los Angeles area?

[ ]  DO FOLLOWING ONLY IF CHECKED

Life Table:  Here is a table that might help you.  It shows the total amount
you would pay for as long as you live in Los Angeles for various amounts of
monthly payments-.

RECORD BID
REVISION IF NECESSARY
     Is there some other payment scheme besides (a monthly charge, an extra
     charge in your utility bill) that you would prefer?  Yes[ ] No[ ]

IF YES:

     What would it be?
     Would you consider moving to a new location in the Los Angeles area
     if air quality were like Picture C everywhere?  Yes[ ] No[ ]

IF YES:

     Where would you move?  (GRID LOCATION ON MAP)	
                                     146

-------
 GENERAL INFORMATION SHEET:  WOULD YOU PLEASE FILL OUT THE FOLLOWING?

 1.   Age	

 2.   Sex:  MaleJ  J  FemaleJ ]

 3.   Marital status:  Single! ]  MarriedJ  .]

 4.   Number of persons- In your household?	
 5.  Your education:	years.  Highest degree obtained:
     High School[ 1  College! ]  Vocational! ]  Advanced! ]

 6.  Address of employment:	
 7.  Location of employer(s) (GRID LOCATION ON MAP)
 8.   Are there any environmental hazards associated with your job, such as
     noise, health, or sight?  Yes! •]  No! 1    IF YES:   What are these hazards?


 9.   What is the percentage of your work time indoors?       %

10.   In a typical work week how much time do you spend on the job?	hours

11.   If you received our pamphlet last week, did you read:

     [  ]  0-5 pages
     [  ]  5-10 pages
     [  ]  more than 10 pages

If you live in area A or B:

     How much would you pay for this same house (apartment) today if it
     were located in an area where the air pollution levels were like those
     shown in Picture C?  $	

If you live in area C:

     Do you believe that any part of the value of your home is because you
     live in a relatively unpolluted part of Los Angeles?  Yes[ ] No[ ]
     IF YES:  How much (% or dollars)  	

If you live in area B or C:

     Would you consider moving if the air pollution problem were as bad as
     A throughout the entire area?  Yes[ ] No! ]
     IF YES:  Where would you most likely move?  ('GRID LOCATION ON MAP)
                                     147-

-------
How much do you think it would cost to clean up air pollution in the Los.
Angeles area to a. condition like that shown in Picture C?
$	

If all citizens- were billed equally, how much do you think, it would cost
each person in order to achieve conditions like that shown in Picture C?
$
                                    148

-------
1.  Characteristics of. home:

     Living Area:	 square feet

     Number of Rooms:	

     Number of Bedrooms
     Number of Bathrooms:
     Other Rooms:  ("PLEASE CHECK)

     [ ] Den
     [ ] Family room
     [ ] Dining room
     [ ] Enclosed porch
     [ ] Attic
     [ ] Basement
          % Easement finished
     Tl"Utility room
     [ ] Other

     Scenic View:  YesI ]  No [ ]     IF YES:  Specify	

     Number of Stories:   (INCLUDE BASEMENT) 	

     Remodeled:  Yes[ ] No[ ]  Don't know[ ]

     IE YES:  Specify previous style and date	

2.   Equipment:   (PLEASE CHECK)

     [ ] Dishwasher
     .[..] Disposal
     [ ] Central Air Conditioning
     [ ] Trash Compactor
     [ ] Central Heating

     Pool:   Yes[ ] No[ ]

     IF YES:  Circle whether heated, enclosed, or other (specify)

     Fireplace:  Yes[ ] No[ ]

     Age of home: 	years (when constructed)
                                     149

-------
  IF YOU LIVE IN AN APARTMENT, GO TO QUESTION 4

  3.   a).  Year of purchase:  	

       b)  Could you please indicate what the purchase price was:  $_

       c)  What are your monthly payments:  $	

       d)  What do you feel your home is worth, in today's- market?  $_

       e)  What are your property tax payments- per year?  $	
       f)  How long have you been living in this house? 	 years

  GO TO QUESTION 5

  4.   a)  How long have you been living in this apartment:  	 years

       b)  Would you indicate your monthly rent?  $	

  5.   What are your insurance payments per year?  $	
  6.  What do you pay monthly for general upkeep around your home (apartment)?
  7.  Why have you chosen to live in this area?  RANK IN ORDER OF IMPORTANCE,
      WHERE ONE IS MOST IMPORTANT.  CHOOSE TOP FIVE.

      [ ] Attractiveness of area in general
      [ ] Close to work
      [ ] Close to recreation activities
      [ ] Close to friends
      [ ] Close to schools
      [ ] Close to services
      [ ] Close to transportation routes
      [ ] Air quality
      [ ] Affordability of home
      [ ] Low crime rate
      [ ] Prestige of area
      [ ] Quiet neighborhood
      [ ] Other

 8.   What are your average expenditures per month for food:  $	
 9.  What are your average expenditures per year for clothing?  $
10.  Please mark the box corresponding to your annual household income.
      I ]  0-$5,000                      I ]  $3Q,QOO-$35,aOO
      [ ]  $5,000- $10,000               I ]  $35,000-$40,000
      [ ]  $10,000-$15,000               [ ]  $40,000-$50,000
      [ ]  $15,000-$20,000               I ]  $50,000-$60,000
      [ ]  $20,000-$25,000               [ ]  $60,000-$80,000
      [ ]  $25,000-$3n,000               I ]  Over $80,000

                                       150:

-------
1.  Do you own or share in the. ownership of a motor vehicle?  Yesf J No[ J

    Type of Vehicle(s)            	Model 	Year

                                  	   Model 	 Year

                                            Model        Year
2.  How many licensed drivers are in your family?
3.  How many miles- per gallon do you get for each car in the city?

    	mpg
            mpg
    How many miles do you and your family typically travel in your automo-
    bile per week?  	miles

4.  How many hours do you and your family spend in a typical week commuting
    to:

    Work or school                	 hours

    Shopping                      	 hours

    Recreational activities       	 hours

5.  Do you participate in a car pool?  Yes [ ] No [ ]

6.  How much do you spend each month on:

    a)  Gasoline costs  $	

    b)  Maintenance costs  $
    c)  Public transportation fares  $	

    d)  Insurance payments  $	

7.  Did you.take a vacation within the last year where you were away from
    home for  more than 4 days?  Yes[  ]  No[  ]

    IF YES:   About how much were your expenditures on this trip?  $	
                                     151

-------
1.   Have you ever had any of the following?  (PLEASE CIRCLE.)

     a)  High blood pressure          a)  Asthma
     b.)  Heart trouble                f)  Chronic nervous trouble
     c)  Strokfc                       g)  Cancer
     d)  Chronic bronchitis           h)  Tuberculosis

2.  Have you ever had trouble with the following?  (PLEASE CIRCLE)

     i)  Pain in the heart or tightness or heaviness in the chest
     j)  Trouble breathing or shortness- of Breath
     k)  Frequent headaches
     1)  Constant coughing or frequent heave chest colds
     m)  Frequent eye irritations
     n)  Allergies
     .o)  Nose and throat irritation

3.  Are any of these (the above) conditions aggravated (or made worse)  by
    heavy air pollution?  Yes[ ] No[ ]

    IF YES:  Which ones?  (LIST LETTERS CIRCLED)
4.  Do .you suffer from any other diesases which could be made worse by poor
    air quality?  Yes[ ]  No[ ]  Specify	

5.  Do you or any member of your family have any physical disabilities which
    limit your activities?  Yes[ ]  No[ ]

6.  Would conditions like those in Picture C, if they occurred' over the
    entire area?

    a)  Make your life more
        pleasant                      Not at all[ ]  To some degree! ]
                                      Greatly[ ]

    b)  Require you to spend
        less money on drug
        items or doctor's             Not at all[ ]  To some degreef ]
        fees                          Greatly[ ]

    c)  Make it easier to             Not at all[ ]  To some degree[ ]
        do your work?                 Greatly[ ]

7.  Do you enjoy doing your leisure activities more during the ^ay or  during
    the night?  Day    Night    Makes no difference

8.  Do you smoke?  Yes I ] No[ J  IF YES:  How many packs per day?	
9.  Do you take medication regularly?  Yes[ ]  NoJ ]
    IF YES:  Monthly expense on this medication?  $	/month
                                     152

-------
10.  How much do you spend monthly on medical problems associated with, air
     pollution effects?  $	/month.

11.  How much, do you spend yearly for doctor's fees?  $	/year

12.  How much, do you spend yearly on medical and life insurance.?
     $_J	/year

13.  Have you purchased any items to reduce your exposure to air pollution
     (such as carbon filters)?  Yes[ ] Not ]
     IF YES:  What items?
                                    153

-------
1.   Which, one, of theae statements applies to you?  (CHECK. ONE)

     [ ] I have not Been bothered by air pollution.
     [ J I have been somewhat bothered by air pollution.
     [ ] I have been bothered quite a lot by air pollution.

2.   Do you believe that air pollution in Los Angeles:   (CHECK ONE)

     [ ] Has become worse since you have lived here.
     [ ] Has stayed about the same since you have lived here.
     I ] Has gotten better since you have lived here.

3.   What do you think should be done about air pollution?  (CHECK ONE)

     [ ] Ignored
     [ ] Reduced

4.   Please rank the following problems in terms of importance (most to
     least) as the major issues facing the community.   (CHOOSE TOP FIVE)

     [ ] Juvenile delinquency         [ ] Nuclear energy
     [ ] Communicable disease         [ ] Alcoholism
     [ ] Unemployment                 [ ] Water pollution
     [ ] Air pollution                [ ] Energy
     [ ] Car accidents                [ ] Congestion
     [ ] Crime                        [ ] Other

5.   Do you believe that air pollution in the Los Angeles area:

     [ ] Cannot be reduced below its present level
     [ ] Can be reduced below Its present level
     [ ] Can be almost completely eliminated

6.   What do you think the words "air pollution" mean  to most  people in the
     Los Angeles area?  Do they mean:

     a)  Frequent bad smells in the air           YesI  J  Nol ]
     b)  Too much dirt and dust in the air        YesI  ]  Nol ]
     c)  Frequent haze or fog in the air          Yes[  ]  Nol 1
     d)  Frequent irritation of the eyes          YesI  ]  No[ ]
     e)  Frequent nose or throat irritation       YesI  ]  Nol ]
     f)  Other                                    YesI  J

7.   Have you read or seen anything in the newspaper recently  about air
     pollution?  YesI ] No! ]

8.   When you read the newspaper, do you generally choose to read articles
     on air pollution?  YesI ]  Nol ]
                                      154

-------
9.   Do you consult the daily air pollution index before engaging in any
     activities?  Yes.f ]  No [  J

     IF YES:   What kind of  activities?
                                    155

-------
1.   If you received our health, pamphlet, what do you think about it?
2.
     [ ] Did not read
     [ ] Very informative
     [ ] Hard to understand
     [ ] Scientific mumbo-jumbo
                                  [ ] Made me more concerned ab.out
                                     health", effects
                                 1 ~] Had no influence on me
Here is a list of words and phrases.  Select two which describe how you
feel about your participation in this survey.

[ ] Stimulating
[ ] Just tolerable
[ ] A waste of time
[ ] Educational
[ j Boring
[ ] An invasion of privacy
'[ ] Interesting
[ ] Kind of fun
[ ] Hard to take seriously
     Here is a different list of words and phrases.
     how you feel about the questionnaire.
                                                Select two which describe
     [ ]  Relevant
     [ ]  Credible
     [ ]  Likely to influence air quality control
     [ ]  Unrealistic
     [ ]  Pretty  flakey
     [ ]  Unlikely to have any effect on air quality control
     [ ]  Irrelevant

     Finally, here is another list of words and phrases.  Select one from
     each column to describe how you feel about your answers to the question-
     naire.
     Column 1

     [ ]  Quite accurate

     [ ]  There was no way I could
         come up with accurate
         answers.

     [ ]  Accurate in a "ball park"
         kind of way.

THANK YOU FOR YOUR COOPERATION.
                                             Column 2

                                             [ ]  A fairly good guide for
                                                 valuing air quality

                                             [ ]  A good guide for valuing
                                                 air quality.

                                             [ ]  A poor guide for valuing
                                                 air quality.
                                     156

-------
                                APPENDIX B

     The following represents  the  health pamphlet  that  was  sent  to  half of
the respondents who were contacted  by  phone  and  agreed  to participate  in
the study.
                                     157

-------
                          AIR POLLUTION AND HEALTH

     This pamphlet will try to answer some questions about air pollution
and human health.  How do the major pollutants affect the body?  What is
known scientifically about these effects?  What kinds of real life studies
have been carried out to test facts learned in the laboratory?  This
infgrmation is provided so that you can draw your own conclusions about
the health effects of air pollution.

     Nearly every day in the Los Angeles area a chemistry of air and sun-
light gives rise to toxic gases known as photochemical oxidants.  These,
together with carbon -monoxide, sulphur dioxide, nitrogen oxide, hydrocar-
bons, aldehydes, and ketones, make up the haze, account for its aroma, and
may impair human health.

     Environmental standards like those in Table B.I
attempt to protect the public.  When the concentration of any pollutant
exceeds the standard, acute, short term, irritating symptoms may be noticed.
These acute effects, such as chest tightness, eye irritation, slowing of
response time, and attention loss, are not experienced by everyone, but
people with pre-existing heart condition and lung disease are particularly
vulnerable.

     In addition to acute health effects, chronic effects of long term
exposure to low and average levels of the oxides, aerosols, particulates,
and other elements of the haze are a particularly challenging question.
Does air pollution cause influenze sometimes, or does it merely make it
more of a problem?  Links between oxides of nitrogen and cancer have been
investigated.  Finally, it is possible that years of continuing exposure
could have some influence on total lifespan.
                                     158

-------
                            The Major Pollutants

Carbon Monoxide

     The major sources of carbon monoxide pollution are automobiles,  trucks,
buses, and,  to the habitual smoker,  cigarettes.  Peak hourly readings of
carbon monoxide  from 1963 to  1970 averaged  10.8 ppm* in the Los Angeles
area, and  exposure in  die California Central Valley was about half  this
figure for  the same period.  [ll]

     Carbon monoxide has a very direct effect on the human body.  Entering
the lungs,  it diffuses into the blood, where it is absorbed by red  blood
cells and displaces and competes with oxygen.  Carbon monoxide reduces the
oxygen carrying  capacity of the blood.  Low concentrations cause tiredness
and listlessness.

     The heart is doubly affected.   Its oxygen supply is reduced, but at the
same time it must exert more  effort  to increase its output if body  oxygen
transport is to  be maintained.  And  so, at relatively low concentrations of
carbon monoxide  (10-50 ppra for one hour exposure), [191 patients with heart
disease may experience adverse effects.

     During heavy muscular exercise, the oxygen consumption rate of the
body increases to as much as  20 times the rest rate.   Consequently, carbon
monoxide exposure reduces maximum exercise performance.

     In controlled experiments with  humans, researchers have projected the
maximum levels and exposure times shown in Table B.2.  Normal healthy
individuals are  unlikely to experience any of the above effects until the
threshold concentration is 21-72.5 ppm. [19] Individuals with emphysema,
bronchitis, and  asthma are more sensitive, perhaps experiencing effects at
17.5-52.5 ppm, and heart patients are extremely sensitive to carbon
monoxide, as noted above. [19]  All  these effects are acut^e, occurring at
high concentrations.   The effects of exposure over long periods to  low
carbon monoxide  levels are not known at this time.

Sulphur Dioxide

     Los Angeles has not had  a deadly pollution episode such as those
observed in Belgium;  Donora,  Pennsylvania; London;  or New York, but an
ingredient of Los Angeles air pollution,  sulphur dioxide,  is held respon-
sible for high death tolls in these places.

     A well known air pollution episode occurred December 1-5,  1930 when
several hundred  persons became ill in the Meuse Valley, Belgium.   There
were 63 deaths.  It was estimated that sulphur dioxide and sulphuric acid,
which may have reached a level of 9 ppm,  were the chief causes of illness.[30]
     *ppm denotes the number of pounds of pollutant for each million pounds
of air.
                                     159

-------
During late October, 1945, Donora, Pennsylvania, was blanketed by dense fog.
Forty-three percent of the population was affected.  Twenty persons died.
Ten percent of the residents were severely affected.  Again, sulphur
dioxide was held to be responsible.
     In London, England during December, 1952, the world's worst air pol-
lution incident occurred, causing about 4,000 more deaths than would be
expected in the Greater London Area  for a month's period.  Marked increases
in deaths both from lung and heart disease were observed.  Detailed invesgi-
gations of 1,280 post-mortem reports of persons who had died before, during,
or shortly after the episode indicated all such fatalities could be explained
by previous health problems among the victims.  The elderly and persons with
already existing lung and heart disease were most susceptible.  During this
time in London, daily sulphur dioxide and smoke measurements were from two
to four times higher than typical winter levels.  [30]

    'Sulphur dioxide, as is well known, has an odor.  It is readily soluble
in water and, when breathed, is absorbed quickly  in the upper airx^ays of the
nose.   In Table B.3 are recorded laboratory observations of throat and lung
effects from sulphur oxides.  In air with small dust particles, sulphur
dioxide is partially converted into  sulphuric acid, which may be a severe
problem of its own.  The Los Angeles area currently has relatively low
levels of sulphur dioxide.

Photochemical Oxidants

     Along with carbon monoxide, gasoline engines produce nitric oxide*
and hydrocarbons.  Secondary products of these emissions, photochemical
oxidants - ozone, nitrogen  dioxide,  and peroxyacetylnitrate (PAN), may be
more toxic than the original compounds.

     Early morning car traffic produces exhaust with large quantities of
nitric oxide and hydrocarbons.   In  the presence of  sunlight these products
react, converting nitric oxide into  nitrogen dioxide but low nitric oxide
levels.  Then  nitrogen dioxide breaks down  into ozone during  the afternoon.
Late afternoon automobile  traffic again emits large amounts of nitric
oxide, which reacts with ozone,  removing most of  the ozone.

     Ozone is  among  the most poisonous of gases.  Relatively  insoluble in
water, when  inhaled, ozone  can damage the central airways and other pas-
sages of the lung.

     Health  studies  of certain occupations  have provided understanding of
the effects  of exposure  to  oxidants. A 51  year old welder who was working
in a poorly  ventilated area, developed  a kind of  pneumonia which lasted
for six days.l9j  A  crane  operator  working  above  a  tank  into  which ozone
was bubbled  developed a  dry cough  and frontal headache after  two hours
      *Nitric  oxide is also a byproduct of natural  gas  combustion and the
 processing  of nitric acid industrially.
                                     160

-------
                                 Table B.I

                       Pollution Levels and Standards
                            (Parts Per Million)**
National
Standard for
one hour
exposure

Average peak
hourly level
in Lennox,
1973-75

Average peak
hourly level
in Costa Mesa
Harbor,
1973-75

Average peak
hourly level
in Pasadena,
1973-75
                     Ozone
0.03
CL04
0.05
0.11
               Carbon
              Monoxide
40.0
12.1
9.53
                Nitrogen
                Dioxide
0.99
                 0.07
0.13
               Sulphur
               Dioxide
                0.50
0.06
                0.03
0,03
Source:  Three Year Summary of Califorjaia_A_ij__r^
         Air Analysis Branch and EDP I'anacement Section  (January 1977),
         State of California Air Resources Board.

     *State of California hourly standard.

    ""'Parts per million denotes the number of pounds of  pollutant found
in a million pounds of air.
                                      161

-------
                                  Table B.2

                                                  o
                           Carbon Monoxide Effects
                      (National Standard:  40 parts per
                          million/one hour exposure)
Concentration ppm
Exposure Time
Effects
  Acute Effects

         50
         53
        100
        500
      1,000
  2 hours
  1.5 hours
  0.5 to 2
  hours
  1 hour
Shortened average time to
a heart attack among
individuals with heart
disease^

Shortened average time to
a heart attack among
individuals with heart
disease.

Loss of physical and
mental coordination among
healthy subjects.

Mild to throbbing headache
among healthy subjects.0

Vomiting, unconsciousness
and death among healthy
subjects.
Sources:    Leung,  Goldstein and Dalkey,  Final Report:	Human Health Damages
          from Mobile Source Air Pollution,  197^,  California Air Resources
          Board.

           W.S. Aronow and H.W. Isbell,  "Carbon Monoxide Effect on
          Exercise-Induced Angina Pectoris," Annals of Internal Medicine 79
          (1973),  392-395.

           J.  Koch-Weser,  "Common Poisons,"  in Harrison (ed.),  Principles of
          Internal Medicine, Ch. 166 (1970), 652-653.
                                     162

-------
                                  Table B.3

                           Sulphur Dioxide Effects
                      (National Standard:  .5 parts per
                         million/one hour exposure)
Concentration ppm          Exposure Time          Effects
  Acute Effects

          1                  0.5 hour             Choking sensation in some
                                                  individuals.

          5                  3-10                 About 80% of healthy
                             minutes              individuals will have
                                                  difficulty in breathing.

       5-10                    —                 Deep gasping feeling,
                                                  severe choking in some
                                                  individuals.
Source:   National Pollution Control Administration, Air Quality for
         .Sulphur Dioxides,  1969,  U.S.  Department of Health, Education
         and Welfare,  Public Health Service,  Consumer Protection and
         Environmental Health Service.
                                    163

-------
exposure.  In attempting to leave the crane, the employee nearly lost
consciousness.  After administration of oxygen, he improved, and 48 hours
later showed no adverse symptoms. [1 I]

     Other studies have confirmed these effects.  A long term study over
the period October, 1961, to June, 1964, recorded daily symptoms in student
nurses in good health from two Los Angeles nursing schools.  The nurses
kept diaries for 868 days on appearance of cough, chest discomfort, and
headaches.  For the same period hourly peak concentrations of photochemical
oxidants, carbon monoxide, and daily temperature were measured at stations
within two miles of the schools.

     Cough and chest discomfort increased with higher hourly concentrations
of ozone.  Headaches had some association with ozone levels but less than
other symptoms.  Eye discomfort, not a direct effect of ozone, although
often associated wtih photochemical oxidants, was the most strongly noted
symptom.   When the oxidant level reached 0.5 ppm, a third of the nurses
reported eye irritation.  Temperature, carbon monoxide, and nitrogen
dioxide levels did not explain  the results found.  Because all participants
were young, healthy adults, relatively free from chronic disease, the
effects on elderly persons or on those with chronic heart or lung disease
could be expected to be more severe. [22]

     During the period 1963-1970 peak hourly readings of oxidants averaged
.104 ppm in the Los Angeles area, and again readings in the California
Central Valley were about half this figure. [19]

     According to a panel of experts,

     "the oxidant threshold in normal individuals ranges from 0.05 to
     0.20 ppm.  The threshold concentration is lowered among young
     and old individuals, and also those with underlying disease.  Those
     with respiratory and chronic obstructive diseases are most sen-
     sitive to the photochemical oxidants, and the threshold levels
     for these population groups range from 0 to 0.20 ppm." [19]

     In addition to discomfort and aggravation of existing lung disorders,
ozone and other photochemicals can cause changes in behavior.   Automobile
accidents, for example, were recorded in each daylight hour of each weekday
in the "high smog" months of August through November for two years.  A
relationship between Los Angeles oxidant concentrations and the number of
car accidents was found. [ 22 ] Attention span and visual performance were
reduced.   Lethargy is reported as well as difficulty concentrating.  0.8 ]

     Because lung function is impaired, evidence suggests photochemical
smog increases individual vulnerability to acute throat or lung infections.
Studies with experimental animal populations have reported changes in the
makeup and working of the lung as well as lung-tumor acceleration.  (25l
Whether ozone is a cancer causing agent is currently an important topic
for research.  Human white blood cells exposed to ozone exhibited chromosome
breakage and genetic abnormality. [12] In Table. B.4 a summary of ozone
effects is given.
                                      164

-------
                                 Table B.4

                Summary of Fxperimental Data on Ozone Effectsc
                     (National Standard:   .OB parts per
                         million/one hour exposure)
Concentration ppn
             Exposure Time
Effects
  Acute Effects

    0.15-0.30


    0-37-0.70
         0.25



         0.5



     0.8-1.7

     1.0-2.0
                2 hours
    Sources:
"ye irritation due to some
photochemical products.

Cough, nose and upper
throat irritation, chest
soreness, chest tightness,
symptoms made worse by
exercise, headache in 50%
of normal subjects.

Less than 6% of asthmatics
may have attacks "hen this
level is reached.

Formation of fluid in the
lungs among healthy
subjects.

Lung congestion.

Incapacitating illness
anonp, normal subjects.
 Leung, Goldstein and Dalkey, Final Report: _Human Health Damage_s
from '"obile Source_Air PollutionTT^lT California Air
Resources Board..

 G.E. Schoettlin and E. Landau, "Air Pollution and Asthmatic
Attacks in the LA Area,'' Public Health Report 76 (1961),
545-543.
                                     165

-------
       Nitrogen dioxide, a by-product of auto emissions, has effects similar
  to ozone but at higher levels of concentration.   Originally,  exposure to
  nitrogen dioxide was known as "silo-filler's syndrome," since extremely
  high condentrations of nitric oxide and nitrogen dioxide are generated
  within farm silos.   Documented deaths from a kind of pneumonia and acute
  throat and lung ailments were traced to this type of exposure.

       Like ozone, nitrogen dioxide's low water solubility allows it to
  penetrate deeply into the lung,  where it damages tissue.  At  low concen-
  trations, it impairs breathing.   At higher levels it increases the risk
  of an individual having a throat or lung ailment.   At 25-100  ppm,  it causes
  acute (but quickly  remedied)  symptoms of pneumonia and bronchitis.

       A study of the environmental health effects of nitrogen  dioxide was
  conducted in four residential areas, each containing three elementary
  schools, in greater Chattonooga, Tennessee.  One  area, close to a large
  TNT-plant (which processes nitric acid), had high nitrogen dioxide and low
  particulate exposure.   Another area had high suspended particulate and low
  nitrogen dioxide concentrations.  The other  two  areas were "clean" and
  used for comparisons.   Careful monitoring of particulate matter,  nitrates,
  sulphates,  and gaseous nitrogen dioxide concentrations was conducted in
  1968 and 1969 in these four areas.

       Two possible health effects of nitrogen dioxide exposure were investi-
  gated:   (1)  difficulty breathing in elementary school, children; and (2)
  increased frequence of respiratory  illness in family groups.   It was
  established  that second grade school children in the high nitrogen dioxide
  area were consistently higher than  those in  the  two control areas  during
  the  study period.   However, the  researchers  could  not establish a  relation-
  ship between chronic bronchitis  and the levels of  nitrogen dioxide.  [26]

     Tn the period 1963-1970, nitrogen dioxide levels in the los Angeles
Area averaged  .28 ppm and therefore constituted a  potential health hazard
given the estimates of effects  in Table B.5

     The air pollution health problem of the Los Angeles Area is far more
complex than brief accounts of  the hazards of carbon monoxide,  sulphur dioxide,
and two photochemical oxidants  can indicate.  For  one thing, literally hun-
dreds of hydorcarbon  compounds  are present in the  Los Angeles air, each with
its own characteristics and products.  A group of  secondary organic aerosols
may be responsible for adverse  health effects and  undoubtedly contribute to
visibility loss.  Little, however, is known of the mechanism and health impact
of these compounds.

     However, there is evidence that  the pollutants discussed have health
effects at levels experienced within  Los Angeles and other urban areas.
These pollutants cause irritation and stress within the lungs and heart.

     Acute effects range from eye, nose, and throat irritation, headache,
chest tightness, difficulty in  breathing to the aggravation of bronchitis,
asthma, emphysema, other lung ailments and heart disease.  A relationship
between episodes of sulphur dioxide and particulate pollution and increased

                                     166

-------
                                   Table B.5

                           Nitrogen Dioxide  Effectsa
                      (National  Standard:   .05  parts per
                         million/one  hour exposure)
Concentration ppm
Exposure Time
Effects
  Acute Effects

  0.7-2.0
    4-5
    4-5
    6-40
  greater than
  25

  150-200

  greater than
  200
 10 minutes
 10 minutes
                            1 hour
 5 minutes
 1  hour or
 less
Difficulty expelling air
from the lungs  increased
by 15% and difficulty
breathing air into the
lungs increased by 50% in
normal subjects.

Half hour after exposure,
difficulty breathing,
increased by 117 to 92%.

Decrease in oxygen in the
blood.

Increased difficulty
breathing by 24% in
normal subjects.

Bronchiolitis and pneunonitis
in normal subjects.
                           c
Disintegration of the lung.

Lunp, fluid formation and
death.
Sources;   Leung.,  Goldstein and Dalkey, Final Report:	Human Health Damages
          from Fobile Source Air Pojlution, 1975, California Air resources
          Board.

           D.V. Bates, "Air Pollution and the Hunan Lunp,,': American Review of
          Respiratory Disorders. 105 (1972), 1-13.

          °H.E. Stokinger and D.L. Cottin, "Biologic Effects of Air Pollutants,1'
          in A.C.  Stern,  ed., Air Pollution and Its Effects, Ch. .13,
          New Yorlr:   Ararlcmin Press£l958~)~j 445-546.       '
                                     167

-------
death rates is accepted.   The chronic effects of photochemical oxidants
lower general resistance  to infections of the respiratory tract and lung
since they cause damage to the lung.  Behavioral changes associated with
carbon monoxide, ozone, and nitrogen dioxide have heen documented.   Activity
levels are depressed and  overall work ability is impaired through visual and
chemical intervention.

    These air pollutants  may also shorten the lifespan by aggravating
existing health problems, particularly those problems associated with the
respiratory tract.
                                     168

-------
                                  BIBLIOGRAPHY

 1.. Abe, Matsuo, "Effects of Mixed NO -SO  Gas on Human Pulmonary Functions,"
       Bulletin, Tokyo Med. Dent. UniversiEy 14 (1967), pp. 415-433.

 2.  Anderson, E.W., R.J. Andelman, J.M. Strauch, N.J. Fortain, and J.H.
       Knelson, "Effects of Low-Level Carbon Monoxide Exposure on Onset and
       Duration of Angina Pectoris:  A Study in Ten Patients with Ischemic
       Heart Disease," Ann. Intern. Med. 79, 46, 50 (1973).

 3.  Aronow, W.S. and M.W. Isbell, "Carbon Monoxide Effect on Exercise-Induced
       Angina Pectoris," Ann. Int. Med. 79 (1973), pp. 392-395.

 4.  Bates, D.V. , "Air Pollution and the Human Lung," American Review qf_
       Respiratory Disorders 105 (1972), pp. 1-13.

 5.  Bates, D.V., G.M. Bell, C.D. Burnham, M. Hazucka, J. Mautha, L.D. Penselly,
       and F. Silverman, "Short-Term Effects of Ozone on the Lung," Journal
       of Applied Physics 32 (1972), pp. 176-181.

 6.  Beard, R.R. and G.A. Wertheim, "Behavioral Impairment Associated with
       Small Doses of Carbon Monoxide," American Journal of Public Health 57
       (1967), pp. 2052-2022.

 7.  "Behavioral Toxicology Looks at Air Pollutants," interview with C.
       Xintaras, Environ. Sci.  Tech. 2 (1968),  pp. 731-733.

 8.  California Department of Public Health,  Clean Air for California, Initial
       Report of the Air Pollution Study Project,  Berkeley (March 1955, March
       1956, and February 1957).

 9.  Challen, P.J.R., D.E. Hickish, and J. Bedford,  "Investigation of Some
       Health Hazards in Inert-Gas, Tungsten Arc Welding Shops," British
       Journal Ind. Med. 15 (1958), pp. 276-282.

10.  Committee on Medical and Biologic Effects  of Environmental Pollutants,
       Division of Medical Science, Assembly of Life Sciences,  National
       Research Council, Ozone and Other Photochemical Oxidants, National
       Academy of Sciences, Washington, D.C.  (1977).

11.  Coordinating Committee on Air Quality Studies,  National Academy of
       Sciences, National Academy of Engineering,  Air Quality and Automobile
       Emission Control, Vol. 3, The Relationship of Emissions  to Ambient Air
       Quality, U.S. Government Printing Office, Washington,  D.C. (1974).

12.  Fetner, R.H., "Ozone-Induced Chromosome Breakage in Human  Cell Cultures,"
       Nature 194 (1962), pp. 793-794.

13.  Haagen-Smit,  A.J. and L.G. Wayne, "Atmospheric Reactions and Scavenging
       Processes,  Ch. 6," in A. C. Stenn,  ed.,  Air Pollution and Its Effects,
       New York:  Academic Press (1968), pp. 149-186.

                                      169

-------
14.  Hammer, D.I., V.  HasselbJLad, 11. Portnor, and P. Wehrle, "Los. Angeles
      Student Nurs.e Study, Daily Symptom Reporting and Photochemical Oxidants,"
      AEH 28 (1974), pp."255-260.

15.  Hazacha, M. ,  F. Silverman, C. Parent, S'. Field, and D.V. Bates, "Pul-
      monary Function  in Man After Short Term Exposure to Ozone," Arch,
      Envinson, Health 27  (1973), pp. 183-185.

16.  Kelly, F.J. and W.E.  Fill, "Ozone Poisoning," AEH 10 (1965), pp. 517-519.

17.  Koch-Weser, J., "Common Poisons," in Harrison (ed.), Principles of
      Internal Medicine, Ch. 166, 6th Edition (1970), pp. 652-653.

18.  Lagerwertt, J.J., "Prolonged Ozone Inhalation and Its Effects on Visual
      Parameters," Aerospace Med. 34 (1963), pp. 479-486.

19. , Leung, Steve, Elliot  Goldstein, and Norman Dalkey, Human Health Damages
      from Mobile  Source Air Pollution:  Final Report, California Air
      Resources Board  (March 1975).

20.  Lowry, T. and L.M.  Schumann, "Silo-Fillers Disease Syndrome Caused by
      Nitrogen Dioxide," JAMA 162 (1956), pp. 153-160.

21.  Mert, T.H., H.A.  Bender, H.D. Kerr, and T.J. Kulle, "OBservations of
      Abberations  in Chromosomes of Lymphocytes from Human Subjects Exposed
      to Ozone at  a Concentration of 0.5 ppm for 6 to 10 Hours," Mutat. Res.
      31  (1975), pp. 299-302.

22.  Proceedings of the  Conference on Health Effects of Air Pollutants,
      Assembly of  Life Sciences, National Academy of Science, National
      Research Council,  Oct. 3-5.  Prepared for the Committee on Public
      Works, U.S.  Senate (1973).

23.  Ramirez, R.J. and A.  R. Dovell, "Silo-Fillers Disease:   Nitrogen Dioxide
      Induced Lung Injury, Long Term Follow-up and Review of the Literature,"
      Ann. Int. Med. 74  (1971), pp. 569-576.

24.  Ruffin, J.B., "Functional Testing for Behavioral Toxicity:   A Missing
      Dimension in Experimental Environmental Toxicology," J. Occup. Med.
      5 (1963), pp. 117-121.

25.  Schoettlin, C.E.  and  E. Landau, "Air Pollution and Asthmatic Attacks in
      the LA Area," Public Health Report 76 (1961), pp. 545-548.

26.  Shy, C.J., J.P. Creason, M.E. Pearlraan, K.E. McClain, F.B.  Benson, and
      M.M. Young,  "The Chattanooga School Study:  Effects of Community
      Exposure to  Nitrogen Dioxide, II, Incidence of Acute Respiratory
      Illness," Journal  of the Air Pollution Control Association 20(9)
      (September 1970),  pp.  582-588.

27.  Shy, C.J., J.P. Creason, M.E. Pearlman, "The Chattanooga School Children
      Study-II, Incidence  of Acute Respiratory Illness," Journal of the Air
      Pollution Control  Association 20 (1970), pp. 582-588.

                                     170

-------
28.  Stokinger, H.E. and D.L. Cot tin, "Biologic Effects of Air Pollutants,"
      in Air Pollution and Its Effects, A.C. Stern (ed.)> New York:  Academic
      Press, Ch. 13 (1968), "pp. 445-546.

29.  U.S. Department of Health, Education and Welfare, Public Health Service,
      Consumer Protection and Environmental Health Service, Air Quality
      Criteria for Sulphur Oxides (January 1969).

30.  U.S. Department of Health, Education and Welfare, Pulbic Health Service,
      Consumer Protection and Environmental Health Service, Air Quality
      Criteria for Sulphur Oxides (January 1969).

31.  Ury, U.K., "Photochemical Air Pollution and Automobile Accidents in Los
      Angeles:  An Investigation of Oxidant and Accidents, 1963 and 1965,"
      AEH 17 (1968), pp. 334-342.

32.  Von Nieding, G.,  "Studies of the Acute Effects of NO  on Lung Function,
      Influence of Differsion, Pertusion, and Ventilation"in the Lungs," Int.
      Trde. Arbeits Med. 31 (1973),  pp. 61-72.

33.  Young, W.A., D.B. Shaw and D.V. Rates, "Pulmonary Function in Welders
      Exposed to Ozone," AEH 7 (1973), pp. 337-340.

-------
                                APPENDIX C

     The following tables have Che following underlying assumptions.

     1.  The mean bids within each area are differentiated with respect to
the sequence that the air quality effects are presented, i.e., whether those
effects are introduced in "Aesthetic -> Acute Health -»• Chronic Health" or
"Acute Health ->• Chronic Health -*• Aesthetic" order.  In the graphs, "Ae"
denotes the mean bid for aesthetic effects; "Ac" denotes the mean bid for
acute health effects; and "Ch" denotes the mean bid for chronic health
effects.

     2.  The mean bids within each "A" area are differentiated with respect
to the range of the hypothetical improvement, i.e., whether the improvement
is from A to B or from A to C.  Since there is only one range of improvement
for the "B" and the "C" areas, i.e., from B only to C and from C only to C*,
no such differentiation is made for these areas.

     Taking into consideration the variations in (1), each A area requires
four different graphs, and each B and C area requires two different graphs.

     A denotes poor air quality

     B denotes fair air quality

     C denotes good air quality

     3.  The mean bids within each area are differentiated with respect to
the proposed completion date of cleanup; i.e., 2 years versus 10 years.

     4.  The bids from each respondent are obtained as follows:  First, his
maximum bid is elicited following a certain hypothetical improvement in the
aesthetic (acute health) effects of air quality.  Second, he is asked how
much he would increase his bid if the acute health (chronic health) effects
are also taken into consideration.  Finally, he is asked to revise his bid
for the additional inclusion of the chronic health (aesthetic) effects.

     The implicit assumption throughout this procedure is the linear addi-
tivity of bids for each effect.

     No differentiation has been made whether a health pamphlet has or has
not been sent to the respondent in advance of the interview.

     No differentiation has been made with respect to the different proposed
vehicles for the collection of bids.

     No differentiation has been made with respect to the different starting
bids offered by the interviewer.

     No differentiation has been made whether a life table has or has not
been shown to the respondent during the interview.  A life table depicts
the "stock" counterparts of the elicited monthly bids for various expected
lifespans.
                                     172

-------
                                                           Table C.I

                                                     Mean Bids by Area by Type*
                                                 (Completion Date of Cleanup:  2 yrs.)


Area**

El Monce
(A - B)

El Monte
(A - C)

La Canada
(A •» B)

La Canada
(A - C)


Koncebello
(A -. B)

Moncebello
(A * C)

Canoga Park
(B ~ C)

Mean Aesthetic Bid
($/month)
Typo I

2.00
(0.95)***
(5)****
13.40
(8.15)
(5)
13.20
(9.37)
(3)
22.60
(13.70)
(5)

1.60
(0:93)
(5)
26.43
(20.69)
(7)
10.00
(5.77)
(3)
Tvpe II

1.00
(1.00)
(5)
0.13
(0.13)
(2)
0.00
(0.00)
(2)
0.00
(0.00)
(5)

3.75
(2.39)
(4)
4.67
(2.60)
(3)
1.20
(0.971
C5).
Mean Acute Health Bid
($/~.onch)
Type I 1 Tvnc II

3.00
(1.55)
(5)
1.40
(0.93)
(3)
1.20
(0.97)
(5)
2.40
(1.12)
(5)

0.20
(0.20)
(5)
3.57
(1-71)
(7)
10.17
(7-58)
(3)

9.20
(4.76)
(5)
7.50
(2.50)
(2)
1.50
(1-50)
(2)
a. oo
(3.74)
(5)
Mean Chronic Health Bid
($/iaonth)
TV£C_I
Tvpe II
Mean Total Bid
(S/!so:ith)
Tv.ic I
1
2.00
(1.22)
(5)
3.60
(2.91)
(5)
2.20
(1.96)
(5)
1.00
(1.00)
(51
j
12.50
(4.79)
(•'•)
10.75
(5.45)
(4)
15.40
(4.71)
0.20
(0.20)
(5)
7.86
(7.06)
(7)
1.67
(1.67)
(5) : (3)
0.40
(0.40)
(5)
2.75
(2.25)
(2)
0.00
(0.00)
(2)
21 .00
(10. 30)
(5)

2.50
(2.50)
(4)
0.25
(0.25)
(4)
3.80
(2. 34)
(5)
7 . DO
(3.00)
(5)
IS. 40
01.75)
(3)
16.60
(9.03)
(5)
26.00
(12.69)
(5)

2.00
(0.84)
(5)
37.56
(27.21)
(7)
21.83
(14.14)
(3)
Tv;>o 11

10.60
(5.45)
(5)
10.33
(4.63)
	 (2)
1.50
(1.50)
<:)
29.00
(9.34)
(5)

18. 75
(6.57)
(4)
11.00
(i.OO)
(3)
20.40
(5.89)
(5)
(continued)

-------
                                                            Table  C.I
                                                            (cone i uued)
Area
Culver City

-------
                                                                Table  C.1

                                                                (continued)
Area
Palos Verdes
(C -* C*)
Redondo Bcarh
(C •* C*)
Mean Aesthetic Bid
($/month)
Ty»c I
/4.31
(2.10)
CO
6.29
(3.39)
(7)
Tyoe II
0.50
(0.50)
(!)
,.29
(4.29)
(7)
Mean Acute Health Bid
($/monthj_
Tvpe I
2.19
(1.29)
•'..86
(2.13)
(7)
Type II
17.75
(10.96)
15.29
(7.85)
(7)
Mean Chr(
Tvpe I
1. 75
O-l-fi)
CO
0. 00
(0.00)
(7)
ronic Heal th Hid



Type II
0.50
(0.50)
CO
', . •'. 3
(2.9-'0
(7)
Mean Total Bid
($/r,onth)
Ty :> e I
8.25
C-.il)
11 .14
(5.07)
(7)
TV we II
1 S . 7 5
(10. '.S)
C-)
(11 .12)
     *The implicit assumption in  this  table has been  that of  strict  addl ::ivity  of  bids  for  each ai.r qualify cffoct..   In obc a i.ning the "'.car.
bids, differentiation has been made vith respect  to:   (1) the completion  i!aue oi  cleanup;  (2)  the biddinc, sequence.   IP. "Type I" questionnaires,
the air quality effects; are  introduced  in  "Aesthetic  -' Acute  iiealtii  -  Chronic Health"  order.   In "Type li" questionnaires, the air quality
effects; are introduced in "Acute  ::eal t'n •+  Ciironlc Health -» Aesthetic"  order.  In  obtaining  the  mean bids,  no differentiation has been !?ace
with rer.pect to:  (1) different: proposed vehicles for  the collection of  bids; (2)  whether a  health pamphlet has or has not been sent to the
respondent in advance of the i:-.terview; and (3) whether a life table has  or  has not  been  shown  to the respondent during the interview.  A life
table depicts the "stock" counterparts of  the elicited monthly bids  for various expected  lifespans.

    **The notation in parentheses represents the  change in air quality for which  the respondents are bidding.   For example, (A -*• 8)  denotes
that the respondent is bidding to change air quality  fro:r. poor to  fair,  (B -*• C) denotes  that  the respondent is  bidding to change air quality
from fair to good, and (C -» C*) denotes that the  respondent is bidding to change air quality  to good across t're entire region.

   ***Scandard error of the nean  bid in all cases.

  ****Sample size of each case in all  cases.

-------
                                                          Table C.2

                                                     Mean Bids by Area by Type*
                                                (Completion Dace of Cleanup:  10 yrs.)
Area**
El Monte
(A - B)
El Nonce
(A H. C)
La Canada
(A -» B)
La Canada
(A -» C)
Montebello
(A - B)
Montebello
(A -» C)
Kean Aesthetic Bid
($/ir.onth)
Type I
7.50
(7.50)***
(2) ****
5.00
(1.53)
(5)
6.33
(2.54)
(6)
16.83
(8.47).
(6)
4.00
(2.45)
(5)
6.00
(1.97)
(5)
Typo II
0.00
(0.00)
(7)
0.00
(0.00)
(1)
0.00
Co. oo)
(4)
0.00
(0.00)
(1)
1.40
(0.98)
(5)
1.67
(1.67)
(3)
Mean Acute Health Bid
(|/cionch)
Type I
2.50
(2.50)
(2)
6.40
(3.53)
(5)
15.83
(14.84)
(6)
4.17
(2.39)
(6)
12.40
(9.58).
(5)
2.20
(1.11)
(5)
Type II
14. 57
(6.49)
(7)
0.00
(0.00)
(1)
15.25
(11.80)
(4)
50.00
(.0.00)
(1)
5.20
(1.56)
(5)
0.00
(0.00)
(3)
Mean Chronic Health Bid
(S/month)
Type I
0.00
(0.00)
(2)
1.40
(0.98)
(5)
O.S3
(0.83)
(6)
0.50
(0.50)
(6)
1.80
(0.97)
(5)
1.20
(0.80)
(5)
Tvnc II
1.43
(1.43)
(7)
0.00
(0.00)
._(!)
22.50
(22.50)
(4)
0.00
(0.00)
(1)
1.60
(1.03)
(5)
0.00
(0.00)
(3)
Mean Total Bid
($/month)
Type I
10.00
(10.00)
(2)
12.80
(2.54)
(5)
23.00
(15.56)
(6)
21.50
(7.96).
(6)
18.20
(11.39)
(5)
9.40
(2.86)
(5)
Tvpe II
16:00
(6.33)
(7)
0.00
(0.00)
(1)
37.75
(23.81)
(4)
50.00
(0.00)
(1)
8.20
(1.80)
(5)
1.67
(1.67)
(3)
(continued)

-------
                                                            Table C.2





                                                            (continued)
Area
Canoga Park
(B •* C)
Culver City
(B H. C)
Encino
(B - C)
Huntington Beach
(B -> C)
Irvine
(B -> C)
Newport Beach
(B - C)
Mean Aesthetic Bid
Typo I
5.58
(1.64)
(6)
12.88
(6.30)
(8)
6.00
(3.21)
(6)
16. 38
(5.33)
(8)
13.08
(5. 12)
(9)
•1.10
(0.51)
(5)
Type II
0.95
(0.95)
(5)
7.50
(4.79)
(4)
0.00
(0.00)
(5)
5.75
(3.43)
OD
4.33
(2.96)
(3)
1.20
(1.20)
(5)
Menu Acute Health Bid
(S/mop.th)
Typo I
0.92
(0.58)
(6)
7.25
(3.51)
(8)
0.50
(0.50)
(6)
7.63
(2.05)
(S)
7.72
(1.97)
(9)
0.20
(0.20)
(5)
Tvne II
3.45
(1.86)
(5)
11.13
(3.23)
CO
11. 20
(6.32)
(5)
12.90
(4.30)
(12)
19.00
(15.63)
(3)
3.80
(1.91)
(5)
Menu Chronic Health Bid
(S/month)
Mean Total Bid
($/510Ht!l)
Type I 1 Tvpe II 1 TVPC I
0.88
(0.58)
(6)
4.50
(3.02)
(8.)
1.00
(0.6.3)
.. . .(&)
6.88
(4.20)
(8)
2. 58
(1. 15)
(9)
0.40
(0.40)
(.5)
o.no
(0.00)
(5)
3.25
(2.36)
i-'O
2.00
(1.22)
(5)
6.82
(2.94)
01)
0.00
(0.00)
(3)
6. 50
(3.83)
(5)
7.38
(1-75)
(6)
24.63
(12.07)
(8)
7. 50
(3.10)
(6)
30.88
(S.31)
(8)
23.39
(6.03)
(9)
1.70
(0.54)
(5)
Tvne II
4.40
(1.69)
(5).
21 .88
(9.21)
(4)
13.20
(7.26)
(5)
26.64
(S.52)
23.33
(18.56)
(3)
11.50
(5.04)
(5)
(continued)

-------
                                                                      Table  C.2

                                                                      (continued)
Area
Pacific Palisades
(C - C*)
Palos Verdes
(C - C*)
Redondo Beach
(C •* C*)
Mean Aesthetic Bid
($/sonth)
Type I
7.40
(4.21)
151
7.29
(1.38)
cn
20. 64
(6.46)
(7)
Tvne II
4.29
(2.30)
(7)
2.00
(1.22)
(4)
1.00
(0.77)
(5)
Htiiin Acute Health Bid
($/mOnth)
Tvnc I
13.60
(5.87)
(5)
7. 71
(5.45)
(?)
9.57
(6. 83)
(7)
Tvse II
15.71
(4.56)
(7)
22.50
(9.46)
(4)
3.30
(1.76)
Moan Chronic Health Bid
($/month)
Tvoe I 1 Tvpo II
139.20
(87.30)
(5)
1.14
(0.77)
(?)
4.14
(2.26)
(7)
2.S6
(1.49)
(7)
5.50
(4.S6)
(•'•)
4.80
(3.01)
(5)
Mean Total Bid
($/r,onth)
Type I
160.20
(92.16)
(5)
16.14
(5.91)
(7)
34.36
(14.35)
(7)
Tv»c II
22.36
(4.21)
(7)
30.00
(20.21)
CO
9.10
13.85)
(5)
. .           *The implicit  assumption in this  table  h.'.s  been  that  of  strict  additivity  of  bids  for  each air  quality  effect.   In  obtaining  the  mean
-j      bids,  differentiation has been made  with respect to:   (1)  the completion  date of cleanup;  (2)  the  biasing  sequence.   In  "Type  I" questionnaires,
W      the air quality  effects  are  introduced in "Aesthetic  •» Acute  Health  -»  Chronic Health" order.   In "Type  II" questionnaires,  the air  quality
       effects are introduced  in "Acuta ilealth •* Chronic  Health •» Aesthetic"  order.  In obtaining  the mean  bids,  no ailferenciation has been  zade  with
       respect to:  (1) different proposed  vehicles for the  collection  of bids;  (2) whether a  health  pamphlet  lias or has  not  been  sent: to  the
       respondent in advance of the interview; and  (3)  whether a  life table has  or  has not been  shown to  the  respondent  during  the interview.  A life
       table depicts the "stock" counterparts of the elicited monthly bids  for various expected  lifespans.

           **The notation  in parentheses represents the change in air quality for which the respondents are bidding.  For example,  (A •+ B)  cenotes
       that the respondent is  bidding to change air quality  from  poor to  fair,  (B •* C) denotes that  Che respondent  is bidding to change air quality
       froa fair to good,  and  (C *  C*) denotes that the respondent is bidding to change air quality  to good across  che entire region.

          ***Standard error of  the  mean bid in all  cases.
         ****Samplc size of  each case In all  cases.

-------
                                APPENDIX D

        Preliminary Regression Relationships for Selected Variables
                        on the South Coast Experiment

Introduction

     The following Tables present a very preliminary set of regression
results on examining the raw data for bid relationships in the South Coast
Air Basin.  These data sets must be viewed as a preliminary set to give the
researchers further guidelines on how to statistically analyze the data set.
They are not meant to be viewed as definitive in either a computational or
final set sense.  However, they should indicate to other researchers the
degree of variation in both the estimates and the sets of relationships
hypothesized for computation.  It is anticipated that it will take a least
four to five months for all such relationships to be adequately statisti-
cally analyzed.

     Table D-l contains a preliminary set of regression equations across all
areas and bid types.  Aesthetic, acute and chronic health bids, along with a
total bid, were regressed against various variables of possible interest.
One of these variables was the interviewer, to find out whether a detectible
bias might exist in terms of the interview selected.  In most instances, no
interviewer bias was discovered; however, for the acute health bid, chronic
health bid, total bid, when related to a small number of variables, there
was an indication of a detectible interviewer bias.  The researchers will
continue to explore this possibility to discover  whether,  in fact, such a
bias is present and how it might be removed from further statistical com-
putations.  A further test was to examine whether years of  education in some
significant way influenced the amount of the bid.  In no circumstances was
a significant relationship (at the 95% level of confidence) discovered.  A
third possible premise was that the duration of years lived in Los Angeles
would influence the bid.  The results here are mixed,  although in almost all
circumstances, statistically nonsignificant.  Both positive and negative
effects of years living in Los Angeles was discovered.   Finally,  as a general
variable to examine, individuals who had read the health pamphlet and those
who had not were examined.  Again,  the results were mixed.   However, in each
circumstance, those who had read the health pamphlet tended to bid signifi-
cantly higher (at the 95% level) than those who had not.  Alternatively, the
bids on aesthetic and chronic health effects appeared to not be related in
any reasonable way to whether the individuals had,  in fact, had access to
additional information on health effects.

     Dummy variables were inserted  for each of the locational sites of the
experiment.  In almost all circumstances, with a few exceptions,  the site-
specific dummy variables were nonsignificant, indicating at least in a
preliminary way that site-specificity would not significantly influence the
bid.  The pollution variable in every circumstance but  one  was insignificant
at the 95% level of confidence.  This would be anticipated  on the basis of
the conceptual research reported in Chapter 2.   That is, when one nets out
all the effects on the various bids with the exception  of pollution, inclusive
of income, then the pollution variable itself may or may not be statistically
significant.   For example, if preferences are nonhomogeneous and  nonidentical,
we could presume that those willing to pay a higher price for clean air

                                     179

-------
                                                                                                                      H'muY trcmsiojt WAIIOXS lea »
CO
O
	 	 . 	 	 . —
,
• i.iU;;+ 1 c**wu*t tc Tt LJI
j

j ; lrl»l . (O.*tl) { .10))
1 &ml«».«!lfcSU -It-lli j on* ' l!J . 711
| * f).6-V») . I 5il) ( . y,tj
' T.ut MJ -W.JJI ! 0.1*0 : .M» - .K'
10.111) i itt) ( .*•»*>
«Mf,*ilc lit 1. tit -O.JJ3 • .!}) - .fit-;



1:11 ' ' CO **> C i**t ( ' Jill




«e»t. B..-,t, in -> oil a 1M -«» • 64[ '


(0,1)0) • ( .JC4) (O.J)»

(o.nn ( .011) .o:i
(t.ilSl


:i


I .i*t-*»
{ .nn>
.»J5-J
( .««-»)

















II


( .i«t- J
( .3*1- )
- .i*i-
( . Jfi- f















	

'„


( .ut- j
t .111- )
< .^'*- t















	



;!:;;!::,
-l.B t-*
O.) £-*)
J.. /.-J
•J.I I-*
(l.J f-JJ















	
d f
»

J.6U-
(t.)iE. >
U.l^t* )
J.Wt-*
(J.t«-4)

















**

J.OH-
(i.e:t- )
-*.»«-
<*.*'!- )
•*.m-
l».tu. )















	

:, 1 ' ,.

.; i- i .jjt-*
(.it-)] < . ilt-H
{.»(-> ( .WI-4J
. 0 t- ' .lit-}
( .1 *- ) j 1 .*it-4)

1



j








'
	

"

-Jt(-*
( .Ut-»)
{ ,!«'*>
.».-.
C .!"-•!









:





	



-Mil-

(t.)lt- }
-LltE-*














t

	

: = , !

( .m-*i
,:oi-*
{ .»»i-»)
I . !>l^*)
.04(_)
















- 	

'

d.i»
:.)i
:. js










10. >9')
- . IM

• , 1M
, ( .Ui>
-.,,.

— .

-7"1





O.'V
(!.*" •"


0 I*

3.J.*








-


"





I.1W

: i )£•!


i.ijoo










	 . 	
r)





i.n*
:. 1:3



*;.)«










—
'













(>.t*r->
:.)Ei-«
<'. \$t- i


I.JJi-*






















1



* '
no '
,,0
no
,,«,

110
1-.6







'"
1
1110
!
mo
1
< !

	 ^— — i
o :* | i; i»
1
„.:> | .: .»
o. ;> .> MJ j
e. )i >c **j 1

o. 10
,.,» .,», j


i

!


o.ii u.ft*f
i

1 I


                                                                          »U.  »  »«M«J» ****. «) »**i

-------
                                                                                        TABU. 0-2
                                                       PRELIMINARY REGRESSION EQUATIONS KOR THE AGGREGATED  "A" AND "B" AREAS
00
Dependent
Variable
Aesthetic Bid
Acute Health
Bid
Chronic Health
Bid
Total Bid
AeftthctU Bid
! Acute Health
Bid
Chronic Health
Bid
Tot.il Bid
log. (Aesthetic
Bid)
log (Acute
Health Bid)
log (Chronic
H.alth Bid)
log (Total Bid)
Constant
8.682
1.327
-0.896
11.112
5.153
11.94;)
0.205
17.301
j.«;s
0.387
1.597
3.447
Independent Variables
LA
-5.66E-2
(B.769E-2)
-2.47E-2
(7.59E-2)
-8.12E-2
(0.111)
-0.162
(0.166)







HP
0.662
(1.161)
1.166
(1.005)
2.745
(1.466)
4.573
(2.203)







1 tf
-5.53E-5
(6.79E-5)
1.49E-4
(5.87E-S)
1.34E-4
(8.57E-5)
2.27E-4
(1.29E-4)







0.145
(0.600)
-0.442
(0.518)
1.231
(0.757)
0.934
(1.136)
5.953
(P1.245)
-21.849
(19.020)
13.246
(26.681)
-2.650
(41.657)



6P2




-0.562
(2.110)
2. OSS
(1.889)
-1.404
(2.650)
0.123
(4.138)



YAP




-9.53E-6
(2.31E-5)
2.53E-8
(2.07E-5)
7.06E-5
(2.90E-5)
6.11E-5
(4.52K-5)



LLA








-0.263
(0.386)
-0.248
(0.179)
-0.343
(0.168)
-n.283
(0. 186)
LHP








-4.72E-?
(0.390)
0.191
(0.377)
0.387
(0.354)
0.223
(0.391)
LY








-0.159
(0.110)
0.152
(0.107)
-5.48L'-:
(0.100)
-4. HE-:
(0.111)
LAP








7.70EI-2
(0.155)
-2.-i3r.-2i
(0.150)
8.82E-2
(0.141)
S.25E-2
'0. 156)












N
75
?5
15
75
75
75
75
75
T,
JS
75
75
R2
0.02
0.12
0.10
0.11
0.00
O.OS
0.09
0.02
0.06
O.OJ
0.03
0.04
S£
10.385
8.986
13.112
19.698
10.400
9.311
13.061
20.392
1.271
1.227
1.153
1.275
                   Independent variables:
                     LA   -  Years  lived  In  L.A.
                     HP   -  Amount  of health  paoplet  read

                     /.P,  -  Ch.lnjc  In pollution  level.  I.e.  ISO
                                                 change  in  the  sqoared
                                                Ion  Itvela,  i.e.  1/2(P^-P
                                                hafige  in pollution level
                                                1th  pnrr.phlet  read)
£P~  -  On.' hoi/ times  th
          values of polio

LLA  -  loj (Vo.irs lived
LHP  •  !(-£ (A.-OU.U of  he

L.'-P  -  log (Chan3f In  pollution  level)
                -0-
S,   -  N-_iber of cases
R"   -  CooJ.ipss of fit
SE   -  Standard error of regression
  (E-n •*  10"";  I.e.  E-2  •> 10"
Observations  are  aggregated  without any differentiation with
  retpect to  1) Bidding  sequence,  2) Starting bid. 3) Vehicle
  used, 4) Health  pamphlet vs.  no  health pamphlet, 5) Ufa
  table vu. no  life  Cable.

Bids for each eir  quality effect are assumed to be strictly separable.

-------
                                                                                                 TABLE 0-3

                                                                       PRELIMINARY REGRESSION  EQUATIONS TOR THE PAIRED AREAS*

                                                                                   (S'EWPOST  BEACH-PACIFIC PALISADES)
CO
Ni
Dependent
Variable
Aesthetic Sid

Acute Health Bid

Chronic Health Btd

Total Bid

log (Aesthetic
31d)

log (Acute Health
Bid)

log (Chronic
Health Bid)


log (Total Bid)


Constant
-49.578

-30.484

179.120

99.057


-21.328


-5.907


0.184


-3.303


1 -
Iniiirpctuient Variable
LA
0.728
(0.735)
0.451
(0.812)
-5.765
(5.386)
-4.586
(5.837)













HP
13.773
(4.634)
13.777
(5.397)
-26.756
(35.810)
0.794
(38.808)

V
7.89E-4
(2.94C-4)
5.713
(3.25E-4)
-1.12E-4
(2.15E-3)
1.25E-3
(2.33E-3)


j




















-M'
-41.709
(26.624)
-44.660
(29.421)
24.615
(195.197)
-61.754
(211.543)













LLA









0.155
(0.861)

-0.118
(.-6.06.1:-?'

-1.657
(1.263)

-0.956
(0.902)

LH? w









0.795
(0.856)

1.144
(0.632)

-O.S50
(1.256)

0.263
(0.69(0










2.067
(1.146)

0.793
(0.847)

0.624
(1.687)

0.346
(J.;OD

UP









3.965
(2.626)

5.468
(l.'J40>

-0.164
(3.854)

3.376
(2.750)







N
13

13

13
1


13


! i3
!
;
R2
0.65

0.57

0.2?

0.27


0.54


13 I 0.66








j
1
13 ! 0.43
,
1
13 j 0.57
1
	 i_


SE
19.466

21.512 .

142.720

154.672


l'.235


0.912


1.812


1.293


                           Independent vari^bleo:
                             LA  -  Years lived  in L.A.
                             HP  -  Anount of health pamphlet  read
                              Y  -  Incoae
                             -P  -  Chnnpe In pollution  level,  i.e. AttOj
                             LLA •  log (Yenrs lived in  L.A.)
                             LHP -  lop (Arount of licalch ponphJet resd)
                             LY  -  log (Incooc)
                             LiP -.  log (Cltun&tt tn pollution  level)
                                             -0-
                             K   «  Xunber of cases
                             R   "  Goodness of fit
                             SE  -  Standard of error regression
Values In paroncheiiCS  arc  coi*f f iclenc standard errora
  (E-n •> 10~n;  I.e.  E-2  •*  10" )

Observations are .ip,j;r«:K)ittd  without any differentiation with
  respect to 1) Bidding  sequence,  2)  Scartinfc bid, 3) Vehicle
  uufcd, A) Health  p.itnphiet vs. no  health paephleC, 3) Life cab
  ve. no life tabli.'}

Bids for each air  quality effect are  assumed to be strictly Be

-------
                                                                                                TABLE D-*

                                                                        PRELIMINARY REGRESSION EQUATIONS TOR THE PAIRED AREAS

                                                                                          (IRVINE-PALOS VERDES)
                  «, >>. c, d
GO
LO
Dependent
Variable
Aesrhe'ic Bid
Acute Health Bid
Chronic Health Bid
Total Sid
Log (Aesthetic Bid)
Log (Chronic Health
Bid)
Log (Total Bid)
Constant
1.086
6.420
0.332
7.838
-8.095
-1.046
-4.534
-6.527
Independent Variables
LA
-8.48E-2
(6.58E-2)
0.187
(0.191)
-1.29E-2
(5.67E-2)
8.90E-2
(0.201)


HP
0.457
(1.014)
3.055
(2.942)
0.197
(0.873)
3.709
(3.102)


Y
1.01E-4
(4.65F.-5)
-2.48E-5
(2.80E-4)
3.13E-5
(8.31E-5)
1.071
(2.95E-4)


OP
2.661
(3.231)
-3.591
(9.375)
2.252
(2.782)
1.322
(9.8861


I.LA




-0.340
(0.205)
-4.48E-2
(0.245)
-0.132
(0.161)
-4.99E-2
(0.183)
LHP




0.145
(0.688)
0.516
(0.823)
0.448
(0.541)
0.468
(0.613)
LY




0.940
(0.602)
0.265
(0.720)i
0.505 i
(0.473):
LAP




-0.517
(1.186)
0.330
(1.418)
-0.506
(0.93J)
0.862 -0.205
(0.536)1 (1.057)







N
26
26
26
26
26
26
26
26
R2
0.15
0.08
0.04
0.07
0:22
0.04
0.13
0.16
SE
4.446
12.900
3.828
13.603
1.083
1.295
0.852
0.965
                             Independent variables:
                               LA  -  Years lived In L.A.
                               HP  -  Aaount of health panphlet read
                                Y  -  Income
                               £P  -  Change in pollution level, i.e. ANO
                               LLA -  log {Years lived In L.A.)
                               LHP -  log (Acount of health pamphlet read)
                               LY  -  log (Incoff.e)
                               L.1P -  log (Change in pollution level)
                                               -0-
                               N   -  Suaber of cases
                               R   -  Goodness of fit
                               SE  -  Standard of error regression
  Values in parentheses are coefficient stAndard errors
    (E-n * Itf"; I.e. t-2 •* 10  )

C Observations arc aggregated without any d I f f erenc lat Ion  uUh
    respect to 1) Bidding sequence. 2) Starting bid,  3)  Vehicle
    used. 4) Health pamphlet vs. no health  pamphlet,  5)  Life  table
    vs. no life table)

  Bids for each air quality effect are assumed to  be  strictly tep*rjbl«

-------
                                                                  TABLE D-5


                                          PRELIMINARY REGRESSION EQUATIONS FOR THE PAIRED  AREAS8'  b'  c>  d

                                                              (LA CANAilA-ENCINO)
Dependent
Variable
Aesthetic Bid
Acute Health
Bid
Bid
Tc>tal Bid
Aesthetic Bid
Acute Health
Bid
Chronic Health
Bid
Total Bid
log (Aesthetic
Bid)
log (AcuLe
Health Bid)
IOR (Chronic
Health Bid)
log (Total Bid
Constant
19.224
2.220
-4.03E-2
21.384
12.099
6.569
-5.559
13.101
5.589
0.770
4.477
6.322

LA
-0.225
(0.229)
-3.70--2
(0.165)
-0.280
(0.372)
-0.542
(0.420)








Independent Variables
HP
-3.122
(2.838)
3.066
(2.049)
6.971
(4.611)
6.915
(5.207)








t
-1.72E-4
(1.14E-4)
2.37E-4
(8.26E-5)
1.30E-4
(1.86E-4)
1.95E-4
(2.10E-4)








AP
0.977
(1.298)
-1.581
(0.937)
3.005
(2.109)
2.4111
(2.382)
-30.4B9
(48.614)
24.025
(40.736)
29.340
(80.346)
22.877
(94.111)




AP2



3.257
(4.872)
-2.589
(4.083)
-2.771
(8.053)
-2.103
(9.433)




YiP



-5.62E-5
(4.27E-5)
1.94E-5
O.53E-5)
5.30E-5
O.05E-5)
1.62E-5
(8.26E-5)




LLA







-0.640
(0.582)
-0.408
(0.572)
-0.732
(0.632)
-0.910
(0.681)
LHP







-1.249
(0.759)
0.717
(0.745)
0.413
(0.82(.)
0.335
(0.887)
LY







-0.246
(0.126)
0.171
(0.124)
-0.160
(0.137)
-0.113
(0.147)
LAP







0.246
(0.287)
r0.247
(0.282)
0.486
(0.312)
0.184
(0.336)












N
22
22
22
22
22
22
22
2:
22
22
22
22
R2
0.19
0.42
0.24
0.23
0.10
0.12
0.13
0.04
0.29
0.20
0.20
0.13
SE
13.241
9.559
21.514
24.292
13.574
11.375
22.435
26.279
1.323
1.300
1.437
1.543
Independent variables:
  LA
  HP
  Y
  t?
LT

YSP
LLA
LHP
LY
LiP
R
SC
     -  I
     -  Ch
     -  0"
           rs lived In L.A.
           nnt of health paciplet read
               e In pollution level, i.e. AUG.
               olf c tuics the cbonp.e in the squared.  .
               ucs of pollution levels,  i.-e. l/2(P^-Pp
               c cir.es the change in pollution level
       -  log (Years  lived in L.A.)
       •  log (Anount of health pamphlet read)
       •  IJR (Incoce)
       -  log (Change in pollution level)
                  -0-
       ••  S'u:i!»cr of cases
       »  Coodneii of fit
       •  Standard error of regrtsalon
Values In parenthu^es are  coeff ieicnt  standard errors
  (E-n * l(f °;  I.e. E-2 •*•  10"  )

Observations are aggregated without  any  differentiation vith
  respect to 1} Bidding sequence,  2) Starting bid,  3) Vehicle
  used/ 4) Health pamphlet vs. no  health pamphlet,  5} Life
  table vs. no  life table.

Bids for each air quality  effect are assumed to b«  atrictly »

-------
                     TABLE n-h


PRELIMINARY REGRESSION EQUATIONS TOR THF. PAIRED  AREAS*'  b>  C'  *

                    OS BEACH-REDONDO BEACH)
Dependent
Variable

• Aesthetic Sid

' Acute Health Bid

Chronic Health Sid

Total Bid

log (Aesthetic Bid)

log (Acute Health Bid)

log (Chronic Health Bid)

log (Total Bid)



Constant

-11.181

-8.864

-2.539

-22.58'.

-8.499

-8.424

-2.636

-8.980


Independent Variables

LA

0.201
(0.367)
0.171
(0.305)
0.120
(0.237)
0.491
(0.70-4)










HP

1.577
(2.992)
-1.180
(2.491)
0.284
(1.934)
0.480
(5.741)







V

5.34E-4
(3.79E-4)
5.47E-4
(3.15E-4)
1. 30E-4
(2.45E-4)
1.21E-3
(7.27E-4)






1
I
1
j

ar

7.166
(11.006)
0.764
(9.212)
9.297
(7.155)
15.700
(21.233)









LLA









0.330
(0.685)
0.793
(0.592)
0.285
(0.645)
0.588

LHP









-3.MF.-2
(0.749)
-0.671
(0.647)
0.332
(0.705)
-0.30B
(0.694) (0.758)
I

LY









O.S'A
(0.618)
0.774
(0.534)
0.240
(0.532)
0.932
(0.626)


LAP









-1.678
(1.578)
-1.096
(1.364)
-1.077
(1.485)
-1.474
(1.593)






















N

73

23

23

23

23

23

23

23
7
R

0.24

0.17

0.14

SE

16.054

13.365

10. J80

0.25 1 30.804
1
0.20 ! 1.522

0.18

0.09

0.19
i



1.316

1.433

1.5U


ndependenc variables:

 LA  -  Years lived  In L.A.
 HP  -  Amount of health paapMet  read

 A?  -  Change in pollution level,  I.e. &KO
 LLA -  log (Years lived In L.A.)
 LH? -  log (Amount of health pamphlet rend)
 LY  -  log (Income)
 LAP -  Ion (Clisr.Ji!  In pollution  level)
                 -0-
 N   -  Kus:ber of cases
 R   "  Coodneus of fit
 St.  "  Standard oi error regression
                                    Vflluca In parL-nthcscs ore coefflcitnt  standard  errors
                                      (E-n -> 10"°; i.e. £-2 •» 10" )

                                    Ob«erv.ir lona arc *ij;r>reSated without  any dif ferentlation  uith
                                      respect to 1) Bidding sequence,  2} Starting bid,  3) Vehicle
                                      used, 4) Health pamphlet vs. no  health  paophlet,  5) Life table
                                      VB. no  life table)

                                    bldtj for  each sir quality effect are assumed to be  strictly sepArable.

-------
                                                                                               TASU
                                                                      PRELIMINARY REGRESSION EQUATIONS FOR HIE PAIRED AREAS
                                                                                        (WNTEBELtO-rULVER CITY)
                                                                                                                        KAS'- "' C'
OO
Dependent
Variable
Aesthetic Bid

Acute Health Bid

Chronic Uraleh
Bid
Total Bid

Aesthetic Bid

Actute Health
Bid
Chronic Health
Bid
Total Bid

log (Aesthetic
Bid)
log (Acute
Health Bid)
log (Chronic
Health Bid)
log (Total Bid)

Constant
14.371

13.877

13.164

41.611

2.018

17.139

6.149

25.306

22.984

8.958

19.392

25.998

Independent
' LA
-3.40E-2
(0,103)
-8.32E-2
(0.141)
-1.27E-2
(4.5E-2)
-0.130
(0.190)
















HP
-1.277
(2.042)
-2.853
(2.800)
-1.229
(0.892)
-5.359
(3.760)
















Y
-2.77E-4
U.52E-4)
-1.55E-4
(2.09K-4)
-2.33E-4
(6.66E-5)
-6.65E-4
(2.B1E-4)
















4P
-0.25*.
(1.042)
-0.313
(1.428)
-1.150
(0.455)
-1.719
(1.91H)
17.058
(19.864)
-55.831
(20.201)
-8.323
(12.024)
-47.096
(39.363)








4P2








-1.532
(1.970)
5.574
(2.003)
0.815
(1.192)
4.857
(3.963)








Variables
YAP








-7.31E-5
(5.92E-5)
-3.41F.-S
(6.02E-5)
-2.48E-5
(3.59E-5)
-1.32E-4
U.19E-4)








1XA
















6.11E-3
(0.639)
-1.189
(0.664)
-8.92E-2
(0.423)
-0.518
(0.683)
LUt
















-0.701
(1.185)
-2.123
(1.232)
-1.349
(0.734)
-1.718
(1.266)
LY
















-2.133
(1.008)
-0.423
(1.048)
-1.773
(0.667)
-2.183
(1.077)
LAP
















-0.294
(0.533)
0.541
(0.554)
-0.540
(0.353)
-0.159
(0.570)

























N
14

14

14

14

14

14

14

14

14

14

14

14

IT2
0.30

0.17

0.66

0.46

0.18

0,47

0.25

0.25

0.36

0.36

0.53

0.40

SE
S.139

7.045

2.244

9.461

5.256

5.345

3.181

10.573

1.078

1'.120

0.713

1.152

                              Independent var 1 ahles :
LA
HP
Y
YAP
LLA
LHP
LY
L.iP
R
SE
                                     •  Y«Mrs lived In L.A.
                                     -  Amount of health pamplet read
                                     •  Income
                                     •  Chang* In pollution level. i.e, QKQ
                                     -  One half timer, the chance In the squared
                                          values of pollution levels, i.e. 1/2(P -P )
                                     -   ncome times the change* In pollution level ^
                                     •   og (Years lived In L.A.)
                                     -   og (Anount of health pamphlet read)
                                     -   og ( Incorio)
                                     -   og (Change in pollution level)
                                                -0-
                                     •  Nuober of cases
                                     -  Coodneaa of fit
                                     -  Standard error of regression
Values In parentheses aro coefficient standard*  erron
  (E-n * 10" ; I.e. E-2 * It)"  )

Observations are aggregated without any differentiation  with
  respect  to 1) Bidding sequence, 2) Starting bid.  3) Vehicle
  used, A) Health pamphlet vs. no health pamphlet,  5) Life
  table vs.  no life table.

Bids for each air quality effect are uat>u
-------
                                                                                        REGRESSION' EIJUATIOSS  VOR Tin; VAIRKD AREAS'
                                                                                                (Kl. MONTt'-CASOCA PARK)
                                                                                                                                    .1.  b.  c. d
CO

Variable
Aesthetic Bid
Acute Health
BU
Chronic Health
KM
Total Bid
Aesthetic Bid
1 Acute Health
B!
-------
years lived in L.A., amount of health pamphlet read, income and change in
pollution indicating a reasonable relationship except for, perhaps, the
sign on the number of pages read of the health pamphlet.  The income var-
iahle is highly significant as is the years of residence in Los Angeles.
However, only after substantial further experimentation on these pairs can
we anticipate that reasonably defensible estimates of coefficients or
elasticities will be forthcoming.
                                    138

-------
Conclusion

     In this Appendix, we have attempted to indicate the. rough, orders of
magnitude of variability of relationships- between ob'served Kids and some
variables of interest.  As yet, the regression results- Rave only roughly
illuminated possible further zones of research.  Both signs and magnitudes
seem to be highly insignificant when the data set is regressed totally.
Thus, it is anticipated that substantial additional research from a statis-
tical perspective and also incorporating x^ell-defined theoretical hypotheses
will need to be developed for this data set to be adequately exploited.  Of
particular importance is the examination of bias effects and disaggregation
down to the paired comparisons.  For our first estimate of the magnitude of
bid in Los Angeles reported in Chapter 6, we selected the last equation
total bid in the preliminary regression equations in Table D.I.  From the
coefficient for pollution and adjusting for the effect of the health pam-
phlet on bids along with adjustments for capital recovery facto.rs and the
length of time to achieve clean air, the numbers reported in Table D.I of
Chapter 6 were obtained.  The researchers believe this is only a preliminary
estimate of the value of the average Bid for Los Angeles.  It is anticipated
that further research will have a highly significant impact on ultimate
calculation of a reasonable, accurate value for citizens' preferences
associated with improved air quality.
                                     189

-------
                                 APPENDIX E.

     Thia appendix presents the variable list for the non-market valuation
experiment in the South. Coast Air Basin.
                                     190

-------
Variable
Name
MVFA
APLV
APWK
AZLV1
AQLV2
AQWK1
AQWK2
ZHAZ
RESCOP
RESEV
RESEN
RESSP
RESUN
RESGM
RESVAL
PCTIN
INTTM
DTINT.
PSI
EQBL
Description
Would you move if like A everywhere (l=yes)
Air pollution influenced where you live (l=yes)
Air pollution influenced where you work (l=yes)
Air quality where live (Good=00; Fair=01;
Poor=10)
Air quality where work (Good=00; Fair=01;
Poor=10)
Are you aware of any health hazards of air
pollution (l=yes)
Respondent cooperative (l=yes)
Respondent evasive (l=yes)
Respondent enthusiastic (l=yes)
Respondent suspicious (l=yes)
Respondent understanding (l=yes)
Respondent playing games (l=yes)
Respondent giving true value (l=yes)
Percent work time indoors
Minutes taken for formal interview
Date of interview (1=365)
Pollution index by location and date
Cost/person/month for air cleanup if all
Column (s)
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55-57
58-60
61-63
64-66

Format
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
F3.0
F3.0
F3.0
F3.0

          billed equally

APCLUP    Total figure for cleanup of pollution
          (100,000's of dollars)
67-72
73-E
F6.0
            F8.0
                                     1.91

-------
Card //I:  Socioeconomic Information and Enumerator Evaluation
Variable
Name
QNM
CRONM
QTYP
QCC
INTCD
TIKE
WORK
AGE
NPER
YRED
LIVLA
PLLVLA
LOCEMX
LOCEMY
LOCLVX
LOCLVY
ADDCON
SEX
MARST
DEC
Description
Questionnaire number
Card number
Questionnaire type: 1 = aesthetic first; no
health; 2 = aesthetic first, health;
3 = acute, no health; 4 = acute, health
City code
Interviewer code
Time spent at leisure (hours per week)
Time spent at work (hours per week)
Age of respondent
Number of persons in household -
Years of education.
Years lived in LA area
Years plan to live in LA area
X-coordlnate, location of employment
Y-coordinate, location of employment
X-coordinate, location of home
Y-coordinate, location of home
Additional conversation time (minutes)
Sex of respondent (l=male)
Marital status of respondent (l=married)
Highest degree obtained (1=K.S.; 2=Coll.;
Column(s)
1-3
4-5
6
7-8
9-10
11-12
13-14
15-16
17-18
19-20
21-22
23-24
25-26
27-28
29-30
31-32
33-34
35
36

Format
F3.0
F2.0
Fl.O
F2.0
F2.0
F2.0
F2.0
F2.0
F2.0
F2.0
F2.0
F2.0
F2.0
F2.0
F2.0
F2.0
F2.0
Fl.O
Fl.O
r* 1 r\
  ENHAZ
  PAMP
3=Voc.; 4=Adv.; 0=no degree)

Environmental hazards associated with job
(l=yes)
                                                               37
                                                               38
How much of pamphlet did you read?  (1=0-5'pages;
2=5-10 pages; 3=10 + pages; 0 = did not receive)   39
                                                                          Fl.O
                                                                          Fl.O
                                                                          Fl.O
                                        192

-------
Card 112:   Bidding Game and Secret Ballot
Variable
Name
QNM
QTYP

CRONM
STBID
CPDT
ZMV1X
ZMV1Y
ZMV2X
ZMV2Y
ZMV3X
ZMV3Y
MAXBD1
MAXBD2
MAXBD'3
VEH
LFCK
MV1
MV2
MV3
ZOTVEH
SB11
SB16
SB21
SB29
Description
Questionnaire number
Questionnaire type (l=aes. + no health;
2=aes. + health; 3=acute + no health;
4=acute + health)
Card number
Starting bid
Completion date of cleanup
Where you would move, 1st stage, X-coordinate
Where you would move, 1st stage, Y-coordinate
Where you would move, 2nd stage, X-coordinate
Where you would move, 2nd stage, Y-coordinate
Where you would move, 3rd stage, X-coordinate
Where you would move, 3rd stage, Y-coordinate
Maximum bid, 1st stage
Maximum bid, 2nd stage
Maximum bid, 3rd stage
Vehicle used (l=monthly charge; 2=utility bill)
Life table checked (1-yes)
Would you move, 1st stage (1-yes)
Would you move, 2nd stage (1-yes)
Would you move, 3rd stage (1-yes)
Is there another vehicle? (1-yes)
Secret ballot; question 1, bracket 1
Secret ballot; question 1, bracket 6

(l=checked ; 0=unchecked)
Column (s)
1-3

4
5-6
7-8
9-10
11-12
13-14
15-16
17-18
19-20
21-22
23-27
28-32
33-37
38
39
40
41
42
43
44
49
50
58
Format
F3.0

Fl.O
F2.0
F2.0
F2.0
F2.0
F2.0
F2.0
F2.0
F2.0
F2.0
F5.0
F5.0
F5 . 0
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
                                   193

-------
Card #2 (continued)
Variable
  Name                       Description                    Coluinn(s)    Format


SB31                                                             59        Fl.O

SB37        Secret Ballot; question 3, bracket 7                 65        Fl.O

SB41        (l=checked; 0=unchecked)                             66        Fl.O

SB46                                                             71        Fl.O

RVBD1       Reverse bid to B from C                            72-75       FA. 0

RVBD2       Reverse bid to A from B or C                       76-80       F5.0

-------
CARDS 3,  5,  7,  9,  .  .  .,  67




ACTIVITIES:   OUTDOOR THEN INDOOR
Variable .
Name
QNM
QTYP
CRDNM
10PT
10IMP
10TMI
10SBB
10SBC
10SBD
10TMSB
10FQSB
10LCSB
100TSB
10PD
10EQX
10HRA
10FQA
10LCXA
10LCYA
10MIA
10DCA
Description Column(s)
Questionnaire number
Questionnaire type
Card number
Activity participation (l=yes)
Importance column checked (l=yes)
Was the activity only time important (l=yes)
Did a substitution occur at 1st stage (l=yes)
Did a substitution occur at 2nd stage (l=yes)
Did a substitution occur at 3rd stage (l=yes)
Was there a time substitution (l=yes)
Was there a frequency substitution (l=yes)
Was there a locational substitution (l=yes)
Was there some other kind of substitution (l=yes)
Percent of activity done during day
Equipment replacement expenditures
Hours per week in activity
Frequency per week in activity
X-coordinate location of activity
Y-coordinate location of activity
Miles travelled
Direct costs
1-3
4
5-6
7
8
9
10
11
12
13
14
15
16
17-19
20-24
25-28
29-32
33-34
35-36
37-40
41-44
Format
F3.0
Fl.O
F2.0
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
F3.0
F5.0
F4.0
F4.0
F2.0
F2.0
F4.0
F4.0
                                       195

-------
CARDS 4,  6,  8,  10,  .  .  .,  68
Variable Name Descripti-on
QNM
QTYP
CRDNM
10HRB
10FQB
10LCXB
10LCYB
10MIB
10DCB
10HRC
10FQC
10LCXC
10LCYC
10MIC
10DCC
10HRD
10FQD
10LCXD
10LCYD
10MID
10DCD
Questionnaire number
Questionnaire type
Card number
Hours
Frequency
X-Location
Y-Location
Miles travelled
Direct costs
Hours
Frequency
X-Location
Y-Location
Miles travelled
Direct costs
Hours
Frequency
X-Location
Y-Location
Miles travelled
Direct costs
Coluran(s)
1-3
4
5-6
7-10
11-14
15-16
17-18
19-22
23-26
27-30
31-34
35-36
37-38
39-42
43-46
47-50
51-54
55-56
57-58
59-62
63-66
Format
F3.0
Fl.O
F2.0
F4.0
F4.0
F2.0
F2.0
F4.0
F4.0
F4.0
F4.0
F2.0
F2.0
F4.0
F4.0
F4.0
F4.0
F2.0
F2.0
F4.0
F4.0
                                    196

-------
Card 69:   Home Characteristics
Variable
Name
QNM
QTYP
CRDNM
LVAR
NRM
NBDRM
NBTRM
DEN
FAM
DIN
PCH
ATTIC
BASE
UTRM
OTRM
SCVW
STOR
REMD
DISH
DISP
CEAIR
TRASH
CEHT
POOL
FRPL
Description
Questionnaire number
Questionnaire type
Card number
Living area (sq. ft.)
Number of rooms
Number of bedrooms
Number of bathrooms
Den (l=yes)
Family room (l=yes)
Dining room (l=yes)
Enclosed porch (l=yes)
Attic (l=yes)
Basement (l=yes)
Utility room (l=yes)
Other room (l=yes)
Scenic view (l=yes)
Number of stories (include basement)
Remodeled (l=yes)
Dishwasher (l=yes)
Disposal (l=yes)
Central air conditioning (l=yes)
Trash compactor (l=yes)
Central heating (l=yes)
Swimming pool (l=yes)
Fireplace (l=yes)
Column(s)
1-3
4
5-6
7-11
12-13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Format
F3.0
Fl.O
F2.0
F5.0
F2.0
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
                                       197

-------
Card 69 (continued)
Variable
Name
AGERM
YRPC
LVHM
PCPR
MTPY
TDYVI
PTYTX
LVAPT
MTRT
INSPY
UPKP
PCTBST
STD
Description
Age of home (years since construction)
Year of purchase (last two digits)
Length of time (years) living in home
Purchase price of home
Monthly payments (rounded to nearest dollar)
Value of home in today's market
Property tax payments per year
Length of time (years) in apartment
Monthly rent
Insurance payments/year
Monthly upkeep around home
Percent of basement completed
Automobile standards (l=increase; 2=decrease;
Column(s)
34-35
36-37
38-39
40-45
46-48
49-54
55-58
59-60
61-64
65-68
69-72
73-75

Format
F2
F2
F2
F6
F3
F6
F4
F2
F4
F4
F4
F3

.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0

            3=same)
76
Fl.O
                                        198

-------
Card 70:   Home  Characteristics  and Transportation
Variable
Name
QNM
QTYP
CRDNM
HC71
HC72
HC713
AVGXFD
AVGXCL
INCOME
PYFC
VLUOTL
NMVEH
LICDR
US TMPC
LWSTMPG
MLTV1.D
HRCMWK
HRCHSP
HRCMREC
CRPOOL
GASCST
MTCST
RTD
AUTOINS
ZVAC
VACX
Description
Questionnaire number
Questionnaire type
Card number
Why have you chosen tc live in chis area
(0=not ranked; other ranked scale of 1 to 5
with 1 the best) 6 if only checked
Average monthly expenditures for food
Average monthly expenditures for clothing
Annual household income (midpoint of groups)
How much would you pay for house if in area
like C
How much of value of your home is due to no
air pollution
Number of vehicles in family
Licensed drivers in family
Highest average miles per gallon
Lowest miles per gallon
Miles travelled per week
Hours per week spent commuting for work or
school
Hours/week spent commuting for shopping
Hours/week spent commuting for recreation
Are you in a car pool (l=yes)
Gasoline costs/month
Maintenance costs/month
Public transportation fares/month
Auto insurance/month
Vacation within last year (l=yes)
Vacation expenditures
Column(s)
1-3
4
5-6
7
19
20-23
24-27
28-33
34-39
40-44
45
46
47-48
49-50
51-54
55-56
57-58
59-60
61
62-64
65-67
68-70
71-73
74
75-80
Format
F3.0
Fl.O
F2.0
Fl.O
Fl.O
F4.0
F4.0
F6.0
F6.0
F5.0
Fl.O
Fl.O
F2.0
F2.0
F4.0
F2.0
F2.0
F2.0
Fl.O
F3.0
F3.0
F3.0
F3.0
Fl.O
F6.0
                                     199

-------
Card 71:   Medical  and Attitudes
Variable
Name
QNM
QTYP
CRDNM
MD1A
MD1H
MD2I
MD20
AGGAP
DSAQ

PHYDIS
LFPL1
LFPL2
DRG1

EASWK1

NGTRDY

SMOKE
PACKS
MEDCTN
HEDX
MEDXAP

DR
MEDINS
Description
Questionnaire number
Questionnaire type
Card number
Medical, question 1, part a
Medical, question 1, part h
Medical, question 2, part I
Medical, question 2, part 0
Conditions aggravated by air pollution (l=yes)
Diseases which could be made worse by air
pollution (l=yes)
Physical disabilities (l=yes)
Life more pleasant (not at all=00; to some
degree=01; greatly=10)
Spend more on drug items (not at all=00;
to some degree=01; greatly=10)
Make it easier to do your work (not at all=00;
to some degree=01; greatly=10)
Prefer night or day (l=day; 2=night ; 3=no
difference)
Do you smoke (l=yes)
How many packs
Do you take medication regularly (l=yes)
Medication expenditures/month
Medical expenditures associated with air
pollution
Yearly doctor's fees
Yearly payments on medical and life insurance
Column(s)
1-3
4
5-6
7
14
15
21
22

23
24
25
26
27
28
29
30

31
32
33
34
35-37

38-40
41-44
45-47
Format
F3.0
Fl.O
F2.0
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O

Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O
Fl.O

Fl.O
Fl.O
Fl.O
Fl.O
F3.0

F3.0
F4.0
F3.0
                                        200

-------
CARD 71 (CONTINUED)
Variable
Name
DEFX

ATT 11
ATT 12
ATT 13

ATT 21
ATT 2 2
ATT23

ATT31
ATT32

ATT 41
ATT A 2
ATT A 3
ATT A 4
ATT45
ATT46
ATT47
ATT48
ATT49
ATT410
ATT411
ATT412
Column (s) Format
Have you ever purchased any item to reduce your exposure
to air pollution? (Such as filter) (1 = yes)
Since living in Los Angeles:
I have not been bothered by air pollution. (1 = yes)
I have been somewhat bothered by air pollution. (1 = yes)
I have been bothered by air pollution. (1 = ves)
Do you believe that air pollution in Los Angeles: (check one)
Has become worse since you have lived here.
Has stayed about the same since you have lived here.
Has gotten better since you have lived here.
What do you think should be done about air pollution?
(Check one)
Ignored.
Reduced.
Rank the following problems in terms of importance
(most to least) as the major issues facing the community.
(Choose top five.)
Juvenile delinquency.
Communicable disease.
Unemployment.
Air pollution.
Car accidents.
Crime.
Nuclear Energy.
Alcoholism
Water pollution
Energy
Congestion
Other
48

49
50
51

52
53
54

55
56

57
58
59
60
61
62
63
64
65
66
67
68
Fl

Fl
Fl
Fl

Fl
Fl
Fl

Fl
Fl

Fl
Fl
Fl
Fl
Fl
Fl
Fl
Fl
Fl
Fl
Fl
Fl
.0

.0
.0
.0

.0
.0
.0

.0
.0

.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
             201

-------
                               CARD 71 (CONTINUED)
Variable
  Name
Column(s) Format
            Do you believe air pollution in the Los Angeles area:
  ATT51     Cannot be reduced below the present level.                     69    Fl.0
  ATT52     Can be reduced below the present•level.                       70    Fl.0
  ATT53     Can be almost completely eliminated.                          71    Fl.O
            What do you think the words air pollution mean to most
            pepple in the Los Angeles area?  Do they mean:
  ATT61     Frequent bad smells in the air.  (1 = yes;  0 = no)            72    Fl.O
  ATT62     Too much dirt and dust in the air.  (1 = yes; 0 = no)          73    Fl.O
  ATT63     Frequent haze or fog in the air.  (1 = yes; 0 = no)           74    Fl.O
  ATT64     Frequent irritation of the eyes.  (1 = yes; 0 = no)           75    Fl.O
  ATT65     Frequent nose or throat irritations.  (1 =  yes; 0 = no)       76    Fl.O
  ATT66     Other.                                                        77    Fl.O
  ATT7      Have you read or seen anything in the newspaper recently
            about air pollution?  (1 = yes; 0 = no)                       78    Fl.O
  ATT8      When you read the newspaper, do you generally choose to
            read articles on air pollution?  (1 = yes;  0 = no)            79    Fl.O
  ATT9      Do you consult the daily air pollution index before
            engaging in any activities?  (1 = yes; 0 =  no)                80    Fl.O
                                          202

-------
                                 APPENDIX F
     This appendix  details  the actual  streets  in  the  paired  areas  from  which
the respondent sample was drawn.
                                     203

-------
                                         PARK
        u
                                   sr
I
I
     w
       COAV
        \JFM.
 I
JS/Y

                   1,1
                   ¥
                                  sr
                                               \"
                         204

-------
                          CUU/ER
^
^
             •+ K^v.w/wvl
                    205

-------
                 ELL
206

-------
        N
207

-------
     Huf/T[NfiTDM
268-

-------
                            IRVINE

STJ\£JTH1STL£
             209-

-------
       LA  CANADA
210

-------
UWCOlN/
                                                 TE
           211

-------
AVEL
         212

-------
RCIFSC PAM SAFES
              KJ
      213

-------
214

-------
H
             REDOMDO BEACH

-------
                                 REFERENCES

Ajzen, I. and Fishbein, M. , "Attitude-Behavior Relations:  A Theoretical
     Perspective and Review of Empirical Research," Psychological Bulletin
     84_  (1977) 888-918.

Anderson, R.J., Jr. and Crocker, T.D.  (1971), "Air Pollution and Residential
     Property Values," Urban  Studies  8, 1971-180.

Babb, E.M., and Scherr, B.A.  (1975),  "Pricing Public Goods:  An Experiment
     with Two Proposed Pricing Systems," Public Choice 23, 35-48.

Bayes, T., "An Essay Towards  Solving  a Problem in the Doctrine of Chances,"
     in  The Royal Society, Philosophical Transactions, Giving Some Account
     of  the Present Undertakings,  Studies, and Labours of the Ingenious in
     Many Considerable Parts  of  the World  for the Year 1763, London:  The
     Royal Society  (1765).

Blank, F., et. al. , "Valuation of  Aesthetic Preferences:  A Case Study of
     the Economic Value of Visibility," Research Report  to the Electric
     Power Research Institute, Resource and Environmental Economics Labor-
     atory, University of Wyoming (1977).

Bohm, P.  (1971), "An Approach to the  Problem of Estimating Demand for Public
     Goods,"  Swedish Journal  of  Economics  73.

Bohm, P.  (1972), "Estimating  Demand for Public Goods:  An Experiment,"
     European Economic Review 3; 111-130.

Bradford, D.F.,  "Benefit-Cost Analysis and Demand Curves for Public Goods,"
     Kyklos,  Vol.  23  (1970).

Brookshire, D.,  et.  al. ,  "The Valuation of Aesthetic Preferences," Journal
     of  Environmental Economics  and Management, Vol. 3  (1976).

Brookshire, D.S. and  Crocker, T.D., "The  Use of Survey  Instruments, in
     Economic Valuations  of  Environmental Goods," Paper  #2, Resource and
      Environmental Economics  Laboratory,  University of Wyoming  (1978).

Brookshire, D.S.,  and Eubanks,  L.S.,  "Contingent Valuation and  Revealing
      the Actual  Demand for Public Environmental  Commodities," Paper Pre-
      sented at Public Choice Society, New Orleans, La.,  1978.   Forthcoming
      in  Proceedings.

Brookshire, D.S.,  and Randall,  A.J.,  "Public Policy Alternatives, Public
      Goods  and  Contingent Valuation Mechanisms," Paper  presented at the
      Western  Economic Association Meeting, Honolulu, Hawaii,  1978.

Brookshire,  D.S.,  and Randall, A., et. al.,  "Economic Valuation of Wildlife,"
      Final report  for U.S. Fish and Wildlife  Service, Phase  I,  Resource
      and Environmental Economics Laboratory, University of Wyoming,  (Nov-
      ember 8, 1977).

                                      216

-------
Clark, E.H.. (1971), "Multipart Pricing of Public Goods.," Public Choice 11,
     17-33.

Clawson, M. and Knetsch, J.L. (19.66), Economics: of Outdoor Recreation, Johns
     Hopkins Press:  Baltimore.

Crocker, T., "National and Regional Models- for the Estimation of Recreational
     Benefits Accruing from PL 92-500,"' Research Paper No. 126, Department of
     Economics, University of Wyoming (1975).

Davis, R.  (1963), "Recreation Planning as- an Economic Problem," Natural
     Resources Journal 3.

Davis, R.K. and Knetsch, J.L. (1966), ""Comparisons of Methods for Recreation
     Evaluation" in Water•Research, A. V. Kneese and S.C. Smith (eds.), Johns
     Hopkins Press:  Baltimore.

Deyak, T.A. and V.K. Smith, "Residential -Property Values and Air Pollution:
     Some New Evidence," Quarterly Review of Economics and Business 14(4),
     (1978), 93-100.

Diewert, W.E., "Applications of Duality Theory," in M.D. Intriligator and D.A.
     Dendrick  (eds.)» Frontiers-of Quantitative Economics, Vol. II, Amsterdam:
     North Holland Publishing Co.  (1974), 106-160.

Freeman, A.M., Ill  (1974), "On Estimating Air Pollution Control Benefits from
     Land Value Studies," Journal  of Environmental Economics and Management 1,
     74-83.

Fromm, G., "Comment," in S.B. Chase, Jr.  (ed.) Problems in Public Expenditure
     Analysis, Blockings Institution, Washington, D.C., 166-176 (1968).

Goldberger, A.S., A Review of Consumer Demand Theory, Social Systems Research
     Institute, University of Wisconsin-Madison, Systems Formulation, Method-
     ology, and Policy Workshop Paper 6703 (Oct. 1967).

Hall, R.,  "The Specification of Technology with Several Kinds of Output,"
     Journal of Political Economy  81  (1973), 387-398.

Hammack, J. and Brown, G.M., Waterfowl and Wetlands:  Toward Bioeconomic
     Analysis, Baltimore:  Johns Hopkins University Press  (1974).

Harrison,  D.,  Jr. and D.L. Rubinfeld, "Hedonic Housing Prices and the Demand
     for Clean Air," Journal of Environmental Economics and Management 5
      (1978), 81-102.

Hicks, J.R., "A Reconsideration of the Theory of Value," Part 1, Economica 1
      (Feb.. 1934), 52-76.

Hori, H.,  "Revealed Preferences, for Public Goods," American Economic Review,-
     Vol.  65  (1975).
                                      217

-------
Horst, R., Jr.  (L978) , Es^jmotion of the      ____
     Application of the Expenditure Function, unpliblTslTefl Ph.D. dissertation,
     University of Wyoming.

Hurwicz, L., "On the  Problem of Integrability of Demand Functions," in J.
     Chipman, et. al. (eds.), Preferences, Utility and Demand, New York:
     Harcourt Brace lovanovich, Inc. (1971)" 174-214.

Johansen, L. (1977),  "The Theory of Public Goods:  Misplaced Emphasis?"
     Journal of Public Economics 7, 147-152.

Knetsch, J.L. (1963), "Outdoor Recreation Demands and Benefits," Land
     Economics.                                                    ~~

Kurz, M. (1974), "Experimental Approach to the Determination of the Demand
     for Public Goods," Journal of Public Economics 3, 329-348.

Lancaster, K. J. , "A New Approach to Consumer Theory," Journal of Political
     Economy 74 (1966),  132-157.                  "    ~	'

Lave and Nagin, "Automobile Related Air Pollutants and Health,  "REF
     Working Paper No. 52-73-74.

Lucas, R.C., "Wilderness Perception and Use:  The Example of Boundry Waters
     Canoe Area," Natural Resources Journal  3  (3), 1964, 394.

Lucas, R.E.B.,  "Hedonic Price Functions," Economic Inquiry 13(2), 1975,
     157-178.

MMler, K.  (1974), Environmental Economics:   A Theoretical Inquiry, Johns
     Hopkins Press:   Baltimore.

Mishan, E.J.  (1971),  Cost-Benefit Analysis:  An Introduction, Praeger
     Publishers:  New York.

Muellbauer,  J., "Household Production  Theory, Quality, and the Hedonic
     Technique," American Economic Review, Vol. 64  (1974).

Pearse, P.H., "A New  Approach to the Evaluations of Non-Priced
     Recreational Resources," Land Economics, 44  (February,  1968).

Pollak, R.,  and Wachter, M., "The Relevence  of the Household Production
     Functions  and  Its Implications for the  Allocation of Time," Journal of
     Political  Economy 83  (1975), 255-277.

Raiffa, H.,  Decision  Analysis, Reading, Mass.:  Addison-Wesley  (1970).

Randall, A., et. al., "Bidding Games for Valuation of Aesthetic Environmental
     Improvements," Journal of Environmental Economics and Management, Vol 1
      (1974).
                                      218

-------
Randall, A., et. al_._, Estimating Environmental Damages from Surface Mining of
     Coal in Appalachia:  A Case Study, U.S. Environmental Protection Agency,
     EPA-600/2-78-003, Jan. 1978.

Randall, A., "Evaluating Non-Market Goods, and Services.:  Some Conceptual
    Considerations," presented to the Symposium on Evaluation o£ Non-
    Market Goods at the Annual Meetings- of the American Agricultural
    Economics Association, San Diego, California  (July 31-August 3, 1977).

Randall, A.: Ives, B.: and Eastman, E., "Bidding Games for Valuation of
    Aesthetic Environmental Improvements-," Journal of Environmental Econo-
    mics and Management 1  (1974a) 132-149.

	, "Benefits of Abating Aesthetic Environmental Damage from
    the Four Corners Power Plant, Fruitland, New Mexico," Bulletin 618, New
    Mexico Agricultural Experiment Station, Las Cruces  (1974b).

Randall, A. and Stoll, J., "Consumer's Surplus in Commodity Space," Working
    paper, Department of Agricultural Economics, University of Kentucky
    (1978).

Ridker, R.B. and Henning, JcA.  (1967), "The Determinants of Residential
    Property Values with Special Reference to Air Pollution," Review of
    Economics and Statistics 49, 246-257.

Roberts, E.M., et. al., "Visibility Measurement in  the  Painted Desert
    through Photographic Photometry," Domes and Moore Engineering Bulletin
    No. 47  (1974).

Rosen,  S., "Hedonic Prices and  Implicit Markets:  Product Differentiation in
    Pure Competition," Journal  of Political Economy, Vol. 82  (1974).

Samuelson, P.A., "The Problem of Integrability  iri Utility Theory," Economica
    17  (1950), 355-385.

 Samuelson,  P.A.,  "The Pure  Theory  of  Public  Expenditures,"  Review of
      Economics and  Statistics,  36  (November,  1954).

Samuelson, P., "Diagrammatic Exposition of a Theory  of  Public Expenditure,"
    Review of Economics and Statistics 37  (Nov. 1955),  350-356.

Schulze, W.D. and d'Arge, R.C.,  "On the Valuation of Recreation  Damages,"
    prepared for the Joint  Session between the  American Economic Association
    and the Association of  Environmental  and Resource Economists, New  York
     (December 28, 1977) „

Steele  W  J   "The Effect of Air Pollution on Value  of  Single Family Owner-
      Occupied Redisential Property in Charleston, South Carolina,  Masters
      Thesis, Clemson University (1972)
                                      219

-------
Thayer, M. and Schulze, W., "Valuing Environmental Quality:  Th.e
     Contingent Substitution and Expenditure Approach," unpublished
     manuscript.  (.1977).

U.W: Bureau of the Census, Census of Population and Housing:   1970,
     Census Tract Final Report.

Wieand, K.F., "Air Pollution and Property Values;  A  Study Qf  the  St» Louis
     Area," Journal of Regional Science, Vol. 13, April 1973,  91-95.
                                      220

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
1. REPORT NO.
   EPA-600/6-79-001b
4.TITLEANDSUBTITLE  Methods Development  for Assessing Air"
 Pollution  Control Benefits:  Volume  II, Alternative
 Benefit  Measures of Air Pollution  Control  in the
 South  Coast Air Basin of Southern  California
                                                           3. RECIPIENT'S ACCESSION NO.
                                                          5. REPORT DATE
                                                            February 19>9
                                                          6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
 David  S.  Brookshire, Ralph C. d'Arge,  William D.
 Schulze,  and Mark D. Thayer
                                                           8. PERFORMING ORGANIZATION REPORT NO
                                                           10. PROGRAM CLEMENT NO.

                                                             1HA616  and  630
. PERFORMING ORGANIZATION NAME AND ADDRESS
Universi'ty of^y
Laramie,  Wyoming
                    82071
11. CONTRACT/GRANT NO.

 R805059-01
12. SPONSORING AGENCY NAME AND ADDRESS
 Office  of Health and Ecological Effects
 Office  of Research and Development
 U.S.  Environmental Protection Agency
 Washington,  DC   20460
                                                          13. TYPE OF REPORT AND PERIOD COVERED
                                                           Interim  Final,  10/76-10/78
                                                          14. SPONSORING AGENCY CODE
                                                           EPA-600/18
15. SUPPLEMENTARY NOTES
16. ABSTRACT
      This volume of a five volume study on  the  economic benefits of air  pollution
 control  includes the empirical results obtained from two experiments to  measure
 the health  and  aesthetic benefits of air pollution control in the South  Coast Air
 Basin of southern California.  Each experiment  involved the same six neighborhood
 pairs, where  the pairings were made on the  basis  of similarities in housing  charac-
 teristics,  socio-economic factors, distances  to beaches and services, average
 temperatures, and subjective indicators of  housing quality.  The elements of
 each pair differed substantially only in terms  of air quality.  Data on  actual
 market transactions,  as  registered in single-family residential  property trans-
 actions, and  on  stated preferences for air  quality, as  revealed
 surveys, were collected.
      Given  various assumptions on income, location, aggregation
 housing characteristics,  and knowledge of the health effects  of air pollution,
 both the survey  and the  property value experiments  yielded estimates of willingness-
 to-pay in early  1978  dollars for an improvement from "poor" to "fair" air quality
 of from $20 to $150 per  month per household.  The results, therefore, indicate  that
 air quality deterioration in the Los Angeles area has had  substantial negative
 effects on  housing prices and that these effects  are comparable in magnitude  to
 what people say  they  are willing to pav for improved air Quality.	
                                                                  in neighborhood

                                                                  by areas, specific
17.
                               KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b.lDENTIFIERS/OPEN ENDED TERMS
                                                                       c.  COSATI Field/Group
 Economic analysis
 Air pollution
 Environmental survays
 Real property
                                              Economic  benefits of
                                                pollution  control
               13B
18. DISTRIBUTION STATEMENT
 Release unlimited
                                             19. SECURITY CLASS (ThisReport)
                                              Unclassified
             21. NO. OF PAGES
               230
                                              20. SECURITY CLASS {Thispage)
                                               Unclassified
                                                                        22. PRICE
EPA Form 2220-1 (9-73)
                               o U S GOVtRNMWI PRfflTINC OFflCC: 1979 -281-147/21

-------
                                                  EPA Library
    •f States
    nomental Protection
Agei
$300
                                                     T 20539
Washington DC 20460
       c

-------