! I' \ KS-7 >- l'»7.»
Socioeconomic Environmental Studies Series
    Economic  Benefits  from an
    Improvement in  Water  Quality
                                  Office of Research and Monitoring
                                  U.S Environmental Protection Agency
                                  Washinaton. DC 20460

-------
            RESEARCH REPORTING SERIES
Research reports of the  Office  of  Research  and
Monitoring,  Environmental Protection Agency, have
been grouped into five series.  These  five  broad
categories  were established to facilitate further
development  and  application   of   environmental
technology.   Elimination  of traditional grouping
was  consciously  planned  to  foster   technology
transfer   and  a  maximum  interface  in  related
fields.  The five series are:

   1.  Environmental Health Effects Research
   2.  Environmental Protection Technology
   3.  Ecological Research
   *».  Environmental Monitoring
   5.  Socioeconomic Environmental Studies

This report has been assigned to the SOCIOECONOMIC
ENVIRONMENTAL   STUDIES   series.    This   series
describes  research on the socioeconomic impact of
environmental problems.  This covers recycling and
other  recovery  operations   with   emphasis   on
monetary incentives.  The non-scientific realms of
legal   systems,  cultural  values,  and  business
systems  are  also  involved.   Because  of  their
interdisciplinary  scope,  system  evaluations and
environmental management reports are  included  in
this series.

-------
                                                      EPA-R5-73-008
                                                      January 1973
        ECONOMIC BENEFITS FROM AN IMPROVEMENT IN
                       WATER QUALITY
                            By

       S.  D. Reiling,  K.  C. Gibbs,  H. H. Stoevener
                     Project 16110  FPZ
                      Project Officer

                   Dr.  Roger Don  Shull
            Implementation Research Division
             Environmental Protection Agency
                Washington, D. C.   20460
                       Prepared for
            OFFICE OF  RESEARCH AND MONITORING
          U.  S. ENVIRONMENTAL PROTECTION AGENCY
                WASHINGTON, D. C.   20460
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.O. 20402
              Price $2,10 domestic postpaid or $1.76 OFO Bookstore

-------
                 Review Notice
This report has been reviewed by the Environmental
Protection Agency and approved for publication.
Approval does not signify that the contents neces-
sarily reflect the views and policies of the Envi-
ronmental Protection Agency, nor does mention of
trade names or commercial products constitute
endorsement or recommendation for use.
                      11

-------
                           ABSTRACT
This report introduces and empirically tests a new methodology for
estimating the economic benefits accruing to society from an improved
recreational facility.  The specific facility under consideration is
Bfpper Klamath Lake, Oregon, which presently has low water quality.
The methodology draws upon previous work done in the evaluation of
recreational demand; however, it focuses upon the individual recrea-
tionist and separates the traditional price variable into on-site
costs and travel costs.  The model is used to estimate the number
of days per visit the recreationist will stay at the site as the
water quality improves.

Data collected at three other lakes with varied characteristics are
used to derive a relationship between the number of visits to a site
and the characteristics of the site.  This relationship is then used
to estimate the increase in visits to Klamath Lake that would be
forthcoming with an improvement in water quality.

The impapt of expanded recreational use of Klamath Lake upon the
local eponomy is also estimated through the use of an input-output
model of the Klamath County economy.

This report was submitted in fulfillment of Project Number 16110 FPZ
under the sponsorship of the Office of Research and Monitoring,
Environmental Protection Agency.
                               iii

-------
                      CONTENTS
Section                                          Fage

  I       Conclusions                              1

  II      Recommendations                          3

  III     Introduction                             5

  IV      A Theoretical Model to Estimate
          Previous Work in Estimating the  ,
          Demand for Recreation            '        9

  V       Empirical Specification of the
          Model                                   29

  VI      Estimating and Applying the Sta-\
          tistical Demand Model                   39

  VII     The Theoretical Model for Esti-
          mating Regional Benefits                61

  VIII    Construction of the From-to
          Model, and An Analysis of the
          Local Economy                           69

  IX      Application of the Input-Output
          Model                                   83

  X       Economic Benefits of Water Quality
          Imp rovemen t                             93

  XI      Acknowledgements                       105

  XII     References                             107

  XIII    List of Patents and Publications       111

  XIV     Glossary                               113

  XV      Appendices                             i!7

-------
                           FIGURES

                                                          Page

 1   The Commodity Space for Commodities 0- and (^           12

 2   An Indifference Curve for Commodities Q.^ and Q2         14

 3   An Indifference Map for Commodities Q, and Q2           15

 4   The Consumer's Budget Constraint                        17

 5   Maximization of Consumer's Utility                      19

 6   Quantities of Q- Purchased at Various Prices            20

 7   The Consumer's Demand Curve for Commodity Q,            21

 8   An Illustration of Consumer's Surplus                   23

 9   The Optional Combinations of Recreation and
     Non-Recreation, Given Travel Costs of k , k-, and
     k  for Given Prices p1 and p», and Fixed Income y       26
10   The Average Individual's Demand Curve, Per Visit,
     for a Lake                                             56

11   The Average Recreationist's Demand Curve, Per
     Visit, for Lake of the Woods                           58

12   Klamath County, Oregon                                 70

13   The Average Recreationist's Demand Curve, Per
     Visit, for Klamath Lake (Step 1)                       95

14   The Average Recreationist's Demand Curve, Per
     Visit, for Klamath Lake (Step 2)                       98
                               vi

-------
                              TABLES
 1  Estimated Use-Intensities for Certain Recreational
    Activities and Lake Size, by Lake                            33

 2  lumber of Interviews Taken at Each Lake, by Block,
    for the Four Time Periods                                    38

 3  The Population, Number of Sample Visits, Percent
    of Sample Visits, Visits Pet Season, Travel Cost,
    and Income, by County, for Each of the Four Lakes            46

 4  A Comparison of the Forest Service Estimates of Visits
    to Each Lake in 1968, and the Estimates Derived from
    the Visits Equation                                          54

 5  A Hypothetical Transactions Matrix                           62

 6  Description of the Sectors in the Klamath County Model       71

 7  Distribution of the Sample Among the Sectors of the Model    74

 8  Transactions Matrix Showing Interindustry Flows in
    Dollars, Klamath County, 1968 (Rounded to Nearest
    $1,000)                                                      77

 9  Direct Coefficients Matrix, Klamath County, 1968             79

10  Distribution of Sales of Each Sector in the Klamath
    County Economy                                               80

11  Direct and Indirect Coefficients Matrix, Klamath
    County, 1968                                                 84

12  Output and Income Multipliers and Income-Output
    Coefficients for Each Sector of the Klamath County
    Economy                                                      85

13  Average and Total Travel Cost Incurred in Klamath
    County, by Component, in 1968                                88
1                                                   i
14  Average and Total On-Site Cost, by Component, for
    klamath Lake in 1968                                         89
                                 vii

-------
No.                                                           Page

15   Total Recreational Expenditures and Percentages,
     in Klamath County, by Sector, Associated with
     Recreation at Klamath Lake in 1968                         90

16   Projected Increases in Final Demand, Total Output,
     and County Household Income, by Economic Sector,
     Associated with Recreation at Klamath Lake in 1968         92

17   Hypothesized Use-Intensity Ratings for Klamath Lake
     at the Present Time, and After Steps 1 and 2               93

18   Net Increase in Expenditures, for Each Component of
     Travel Cost, Associated with Improvements of Water
     Quality at Klamath Lake                                    99

19   Net Increase in Expenditures for Each Component of
     On-Site Cost Associated with Improvements of Water
     Quality at Klamath Lake                                    99

20   Projected Increases in Final Demand, Total Output,
     and County Household Income, by Economic Sector,
     Associated with Klamath Lake After the Completion
     of Step 1                                                 101

21   Projected Increases in Final Demand, Total Output,
     and County Household Income, by Economic Sector,
     Associated with Klamath Lake After the Completion
     of Step 2                                                 102
                                 viii

-------
                             SECTION I

                            CONCLUSIONS

1.  Improved water quality in Upper Klamath Lake, Oregon, would result
in expanded use of the lake for recreational activities.

2.  A new conceptual model is developed and used to estimate the demand
for recreation at Klamath Lake.  The traditional price variable is sepa-
rated into travel costs and on-site costs, because recreationists react
differently to changes in the two costs.  When both costs are combined
in a single variable, the effects are dampened and only the stronger
effect is observed.

3.  The methodology also focuses upon the individual recreationist in-
stead of the population of recreationists.  This avoids the necessary
assumptions made in earlier studies concerning the characteristics and
homogeneity of the population groups.

4.  A relationship is developed to explain the number of visits to a
recreational site with certain characteristics.  This is accomplished
by utilizing data gathered at three other lakes which possess different
characteristics.  The relationship is used to estimate the number of
visits to Klamath Lake after its characteristics are improved.  The man-
ner in which the site characteristics are specified in the model is not
considered to be entirely satisfactory*  The use-intensities for the
various water-related activities were used because the level of these
activities are, to some extent, dependent upon the water quality and
other physical features of the lakes.  However, it would be more satis-
factory to specify the model with respect to the biological and physical
parameters of the lake directly, since the use-intensities may not depend
entirely upon the changing parameters of the lake.  Unfortunately, the
necessary data were not available to specify the model in this manner.

5.  It is estimated that, if the algae were removed from Klamath Lake,
the net economic value of the lake would increase by $1,200,000 per year.
In addition, the income of households in the county would increase
$194,000 per year.

6.  If, in addition to removing the algae, the water temperature of the
lake were lowered and the beaches improved, the net economic value of the
lake would increase by an additional $2.65 million.  Thus, the value of
the recreational benefits associated with the two-step improvement is
estimated to be $3.85 million per year, while household incomes would be
$542,000 higher annually.

-------
                            SECTION II

                          RECOMMENDATIONS

Because of the unique shape of the demand curve derived  in  the  study,
more than one value can be interpreted to represent  consumer's  surplus.
It is not clear at this time which value is  correct.   Since the interpre-
tation used has such a significant influence on the  final results,  addi-
tional work on this problem is strongly recommended.

Throughout the analysis it is assumed that the expanded  use of  Klamath
Lake for recreation will not affect the level of use of  other lakes in
the vicinity.  This ignores the possibility  of recreationists substi-
tuting one recreational site for  another, and implies  that  the  postulated
increase in recreation at Klamath Lake represents  a  net  increase of
recreation in the area.  This is, to some extent,  an invalid assumption.
Additional research is warranted  to determine the  degree of substituta-
bility among recreational sites.

An attempt to incorporate the value of recreational  equipment in the model
as a substitute for the recreationist's current income was  not  success-
ful.  This effort was made because current income  may  not accurately re-
flect the recreational budget of  the recreationist.  Further work in this
area is suggested.

Future use of the model to predict the demand for  recreation at a site
after its characteristics have been altered  will require a  better speci-
fication of the model with respect to these  characteristics. Multi-
disciplinary work is clearly needed to accomplish  this goal.

Most of the work done in determining the demand for  publicly provided
outdoor recreation has focused on the estimation of  the  number  of days
per visit "consumed" at a site.   The problem of determining the number
of visits an individual will make to a site  in a specified  time period
has been ignored.  Work in this area is needed because of the relation-
ship that exists between days per visit and  the total  number of visits.
A model which could determine both variables simultaneously would be
very useful.

This study represents only the initial test  of the new methodology. Re-
finements and further tests are warranted, and are presently being  con-
ducted.

-------
                             SECTION  III

                            INTRODUCTION

Extensive research has been conducted by  the  staff  at  the  Pacific North-
west Water Laboratory of  the Environmental  Protection  Agency,  on the
physical and biological aspects  of the water  quality problem of Upper
Klamath Lake.  Aside from the strictly scientific benefits resulting
from this work, it is directed toward finding techniques which would
permit alteration of the  biological  processes in such  a way as to im-
prove the lake's water quality and its usefulness to man.

Upper Klamath Lake is located in south-central Oregon, near the city of
Klamath Falls, which has  a population of  about 38,000  in its immediate
vicinity.  It is the largest body of fresh  water in the State, being more
than 30 miles in length and comprising a  total area of more than 130
square miles.  U.S. Highway 97 follows the  eastern  shore of the lake
north of Klamath Falls for about 15  miles.  The highway is used by  tour-
ists during the summer, since it is  one of  the principal routes to  Crater
Lake National Park.  Usually one would expect such  an  accessible body of
water to be a popular site for water-based  recreation. However, this is
not the case at Upper Klamath Lake.  Although the lake now supports a
limited amount of recreational activities,  poor water  quality  renders
the lake undesirable for  large-scale recreational use. The poor water
quality stems primarily from large concentrations of algae and warm water
temperatures during the summer months.

A brief examination of some of Klamath Lake's physical characteristics
may be helpful in identifying the present water quality situation.  First,
mud deposits, ranging in  depth from  a few inches to more than  150 feet,
have accumulated on the bottom of the lake  [Bartsch, 1968].  These  de-
posits contain large concentrations  of the  primary  nutrients necessary
for algae growth, particularly nitrogen and phosphorus.

The accumulation of mud deposits has also decreased the depth  of the lake.
At the present time its average  depth is  less than  ten feet.  This  pre-
cludes the formation of temperature  stratifications and allows low  wind
velocities to cause sufficient water movement to keep  the  needed nutri-
ents suspended in the water, where they can be utilized by the algae.
The warm water temperatures also limit the  size of  the sport fishery of
the lake.  All of these factors  adversely affect the water quality  of
Klamath Lake and retard its use  as a recreational facility.

                       The Economic  Problem

The ultimate problem facing the  economist is  to estimate how much it is
worth to society to improve the  water quality of Klamath Lake. Or, to
restate the problem in a  more general way,  what are the benefits that
would result from a water quality improvement at Klamath Lake?

-------
The benefits associated with water quality improvements at Klamath Lake
are primarily recreational benefits.  Although recreation is only one  of
many possible uses of land and water resources today, its importance is
increasing as leisure time and the level of family income continue to
increase*  People appear to be willing to make considerable financial
sacrifices, in terms of travel costs to the site and equipment purchases,
to participate in outdoor recreation.  Thus, the estimation of recrea-
tional benefits has become a very important factor in the evaluation of
natural resource projects.

Of the various competing uses for natural resources, the evaluation of
the economic benefits associated with outdoor recreation is especially
difficult.  Unlike many of the other uses for natural resources, an ade-
quate market does not exist for outdoor recreation.  That is, "outdoor
recreation" is not a commodity that is purchased and sold at a given
price.  Instead, recreational resources have traditionally been con-
sidered a "free" commodity.  This has necessitated the development of
other methods for evaluating recreational benefits.  Several methods
have been used to estimate the demand for recreation with varying de-
grees of success.  This study presents a method for estimating the de-
mand for a recreational site which is new in several important aspects.

Another problem confronting the economist is the determination of which
benefits should be included in the evaluation of a project.  Other bene-
fits, in addition to the recreational benefits already mentioned, can
also be attributed to water quality improvements at Klamath Lake.  For
example, as the quality of water improves and Klamath Lake is used more
extensively as a recreational facility, the local economy will be the
recipient of some economic benefits.  As more recreationists use the
lake, additional goods and services will be purchased from the local
community.  As business activity in the community increases, household
income will also rise.  Thus, the community benefits indirectly from the
improved water quality.  The indirect or "secondary" benefits differ from
the recreational benefits in that the latter accrue to the nation as a
whole, while the former, as measured in this study, accrue only to the
local region.  Depending upon the prevailing conditions in the local and
national economies, all or a part of the secondary benefits may also
accrue to the nation as well as to the local region.  However, it is
difficult to determine the magnitude of the secondary benefits that should
be included in the national accounts.  The theoretical and empirical issues
in the determination of secondary benefits, given the national viewpoint,
are discussed in Seattle [1970, pp. 76-93], and will not be reconsidered
in this study.

Estimation of the regional benefits may provide other useful information.
For example, the local community may be willing to pay part of the cost
of the project.  The magnitude of secondary benefits would set the upper
limit to the local community's willingness to participate in such cost
sharing.  Also, estimation of the regional secondary benefits is a pre-
requisite for estimating the level of these benefits that may be relevant
from the national viewpoint.

-------
                            Objectives

The specific objectives of this report are  four-fold.  The  first  is  to
present a methodology appropriate for determining  the  economic benefits
accruing to society from the development of a  recreational  facility.
The second is to determine the relationship between water quality and
recreational use of a facility by using the new methodology.  This mafyes,
it possible to predict the change in recreational  use  that  would  accom-
pany a substantial improvement in water quality.

The third objective is to determine the economic benefits accruing to
society in general as a result of water quality improvements  and  the
associated increase in recreational use of  Klamath Lake.  The final  ob-
jective is to estimate the benefits that would accrue  to the  local econr
omy as a result of the hypothesized increase in recreation  at Klamath
take.

As the objectives indicate, the research presented in  this  report serves,
two purposes.  First, it provides specific  information concerning the
feasibility of allocating public funds for  water quality improvement £n
Upper Klamath Lake.  Second, and perhaps more  important, is the evalua--
tion of the methodology presented herein, so that  it may be applied  to
the estimation of recreational benefits resulting  from water  quality
improvements at sites other than Klamath Lake.

                    Organization of the Report

The next two sections consider the theoretical concepts involved  in  the
estimation of demand relationships for outdoor recreation.  Conventional
economic demand theory, and some of the previous work  done  in estimating
the demand for outdoor recreation, are reviewed in Section  IV.  These
topics provide the background necessary to  introduce the new  model used
in this study.  The theoretical model is also  presented in  Section IV.

Section V contains a discussion of the statistical model for  recreational
demand.  Each variable in the model is discussed.  Sampling techniques
and measurement problems are also described.   The  application of  the model
is made in Section VI.

Section VII explains the procedures used to estimate the local economi^
impact of increased recreational use of Klamath Lake.  Since  input-ou|put
techniques are used to quantify the regional benefits, the  theory of
input-output analysis is explained briefly. The model of the Klamath
County economy is also presented in Section VII.

Sections VIII and IX contain the empirical  analysis and the results  qf
the input-output study.  Finally, the effects  of two hypothetical impirovew
ments in water quality of Klamath Lake are  estimated in Section X.  This
section utilizes the results obtained from  the demand  analysis as well
as those resulting from the input-output study.

-------
                            SECTION IV

         A THEORETICAL MODE7! TO ESTIMATE RECREATION DEMAND


       Previous Work in Estimating the Demand  for  Recreation

Many outdoor recreational services are provided publicly.  As  the supply
of these services has not relied upon the market system,  the economic
evaluation of the benefits derived from the  consumption of outdoor  recre-
ation cannot depend upon the methods relevant  for  evaluating market
goods.  A different set of techniques has to be developed.

The estimation of the value of outdoor recreation  has proceeded  in  two
general directions.  Both seek to determine  the amount recreationists
are willing to pay to be able to use a certain facility.  One  approach
is the "direct" method of estimating the consumer's willingness  to  pay
[Knetsch and Davis, 1966].  The recreationists are asked, by means  of a
personal interview, to state how much they would be willing to pay  for
the use of the recreational facility, rather than  be excluded.   The de-
mand estimates obtained in this fashion are  defensible on theoretical
grounds, but the degree of reliability which can be placed upon  the re-
spondent's answers is often difficult to determine.  Biases are  likely
to exist, especially when questions are asked  which deal with  matters of
opinion concerning a person's activity that  has customarily been regarded
as "free".

One source of bias is the consumer's understatement of his preference for
a facility if he fears that a charge may be  levied for future  use of the
facility.  He may feel that understating his level of benefits may  enhance
his possibility of enjoying the facility in  the future while paying some-
thing less than his actual willingness to pay.  On the other hand,  if the
recreationist feels that the recreational service  in question  will  con-
tinue to be publicly provided without the levying  of user fees,  he  may
overstate his willingness to pay.  In doing  so, he may expect  to combat
competitive pressures for non-recreational uses of the facility  and thus
support the case for improvement and/or preservation of the recreational
site.

One advantage of the direct method is that it  is not restricted  to  esti-
mating only the effective demand for outdoor recreation.  It can also
estimate the demand of those people who are  not participating  in the
activity at the present time, but who may wish to  participate  at a  later
date.  These people would be willing to pay  to preserve the option  of
participating in the future.  Unfortunately, measurement  problems are
also present here because of the hypothetical  nature of the questions
posed to the prospective recreationists.

The second approach for determing the amount recreationists are  willing

-------
to pay for the use of recreational facilities relies upon the reactions
of recreationists to changes in costs of participating in recreation  at
the recreational site.  Willingness to pay can be computed from this
"indirect" evidence.  The procedure is limited in that inferences must
be made to the entire population consisting of recreationists and non-
re creationist s ;  only the effective demand can be analyzed.  This approach
is utilized in this study.

Hotelling, in a letter published by the U.S. National Park Service  in
1949, defined concentric zones around a recreational site in such a way
that the cost of traveling to the site from a given zone would be approxi-
mately constant.  His suggestion was to use the travel cost that existed
within each zone as the price variable to be compared to the number of
visitors from each zone.  A demand function for recreation could then be
obtained.

Although Clawson [1959] recognized problems in this formulation, he used
the basic idea underlying Hotelling's approach and gave it an interpreta-
tion that further facilitated the measurement of recreational values.
Clawson envisioned deriving two demand relations.  The first was expressed
as a relationship between the level of travel cost and the rate of  parti-
cipation of a population group derived in the manner suggested by Hotel-
ling.  From the first demand relation he derived a second demand schedule.
The rate of participation from a given population group, for a certain
fee increase, was predicted by referring to the observed participation
rate of another population with travel costs equal to the travel cost of
the group in question, plus the fee increase.  A demand schedule for  a
population group was obtained by relating various fee increases to  the
resulting participation rates.  The participation rates were multiplied
by the number of people in the group to estimate the quantity of use  from
that population group.  This procedure was repeated for each zone,  and the
resulting demand schedules were added horizontally to obtain the aggre-
gate demand schedule for the recreational resource.

Brown e£ al. [1964] and Stevens [1966] further refined the Clawson  model
by incorporating income and the quality of the recreational experience
into the model.  Angling success per unit of angling effort was added as
an independent variable.  Work was also done on the inclusion of distance
as a separate explanatory variable.  However, it was found to be highly
correlated with transfer costs, due to the manner in which transfer costs
are defined.

Many important and restrictive assumptions are involved in the indirect
approach discussed up to this point.  First, it is implied that the re-
actions of recreationists to a fee increase would be identical to an
equal addition in the cash cost of travel.  No distinction is made  between
a recreationist's reaction to increased variable costs (daily costs while
at the recreational site) and increased fixed costs (costs of travel).
Secondly, in predicting a group's response to a change in costs by  ob-
serving other groups, it is assumed that all of the groups, stratified by
distance, face identical alternatives to the recreational resource  in
                                 10

-------
question.  Thirdly, it is assumed that recreationists  are  the same  in all
other respects, in all distance zones, so  that  they will react in the
same way to increases in the cost of  travel.

This study presents an extension of the  indirect  approach  to  the evalua-
tion of recreational resources.  It should be noted that,  independent of
the construction of the theory found  in  this section,  Fearse  [1968]  de-
veloped a very similar approach.  His approach  differs slightly in  the
technical development, and his application is entirely different.  The
theoretical development of the procedure used herein will  be  examined,
with the focal point being the individual  recreationist instead of  a
population group.  This eliminates some  of the  restrictive assumptions
made about the population groups in the  Clawson method.

                   Conventional Demand Analysis

Before presenting the theoretical model  used in this study to estimate
the demand for recreation, it may be  helpful to review the fundamental
concepts of demand analysis.  Economists and others familiar  with demand
theory may wish to proceed directly to the subsection  entitled "The Theo-
retical Model Applied to Outdoor Recreation".   Let us  begin by assuming
that there is one consumer and two commodities, Q- and Q.. It is also
assumed that the consumer prefers to  have  as much of each  commodity as
possible;  that is, both commodities  are considered desirable by the
consumer.  The objective of the consumer is to  maximize his satisfaction,
subject to a constraint - the amount  of  money he  has available to pur-
chase the two commodities in a given  time  period.

All possible combinations of commodities Q, and Q~ the consumer could
conceivably choose, if he were not constrained  by a budget, can be  en-
visioned in Figure 1 as lying above and  to the  right of the origin.   This
is referred to as the "commodity space".  Every point  within  the commo-
dity space represents a combination of commodities Q.  and  Q~  that may be
consumed.

The consumer is confronted with an infinite number of  combinations  in the
commodity space.  It is assumed that  the consumer has  the  ability to rank
all of the combinations in the commodity space  according to the level of
satisfaction that he receives from each  of them.  All  of the  information
concerning the consumer's satisfaction is  contained in his utility  func-
tion.  A utility function expresses the  relationship between  satisfaction,
or utility (U), and the quantity of commodities Q- and Q2  consumed:

                             U = U(q;L, q2)                    (1)

where q, and q~ refer to the amounts  of  Q. and  Q2 consumed.  It is  also
assumed that the consumer is consistent  in his  rankings;   if  he prefers
Combination B to Combination C, and C to A, he  must also prefer B to A.

Indifference curves are a helpful analytical device.   They are obtained


                                 11

-------
                                 B
                   12           3
Figure 1.  The Commodity Space for Commodities Q, and
                                     12

-------
by holding the level of utility  constant,  and  observing the  various
combinations of the commodities  that  are consistent with the fixed level
of utility.


                             Uo  = U(V q2>                    (2)

where U  refers to a constant  level of  utility.   The  indifference  curve
depicted by Equation (2)  is  shown in  Figure  2.

The indifference  curve shows the different combinations of Q,  and  Q~
that yield equal  satisfaction  to the  consumer;   he is indifferent, or
has no preference, between  the combinations  of Q, and Q2 that lie  on  the
indifference curve U .
                    o

It should also be noted that all combinations  of Q. and Q. which lie
above and to the  right of U are preferred to  those combinations along
U  , since the consumer enjoys  more  of at least one of the commodities.
For example, Points 8 and C in Figure 2 yield  a higher level of satis-
faction than Point A, since the consumer has more of  one commodity and
the same amount of the other.   Therefore,  Points B and C must lie  on
higher indifference curves.  Likewise,  all points below and  to the left
of U  are less preferred  combinations of Q,  and Q7 than those located on
U  . °                                                     ,
 o

The indifference  curve U  in Figure 2 represents only one of an infinite
number of indifference curves  for  the consumer.   Figure 3 contains a
portion of the indifference curves  defined by  the consumer's utility
function.  This is referred to as  an  indifference map.  Again, the curves
above and to the  right of another  indifference curve  represent higher
levels of utility or satisfaction  to  the consumer.

Some of the more  important  properties of indifference curves should be
noted.  The shape of the  indifference curve  can be ascertained from the
total derivative  of Equation (2):

                             au  .    . au   .                    ,0,
                      dUo = 3^ dql + 3q^  dq2                  (3)

Setting Equation  (3) equal  to  zero  and solving for dq^/dq., yields:

                      j   /j       3u   / 3u                    (,\
                      dq2/dql  = "  3^ ' 3q^                   (4)

Equation (4) represents the slope  of  the indifference curve.  Since the
consumer prefers  to have  more  of both commodities,  9U/3q. and 3U/3q  are
positive.  Therefore, the indifference curve has a negative  slope.  The
second derivative of the  utility function  is positive, which means the
curve is convex to the origin.

Another point worth noting  is  that  indifference curves cannot intersect.


                                 13

-------
                   1234
Figure 2.  An Indifference  Curve  for Commodities Q.. and
                                                                     >-  Qi
                                      14

-------
Figure 3.  An Indifference Map for Commodities Q. and Q-,
                                  15

-------
Since the curves reflect different levels of satisfaction,  two  curves  can-
not share a single combination of commodities.  Two curves  passing  through
the same point indicate that one combination yields two different levels
of satisfaction to the consumer.  Since it is assumed that  the  consumer
has the knowledge to rank all of the combinations of commodities in the
commodity space, such inconsistencies cannot exist.

Now that the concept of indifference curves has been introduced, let us
return to the commodity space.  It contains all possible combinations
of commodities Q- and Q« that the consumer could conceivably  consume.
However, it may not be possible for the consumer to choose  all  of the
points in the space, since he is constrained by the amount  of money he
has available to purchase the two commodities.

To determine the consumer's attainable set, assume that he  has  M^ dollars
to spend on commodities Q., and Q2-  If the consumer chooses to  purchase
commodity Q. only, he can obtain M./p1 units, where p. represents the
price of commodity Q-.  Similarly, if he chooses to purchase  Q. only,  he
can obtain M /p  units.  The budget constraint faced by the consumer is:


                        Ml = qlpl + q2p2                      (5)

Equation (5) is illustrated in Figure 4.  The consumer is able  to pur-
chase any combination of the two commodities in the set bounded by  the
horizontal and vertical axes and by the line representing Equation  (5).
This portion of the commodity space represents the consumer's attainable
set.  The slope of the budget constraint can be determined  from Equation
(5):
                        dqjL     P2
From Equation  (6) and Figure 4, it can be seen that a change in  the
price of either commodity will change the slope of the budget constraint
line.  However, a change in the consumer's money income will cause a
parallel shift of the budget constraint.

We are now ready to determine the "most preferred" combination of commo-
dities Q.^ and Q2-  The consumer will choose the combination of commodi-
ties Q., and Q2 to maximize his utility, given his income constraint.   This
is equivalent to maximizing the following function:


                                       " qlpl " q2p2}         (7)

where X is the Lagrangian Multiplier.  The conditions for maximizing  con-
strained utility are fulfilled when the partial derivatives of Equation
(7), with respect to q^, q2> and X are set equal to zero:

                        9y    3u
                        K   ** n   — Ap_ = 0
                        3*!   3qx     i
                                  16

-------
                                   qlpl + q2P2
Figure 4.  The Consumer's Budget Constraint,
                                                       M,
                                 17

-------
                        \/ V    v ^J    ^     /%
                        3q0   3q_    P2
                          /     f.

                        3V=M_      _qp=0
                        3X    1    11    22

The consumer maximizes his constrained utility by consuming those quanti-
ties of 0^ and Q2 that satisfy Equation (10) and the following  first-
order condition:

                        3U   /3U    Pl
                               3q2 ~ ?2
That is, Q- and Q2 will be consumed until the ratio of their marginal
utilities equals the ratio of their respective prices, and all income
is spent.

The maximization of constrained utility is illustrated graphically  in
Figure 5.  The consumer's budget constraint is superimposed on his  in-
difference map.  Money income is M-, and the price of commodities Q^
and Q2 are p. and p2, respectively.  Indifference curve U2 represents
the highest level of utility that is attainable with the given budget ,
constraint.  Therefore, the consumer will maximize his utility by con-
suming q? units of Q., and q? units of 0_.

The Demand Curve

The consumer's demand curve for a commodity is derived from his  indif-
ference map.  A demand curve is a schedule that shows the various quan-
tities that the consumer will purchase at various prices.

Assume the price of commodity Q_ and money income are fixed at p? and
M. , respectively, while the price of commodity Q- is varied.  As shown
in Figure 6, the quantity of Q., purchased will vary as its price varies.
When the price of Q- is p9, the consumer will purchase q9 units  of  the
commodity.  When the price of QI increases to p' and p", the individual
will purchase only q! and q2 units, respectively.  Thus, as the  price of
Q. increases, the consumer purchases less of the commodity.  The indi-
vidual's demand curve is derived by plotting the various prices  and the
respective equilibrium quantities purchased, as shown in Figure  7.

It should be noted that the derivation of the demand curve is contingent
upon the continued optimizing behavior of the consumer.  This is illus-
trated by the tangencies at Points A, B, and C in Figure 6.  Any changes
in the consumer's utility function, income, or the price of commodity
Q2 will shift the individual's demand curve for 0^.  A market demand
curve for a commodity is obtained by horizontally adding the demand curves
of all individuals in the market.

One other concept needs to be discussed before introducing the demand
model f)br recreation.  It is "consumer's surplus".  Let us begin with


                                 18

-------
                                                         u.
Figure 5.  Maximization of Consumer's Utility.
                                 19

-------
                                                 u.
1 \ V-^X
1 \ «\^
1 l\ 1 \
I 1 \ 1 \
	 	 . U2
V— ui
\
\
*i qi!i q° !i ^L
—ii •*• _ i "_
Figure 6.  Quantities of Q.  Purchased at Various Prices.
                               20

-------
  o
 PI
                                                     Quantity of Q,

                                                     purchased
Figure 7.  The Consumer's Demand Curve for Commodity Q,.
                                 21

-------
the notion that a consumer pays a price for a commodity  that  is  less  than
or equal to the benefit he receives from the commodity.  This  can be
illustrated by an individual's demand curve DD, as shown in Figure  8.
Assume that the market price for the commodity is two dollars.   At  that
price, the consumer demands five units of the commodity.  He  pays two
dollars for each of the five units of the commodity.  However, the  con-
sumer's demand curve shows that he would be willing to pay more  than  two
dollars for the first unit of the commodity.  In fact, the consumer would
be willing to pay six dollars.  However, since the first unit, like all
other units, is sold at the market price of two dollars, the  consumer
receives six dollars worth of satisfaction for only two  dollars.  Thus,
he enjoys a "surplus" by receiving excess benefits from  the first unit.
The same situation exists for the second, third, and fourth units of  the
commodity;  the consumer is paying less for those units  than  he  would  be
willing to pay.  Therefore, those units of the commodity result  in  a
surplus to the consumer.  The difference between the price the consumer
is willing to pay for the units of the commodity and the price he actually
has to pay for them is called "consumer's surplus".  Consumer's  surplus
is used later in this study to estimate some of the values associated
with the demand for outdoor recreation.

        The Theoretical Model Applied to Outdoor Recreation

To analyze the economic behavior of recreationists, it is necessary to
derive a model which will account for the number of visitor-days recrea-
tionists will take at various levels of expenditures.  This will require
a slight alteration of the conventional demand analysis  outlined above.
The theory conceptualized in this section comes from an  unpublished paper
by Dr. John A. Edwards, Department of Agricultural Economics,  Oregon
State University.

In the presentation above, it was assumed that to purchase a  unit of
either commodity, the consumer had to pay the relevant prices, p- and  p...
Now suppose the consumer must pay, in addition, a certain charge, or  cost,
referred to as k, in order to buy any units of Q .  However,  the value
of k does not depend on the amount of Q. purchased.

Recreation is a good example where k is relevant.  In order to enjoy any
amount of recreation, Q., the recreationist must incur a certain price
per day, p., while recreating.  However, he must also travel  to  the re-
creational site.  The travel costs, including the transportation cost,
food, lodging, camping fees, etc., that occur while enroute to and  from
the recreational site, do not depend upon the quantity of Q.  consumed,
and will be referred to as k.                              *

The recreationist will allocate his income in order to maximize  his util-
ity, U:
                 U = U(qr q2)       qlf q2 >_ 0               (12)

where q-j^ indicates the number of recreation-days the recreationist  en-
joys at a particular site per visit, and q2 represents the amount of  all


                                 22

-------
Price ( )
 M
                                                              Quantity
 Figure  8.
  1234567
An Illustration of Consumer's Surplus.
                                  23

-------
other goods and services the recreationist could purchase with his  in-
come.  The recreationist is again limited to a fixed budget, which  im-
poses a constraint upon him.  It is recognized that the recreationist
also faces a time constraint and, in some cases, this may be more severe
than his budget constraint.  However, the time constraint is not con-
sidered explicitly in this analysis.
The income, y , allocated to the consumption of the  two commodities  must
just equal the total amount spent for the recreation-days  commodity,
p.q. + k, plus the total expenditures for the non-recreation-days  commo-
dity, Pq:
   yo » p^ + k + p2q2    yo, PI, P2 >_ 0;    q.^ q2» k > 0    (13)

Thus, the recreationist will maximize the Lagrangian function

                 V - U(qlf q2) -I- X(yo - p^ - k - p^)       (14)

by consuming those amounts of q^ and q2 that satisfy the budget  constraint
and the following first-order condition:
The consumer will consume recreation and non-recreation up to  the point
where the ratio of the marginal utilities associated with recreation  and
non-recreation is equal to the ratio of their respective prices.

The budget constraint can be rewritten as:

                      yo - k - p^ + p2q2                     (16)

Equation (16) more clearly illustrates the importance of k.  It reduces
the income available to the consumer for purchasing the commodities.   It
is assumed that k can be equal to zero if, and only if, no recreation is
consumed :

                         k - 0 <==> qx • 0.

This assumption points out a unique characteristic of the budget con-
straint.  If k - 0, the budget constraint takes on a new form:
yo * P2q2     for V P2' - 0;    q2 >
                                              0;
If faced with the budget constraint given in Equation  (17), the  consumer
would maximize his utility by taking as many non-recreation units as his
income would allow;  he would consume q9 * y /p7 units of Q-, and no units
of recreation, Q..                          °
                                 24

-------
Variations in Travel Cost  (k)

The indifference map and budget  constraints  of a typical consumer are
presented in Figure 9.  The  two  prices,  $i and p2,  and the level of income
are held constant.  The only variation in the budget constraints is due to
changes in k.  The budget  constraint BC  is  one for which k = k , a large
positive value.  The point on  the vertical axis at  q2 - y /p° illustrates
the discontinuity in the budget  constraint.   If the consumer allocated
all of his income to the consumption of non-re ere at ion, he could attain
the position y /p«, since  he will not have to withstand the additional
              Q  £                 	il-.l-.L-l-L
travel cost required to enjoy  recreation, as he must at all other points
on budget line BC  .
Given the set of  indifference curves and the budget constraint BC  in
Figure 9, the consumer will prefer the combination {q..  = 0,  q_ - y /p°}
over any other  attainable combination.  He will enjoy a level of utility
of U-.  If he chose  any  other attainable set of the commodities, he could
attain only  a level  of utility of U .

If the travel cost were  less than previously considered, say k - k , the
consumer would  be faced  with the budget line BC.  in Figure 9.  Shifting
the budget line to the right indicates the availability of more income
to allocate  to  the two commodities;  k  - k. more dollars are available.
Now the consumer, in maximizing his utility, has two alternatives:  (1)
he can take  the combination {q1 » 0, q_ = y /p9), or (2) he  can consume
                (ci.)         (ci) i
the set lq,  » q}   ,  q, • 

Based on the theoretical concepts developed up to this point, recreation-
ists will tend  to spend  fewer days at a site, per visit, as  costs of
travel increase.  There  is, however, a limit as to how high  the travel
costs can become, beyond which the consumer will not recreate.  In Figure
9, when k «  k., the  consumer is indifferent between recreating and not
recreating,  while if k < k., he prefers to recreate a certain amount,
depending on how  much  smaller k is than k^  If k > k^ the  consumer would
maximize his utility by  not recreating, as when k « k .  The travel cost
of k » k.  will  be referred to as the "critical" travel cost, and will be
        .,
denoted k*.   In  this  case,
                            *
                                       for PI = p°             (18)


                                  25

-------
  Non-

recreation
                                                  y -k   Recreation (Q )
                                                   O  2            1
  Figure  9.   The Optimal Combinations  of Recreation and Non-Recreation


              Given Travel Costs of k  ,  k.,  and k,, for Given Prices p?
                   Q                 O    1       L                    J.

              and p_, and Fixed Income  y .
                                  26

-------
The effect on the number of days of recreation  demanded, per  visit,  is
zero for decreases in k when:

                           k - Ak  > k*

That is, if the value of k decreases, but  the resulting travel  cost  is
still larger than the critical travel cost, k*,  the person will not  re-
create.  For any decrease in k such that:

                           k - Ak  < k*

the effect on the quantity demanded for  recreation will be the  same  as an
increase in income.  The effect on the amount of recreation consumed de-
pends on how much smaller k is than k*,  that is, upon  the size  of  (k* - k)

The value of the critical travel cost can  be expressed as a function of
four independent variables,  k* depends  upon on-site costs, the cost of
one unit of q-, the  level of income, and the utility function:

                        k* - k*(Pl, p2,  y, U)                 (19)

As p^ or p2 increase, the value of k* will decrease, due to the fixed
budget available:  3k*/3px < 0, 3k*/3p2  <  0.  As the level of income in-
creases, ceteris paribus, so will  k*;  3k*/3y > 0, due to the nature of
recreation;  that is, it is a normal good. The effect of U on  k* will
be discussed later,  in conjunction with  the utility variable.

The conclusions drawn in this analysis hold for any system of indiffer-
ence curves in which both commodities have positive marginal  utilities
and that possess the general shape consistent with maximization of util-
ity, that is, they satisfy the first- and  second-order conditions.   The
commodities Q, and Q. must be defined in such a way that the  indiffer-
ence curves intersect the q_ axis, that  is, it  must be possible to have
positive utility while consuming no Q..

Changes in On-Site Costs (p.)


Changes in the relative prices of  recreation and non-recreation commodi-
ties have a different effect on the demand for  recreation, q.,  than  did
the travel cost considered above.  A change in  either  of the  prices, p^
or p., will directly affect the optimal  budget  allocation of  the consumer.
The indifference map and budget constraints of  an individual, similar to
the one presented in Figure 9, may be analyzed  analogously.

By varying the on-site costs, p^  the resulting changes in q^^ can be
observed, and a critical on-site cost, p*, can  be defined.  The value of
the critical on-site cost depends  upon travel costs, the level  of  income,
the price of other commodities, and the  utility function:

                   Pf - pj(k, 7, P2. u>


                                 27

-------
The effect of k on p. can be illustrated.  As k Increases, the critical
price of recreation decreases, 3p./3k < 0.  On the other hand, as the
level of income increases, it is expected that the value of the critical
price would also increase, 3p-/3y > 0, due to the fact that recreation is
a normal good.  As the price of q. increases, the critical on-site costs
will tend to decrease:  3pj/3p2 < 0.

In general, for given conditions relating to preferences, income, and
prices of other commodities, there exists a critical value of p.^ for every
value of k, and a critical value of k for every value of p..

The model used in this study to determine how many days of recreation will
be taken, per unit of time, can be written in three structural equations:

                  qL = qi[(k* - k), (p* - PI)]                (21)


                  k* = k*(?1, p2, y, U)                       (22)

                  p* - p*(k, y, p2, U)                        (23)

In Equation (21) it is hypothesized that the quantity of recreation de-
manded per visit will increase as the difference between the travel cost
and the critical travel cost become greater, and as the difference between
on-slte costs and critical on-site costs become greater:
                             and - -  - > 0.
               3(k  - k)

The empirical specification of this model will be discussed in the next
section.  The empirical content of each variable will be specified in
order to determine the statistical model.  Sampling procedures are also
discussed.
                                  28

-------
                            SECTION V

               EMPIRICAL SPECIFICATION OF THE MODEL
Some of the variables in the  theoretical model  cannot be measured  directly.
A good example is the utility function of  the recreationists.   Before  the
hypotheses developed in the previous  section can be  tested,  a  statistical
model must be developed from  the  theoretical model.  In this section,
empirical implications of each variable in the  theoretical model are dis-
cussed .

A general comment should be made  before discussing specific  variables.
Several of the variables share a  common property that cause  measurement
problems.  The magnitudes of  many recreational  expenditures  increase near-
ly proportionally to increases in the size of the recreational group.
Examples are food, lodging, and other expenses  related to individual con-
sumption.  To reduce the variation in the  variables  that is  due to the
size of the group, each variable  is expressed on an  individual basis.  The
expenses that a group incurred are divided by the number in  the party.

A disadvantage of expressing  the  cost variables on an individual basis is
that economies of size may be realized when they are not relevant. For
example, automobile expenses  remain nearly constant  as the size of the
group increases, as long as the number of  autos does not increase. How-
ever, the average automobile  cost per person would decrease  as the size
of the group increases.  The  low  cost-per-person is  considered unimportant
in the decision making if the recreational group is  a family unit, since
all expenses are paid by the  family.

The biases that are forthcoming as a  result of  expressing the  variables on
a per-person basis are not thought to be serious.  It is believed  that the
economies of size bias are less serious than the biases which  would be
introduced by using group observations.

                    Discussion ofthe Variables

Days of Recreation per Visit  (q.)

The questionnaire used for data collection was  designed to determine when
the recreationist arrived at  the  site, and when he planned to  leave.   From
this information, the number  of days  the recreationist stayed  at the site
can be determined.

Travel Cost Ck)

The travel cost the recreationist incurs to recreate consists  of  the  cost
of transportation, food expenditures, lodging,  camping  fees, and  any  other
                                  29

-------
expenses encountered while  traveling  to  or  from the recreational site.
Some  of these  components will be  studied separately.

      Cost of Transportation;  The transportation cost  is  the
      amount it costs the recreationist to drive to  and from
      the recreational  site.  This includes  gasoline, oil,  de-
      preciation, insurance, repairs,  and any  other  miscellaneous
      items involved with operating an automobile.   These  costs
      were not  specified in  the questionnaire.   The  transporta-
      tion cost was computed by multiplying  the  total number of
      miles traveled, in both directions,  by a cost-per-mile of
      five cents.  The  cost-per-mile figure  was  determined  from
      previous  studies  as well as  current research,  in  which
      recreationists were asked to enumerate their transporta-
      tion costs  [Guedry, 1970, and Stevens, 1966].  Although
      none were encountered in the study,  other  means of trans-
      portation can be  accommodated in the model.

      Food;  The recreationist who spends  time away  from home
      must make arrangements for his food consumption.   He  may
      prepare food at home to take with him, or  purchase the
      ingredients at grocery stores to prepare along the way,
      or he may patronize restaurants  and cafes.  In any case,
      the recreationist may end up spending  more, less,  or  the
      same amount of money for food while traveling  as  he would
      have spent if he  had stayed  home.

      The relevant food expenditure made  by  the  recreationist
      is not the total  amount spent for meals while  traveling.
      The amount that he would have spent  at home, had  he
      chosen not to recreate, should be deducted from his food
      expenditures associated with the recreational  visit.   In-
      formation was not collected  to determine how much the
      recreationist would spend for food while at home.  How-
      ever, U.S. Department of Agriculture data  [1968,  Table
      2, p. 7]  was adjusted to indicate the  average  daily food
      costs per person  for various levels  df income  in  1967.
   '   The appropriate average expenditure  for  food at home  was
      subtracted from the average  expenditure per person for
      food while traveling.  The difference, which,  in  some
      instances was negative, was  added to the calculations
      of the travel cost variable.  If no  expenditure was made
      for meals while enroute, this adjustment was not  made.

      Other Costs of Travel;  The  cost of  lodging, camping
      fees, and any miscellaneous  expenses while  traveling
      were obtained in  the personal interview.

On-Site Costs  (p.)
                                      f

On-site costs  are the  total costs  incurred by the recreationist  per  day
while visiting the recreational site.  They include the costs of lodging,


                                  30

-------
camping fees, equipment rentals, meals,  and  other miscellaneous  expenses
incurred at the site.  The total expenses  that  the recreational  group
incurred per day are divided by the number of persons  in  the group to
obtain the daily on-site cost per person.

The daily food cost per person, while  recreating at the site,  is computed
from information obtained in the interview.  For reasons  mentioned above,
the average daily expenditure per person for food while at  home, within
the corresponding income groups, is subtracted  from the calculated on-site
food cost per person per day.  The difference in the two  figures, whether
positive or negative, is used in the calculations to determine the p.
variable.  All of the remaining components of the on-site cost variable
are accounted for in the questionnaire.
                                                         j
                                                          !•
Income (y)

In this analysis the family income of  the  recreationist,  after taxes,  is
used.  The questionnaire was used to obtain  data about the  size  of the
recreationist's family, the age of the household head, and  the gross
family income.  It was then possible,  using  a Federal  Income Tax table,
to estimate the amount of federal taxes  paid in 1967.  It was  not feas-
ible to estimate the amount of state taxes paid, since recreationists
came to Oregon's lakes from several states.  Thus, the income  used in
this analysis is biased upward.  However,  the error is not  considered
significant.  Federal income taxes are usually  much larger  than  state
taxes.  The majority of the relevant taxes have been subtracted.
                                                        )
Price of Other Commodities (p.)

The price of other commodities, in theory, is equal to the  weighted aver-
age of the prices of all other commodities that the consumer may choose
to purchase:

                      P2 = V3 + V4 + — +«n-2*n
where p. is the price of the i   commodity,  and a is the  weighting fac-
tor of the i   commodity.  In practice,  however, the commodities that
comprise the alternatives to recreation  are  so  numerous and so diverse
among individuals that it is impossible  to specify them.  Therefore,  the
P2 variable was deleted from the statistical model.

Utility  (U)

The utility variable in Equations (22) and (23) is important in  deter-
mining the critical values of travel cost  and on-site  costs.  A  change
in the shape of the indifference curves, through a change in the utility
                                  ^t      ^t
function, will cause a change in k  and  p..  Unfortunately, the  utility
function is not measureable.  However, it  can be represented,  in part, by
several other variables.  Some of the  variables considered  important  in
this respect are the characteristics of  the  recreationist,  jthe character-
istics of the site, and the value of recreational equipment.
                                  31

-------
Characteristics of the Recreationist;  The recreation-
ist's characteristics influence his decisions concern-
ing the consumption of recreation.  The background of
the person, his stage in the family cycle, his age, and
so on, influence his utility function.  However, with
the exception of income, these personal characteristics
of recreationists were not taken into account in this
study.  In another study conducted at Oregon State Uni-
versity, Guedry [1970] specified the relationship be-
tween the characteristics of the individual recreation-
ist and the demand for recreation.

Site Characteristics;  Recreationists choose a site be-
cause of the "desirability" of its characteristics.  If
a study is conducted to determine the recreational use
and value of a particular site, and information is gath-
ered only at that site, the characteristics of the site
are fixed and do not need to be analyzed explicitly.  How-
ever, if data are collected at more than one site, and
each of the sites has different characteristics, the char-
acteristics may explain the variation in quantity of recrea-
tion days observed at the sites, and the characteris-
tics that are desirable to recreationists could be iden-
tified.  Determination of the characteristics that have
a substantial effect upon the use and value of the site
can help decision makers plan site alterations to gain
the user benefits from favorable characteristics.

The specific site characteristics of interest in this
study are the size of the lake (measured in acres),
and the level of use of each lake for various recrea-
tional activities.  The activities considered are swim-
ming, boating, water skiing, fishing, and camping.  Dis-
crete values were used to indicate the amount each of
the lakes was used for each activity.  If an activity
was non-existent at a lake, it was given the value of
zero.  Low, medium, and high uses of the lake for a given
activity were assigned the values one, two, and three,
respectively.  The use-intensities were estimated by U.S.
Forest Service personnel and employees of the E.P.A. who
were knowledgeable about the lakes' characteristics.
It is recognized that the values are subjective.  Table
1 contains the estimated use-intensities for each activ-
ity at each lake in the study area.  The selection of
the four lakes used in the study will be discussed later
in this section.

Investment in Recreational Equipment;  It is hypothesized
that the value of recreational equipment owned by the
individual may serve as a proxy for a person's utility
function with respect to recreation.  An individual with
                            32

-------
       Table 1.  Estimated Use-Intensities  for Certain Recreational
                 Activities and Lake  Size,  by  Lake
CO
Use- Intensities
Lake
Lake of the Woods ....
Odell 	
Willow 	

Swimming
3
0
1
0
Water
skiing
3
0
1
1
Boating
3
1
1
1
Fishing
1
3
3
1
Camping ,
picnicking
3
2
2
0
Lake
size
(acres)
1,055
3,500
320
98.560

-------
     a large quantity of recreational equipment, such as boats,
     campers, trailers, and so forth, probably has a stronger
     preference for recreation than a person with less equip-
     ment.  A qualification is necessary, however.  This pheno-
     menon would be expected to be observed only when a compari-
     son is made among similar activities.  Wilderness-type activ
     ities cannot be compared to those at a more accessible site,
     since each requires a unique combination of equipment.

     Investment in recreational equipment may also be a useful
     substitute for income as an explanatory variable.  Two rea-
     sons seem apparent.  First, investment is more correlated
     with the permanent income of the recreationist than with
     current annual income.  This is important, since permanent
     income is considered more closely associated with the quan-
     tity of recreation taken than is current income.  A case
     in point is a retired person.  He may have a low current
     income, but the amount of equipment he owns is related to
     his previous permanent income.  On the other hand, a young
     person beginning his career probably purchases equipment
     with the future in mind.  Even though he may have a low
     current income, it may be his expected permanent income
     that determines his purchases of equipment.

     There is another closely related reason why the amount of
     recreational equipment may be more useful as an explanatory
     variable than is current income.  While the recreationist*s
     income stream may be stable over time, his expenditures
     for items other than recreation may fluctuate widely.  These
     fluctuations may be due to the "lumpiness" of certain house-
     hold expenditures, as well as to the occurrence of events
     in a person's life which necessitate unusually high expendi-
     tures during a given period for which no plans had been
     made.  Cash outlays for outdoor recreation during any one
     period may be subject to these fluctuations in expendi-
     tures for other commodities.  On the other hand, the con-
     sumer's investment in outdoor recreational equipment may
     reflect more nearly his long-run budgeting plans.
                                                   ^

                       The Statistical Model

The statistical model, when more than one site is analyzed, is:


   qi= qi[(k  ~k)> (pi" V1    for (k  ~k)> (p* ~ pi) - °

   k  = k (p^ y, Sw, Ws, B, F, C, I, Si)                        (25)

    *  ' *,  '
   pl = pl*k» y> Sw' Ws> B» F> C> I) Si)                         (26)

where Sw, Ws, B, F, and.C represent the site's use-intensity for swim-
ming, water skiing, boating, fishing, and camping, respectively,  i
                                 34

-------
indicates the value of recreational equipment  and  Si  indicates  the  size
of the lake, in acres.  All other variables  are  as defined  above.

The demand function of interest in the  study is  the aggregate demand
function for the total number of visitor-days  per  season.   The  preceding
model is appropriate for explaining the number of  visitor-days  per  visit.
That is, it pertains to an individual,  not to  the  population of users.
The total number of visitor-days can be obtained by multiplying Equation
(24) by the estimated number of visits  (V) to  the  site  during the season.
The appropriate aggregate model is:

   Vq-L =* f[(k* - k), (p* - PI)]     for (k*  -  k),  (p* - PI) > 0  (27)

    k* = k*(pr y, Sw, Ws, B, F, C, I,  Si)                        (28)

    p* = p*(k, y, Sw, Ws, B, F, C, I, Si)                         (29)

where V indicates the number of visits  to the  site during the season.
The total number of visits at each site during the relevant season  was
obtained from the U.S. Forest Service,  since the lakes  studied  had  facili-
ties maintained by the Forest Service.

                 The Number of Visits Relationship

One other relationship must be discussed.  The aggregate demand model
contains the V variable, which represents the  number  of visits  to the
site.  This variable can be estimated in two possible ways.

The first approach is to use the estimates of  the  number of visits  re-
ported by the U.S. Forest Service, and  have  it remain fixed with respect
to any changes in the independent variables  in the model.   The  second
method is to use the estimated number of visits, as above,  but  then ex-
press V as a function of some appropriate explanatory variables.  In this
way, when a postulated change occurs in the  independent variables,  the
effect upon the number of visits, as well as upon  the length of stay per
visit, can be observed.  The second approach gives a  more meaningful
estimate of the total number of visitor-days.

In line with the second approach, it is hypothesized  that the number of
visits forthcoming at a site for an individual can be expressed as  a
function of the travel cost, the income of the recreationist, and the
characteristics of the site.  Information was  not  obtained  from recrea-
tionists to determine the number of visits they  would make  during the
season.  Thus, it is impossible to focus attention on the individual
recreationist.  An alternative does exist, however.  The total  estimated
visits, V, can be allocated to regional population groups,  and  the  re-
lationship between the number of visits from the i   region and the ex-
planatory variables can be computed.  It is  hypothesized that the number
of visits forthcoming from an area, V,  is functionally  related  to the
average travel cost, k, of recreationists residing in the area, the
                                  35

-------
the average income of recreationists from the area, y, the total number
of persons living in the area, pop, and the characteristics of the  site
under consideration:

              V - V(k, y, pop, Sw, Ws, B, F, C, Si)            (30)

                        Sampling Procedures

To estimate the statistical demand model, primary data had to be gathered
from recreationists.  Since the primary objective of the study is to
estimate the recreational use of Klamath Lake with improved water qual-
ity, it is necessary to collect the data at lakes that have different
characteristics.  By observing the various characteristics of these
lakes and the intensity of each lake's use, it is possible to predict
the increase in use of Klamath Lake as its characteristics, such as
water quality, improve.

Three lakes, in addition to Klamath Lake, were chosen for the data  col-
lection process.  They are Odell Lake, Lake of the Woods, and Willow
Lake.  These lakes were chosen for several reasons.  First, many of the
recreationists living in the Klamath Falls area are now using the three
lakes, or other lakes in the vicinity of these lakes, for recreation.
Also, these lakes possess many of the characteristics, such as the  size
and the water recreation activities, that are of interest in this study.
Only lakes that were easily accessible and had overnight facilities were
considered, since these attributes were deemed a necessity for the  aver-
age recreationist to consider substituting that lake for Klamath Lake.

For sampling purposes the population was defined as the total number of
recreationists that visited the four lakes during 1968.  The sampling
unit is the recreation unit;  that is, the group that recreates together.

The appropriate sampling frame would be a list of all the recreation
units found in the area of study in 1968.  Of course, such a listing is
impossible to obtain.  Therefore, stratified random sampling techniques
were used to obtain the sample.

First, standard statistical methods were used to estimate the sample
size necessary to obtain the precision desired for the study.  A sample
size of 300 was estimated.  A complete discussion of the estimation of
the sample size is contained in Gibbs  [1969].

The total sample of 300 was then allocated to the four lakes on the basis
of the estimated number of recreation days taken at each lake.  Data
available from the U.S. Forest Service and other agencies provided  esti-
mates of the number of recreation-days taken at each of the four lakes
in 1967.  These were summed to obtain the total number of recreation-
days taken at all four lakes.  If a lake accounted for 30 percent of  the
total recreation-days, 30 percent of the sample size of 300 was allocated
to that lake, and so forth.
                                 36

-------
The area immediately surrounding  each  lake was  also divided into several
geographic blocks.  For example,  all camping areas, boat ramps,  picnic
areas, lodges, cabins, etc., were placed in  separate blocks.   The sample
size allocated to each lake was again  distributed to the various geo-
graphic blocks on the basis of  frequency-of-use of each of the blocks.
The same procedure was used at  each of the four lakes.

The proper time to conduct the  interviews also  had to be determined.   To
accomplish this, all possible weeks in the summer season were numbered,
and two weeks were chosen randomly. The total  number of interviews needed
at each block at each lake was  divided equally  between  the two weeks.
Estimates were then obtained from the  appropriate agencies as to the  rel-
ative use of the lakes during  the week and weekends. The sample was  then
allocated by this factor.  If  twice as many  people recreated on  weekends
as during the week, two-thirds  of the  sample data were  collected on  the
weekend.  Table 2 shows  the  final distribution  of the sample among the
sites, weeks and days for each  lake.

The recreationists within each  block  at each lake were  then selected  at
random to be interviewed.  Care was taken that  the sample was representa-
tive of  the activities at  the  site.  That is, if approximately two-thirds
of the people at a site were picnicking, and one-third  were water skiing,
the sample contained  twice as many picnickers as water  skiiers.   Care
was also taken to prevent  interviewing recreationists on the basis of the
type or  amount of equipment  they  were  using.  The questionnaire  used  in
the study is included in Appendix A.
                                  37

-------
Table 2.  Number  of Interviews Taken at Each Lake, by
          Block,  for the Four Time Periods.
                                            First  Week
                                                   Second Week
      Lake
    Block
Week
days
Week
end
Total
Week
days
Week
end
Total
Two-week
  total
 WILLOW:
KLAMATH:             Recreation Creek	  1      2
                     Odessa.	  0      1
                     Rocky Point Resort	  3      11
                     Moore Park	  2      7
                     Pelican Marina	  0      1
                     Yacht Club	  0      1
                     Public Boat Launch
                       & Bank Fishermen	  0      0
                     TOTAL KLAMATH LAKE	  6      23

LAKE OF THE WOODS:  Rainbow Bay	  5      17
                     Aspen Point	  3      13
                     White Pine	  0      2
                     Lake of the Woods
                       Resort	  3      11
                     TOTAL LAKE OF THE WOODS 11      43

ODELL:              Princess Creek	  6      7
                     Sunset Cove	  3      3
                     Trapper Creek.r........ 10      11
                     Pebble Bay	  0      0
                     Odell Creek....	  1      2
                     Odell Summit Lodge	  1      2
                     Odell Lake Resort	  1      1
                     Shelter Cove Marina....  0      1
                     TOTAL ODELL LAKE	 22     27
County Campground......  5	16

TOTAL	 44    109
                     3
                     1
                   14
                     9
                     1
                     1

                     0
                   ^tn^ivgak^
                   29

                   22
                   16
                     2

                   14

                   54

                   13
                     6
                   21
                     0
                     3
                     3
                     2
                     1
                   (•••^VO^
                   49

                   21

                  153
                                                     1
                                                     0
                                                     3
                                                     2
                                                     0
                                                     0

                                                     0
                                                    mmillitimm
                                                     6

                                                     5
                                                     3
                                                     0
          11

           6
           2
          11
           0
           1
           1
           1
           0
          M^Mtov
          22
                                                                         44
                              2
                              0
                             11
                              6
                              1
                              1

                              1
                            ••••••••HH
                            22

                            17
                            12
                              2

                            10
                            •^•M^MHUVI
                            41

                              7
                              3
                            10
                              0
                              1
                              1
                              1
                              1
                            ^^•••^••v
                            24

                            16

                           103
                                                                    46
                          21

                         147
                             42
                            300
                                                CO

-------
                             SECTION VI

       ESTIMATING AND  APPLYING  THE STATISTICAL DEMAND MODEL


The data collected  from  recreationists  at  the  four  lakes  in  the  study
area are used  to estimate  the equations of the hypothesized  statistical
demand model.  Four relationships  are estimated.  They  are:  p*   the
critical on-site cost  relationship;  k*, the critical travel cost
tionship;  q  , the  demand  relationship; and V, the number of visits
relationship.

Before each of the  relationships are discussed, a problem which  was en-^
countered in  three  of  the  equations will be discussed.  When the equa-
tions for critical  travel  cost, critical on-site cost,  and the number
of visits were estimated,  multicollinearity was observed  among the site
characteristic variables of swimming, water skiing, and boating  use-
intensities.  When  one of  the three use-intensity variables  entered int;p
the regression equation, neither of the other  two variables  could explain
enough additional variation in  the dependent variable to  be  significant.
Since the simple correlation coefficients  between the three  variables
ranged from .957 to .980,  it was not possible  to estimate the separate
influence of each of the three  variables.   Consequently,  the three vari-r
ables were combined into one variable.   The use-intensity rankings for
swimming, water skiing,  and boating were summed to  represent a single
variable for each lake,  denoted as W.

Another case of multicollinearity  was observed between  the camping in-
tensity variable and income.  This was  considered to  be a statistical
problem with no economic significance.   Because of  their  differing na-
ture, it was not feasible  to combine the two variables.   Instead, the    .
camping intensity variable was  deleted  from the model,  since camping inr
tensity may be more closely related to  the man-made characteristics of
the site than  the natural  characteristics  being considered in this study.

Statistical problems also  prevent  the use  of the recreational equipment
variable (I) in the study.   The problems can be illustrated  by discuss-*
ing the work of Guedry [1970].  In his  equation to  determine the number
of recreation days  taken,  per capita, in the Bend Ranger  District in
Oregon, the "investment  in recreational equipment"  variable  was  signifi^-
cant at the 2 percent  level.  However,  the sign of  the  coefficient was
negative.  The negative  sign is somewhat confusing, since it suggests
that the average length  of stay at the  recreational site  decreases as
the amount of investment in recreational equipment  increases.  This is
opposite from the effect that one  would expect.  To determine a  reason
for this, all variables  in the  equation that had a  simple correlation
coefficient greater than or equal  to .7 with the recreational equipment
variable were removed, and a new equation  was  estimated.  The new par-
tial regression coefficient for the equipment  variable  in the new equa-
tion was significant only  at the 50 percent level.
                                  39

-------
In a further attempt to clarify the problem, Guedry removed  all variables
with a simple correlation coefficient greater than or equal  to  .6 with
the equipment variable.  A new equation was then estimated,  and the  par-
tial regression coefficient of the recreational equipment variable was
significant at the 10 percent level, and the sign of the coefficient was
positive.  The erratic performance of the recreational equipment variable
makes it difficult to form any conclusions as to its role in the demand
equation.  Since the same type of problems were experienced  in this  study,
investment in recreational equipment is omitted as an explanatory vari-
able.  The variable appears to be too highly correlated with other vari-
ables in the model to be of significant value.                           „

The above changes leave the statistical model to be estimated as:

     q;L - q]L[(k* - k), (p* r PL)1  for (k* - k) and (p* - PI> >, 0    (31) .

     p* - p*(k, y, W, F, Si)                                         (32)


     k* = k*(pr y, W, F, Si)                                        (33)

     V  = V(k, y, pop, W, F, Si)                                     (34)

Each relationship will be discussed below.

                Critical On-Site Cost Relationship

Critical on-site cost is related to travel cost, level of income, and
site characteristics.  Since the estimation procedure is performed on
more than one recreational site, the observations were categorized by
sites, and placed in homogeneous groups determined by income level and
travel costs.  The income groups were determined as follows: the in-
comes of all recreationists were listed in descending order  of magnitude.
Consideration was given to the "natural" breaks in the list; but, due
to the small number of observations, only two cutoff points  were chosen.
The cutoff points were chosen to ensure that each group would have
approximately the same number of observations.  For all lakes, the three
income groups are:  (1) less than or equal to $8,000, (2) between $8,000
and $10,000, and (3) greater than or equal to $10,000.

The distribution of travel costs is different for each lake. For this
reason the groups were determined individually for each lake. The cut-
off points for Klamath Lake are $1 and $19.  That is, the three groups
are:  (1) less than or equal to $1, (2) between $1 and $19,  and  (3)
greater than or equal to $19.  For Lake of the Woods, the cutoff points
for k are $2 and $20;  for Odell Lake, $5 and $10;  and for  Willow Lake,
$2 and $10.

To estimate the maximum PI within each subgroup, determined  by  income
and travel costs, the variable on-site costs in each subgroup were ar-
ranged in ascending order of magnitude.  The last (or n  ) p.. is the


                                 40

-------
maximum p. observed  in  the  subgroup,  and is referred to as the nth order
                 4*^t
statistic.  The n    order statistic is used as a reliable estimate of the
maximum p  in the distribution [Hogg  and Craig, 1965, ch. 6],   The use
         •L*                                      j^
of the maximum observed p^  as  an estimator of p. does not imply that all
other p^'s are ignored.   The maximum p-  is chosen only after it has been
compared to all other p's  in  the subgroup.  Thus, all the sample informa-
tion is used.

As a result of the above classification technique, the number  of observa-
tions varies in each subgroup.  The reliability of the estimate of p..
depends upon the number of  observations in the subgroup.   Reliability
refers to the size of the variance.  If the variance is large, the esti-
mator is less reliable.   In utilizing regression analysis, it  is assumed
that the diagonal elements  in  the variance-covariance matrix are con-
    .'j....              ,              ^
stant, or nearly so.  In this  case, the p's of some subgroups have
higher variances than others;   thus,  the constancy of variances assump-
tion is violated.  The  coefficients estimated by ignoring this type of
problem will'be unbiased,, but  will not have minimum variances.  It is
necessary,'then, to  make adjustments  in the analysis to ensure unbiased,
minimum variance estimators.   Weighted least squares was  used  to correct
the problem.  The technique involves  assigning larger weights  to the
reliable estimates than to  the unreliable ones [Draper and Smith, 1966].
The weights are equal to the square root of the number of observations
in the subgroup.

The level of income  and the travel cost, from each subgroup, are used as
observations and are then regressed on the resulting p, *s (for each sub-
group) to obtain an  estimate of the relationship between the independent
variables and p?.  The  critical on-site cost relationship estimated by
weighted least squares  is:

     p* = -7.263 + 7.80 W### + 2.630  F### +  .000067 Si#// -  .004 k2#
      1            (.197)      (.815)       (.000025)       (.002)
                  *                 ")&       ")       ftiHt              (35)
          +  .269 k   +   .000000017 y V      R  - .684*
           (.143)      (.0000000094)        Degrees of freedom =27

where W, F, and Si are  the  site characteristic variables, as previously de-
fined, k is the average travel cost of the subgroup, and y is  the aver-
age income for the subgroup.   The standard error of the coefficients is
shown in parentheses below  the coefficients.  Each coefficient was tested
to determine if it was  significantly  different from zero.  The test is
made by dividing the coefficient by its standard error and comparing the
resulting value with the values in a student's t table.  A 1 percent
level of significance is referred to  by ###, a 5 percent  level of signi-
ficance by ##, and # indicates the coefficient is significant  at the 10
percent level.  If no marks are listed, the coefficient is sufficiently
different from zero  at  a level of probability greater than 10 percent.
                                  41

-------
Only two variables, W, and F, are significantly different  from zero  at
the 1 percent level.  The size of the lake variable is significant at
the 5 percent level, while travel cost, travel cost squared,  and  income
squared are all significant at the 10 percent level.

It was hypothesized that the site characteristics should have a positive
effect on the critical on-site costs.  The obtained results are consis-
tent with this hypothesis.  The conceptual framework of this  study also
suggested a negative relationship between travel cost and  critical on-
site costs.  Equation (35), however, indicates that the relationship is
positive.  The analysis suggests that as travel costs increase, the  crit-
ical on-site costs will also increase, but at a decreasing rate,  since
the k-coefficient is positive but the k  coefficient is negative.
                                                           *
Income has, as was hypothesized, a positive influence on p.^.   The relia-
bility of this relationship is not high, since the coefficient is signi-
ficant only at the 10 percent level.  The evidence indicates,  however,
that as income increases, the critical on-site costs will  increase at  an
increasing rate.  The difficulties of defining and obtaining  data for.the
appropriate income variable were discussed earlier.  Perhaps  these ex-
plain the failure of the empirical model to produce a statistically'more
significant result with respect to the income variable.
     2
The R  for the critical on-site cost equation is .684, which  is signi-
ficant at the 1 percent level.

From the estimated equation it is possible to see the predicted effect
on p, of changing the characteristics of the site.  For example,  if  fish-
ing intensity were to increase one unit, perhaps due to a  change  in  the
water quality, the recreationist would be willing to increase  the maxi-
mum willingness to pay at the site by $2.63 per day.

                 Critical Travel Cost Relationship

The critical travel costs, k fs, were also estimated by using  the n
order statistic after the observations had been separated  into subgroups
on the basis of lakes, income, and on-site costs.  The income  groups are
the same as those used in the critical on-site cost relationship, while
the subgroups defined by p.., for all lakes, are:  (1) < $1.50,  (2) be-
tween $1.50 and $2.50, and (3) >_ $2.50.  Average values" of income and
on-site costs were determined for each group.  The predicted k* equation
is:

     k* - -36.711 +  6.248 W * +  3.779 F +  .0003 Si +  .0020 y
                    (2.322)      (7.800)    (.0002)     (.0018)

          + 10.435 p/##                 R2 . .616*"             (36>
            (3.349)                      Degrees of freedom -  28

In Equation (36) only the on-site costs variable is significant at the
1 percent level, while the "water" variable, W, is significant at the
                                 42

-------
5 percent level.  All other variables  fail  to obtain  the  10  percent  level
of significance.

It was hypothesized that the  characteristics  of  the site  have  a  positive
effect on k .  That is, as the  characteristics of  the site improved,  re-
creationists would be willing to pay more in  travel costs.   The  sample
data supports this hypothesis,  since the W, F, and Si variables  are  all
positive.  However, due to the  low  significance  of the fishing intensity
and the lake size variables,  few conclusions  can be stated with  much  re-
liability concerning the existence  of  a relationship  between these vari-
ables and k*.

The sample data also suggests a positive relationship between  income  and
critical travel costs.  This  had been  hypothesized.   Again,  however,  the
low level of significance of  the income coefficient makes it impossible
to place much reliability upon  the  relationship.

The'lack of significance of the income variable  may be due,  in part,  to
the broad income groups used  to estimate k*.   Since only  three income
groups were used, the income  range  within each group  was  quite large.
More data would have made it  possible  to define  smaller ranges of income.
Then the variation in income, and its  effect  upon  k*, could  have been
more clearly observed.

On-site costs, statistically  the most  significant  variable in  the equa-
tion, exhibits a positive relationship with k .  However, the  theory  sug-
gested that as a person was required to spend more at a recreational
site, he would be willing to  pay less  in travel  costs, due to  his fixed
budget.  This hypothesis is rejected by the sample data.  Possible ex-
planations may include the fact that on-site  costs usually comprise  a
much smaller portion of the recreationist's budget than do travel costs.
Because of this, other effects  may  be  more  important  in the  decision-
making.  For example, in this study the more  desirable sites may have a
higher daily on-site cost.  Recreationists  may be  willing to pay higher
travel costs to visit the more  desirable site, even if on-site costs  are
slightly higher.

The coefficient of determination, R ,  for the k  equation, is  .616.   It
is significantly different from zero at the 1 percent level.

                    Days Per  Visit  Relationship
                                                       *     *
The above discussion refers to  the  estimation of the  k and  p^ values.
Each questionnaire contains data that  made  it possible to compute the
length of stay per visit, q-, the actual travel  cost, k,  and the actual
on-site cost, p,.  Therefore, the independent variables for  the  q1 equa-
tion, (p? - p,) and (k* - k), are obtained  by subtracting the  observed
        1    i                                 ju       A
p.. and k values from the appropriate group's  p..  and k  values.
                                  43

-------
 Since 304 observations  were collected,  there are 304 possible sets of
 values of (p? - p,)  and (k* - k)  to use in the regression equation.  How-
,ever, two distinct groups  of recreationists were observed in the sample.
 One group was characterized by spending less than three weeks at the site
 and having relatively high values of (p* - p.^.   The other (much smaller)
 group spent a longer time  at the  site (up to 169 days) and had relatively
 low values of (p* - p..).   This group included retired people spending
 most of the season at the  site, and persons such as teachers, who did
 not have summer job commitments.

 Further analysis of the recreationists  in the second group revealed that
 they did not respond significantly to changes in p1 and k.  That is^ there
 was no relationship between the length  of stay per visit and the (p. - p^)
 and (k* - k) variables.  Therefore, the group of recreationists that
 stayed at the site 20 days or more was  not used to estimate the q^ equa-
 tion.  The remaining 282 observations were used to estimate the follow-
 ing q.. relationship:

      Inq-  = .759 -  .0064  (k* - k)*** +  .0637 (p* - p.)***     (37)
         1          (.0018)              (.0189)   L

        R2 = .113***
        Degrees of freedom  =279

 The average length of stay per visit for recreationists in this group is
 3.4 days.

 It was hypothesized that as the difference between the critical travel
 cost and the actual travel cost increased, the number of days the recrea-
 tionist would remain at the site  per visit would also increase.  In
 Equation (37) a negative coefficient is estimated which is significant
 at the 1 percent level. In other words, evidence suggests that, for a
 given critical travel cost, as a  person's actual travel costs increase,
 he will tend to recreate more days per visit.  As was mentioned earlier,
 the appropriate income  in  the budget constraint depicted in the theory
 is (y - k).  As k increases, less income is available to be spent at the
 site after arrival - thus  fewer days will be forthcoming.  Apparently
 travel costs are so small, relative to  income, that they do not signi-
 ficantly reduce the income available for expenses at the site.

 It was also hypothesized that as  the daily on-site costs increased,
 given a fixed p^, fewer days per  visit  would be observed.  This hypothesis
 is substantiated by the analysis.  The  (p* - p ) variable is significant
 at the 1 percent level.
      2
 The R  for Equation (37) is .113, which is significant at the 1 percent
 level.  The R  is low;   yet it must be remembered that in contrast to
 the previous equations, the number of observations is very large.


                                  44

-------
                  The Number of Visits  Relationship

Counties were used to represent the  regions  in  the visits  equation  dis-
cussed earlier.  Information in the  questionnaire indicates  the  county
in which the recreationists resided.  Record was  made  of the number of
recreationists that originated from  each  county.  These were expressed
as a percent of the total  sample  compiled at each lake, and  then multi-
plied by the estimated  total visits  for the  season to  estimate the  number
of visits from each county for the season.

The travel cost and income of all persons in the  sample from each county
had to be averaged to obtain average values  for the  regions. The popula-
tion of each county was determined from the  U.S.  Bureau of the Census
Population Estimates, 1966.  The  relevant counties,  the population  of
the counties, as well as the information  on  visits,  travel cost, and
income, are listed in Table 3.  Each lake is listed  separately.   It
should be noted that the recreationists in the  sample  came from  only
four states:  Oregon, California, Nevada, and  Idaho.  The  sampled lakes
are not well known outside the State of Oregon.
                                                        'i

All of the data collected at the  four lakes  are used to estimate the
visits relationship.  For each of the counties  observed at Klamath  Lake,
for example, the values assigned  to  the characteristics of Klamath  Lake
are used, along with the values of the  other independent variables  for
that county group.  There are data from 12 counties  for Klamath  Lake,
27 counties for Lake of the Woods, 29 for Odell Lake,  and  9  for  Willow
Lake.  Thus, 77 observations are  available.

Since the independent variables k and y were aggregated within each
county, some were more  reliable  than others, due  to  the number of obser-
vations within each county.  Weighted least  squares, similar to  that
                               JU       JL
discussed for determining the p.,  and k  relationships, was used  to  cor-
rect this problem.  The estimated visits  equation is:
         -67,947.046 + 7,312.442 W™" + 21,024.198 F1M"r       (38)
                         (842.177)       (2,366.595)
       -I-  .648 Si*" - 149.953 k - .118 y - .003 pop"
         (.068)        (167.595)   (.824)   (.0019)

    R2 =  .874*"
    Degree  of freedom * 70.

The site  characteristics,  referred to by W, F, and Si, are as defined
earlier;  k refers  to the  average travel cost of recreationists from each
county;   y  represents the  average income of recreationists from each
county;   and pop refers to the total population of each county.

It should be noted  that the  visits equation is redundant if one is not
interested  in changing some  of the independent variables and predicting
a change  in total visitor-days.  That is, if no changes are considered,
                                  45

-------
Table 3.  The Population, Number of Sample Visits, Percent of Sample Visits,
          Visits per Season, Travel Cost, and Income, by County, for each
          of the Four Lakes
                                                         KLAMATH LAKE
      County
Population
 of county
   (pop)
Sample
visits
Percent of
sample from
each county
                                                                            Visits
                                                                             (V)
Average
travel
 cost
  (k)
Average
income
  (y)
       Oregon:
          Deschutes	         27,600       1
          Jackson	         91,300       5
          Josephine	         37,000       2
          Klamath	         49,600      29
          Lane	        200,700       1
£         Multnomah	        534,900       2
          Washington	        126,100       1

       California.:
          Orange	      1,171,400       1
          Los Angeles	      6,814,500       3
          San Bernardino	        628,900       1
          Santa Clara	        929,800       1
          Trinity	          8,200       1

                   TOTAL	                     48
                                                       2.1
                                                      10.4
                                                       4.2
                                                      60.4
                                                       2.1
                                                       4.2
                                                       2.1
                                                       2.1
                                                       6.3
                                                       2.1
                                                       2.1
                                                       2.1

                                                     100.2
                                          3,070
                                         15,205
                                          6,140
                                         88,305
                                          3,070
                                          6,140
                                          3,070
                                          3,070
                                          9,211
                                          3,070
                                          3,070
                                          3.070

                                        146,491
                                     (dollars)


                                       4.67
                                       2.16
                                       3.83
                                       0.26
                                       4.70
                                      51.33
                                      80.32
                                      28.63
                                      80.26
                                     119.39
                                      19.29
                                      63.17
                                            (dollars)


                                            18,372
                                             7,547
                                            10,329
                                             8,385
                                             3,956
                                             5,342
                                             3,860
                                            13,440
                                            13,355
                                             5,228
                                            15,315
                                             7,414
                                                                                 Continued

-------
Table 3.  (Continued)
                                                       LAKE OF THE WOODS
      County
Population
 of county
  (pop)
Sample
visits
Percent of
sample from
each county
Visits
  (V)
Average
travel
 cost
  (k)
Average
income
  (y)
Oregon;
   Coos	      54,100       1
   Des chutes	      27,600       1
   Douglas	      72,600       2
   Jackson	      91,300      46
   Josephine	      37,000       6
   Klamath	      49,600      19
   Lane	     200,700       2
   Lincoln	      25,500       1
   Linn	      65,600       2
   Multnomah	     534,900       2

California;
   San Francisco	     714,600       2
   Monterey	     229,900       3
   Contra Costa	     514,400       1
   Orange	   1,171,400       3
   Alameda	   1,030,400       4
   Los Angeles	   6,814,500      10
   Humboldt	     101,300       1
   Santa Barbara	     253,400       1
   Siskiyou	      35,000       3
   San Mateo	     519,100       1
   San Bernardino	     628,900       1
   Santa Clara	     929,800       9
   San Joaquin....	     283,500       1
   Sacramento	     597,700       2
   Del Norte	      16,700       1
                            .8
                            .8
                           1.6
                          36.2
                           4.7
                          15.0
                           1.6
                            .8
                           1.6
                           1.6
                           1.6
                           2.4
                            .8
                           2.4
                           3.1
                           7.9
                            .8
                            .8
                           2.4
                            .8
                            .8
                           7.1
                            .8
                           1.6
                            2,130
                            2,130
                            4,260
                           96,388
                           12,515
                           39,940
                            4,260
                            2,130
                            4,260
                            4,260
                            4.260
                            6,390
                            2,130
                            6,390
                            8,254
                           21,035
                            2,130
                            2,130
                            6,390
                            2,130
                            2,130
                           18,905
                            2,130
                            4,260
                            2,130
                           (dollars)


                              14.36
                               9.17
                              11.51
                               1.80
                                .78
                               2.84
                               8.66
                              91.30
                               9.96
                               4.58
                              58.63
                              50.86
                              79.38
                              33.85
                              36.71
                              54.78
                               6.48
                              19.20
                               3.92
                              17.01
                              26.76
                              28.92
                              35.28
                              26.20
                               9.10

                              Continued
                          (dollars)

                             7,528
                           12,048
                             7,131
                           10,471
                           10,589
                             8,327
                           10,335
                           10,746
                           11,850
                           12,506
                           13,106
                            7,057
                           13,740
                           10,790
                           15,570
                           14,942
                           10,215
                            5,646
                           11,893
                           11,010
                           13,440
                           15,412
                           11,010
                            6,492
                            7,414

-------
Table 3.  (Continued)
LAKE OF THE WOODS
County
Nevada :
Clark 	 	
Idaho:

TOTAL 	
Population
of county
(pop)
233,700
60,400

Sample
visits
1
1
127
Percent of
sample from
each county
.8
.8
100.4
(Continued)
Visits
(V)
2,130
2,130
267.327

Average
travel
cost
(k)
(dollars)
46.22
8.68

Average
income
(y)
(dollars)
13,590
10,620
                                                                                                               oo
                                                                                 Continued

-------
Table 3.  (Continued)
                                                              ODELL LAKE
      County
Population
 of county
  (pop)
Sample
visits
Percent of
sample from
each county
Visits
 (V)
                                                                                    Average
                                                                                    travel
                                                                                     cost
                                                                                     (k)
Average
income
  (y)
   Oregon;
      Benton	         49,100       3
      Clackamas	        146,100       4
      Coos	         54,100       2
      Crook	         10,100       1
      Des chutes	         27,600       2
      Douglas...	         72,600       2
      Jackson	         91,300       3
      Klamath	         49,600       4
      Lane	        200,700      30
      Lincoln	         25,500       3
S     Linn	         65,600       1
      Marion	        141,700       5
      Multnomah	        534,900       7
      Polk	         30,900       2
      Washington	        126,100       4

   California;
      Contra Costa	        514,400       2
      Orange	      1,171,400       1
      Alameda	      1,030,400       2
      Los  Angeles	      6,814,500       6
      Shasta	         75,500       1
      Siskiyou	         35,000       1
      San  Mateo	        519,100       1
      Santa  Clara	        929,800       3
      Modoc	          7,500       1
      Kern	        325,200       1
                                                        3.1
                                                        4.1
                                                        2.1
                                                        1.0
                                                        2.1
                                                        2.1
                                                        3.1
                                                        4.1
                                                       30.9
                                                        3.1
                                                        1.0
                                                        5.2
                                                        7.2
                                                        2.1
                                                        4.1
                                                       2.1
                                                       1.0
                                                       2.1
                                                       6.2
                                                         .0
                                                         .0
                           1.
                           1.
                                                        1.0
                                                        3.1
                                                        1.0
                                                        1.0
                                          5,595
                                          7,400
                                          3,790
                                          1,805
                                          3,790
                                          3,790
                                          5,595
                                          7,400
                                         55,769
                                          5,595
                                          1,805
                                          9,385
                                         12,995
                                          3,790
                                          7,400
                            3,790
                            1,805
                            3,790
                           11,190
                            1,805
                            1,805
                            1,805
                            5,595
                            1,805
                            1,805
                            (dollars)


                               6.45
                               7.22
                              11.29
                               3.87
                              12.14
                               2.40
                               8.17
                               2.32
                               4.66
                               6.45
                                ,00
                                .67
                                ,86
                              10.58
                              11.77
              5,
              7.
              9,
                              23.11
                              44.53
                              44.04
                              28.79
                              12.37
                              15.09
                              12.87
                              29.44
                              10.05
                              24.97
                                                     (dollars)


                                                       8,318
                                                       5,259
                                                       5,280
                                                       7,414
                                                       5,373
                                                       7,748
                                                     11,194 ,
                                                       6,822
                                                       7,256
                                                     11,804
                                                       8,472
                                                       6,688
                                                     11,747
                                                       9,005
                                                     12,567
                            8,669
                           18,540
                            6,940
                           15,502
                           11,784
                            7,114
                           12,048
                           11,826
                            9,519
                           10,746
                                                                        Continued

-------
Table 3.  (Continued)

County
California (Continued) :
Marin 	



TOTAL 	


Population
of county
(pop)

188,600
1.188.000
283,500
193,700



Sample
visits

1
2
1
1
97

ODELL LAKE
Percent of
sample from
each county

1.0
2.1
1.0
1.0
99.8

(Continued)
Visits
(V)

1,805
3.790
1,805
1,805
180 , 304


Average
travel
cost
(k)
(dollars)
11.72
40.36
8.39
26.59



Average
income
(y)
(dollars)
21,804
8.330
15,615
1,750


                                                                                                               o
                                                                                                               in
                                                                          Continued

-------
    Table 3.   (Continued)
                                                               WILLOW LAKE
          County
Population
 of county
  (pop)
Sample
visits
Percent of
sample from
each county
Visits
  (V)
Average
travel
 cost
  (k)
Average
income
  (y)
    Oregon;
       Douglas	         72,600        1
       Jackson	         91,300       17
       Josephine	         37,000        3
       Klamath	         49,600        1
i-      Lane	        200,700        1
       Lincoln	         25,500        1
       Marion	        141,700        1

    California:
       Los Angeles	      6,814,500        5
       Siskiyou	         35,000        2

                TOTAL	                     32
                           3.1
                          53.1
                           9.4
                           3.1
                           3.1
                           3.1
                           3.1
                          99.9
                            3,397
                           58,187
                           10,301
                            3,397
                            3,397
                            3,397
                            3,397
                           17,094
                            6,904

                          109,471
                             (dollars)

                               2.77
                               1.90
                               1.94
                               2.30
                               8.65
                              28.53
                              14.53
                                                     78.60
                                                      9.83
                            (dollars)

                            6,700
                            8,127
                            9,589
                            8,586
                            7,870
                           10,614
                            1,750
                                             6,299
                                             3,580

-------
Equation (38) would merely predict the total number of visits to  the  site,
which is the information from which the equation was derived.  However,  in
this study the main interest lies in predicting the results of a  change  in
the existing situation resulting from a change in water quality on  the
number of visitor-days.  Here the relationship is important because "visi-
tor-days" is made up of two components - the number of visits, and  the
length of stay per visit.  Without this equation, only the length of  stay
per visit would have been considered.

It should be noted that the population variable in the visits relationship
is not a population of recreationists but, instead, the population  of a
county.

Many of the relationships that were hypothesized earlier are substantiated
by the sample data.  The positive coefficients of the site-characteristic
variables indicate that the number of visits from a given county  will in-
crease as the characteristics improve.  Also, the negative coefficient for
travel cost indicates that the number of visits from a county decrease as
the travel costs increase.  These reactions were hypothesized.

However, two statistical problems were encountered in Equation (38).  First,
the population and travel cost variables are highly correlated.   The  simple
correlation coefficient between the variables is .711.  This relationship
is easily explained.  The counties near the recreation sites have a lower
average travel cost than those counties farther away from the sites.  Also,
those counties close to' the site have smaller populations than those  far-
ther away.  Therefore, the population figure increases as the value of k
increases.  Johnston and Pankey [1968] observed the same phenomenon in
their study.

Although the correlation between travel costs and population is easily
explained, it creates difficulties.  The interrelationship causes each of
the variables to be relatively insignificant in the estimated equation.
The population variable is significant only at the 10 percent level,  while
the k variable is not significant at the 10 percent level.

The other problem encountered in Equation  (38) is a high level of corre-
lation (.694) between the W variable and income.  Once the W variable
enters into the equation, the entry of the income variable fails  to reduce
significantly the variation in V.  It is believed that the high correla-
tion is the cause for the low significance levels of the coefficients and
the negative coefficient estimated for income.  The*"negative coefficient
for income indicates that the number of visits from a county will decrease
as average county income increases.  This  is not the type of relationship
that was hypothesized.
                                             /
To solve the two statistical problems, a new visits equation was  estimated
with the income and population variables deleted, since they did  not  ex-
plain much of the variation in the previous equation.  The new equation
is:
                                 52

-------
     V - -71,166.121 +  7,141.764 Wffffff + 19,825.384 r"       (39)
                         (460.199)        (1,651.425)

       +   .641 Siff* -  379.786  k###
         (.059)        (115.473)             R2 = .868*"
                                            Degrees of freedom = 72.

All variables in Equation  (39), including the travel cost variable,  are
significant at the  1 percent  level.  The R2 of the new equation is only
slightly lower than the R2 of Equation (38).

In order to test its ability  to predict accurately, Equation (39) was
used to estimate the original visits  data received from the U.S. Forest
Service.   This is done  in  the following manner.  Substituting the values
of W, F, and Si variables  for Klamath Lake into Equation (39) gives:

              V - 26,119.751  -  379.786 k                      (40)

where W, which is the  sum  of  the  swimming, boating, and water skiing use-
intensities listed  in  Table 1,  is  equal to 2;  the fishing use-intensity
(F) is 1;  and the  lake size, in  acres, is 98,560.

Substitution of the recreationists' average travel costs from each county,
which are  listed in Table  3,  gives an estimate of the number of visits  to
Klamath Lake from each  county.  For example, using Klamath Lake and
Deschutes  County, Equation (40) yields the following estimate:

              V = 26,119.751  -  379.786 ($4.67)

              V - 26,119.751  -  1773.601
              V - 24,346.

Determination of the number of  visits from each of the other counties in
the same manner, and summing  over all of the counties, gives an estimate
of 243,211 visits to Klamath  Lake.  The same method is used to obtain
the estimates of total  visits to  the three other lakes in the study.  The
number of  visits estimated by Equation (39) for each lake are listed in
Table 4.

The total  number of visits to the  study area, that is, the sum of the num-
ber of visits to each  lake, is  accurately explained by the equation
(702,593 vs. 703,871).   However,  the distribution of the total visits among
the four lakes is not  accurately  predicted.  The number of visits to Klam-
ath Lake and Lake of the Woods  are over-estimated, while the number  of
visits to  Odell Lake are under-estimated.  The estimate for Willow Lake is
quite accurate.

The particular manner  in which  the site characteristics are represented in
Equation (39) is believed  to  be responsible for the lack of predictive  accu-
racy among individual  lakes.  Large variations exist in the values  assigned
to the site characteristics in  the equation.  For example, lake size varied
                                  53

-------
Table 4.  A Comparison of the Forest Service Estimates of
          Visits to Each Lake in 1968, and the Estimates
          Derived from the Visits Equation

                          U.S. Forest Service       Number of visits
                          estimates of number         estimated by
Lake


Odell 	 	
Willow 	
TOTAL 	

of visits
146,491
266,327
180,304
109,471
702,593

V « f(W,F,Si,k)
243,211
335,331
14,968
110,361
703,871

from 320 acres to 98,560 acres;  swimming, water skiing, and boating  in-
tensity ranged from 1 to 9, and so forth.  Since a direct and constant
relationship between visits and characteristics is Implied in Equation
(39), one would expect the magnitude of the site characteristics variables
to influence the estimated number of visits proportionately.  As very high
values of the characteristics are observed, a very large estimated number
of visits will be obtained.  If the values of the characteristics are gener-
ally low, a very low estimate of the number of visits will be forthcoming.
This phenomenon can be seen to exist in this study.  To avoid this prob-
lem, the original U.S. Forest Service estimates are used to obtain the
aggregate demand functions and net economic value of each site.  That is,
when q. is multiplied by V to obtain the aggregate demand model, Vq-, the
Forest Service's estimates of V for 1969 are used.

However, the estimated equation is used in another context.  It is employed
to predict the number of visits subsequent to a change in water quality at
Klamath Lake by assigning new values to the site characteristics in  the   *
equation.  Since the new values of the characteristics are larger than the
original values, and since the coefficients of these variables are positive
in the visits relationship, an increase in the number of visits is forth-
coming.  However, the increase in the estimated visits should be expressed
as a percent change in total visits.  Since Equation (39) predicts the    •.-".
total number of visits to all four lakes quite accurately, it is assumed
that it can adequately predict a percentage change in the total number of
visits.  This allows us to estimate the number of visits to Klamath  Lake  »
after the water quality has been improved.

                 Estimation of Recreational Value

The equations estimated in the previous section are used to estimate the
demand for recreation at each of the four lakes in the study area.   The
                                 54

-------
recreationist's demand model for a lake consists of the k ,  p., and q.
equations and the values  of  the site characteristic variables of that lake.

An average individual's demand curve is illustrated in Figure 10.  In the
figure, ^ is the critical daily on-site costs for the average person;
P! represents the average on-site costs per person per day;   q. is the
length of stay per  visit  when daily on-site costs are at the critical
level, and ^ is the number  of days the recreationist will stay, per visit,
if on-site costs were pT . It is here that the concept of consumer's sur-
plus is used.  It is recalled that consumer's surplus is the difference
between the  amount  a consumer is willing to pay for a commodity and the
amount he actually  has  to pay for it.  That is, consumer's surplus can be
measured as  that portion  of  the area under the demand curve which lies
above the price of  the  commodity.  This technique is straightforward when
the demand curve is shaped as the one illustrated in Figure 8.  However,
the demand curve in Figure 10 has a unique property in that it does not
exist for quantities of recreation less than q...  That is, the average
recreationist will  either stay at the site a minimum of q. days per visit
or he will^not recreate at all.  Therefore, it becomes a question of whether
the area ifL  AB ~p~ or the  area ABC is the appropriate measure of consumer's
surplus for  a demand curve of this type.  For the purpose of this study,
the area of  the  triangle, ABC, will be used to measure the net economic
value per visit.  Work  is continuing on the interpretation of the consumer's
surplus for  the  case of a truncated demand curve.  This solution should be
viewed as one to  facilitate  the derivation of the remaining empirical esti-
mates.  The  area ABC can be  determined mathematically:
value per visit.    (40)
 The  total value of the site is determined by multiplying the value per visit
 by the estimated number of visits to the site during the season.

 The  procedure described above is used to estimate the value of each of the
 four lakes considered in this study.

 Lake of the Woods

 The  use-intensities of the activities and the size of Lake of the Woods
 were substituted into the k*, p*, and q^^ equations.  The demand model for
 Lake of the Woods is represented by the following equations:

           k* - 23.617 + .0020 y 4- 10.435 PI                     (41)

           p* -  2.458 4- .269 k  -   .004 k2 + .000000017 y2     (42)

                 .759 - .0064 k* 4- .0064 k 4- .0637 p* - .0637 p  (43)
           qx • e                                   11


                                  55
              [f(Pl)  dP;L] - [(px -

-------
          —*
          pl I-
Figure 10.   The Average Individual's Demand Curve, Per Visit,
            for a Lake.
                               56

-------
Recreationists stayed  at  Lake  of the Woods an average of 2 days,  and spent
an average of $2.01 per day.   Travel costs to Lake of the Woods  averaged
about $15, while  the average income of recreationists was just over
$10,500.  Critical on-site cost was estimated to be $6.10, and critical
travel cost was about  $66. The demand function estimated by incorporating
the above data into the demand model is:

                        qj . ..Ml - .0637 PI

This function is  illustrated graphically  in Figure 11.  The value per
visit is:

       6.10

     f     (e-821 "  -°637 Pl) dp1 - [(4.09X1.54)]           (45)
  2.01
          Value per visit • $7.17 - $6.30 - $0.87.

The 1968 seasonal value of Lake of the Woods is obtained by multiplying
the per-visit value by the estimated number of Visits in 1968:
                                •j

                       ($0.87)(266,327) -  $231,704.

The derived value of Lake of  the Woods refers to the 1968 season.  Any
inferences concerning  estimated value for subsequent years requires addi-
tional assumptions.  It must be assumed that the reactions of recreation-
ists remain constant,  and the  measured variables are not altered, or the
exact change  can  not be specified.  Use of the lake in seasons other than
the summer season was  not considered in this value estimate.  For example,
ice fishing and duck hunting performed at Lake of the Woods during the
fall or  winter are excluded.   The above limitations also apply to the
value estimates of the remaining lakes.

Odell Lake

The following equations were  estimated for Odell Lake:

     k*  = -18.076 +  .0020 y + 10.435 pl                       (46)

     p*  -   1.641 +  .269  k -  .004 k2 + .000000017 y2          (47)

            .759 - .0064 k* +  .0064 k + .0637 p. - .0637 p,    (48)
     qi  - e                                    l          J-

Recreationists stayed  at  Odell Lake an average of 2.4 days, spent $2.34
daily for on-site costs,  and  about $11 in travel costs.  They had an aver-
age income of $9,063,  and were estimated  to have critical on-site costs,
p*, and  critical  travel costs, k*, of $5.57 and $24.71 respectively.  The
incorporation of  the above information into the demand model yields the
following demand  function:

                     q « el-025 - .0637  PI                   (49)



                                  57

-------
      $6.10
      $2.01
       0.00
                        1.54  2.0
                                     .821 -  .0637P
                                                K
Figure 11.  The Average Recreationist's Demand Curve, Per

            Visit, for Lake of the Woods.
                               58

-------
The value per.visit  for  1968  is  estimated as:


    f      (e1'025 *  '°637  Pl) dPl - [(1.96)(3.23)]          (50)
  2.34
              Value per visit  - $7.17 - $6.33 - $0.84.

When this figure  is  multiplied by the number of visits for the 1968 season,
a total seasonal  value of  Odell Lake in 1968 is obtained:

                   ($0.84)(180,304)  = $151,455.

Willow Lake

Willow Lake  is  the smallest of the four lakes in this  study.   However,  the
lake has approximately  average value of the use-intensities for the vari-
ous activities.   When these characteristics are substituted into the k  ,
p.., and q_ equations, the  demand model applicable to Willow Lake becomes:

     k* • -6.534  + .002  y  + 10.435 Pj^                         (51)

     p* -  2.988  + .269  k  - .004 k2 + .000000017 y2           (52)

        = £.759 - .0064  k* +  .0064 k + .0637 p* - .0637 PI    (53)


The recreationists at Willow  Lake recreated an average of  2.18 days per
visit, and spent  $1.72  per day.  They had an average income of $7,790,
and incurred approximately $6 in travel cost per visit. The estimated
critical travel cost averaged nearly $31, while the estimated critical
daily cost averaged  $4.55. When p*, k*, and k are assigned average values
applicable to Willow Lake, the demand function for an individual visit  is:
                            m  e.888 - .0637 pi                 (54)

The average  value per visit is:

       4.55
       r      (e.888  -  .0637 PI) d   . [(2.83X1.82)]          (55)
     J                            *
  1.72
              Value per  visit  - $5.64 - $5.15 - $0.49.

The seasonal economic value for Willow Lake in 1968 is:

                     ($0.49)(109,471) - $53,641.
                                  59

-------
Klamath
The demand model for Klamath Lake, given its present characteristics  and
use-intensity values , is :

     k* = 9.132 4- .002 y + 10.435 p.,^                           (56)

     p* = 3.531 + .269 k - .004 k2 + .000000017 y2             (57)

        = e.759 - .0064 k* + .0064 k + .0637 p^^ - .0637 PI     ^5g)


The average stay per visit was 1.97 days, while average daily  expenses  at
the site were $1.84 per person.  The average recreationist had an annual
income of $8,900, and allocated about $6.80 to travel costs per visit.
Average critical values of travel cost and on-site costs were  estimated at
approximately $55 and $5.54 respectively.  Using these values, the  demand
function for Klamath Lake, as it now exists, is:

                        qi=e.801- .0637 PI


The economic value of Klamath Lake for the summer of 1968 is computed in
the same manner as for the other lakes:
       mf w *f *

      r     (e.801 - .0637 p       _ I(3.70)(1.57)]
    »/                            J.
  1.84

             Value per visit = $6.37 - $5.81 = $0.56.
The economic value of Klamath Lake, given its present characteristics,  is
obtained by multiplying the value per visit by the number of visits to  the
lake in 1968:
                       ($0.56)(146,491) - $82,035.
                                 60

-------
                            SECTION VII

       THE THEORETICAL MODEL FOR ESTIMATING  REGIONAL BENEFITS


The three preceding sections present methods for estimating the demand for
recreation and the economic value  of four  recreational sites.  Before the
model is used to estimate  the benefits resulting from water quality im-
provements at Klamath Lake, let  us look  at the procedures used to estimate
the local economic impact  resulting from water quality improvements at the
lake.

As pointed out, the local  community is often the recipient of indirect
benefits as the result of  a project which  increases the economic activity
of the region.  To measure the magnitude of  these  benefits, it is necessary
to determine the economic  structure of the region.  Because of its unique
ability to portray the structure of an economy, a  regional input-output
model was selected for use in this study.

An input-output model portrays the flow  of goods and services throughout
the economy.  It provides  a means  to measure the impact of changes in the
overall economic activity  of  the economy,  resulting from a change in the
activity of any one segment of the economy.   Two reasons have been cited
[Stoevener and Castle, 1965] why the interdependences portrayed by an
input-output model are important.  They  are: (1) the determination of the
aggregate level of regional secondary benefits, and (2) the distribution
of these benefits.  Knowledge of the distribution  of the regional benefits
among the various sectors  of  the economy may be especially important if
institutional arrangements for cost-sharing  are to be made among the vari-
ous beneficiaries.

                         Input-Output Theory

Although some of the basic concepts had  been developed earlier, Wassily
Leontief is usually credited with  the introduction of input-output analy-
sis.  In his first publication on  the subject in 1936, Leontief introduced
input-output analysis as an empirical tool.   Since Leontief's first publi-
cation, several extensions and alternative uses of the model have been de-
veloped.  The input-output method  has spread rapidly, and is a widely used
economic tool in the world today.  The volume of literature relating to
input-output analyses is so large  that several bibliographies have been
compiled (Riley and Allen  [1955],  Taskier  [1961],  and Input-Output Biblio-
graphy 1960-63 [1964]),  and international  conferences relating to input-
output analysis have been  held.

An input-output model is illustrated by  a  "transaction matrix".  It de-
scribes the flow of goods  and services between the sectors of the economy
in a given time period.  A sector  is made  up of a  group of firms that carry
on similar types of business.  The criteria  used to define a sector will
                                  61

-------
be discussed later in this section.  A simplified transactions matrix is
presented in Table 5.  The hypothetical economy is divided into three sec-
tors.  Each of the sectors is listed twice in the matrix - once as a "pro-
ducing sector", indicated by the row heading, and again as a "purchasing
sector", shown by the column heading.  The first three rows of the matrix
describe how the total output of Sector i (i - 1, 2, or 3) is distributed
among the various sectors of the economy.  The first three columns describe
Sector j's (j « 1, 2, or 3) purchases of inputs from the various sectors
of the economy.  Therefore, the x's, which represent the intersectoral
flows, represent the dollar value of goods and services sold by Sector i
to Sector j.
Table 5.  A Hypothetical Transactions Matrix
CO
|4
s
o
0)
OT
I
o
•3
0
M
fe


1
2
3
Value
added
Total
purchases
Purchasing Sectors
1
xll
X21
X31
Vl
Xl
2
X12
X22
X32
V2
X2
3
X13
X23
X33
V3
X3

Final
demand
Yi
*2
T3
V£


Total
output
Xl
X2
X3


The Final Demand column of the transactions matrix, represented by the
Y.'s (i • 1, 2, 3) shows how much of the producing sector's output is con-
sumed as a "final good".  Outputs allocated to final demand are not re-used
as intermediate inputs in producing other goods and services within the
economy under study.  Final demand is referred to as the exogenous or
autonomous portion of the input-output system, because the level of demand
is not dependent upon the level of economic activity within the economy
being studied.  The components usually assigned to final demand in an open
input-output model are consumption, investment and inventory accumulation,
purchases of the various levels of government, and exports.

The last column of the transaction matrix represents Total Output.  Total
output of Sector i (X.) is the sum of its sales to other sectors and to
final demand:
                                 62

-------
          Xi =     Xij + Yi           (i  =  1'  2'  or  3)


The Value-Added row of the transactions  matrix'is composed  of  the payments
made by the various sectors  for  "primary inputs" used  in  their production
processes.  The components of  the  Value-Added row are  payments to all
levels of government  for services  rendered, payments to households  for
labor and entrepreneurial services,  imports,  depreciation of capital equip-
ment used in the production  process, and depletion  of  inventories*  A sec-
tor purchases produced inputs  from the other  sectors and  produces a new
product of greater value.  The increase  in value is reflected  in the pay-
ments made for the primary factors of production, which are recorded in
the Value-Added row.

At first it may appear that  imports  should not be included  in  the Value-
Added row, since they may consist  of intermediate goods produced by other
economies.  However,  imports represent primary inputs  into  the economy
being studied, and they are  being  used in  a production process within that
economy.  Therefore,  their value is  reflected in the producing sector's
product, and must be  included  in the Value-Added row.

It should be noted that some of  the components of final demand may  also
purchase primary inputs.  For  example, wages  paid to government employees
represent purchases of primary inputs.   The symbol  Vf  in  the transactions
matrix represents the value  of all purchases  of  primary inputs made by
the components of the final  demand sector. Although the  V. portion of
the matrix adds additional information,  it is not necessary to know these
values to solve the model.

The final row of the  transactions  matrix represents Total Purchases.  Total
purchases of Sector j (X.) are the sum of  all inputs used in the produc-
tion process of that  sector:

               3
         X. =  I  x.. + V.            (j  -  1,  2,  or  3)          (62)
          3   i=1  13   3
                                   ' t                   -           .    '
The model used in this study is  referred to as an "open"  model.  However,
the model may be "closed" with respect to  the household or  government
component of final demand by including either or both, as  an additional
sector in the processing portion of the  ecqnomy. For  example, the  model
can be "closed" with  respect to  .households by removing consumption  from
final demand, and payments to, households from value added,  and forming,a
"Households" sector in the endogenqus portion of the model. Formation  of
a households sector gives a  more complete  picture of  the  economy, because
it shows the relationship that exists between household income and  expendi-
tures.  However, the  cost of obtaining  the additional  necessary  informa-
tion prevented the construction  of a household sector  in  this  s£udy.  The
model constructed in  this study  is,  however,  closed with  respect  to the,,,
local governments in  the region.                                        ;
                                  63

-------
A final characteristic of the transaction matrix should be noted.  The
total output of each sector is equal to its total purchases :

                       Xi = X.          where (i - j)

This characteristic is a manifestation of Euler's Theorem, which states
that, under conditions of constant returns to scale  (linear, homogenous
production functions), "total product is equal to the sum of the marginal
products of the various inputs, each multiplied by the quantity of its
input"  [Stigler, 1966, p. 152],

Direct Coefficients Matrix

The transactions matrix is used to derive the direct coefficients matrix.
This step relies upon one of the basic assumptions of input-output analy-
sis.  It is assumed that the level of inputs purchased by a sector is de-
pendent upon the level of total output of the purchasing sector.  More-
over, inputs are purchased in a direct and constant relationship to  the
output of the sector.  From this assumption we can derive the direct co-
efficient, a  .  We recall from the transactions matrix that the x
represents Sector j's purchases -from Sector i, and that the total output
of Sector j was recorded in the Total Output column.  It follows directly
from the assumption that:
                              XH
                        *ij ' xf                             (63)
where      a.. » the direct coefficient, and

           X.  - total output of Sector j .

By solving Equation (63) for x.., and substituting it into Equation  (61),
a new equation for total output is obtained:
             3
       X  -  I  a   X  + Y            (i - 1, 2, or 3)        (64)
Direct coefficients indicate the value of inputs Sector j must purchase
from Sector i to produce one dollar of output.  They illustrate the  direct
interdependencies that exist between the sectors of the economy.  For ex-
ample, an increase in the output of Sector 1 will lead to increased  output
in Sectors 2 and 3 (providing a. . > 0) , because Sector 1 will require
additional inputs from Sectors 2 and 3 to produce its increased output.

The Direct and Indirect Coefficients Matrix

The direct coefficients do not, however, explain the full addition to
total output caused by an increase in demand of the Sector 1 product.
As Sectors 2 and 3 increase output tp satisfy the additional requirements
of Sector 1, they must also purchase more inputs from the various other
sectors to produce their increased 6utput.  This causes another increase
                                 64

-------
in the level of demand within  the  economy.   Therefore,  a change  in  the
output of one sector will  cause  direct  and  indirect  increases  in the  out-
put of other sectors.  The matrix  of direct and  indirect coefficients is
used to describe the total effect  an exogenous change in the demand of  one
sector will have upon the  output of the entire economy.

The matrix of direct and indirect  coefficients,  or the  "R" matrix,  is de-
rived from Equation (64).  Rearranging  Equation  (64) gives:


                         _ S±j  X.j  - Yr                      (65)


Rewriting Equation  (65) in matrix  notation  yields:

                        X  -  AX'- Y,                           (66)
where       X is the column  vector of total output,

            A is an nxn  (where n = 3 in our example)
              direct coefficients  matrix, and

            Y is the final demand  column vector.

Factoring X out of the left  side of the equation yields:

                        X(I  -  A) = Y                          (67)

where       I is an identity matrix of  the  same
              dimensions as  the  A  matrix.

The identity matrix has the  characteristic  that,  when multiplied by another
matrix of the same dimensions, it  does  not  change the original matrix.
It serves the same purpose in  matrix algebra that multiplying by unity
serves in arithmetic algebra.  The main-diagonal  cells  of the  identity
matrix contain unity, while  all  other cells in the matrix are  zero.

Solving Equation (67) for  X  yields:
                        X  *  (I - A)"1 Y.                      (68)

The (I - A)   or "R" matrix  is also called  the matrix of direct  and indi-
rect coefficients.  Each coefficient shows  the output of Sector  i needed
by Sector j to deliver one dollar  of its output  to final demand.  It  takes
into account the direct and  indirect effects caused by  a change  in  final
demand of one of the processing  sectors.  The availability of  digital com-
puters has made the method of  matrix inversion the most  popular  means of
obtaining the interdependence  coefficients, since other  methods  using
simultaneous equations are more  cumbersome.

                     Input-Output  Assumptions

Three general assumptions  are  used in input-output analysis:

     (1) Each commodity is supplied by  a single  sector.
                                 65

-------
     (2) The inputs purchased by each sector are dependent
         upon the level of output of that sector.

     (3) The total effect of carrying on several types of
         production is equal to the sum of the separate
         effects *. .     -

The first assumption implies that each sector uses only one method of pro-
duction, and that only one primary output is produced by each sector.  This
assumption requires satisfactory criteria by which to aggregate the numer-
ous economic activities in the economy into sectors.  Ideally, two criteria
should be used in defining sectors.  They are to aggregate industries with
similar input structures, and/or industries that produce strictly comple-
mentary outputs  [Chenery and Clark, 1959].  It is seldom possible to
strictly follow these criteria when constructing a model, but they should
be adhered to as closely as possible, to insure more stable input coeffi-
cients.

Another point should be considered in the aggregation process.  It should
be clear to the researcher how the model is to be used after its comple-
tion.  More aggregation in some industries may be possible if they are
not expected to be important in the model's final use.  Likewise, increased
specialization may be beneficial in the industries that are of major inter-
est in the analysis.  For example, in this study it was felt that the
major portion of recreational spending would be spent on gasoline, grocer-
ies, prepared food, and lodging.  Therefore, a sector was designed for
each of these categories, to obtain a more detailed analysis.

The second assumption, that the amount of inputs purchased by a sector
is a function of the level of the output of that sector, was mentioned
earlier.  This is the most important assumption of input-output analysis.
However, it has drawn criticism from sources who feel it does not apply
in the real.world.  These criticisms will be mentioned in the next sub-
section.

The third assumption of additivity disallows external economies and dis-
economies.  It states that all the production processes carried on with-
in the economy are independent of each other.  One production process
has no effect - either beneficial or detrimental - upon any other produc-
tion process.,

                      Limitations of the Model

Some doubt has been expressed concerning the ability of input-output
models to predict accurately.  Many feel that the assumptions upon which
the model are based are too restrictive and unrealistic to give meaning-
ful predictions.  Most of the criticisms have centered around the assumed
fixed direct coefficient because it ignores three changes that can occur
in an economy [Chenery and Clark, 1959].  They are:  (1) changes in the
composition of demand, (2) changes in the relative prices of inputs, and
(3) changes in production technology.  Of the three, a change in technol-
ogy is usually considered to have the greatest effect upon the direct
coefficients.
                                 66

-------
Although the changes  listed above will certainly affect  the  input  coeffi-
cients in the long  run,  their  effects  may be minimal over shorter  periods
of time.  Even  though new production processes  become available, exist-
ing techniques  are  often used  until capital items have depreciated and
new plants are  built.   In the  same  vein,  changes in the  composition of
demand and input substitutions usually occur gradually.   Therefore, it
would seem that the assumption of fixed input coefficients would be justi-
fiable in the short run.

Cameron  [1953], in  a time series  analysis of selected input  coefficients
from the Australian model, concluded that his results generally supported
the assumption  of fixed input  coefficients in the short  run.   The  most
important input coefficients appeared  to  remain relatively constant for
a period as long as a decade.   The  substitution that did occur between
inputs appeared to  be caused by a change  in the product-mix  of the indus-
try, and not changing technology  or price ratios.   However,  caution
should be exercised.  One must be aware of the  limitations of  the  model
and restrictions of the assumptions, to prevent misusing the model.

                     Regional Input-Output Models

Regional input-output models,  such  as  the one constructed in this  study,
have become a popular analytical  tool.  A wide  variety of types of re-
gions have been analyzed.  In  most  cases  regions have been defined by
using political boundaries - that is,  counties, states,  or multi-state
regions.  However,  geographic  characteristics,  such as river basins,
have also been  used to  define  the boundaries of study areas.

A regional model is constructed in  the same manner as a  national model.
Thus, the regional  model is open  to the same criticisms  discussed  earlier.
However, obtaining  the  data necessary  to  construct a regional  transac-
tions matrix has been the dominant  problem faced by the  regional re-
searcher *  Although it  is expensive, extensive  interviewing  is often the
only available  means  to obtain the  data required to construct  a model.

                  Modifications of  the Basic Model

The model constructed for this study is a regional input-output model,
but it differs  slightly from the  basic input-output model that was just
presented.  The previous model was  concerned with the technical input
structure of the various sectors.  This gives the model  a technical
orientation in  that the input  structure of each sector is determined by
the state of technology used by the sector.  In the Klamath  County model
the trade flows of  the  sectors of the  economy are studied, instead of
the more technical  input-output relationships.   Therefore, the a±. *s
may more accurately be  termed  "trade"  coefficients rather than technical
coefficients.   A model  of this type has been called a "from-to" model
[Leven, 1961].

There is one other  conceptual  difference  between the Klamath County model
and the basic input-output model.  In  the latter model,  all  transactions
                                  67

-------
involving capital items are removed from the endogenous flows.  They are
then included in the investment component of final demand.  This is neces-
sary, since investment purchases usually are not a function of the current
level of output.

However, in a from-to model, where the a  's are interpreted as trade
coefficients rather than input coefficients, it is not necessary to re-
move capital item purchases from the interindustry flows.  Only the gross
flow of goods and services between the various sectors is relevant in the
from-to model.

The inclusion of investment purchases in the endogenous flows of the model
does, however, raise an important question:  What effect does their in-
clusion have upon the stability of the trade coefficients?  At first it
may appear that including investment purchases in the interindustry flows
would decrease the stability of the coefficients, since investment decis-
ions are not determined entirely by current output trends.  That is, in-
vestment purchases are cyclical, and reflect conditions other than those
portrayed in the model.  Therefore, the interindustry flows would vary
as investment purchases varied.
                                                                       ~ "i - *
To determine if the inclusion of investment purchases reduces the stability
of the coefficients, two types of investment cycles must be considered.
The first is the cyclical investment decisions of the individual firm.  A
firm usually does not have a constant rate of investment over a given
period of time.  Instead, a large investment is usually made during some
year, while smaller investments are made during other years.  This type
of investment cycle may not significantly affect the stability of the
trade coefficients.  One firm may purchase capital items during one year,
and another firm may invest the following year.  Therefore, if the firms
are sampled in a random fashion, the observed level of investment may re-
flect a relatively stable yearly estimate for all of the firms in a sec-
tor.

The ot,her investment cycle of importance is that observed for the entire
economy.  Aggregate investment varies from year to year^  This variation
is due to several factors, including the expectations of businessmen and
the availability of investment funds which may be subject to changes in
national fiscal and monetary policies.  If data were gathered during a
year when investment spending was significantly above or below its usual
level, inclusion of investment purchases in the interindustry flows would
distort the coefficients.  To determine if this had occurred in this
study, national aggregate investment figures were studied [Board of Gover-
nors of the Federal Reserve System, 1968].  The data indicated that in-
vestment spending at the national level increased sharply in the second
half of 1968.  However, whether the Klamath County economy experienced
the same general increase in investment at that time cannot be determined.
Therefore, one can only conclude that the trade coefficients estimated
for the Klamath County model may have been subject to cyclical investment
fluctuations, but the extent of that effect, if any, could not be deter-
mined.
                                 68

-------
                            SECTION  VIII

               CONSTRUCTION OF THE  FROM-TO MODEL AND
                  AN ANALYSIS  OF THE LOCAL ECONOMY


The first step in the  construction  of the from-to model  is  to  define  the
study area.  Klamath County was selected as  the study  area, primarily
because it includes the  communities that would benefit from expanded  use
of Klamath Lake.  Defining the county as the study area  also offers addi-
tional benefits.  Secondary data that would  aid in the study are  avail-
able on a county basis.   Also, defining the  study area to  coincide with
political boundaries is  advantageous, since  the model  may  provide useful
information to county  planning and  development groups.  A  map  of  Klamath
County is provided in  Figure 12.

                         Sampling Procedures

Since sampling techniques were to be used to obtain the  data required to
construct the model, a listing of all the business firms in Klamath County
was obtained.  Three secondary sources were  relied upon  to  compile the
population.  They are:  (1) the 1968 telephone directory of Klamath Falls
and surrounding communities, (2) a  listing of business firms in Klamath
County, received  from  the Klamath County United Good Neighbors, and (3)
the 1967 Klamath  Falls City Directory.  A final population  included
1,840 business firms.

Each firm in the population was then placed  in the appropriate sector of
the model.  Table 6 contains the sectors of  the model  and  examples of
the types of firms in  each sector.   The Household sector is also  defined
in the table, although it is not included as a sector  in the processing
portion of the matrix.  A firm with multiple economic  activities, such
as selling and servicing, was  placed in the  sector that  described its
largest income-producing activity.

After each firm had been assigned to the appropriate sector of the model,
most of the sectors in the model were stratified so that the firms in
each sector could be grouped into more homogenous categories than a sec-
toral grouping alone could provide.  By increasing the homogeneity of a
group of businesses, a smaller sample could  be used to obtain  informa-
tion about the businesses in each stratum.

Two types of stratification were used.  All  sectors except  Agriculture,
Professional Services, Service-Oriented, Resorts and Marinas,  and Local
Government were stratified according to size.  That is,  each firm in  the
sector was placed into a size  group, usually large, medium, or small,
within the sector.  For  example, if the gross sales of a firm  were thought
to be large, relative  to the other  firms in  the sector,  the firm  was
placed in the large-firm stratum.


                                  69

-------
        Odell
        Lake
 Crater Lake
National Park
                       Klamath   /" Highway 140
                         alb
     Figure 12.   Klamath County, Oregon.
                                         70

-------
Table 6.  Description of the Sectors in the Klamath
          County Model
Sector
number      Sector  title
         Sector description
  1,    Agriculture

  2.    Agricultural Services
   3.    Lumber

   4.    Manufacturing and
        Processing
   5.    Lodging
   6.     Cafes  and Taverns
   7.     Service Stations
   8.     Construction
   9,     Professional Services
  1Q.     Product-Oriented
  1       (wholesale and retail)
 11.     Service-Oriented
Farms, ranches, and feedlots.

Farm implement dealers, farm coopera-
tives, feed, seed, and fertilizer stores,
livestock auction yards, and irrigation
pump dealers.

Logging, log hauling, lumber and ply-
wood mills.

Potato processors, creameries, bottling
companies, meat and poultry processors>
machine manufacturing, trailer manu-
facturers, and stone, clay, and glass
manufacturers.

Hotels, motels, trailer parks, and
apartments.                           •»
                                   {-,
Businesses that sell beverages and pre-
pared food that may be consumed on the
premises.

Gasoline bulk plant distributors, ser-
vice stations, and heating fuel dis-
tributors.

General building contractors, electriea^
and plumbing contractors, sand and gravel
operations, asphalt paving contractors,
carpenters, concrete manufacturers, exca-
vators, land levelers, road and highway
contractors, roofing and painting contrac-
tors, masonries, and well-drillers.

Doctors, dentists, lawyers, accountants,
architects, surveyors, engineers, hos^
pitals, veterinarians, ambulance ser-
vices , and nursing homes.
All firms that receive the largest part
of their income from the sale of pro-
ducts at the wholesale or retail level
that are not included in other sectors.
Examples: utility companies, department
stores, specialty stores, drug stores,
and bottled beverage distributors.    j

Firms that receive the largest part of
their income from the sale of services.

       (Continued on following page)
                                  71

-------
Table 6. (Continued)
Sector
number      Sector title
         Sector description
 12.    Communications and
        Transportation
 13.    Financial
 14.    Grocery
 15.    Resorts and Marinas
 16.    Automotive
 17.    Local Governments
        Households
Examples:  beauty and barber shops,
insurance and real estate agencies,
repair stores, laundries, churches,
social organizations, and labor unions.

Trucking firms, railroads, airlines,
buses, radio, television, telephone,
telegraph, newspapers, and television
cable.
Banks, savings and loan associations,
loan  companies, and securities invest-
ment businesses.

Firms which sell food for pff-premise
consumption.  Examples:  grocery;,
stores, seafood stands, meat stores,
and fruit stands.
Stores at recreational sites, marinas,
and boat dealers.

Auto  and trailer sales, tires, parts
and accessory stores, and auto repair
shops.

County and city governments, school
districts, and special taxing districts.

All private individuals.
The Professional Services, Service-Oriented, and Resorts and Marinas
sectors were stratified on the basis of the economic activities within
the sector.  For example, in the Professional Services sector, physi-
cians were put in one stratum, dentists in another stratum, accountants
in a third, and so forth.  Alternatively, each of these strata could
have been considered as an additional sector.  This would have led to
a much larger number of sectors than were used in the study.  However,
it was felt that the benefits from increased homogeneity in the smaller
sectors were not sufficient to compensate for the increased difficulty
of collecting data for a more highly disaggregated model.

Data for the Agriculture and Local Government sectors were not collected
by the same procedure.  They will be discussed later in this section.
                                 72

-------
Determination of the Sample  Size

To estimate the sample  size, the  variances of the variables  to be esti-
mated must be known.  Since  estimates of the variances  were  not avail-
able, other techniques  had to be  used to determine the  sample size.   Two
factors were considered.   The first was the amount of funds  available
for collecting the  data;   the other was the sampling rate used in pre-
vious input-output  studies in Oregon.  In previous studies by Bromley
[1967] and Stoevener  [1964], the  sampling rates were between 25 and  30
percent of the total population.

After considering these factors,  a sample size of 500 was selected.   This
is equal  to about 27 percent of the population.  A 10 percent oversample
was drawn, to  allow for incomplete questionnaires and refusals.  There-
fore, the total  sample  consisted of 550 firms.

Allocation of  the  Sample

Three criteria were considered to allocate the total sample among the
various  sectors.   They  are:   (1) the number of firms in each sector, (2)
an estimate  of the gross sales of each sector, and  (3)  the amount of
variability  in size and types of economic activity in each sector.  In
the  case of  the first criterion, an  allocation of the sample was made,
based entirely upon the number of firms in each sector.  If a sector
 contained 10 percent of the businesses in the county, 10 percent of the
sample was allocated to that  sector.

The  second criterion was used to give  a weighting factor to each sector,
based  upon its total output  relative to the total output of the entire
economy.  In order to do this, it was  necessary to  obtain an estimate
 of the total output of each  sector.  Secondary sources were used to ob-
 tain these estimates.  Even  though  complete sales data were not avail-
 able for all the sectors, it  was possible  to  get  a  general indication of
 the volume of sales.  These  estimates  were  then used to  allocate the
 sample on the basis of the  total output of  each sector.

 As mentioned previously, data were  not available  to estimate  the vari-
 ances  of the parameters to  be estimated.   However,  a subjective measure
 of variability was considered in the allocation of  the sample.  Two  types
 of variability were studied.  The  first was  the variation in  the  size
 of firms in each sector.  The other dealt  with  the  amount of  diversifi-
 cation as to the types of economic  activity in the  sector.

 The two sample allocations,  based  on the  number of  firms in the  sector
 and the total sales of the  sector,  provided a range for  the sample  size
 for each sector.   The  variability within  each sector was then considered.
 If the variability was considered  to be large,  the  actual Cample size
 allocated to the sector was taken  from the upper portion of the  range
 that had been determined.   Table 7  lists  the sectors and the allocation
 of the sample.


                                   73

-------
Table 7.  Distribution of the Sample Among
          the Sectors of the Model
                                              Number of Firms
            Sector                     Population         Sample
 1. Agriculture	         *                 *
 2. Agric. Services	        25               17
 3. Lumber	        39               17
 4. Manufacturing & Processing	        36               20
 5. Lodging	       165               36
 6. Cafes & Taverns	       119               26
 7. Service Stations	       135               34
 8. Construction	       156               40
 9. Professional Services	       160               33
10. Product-Oriented	       259              100
11. Service-Oriented	       442              123
12. Communications & Transportation.        60               23
13. Financial	        30               15
14. Grocery	        95               29
15. Resorts & Marinas	        15                 4
16. Automotive	       104               34
17. Local Governments	         *                 *
         TOTAL ECONOMY	     1,840              551
—                :                                            _   _
  Secondary data was utilized in these sectors.
Once the sample size had been determined for each sector, it was neces-
sary to allocate the sector sample size among the various strata in  each
sector.  All of the firms classified as "large" were included  in the
sample.  A random sample was then drawn from each remaining strata of
each of the 15 sectors that were sampled.  The sample drawn included 551
firms.
Local Governments
Since data concerning local governments were readily available, it was
not necessary to include the various units of government in the county

                                 74

-------
in the interviewing  process.   The expenditures of all governmental units
were obtained  for  the  1967-68 fiscal year.   These expenditures were then
allocated to the various  sectors in the model, with the assistance of the
bookkeeper or  purchasing  department of each unit of government.  These
data were used to  fill in the Local Governments column in the transactions
matrix.

Agriculture

Agriculture, being the second largest industry in Klamath County,  is  an
important sector in  the model.   The 1964 Census of Agriculture estimated
that there were more than 1,000 farms in Klamath County.   However, more
secondary data were  available for agriculture than for any other sector.
It was felt  that the necessary estimates could be determined from the
available data, without the collection of primary data.

                          Sampling Results

Personal interviews  were  conducted with each firm in the sample.  It  be-
came obvious in the  early stages of data collection that 500 responses
would not be obtained from the 551 firms in the sample.  Therefore it was
necessary to draw  replacements for 175 non-responding firms.  The alloca-
tion of,the  second sample was determined by the number of firms in each
stratum that did not respond in the first sample.  If four firms did  not
respond in a particular stratum, four additional firms were drawn from
that stratum to replace the non-responding firms.  Completed question-
naires were  eventually obtained from 438 firms.

Several factors contributed to the low response rate.  First, relying upon
secondary sources  to compile the list of businesses resulted in including
firms in the sampling frame that were no longer in business.  Lists of
businesses become  obsolete very rapidly, due to the rate of business  turn-
overs .

A second factor contributing to the non-response rate was the type of data
being collected.   In order to construct the model, detailed financial
data were required from each firm in the sample.  Most of the refusals
resulted from  firms  who considered the data to be confidential and, there-
fore, would not release it.

The type of businesses in the study area also contributed to the poor re-
sponse.  Many  of the businesses in the county are sub-divisions of larger
companies.   In many  cases they did not have the detailed information
available at their Klamath County office.  The head offices of these
companies were contacted  by mail and telephone.  However, the interview-
ing procedures used  in the study were not well designed for these non-
personal contacts, and many firms did not reply to the requests for in-
formation.  Other  firms indicated they did not maintain the type of data
requested for  individual  branch offices of the company, and therefore
could not respond.  Furthermore, it was not possible to determine the
owners of some businesses such as self-service laundromats and car washes.
                                  75

-------
 Therefore,  some firms  of this type could not be contacted.  The 438 com-
 pleted interviews  represent nearly 24 percent of the firms in the county.
 The questionnaire  used to obtain the data is included in Appendix B.

                     The Klamath County Economy

 Transactions  Matrix

 The transactions matrix, presented in Table 8, was constructed from the
 collected data.  As mentioned earlier,  each sector of the economy is
 listed at the left, and the numbers at the top of the matrix correspond
 to the numbered sectors at the left.   Those listed at the left represent
 the selling sectors, while those across the top indicate the purchasing
 sectors.  The figures  in the cells of the matrix indicate the value of
 goods  and services sold by the sector at the left to the sector at the,
 top.   For example, reading across  the Agriculture row shows that it sold
 $6,189,800  worth of goods and services  to Agriculture (intraindustry
 sales), $602,220 to Agricultural Services, zero to Lumber, $9,484,110
 to Manufacturing and Processing, and so on across the row.  The distribu-
 tion of sales of each  of the other 17 sectors can be determined in the
 same manner.

 The purchases of each  sector of the economy can also be  determined from
 the transactions matrix.   This is  done  by reading down the column of a
 sector.   For  example,  reading down the  first column shows that Agricul-
 ture purchased $6,189,600 of goods and  services from itself, $4,445,044
 from Agricultural  Services,  $75,660 from Lumber,  $168,712 from Manufac-
 turing and  Processing,  and so on down the column.

 The first 17  rows  and  columns comprise  the endogenous portion of the
 transactions  matrix.   Row 18 is  the sum of the first 17  rows in each
 column.   The  figures indicate the  total value of  goods and services the
 sector purchased within the  local  economy.   The figures  provide a general
 indication  of how  much  that  sector depends upon the local economy.   The
 larger the  figure,  relative  to the total purchases  of the sector (Row 23),
 the greater the  magnitude of dependence of that sector,  in terms of its
 inputs, upon  the other  sectors of  the local economy.

 Rows 19-22  comprise  the value-added portion of the  matrix.   The House-
 hold row  (19)  indicates  the  value  of  services purchased  from private
 individuals.   It includes  such payments  as wages,  returns to entrepre-
 neurial services,  interest,  dividends,  and rent.   This figure indicates
 another characteristic  of the various sectors - the larger the payments
 to  Households, relative to the total  purchases of the sector, the more
 labor-intensive  is  the  industry.

 The Import  row  (20) depicts  the  purchases  of goods  and services from out-
 side the Klamath County economy.   This  provides a measure of the degree
 of  self-sufficiency of  the economy.   The large volume of imports purchased
by  the sectors in  this model  indicates  that the economy  of Klamath Countv
 is not nearly self-sufficient.
                                 76

-------
Tabu 8.  Transactions Matrix showing  Interindustry  Flova  in  Dalian,
          Klamath County, 1968  (Rounded  to Nearest $1,000)

1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17

18.

19
20.
21.
22.
23.

Agriculture. * . . * .
Agric. Service*..
Manufacturing 6
Processing 	
Cafes & Taverns*.
Service Stations.
Construction 	
Professional
Services • • * * f , .
Product-Oriented .
Service-Oriented .
Communications &
Transportation .

Resorts &

Local Gov* t ...

Summation^-1'




Depreciation &
Negative Inven-
tory Changes...
Total Inputs5'...
1
6190
4445
76
169
0
0
931
175
132
918
430
244
665
26
5
218
1363

15985

9629
61
573
2607
28854
2
602
204
o
*
o
1
83
0
15
371
57
760
128
o
o
77
40

2338

1167
7203
130
289
11129
3
o
37
9079
149
*
19
660
260
39
1960
169
12880
620
48
o
1000
1204

28122

22774
15634
8166
5123
79819
4
9484
203
143
A
o
*
149
228
626
46
2584
146
o
o
68
96

13799

3582
4539
645
591
23155
5
Q
3
0
82
o
0
47
33
17
773
162
331
96
o
o
o
96

1636

1590
502
137
461
4326
6
35
3
Q
720
o
0
561
104
7
1871
203
54
17
174
0
o
41

3285

2213
344
327
294
6464
7
o
0
0
o
0
1685
64
1 C
346
100
146
49
o
o
326
51

2765

2493
10283
569
284
16394
P U ]
8
Q
12
425
294
g
18
437
2214
37
971
292
322
148
o
o
193
70

5439

4760
7383
504
527
18613
1 C H A
9

3
Q
26
*
0
52
28
261
134
230
59
0
0
58
79

951

8147
4108
758
776
14740
SING
10

0
375
375
Q
0
213
927
1ft
993
257
1887
82
g
o
445
986

6765

8496
27114
4325
863
47564
SEC
11

46
3
21
*
0
95
231
815
849
424
48
0
0
128
297

3035

6103
5937
917
488
16480
TOR
12

12
Q
1
o
0
362
7
1Q
242
15
100
17
19
0
92
810

1697

8660
13514
3707
2475
30053
13

0
0
7
*
1
44
173
305
109
65
o
0
5
42

762

1526
3354
291
147
6079
14
160
0
g
3887
Q
0
35
22
1743
110
429
17
1009
0
7
57

7466

2395
13419
482
269
24060
15
Q
6
Q
26
Q
0
92
2
2
64
23
0
17
0
12
12
29

274

109
281
22
17
703
16
Q
3
0
ft
0
41
23
20
224
254
182
433
o
o
1064
60

2324

3338
13267
378
629
19936
17
j
82
Q
314
*
1
161
565
11 5
782
314
103
0
43
1
343
75

2899

8466
907
958
0
13230
19
88
1574
164
1886
4145
6389
10505
6232
12554
25394
11768
4833
3325
21822
684
14700
1892

127953





127953
20
12223
4108
66842
11621
117
7
644
629
411
5895
668
2577
148
651
3
379
0

106924





106924
21
58
703
1589
3363
50
7
108
6818
1158
2106
253
1844
0
176
1
458
5942

24002





24002
22
0
316
912
222
0
21
37
28
53
1037
72
13
0
84
9
344
0

3147





3147
23
28854
11129
79819
23155
4326
6464
16394
18613
14740
47564
16480
30053
6079
24060
703
19936
13230

361597

95449
127849
22889
15840
623623
    Hay not SUB,  due to rounding error.

    Number in cell rounds to zero.

-------
Columns 19 through 22 comprise the final demand for goods and services
sold by the sectors listed at the left of the matrix.  The four  compo-
nents of final demand indicate the major uses of the goods and services
sold.  The Household column (19) represents the value of goods and  ser-
vices sold to individuals, while the Government column (21) shows the
value of sales to state and federal governments.  Column 20 represents
exports from Klamath County by the various sectors, while Column 22
indicates positive inventory changes.  Column 23 summarizes total sales
by each sector.  The remaining rows and columns of the transactions ma-
trix are self-explanatory.

Direct Coefficients Matrix

The direct coefficients, or "A" matrix, is presented in Table 9.  The
matrix is utilized by reading down the columns in order to determine the
input structure of the sectors of the economy.  In most cases, the  a   's,
which represent Sector j's purchases per unit of output, reveal  more
about the structure of a sector than the absolute magnitude of inter-
industry sales depicted in the transactions matrix.

The first column of the "A" matrix shows that Agriculture must purchase
21 cents worth of goods and services from itself if it is to increase
its output by one dollar.  Continuing down the column, Agriculture  must
also purchase 15 cents worth of goods and services from Agricultural Ser-
vices;  less than one cent from Manufacturing and Processing, and Lumber;
zero from Lodging, and Cafes and Taverns;  and so on down the column.
The Summation row (18) shows that Agriculture purchases over 55  cents
worth of goods and services from the local economy for every dollar of
output.  Row 19 shows that Household incomes increase 33 cents per  dollar
increase in sales of the Agriculture sector.

                  Characteristics of the Economy

Some of the characteristics of the local economy can be studied, using
the numbers in the transactions and direct coefficients matrices.   A
breakdown of the sales of each of the 17 sectors is given in Table  10.
The dollar value and percentages of sales are separated between  the local
economy and final demand.  It is interesting to note that only the  Agri-
culture and Communications and Transportation sectors sell more  than one-
half of their total output to the businesses in the local economy (57 and
69 percent, respectively).  Most of Agriculture's local sales (95 percent)
are to itself and Manufacturing and Processing;  the Lumber sector  pur-
chases almost 62 percent of Communications and Transportation interindus-
try sales.  This is primarily due to the use of transportation facilities
for shipping lumber and wood products out of the county.

Five sectors sell less than 10 percent of their total output to  the sec-
tors of the local economy (see Table 10).  They are: Lodging (0.33  per-
cent), Cafes and Taverns (0.60 percent), Professional Services (3.83 per-
cent), Grocery (5.52 percent), and Resorts and Marinas (0.93 percent).
                                 78

-------
                         Table 9.  Direct Coefficients Matrix, Klamath County,  1968
VO

1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14,
15.
16
17.

18.

19
20.
21
22.


Agric. Services.. .
^•"fflbflr. .!•...... >
Manufacturing &
LodfeinK ........
Cafes & Taverns . . .
Service Stations..
Professional
Services ........
Product-Oriented. .
Service-Oriented. .
Communications &
Transportation. .

Resorts &



Summation— 	 	




Depreciation &
Negative Inven-
tory Changes ....
I
2145235
.15*0545
.0026222
.0058472
0
0
.0322691
.0060485
.0045879
.0318111
.0149091
.0084421
.0230436
0009011
.0001560
0075415
,0472506

.5540078

.3337123
.0020977
0198452
.0903370
2
0541127
.0182977
o
.0000182
o
.0000496
.0074740
o
0013681
.0333225
.0051041
.0682645
.0115460
o
0
0069620
. 0036119

.2101313

. 1048810
.6472559
0117208
.0260110
3
o
.0004619
1137430
.0018617
0000018
.0002331
.0082637
0032604
0004858
.0245554
.0021204
.1613647
.0077651
.0006067
o
.0125246
,0150783

. 3523266

.2853242
. 1958690
.1023003
.0641799
4
4095844
.0087807
0061852
.0000057
Q
.0000158
.0064210
0098342
0010641
.0270465
.0020055
.1115995
.0063175
o
o
.0029208
. 0041350

.5959160

.1546968
. 1960159
.0278680
.0255033
5
o
.0007046
o
.0188407
o
0
.0108871
0075774
0028599
.1787581
.0374578
.0766000
.0222967
0
o
0
.0222105

. 3781929

.3675640
. 1159689
.0316050
.1066693
6
0054769
.0004716
o
.1113510
0
0
.0086890
0160701
0010932
,2894551
.0313445
.0084104
.0025868
0269123
0
o
.0063460

.5082068

. 3423530
.0532527
.0506385
.-0455490
P U R C H
7
o
0
0001830
o
o
0
.1028092
.0026576
.0009287
.0211160
.0060842
.0089191
.0029653
o
o
.0198870
.0030832

.1686333

.1520755
.6272337
.0347320
.0173256
A S I N G
8
0
.0006706
0228384
0157952
0003481
.0009698
.0234820
1189359
0019959
.0521561
.0156754
.0172840
.0079256
o
o
.0103490
.0037769

.2922030

.2557389
.3966703
.0270772
.0283106
SECT
9
0008005
.0002068
o
0017970
0000293
0
.0035185
0019020
0005697
.0176928
.0090787
.0156299
.0040125
0
0
0039508
,0053372

.0645257

.5527168
.2786876
.0514301
.0526398
0 R
10
o
0
0120870
.0078912
o
0
.0044844
0194974
0003736
.0208713
.0053941
.0396769
.0017154
0001717
0
0093467
,0207205

.1422302

.1786292
.5700667
.0909371
.0181367
11
o
.0027971
0001820
.0012742
000010n
0
.0057526
.0140374
.0046699
.0494391
.0515277
.0257284
.0029308
0
0
.0077758
.0180357

.1841608

.3703383
. 3602376
.0556542
.0296090
12
o
.0004057
o
.0000479
o
0
.0120488
.0002456
.0006469
.0080421
.0005024
.0033210
.0005513
. 0006252
o
.0030664
.0269524

.0564557

.2881784
.4496803
.1233446
.0823409
13
o
0
o
0
0010795
.0000604
.0002445
.0072518
.0017290
.0285294
.0501277
.0178865
.0106717
0
0
.0007480
.0069547

.1252831

.2509756
.5517197
.0478354
.0241862
14
0066699
0
.0003325
.1615655
0
0
.0014430
.0009029
.0004641
.0724594
.0045806
.0178337
.0006948
.0419426
0
.0002876
.0023593

.3115359

.0995215
.5577133
.0200398
.0111896
15
o
.0086761
0
.0374943
0
0
.1311882
.0021342
.0025126
.0907160
.0327138
0
.0237859
0
.0021342
.0173536
.0417669

.3904757

.1550822
. 3992442
.0312953
.0239026
16
o
. 0001529
o
0
.0000072
0
.0020521
.0011785
. 0009 786
.0112363
.0127375
.0091382
.0217047
o
o
.0543582
.0030266

.1165707

.1674407
.6654813
.0189422
.0315651
17
.0000548
.0061895
0
.0237040
.0000363
.0000654
.0121899
.0426874
.0087201
.0590806
.0237012
.0077579
o
0032325
. 0000401
0259638
00566 76

2190911

.6399147
.0685791
0724151
0
                          —  May not  sum,  due  to  rounding error.

-------
Table 10.  Distribution of  Sales  of  Each Sector in the Klamath County Economy



1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.

13.
14.
15.
16.
17.


Sector



Manufacturing & Processing...
Lodging 	




Product-Oriented 	 * ...

Communications & Trans-

Financial 	 	





Interindus t ry
sales
$16,484,515
5,050,809
10,311,758
6,064,412
14,407
38,804
5,100,946
4,906,159
563,723
13,132,289
3,718,892

20,785,067
2,605,379
1,327,255
6,530
8,054,992
5,395,930

Sales to
final demand
$12,369,200
6,068,819
69,507,488
17,091,036
4,311,338
6,424,758
11,292,742
13,707,191
14,176,275
34,431,310
12,761,510

9,267,485
3,473,132
22,733,102
696,323
15,880,679
7,833,912
Percent of
sales sold to
local economy
57.13
45.47
12.92
26.19
0.33
0.60
31.12
26.36
3.83
27.61
22.57

69.16
42.86
5.52
0.93
40.40
40.79
Percent of
sales sold to
final demand
42.87
54.53
87.08
73.81
99.67
99.40
68.88
73.64
96.17
72.39
77.43

30.86
57.14
94.48
99.07
59.60
59.21
                                                                                                           o
                                                                                                           CO

-------
The reason for  the  low percentages is that each of these sectors primar-
ily serves the  needs  of the Household sector,  which is exogenous in this
model.

As mentioned earlier, Row 18 of the direct coefficients matrix (Table 9)
indicates the value of each sector's purchases, per dollar of sales,
that are obtained from the local economy.   Only three sectors purchase
more than one-half  of their goods and services from the local economy.
They are: Agriculture (55.4 percent);  Manufacturing and Processing
(59.6 percent); and  Cafes and Taverns (50.8 percent).  The Communica-
tions and Transportation sector is least dependent upon the local econo-
my, as it purchases only 5.6 cents worth of goods and services,  per dol-
lar of output,  from local businesses.

In absolute terms  (Table 8) , Lumber purchases  more goods and services
($28,122,000) from  the Klamath County economy  than any of the other sec-
tors.  Agriculture  ranks second ($15,985,000), followed by Manufacturing
and Processing  ($13,799,000).  This gives  some indication of the impor-
tance of these  sectors in the economy.
The Household coefficients (a^'s) in Row 19 of the direct coefficients
matrix  are  important because they indicate the amount household incomes
will  rise if  the output of Sector j increases one dollar.   They repre-
sent  the direct effect a one-dollar change in output will  have upon
household incomes.   It is interesting to note that the Local Government
sector  has  the largest a, . (.64) in the economy.  This indicates that
payrolls and  other  payments to households account for a large portion of
the budgets of the  various units of local government.  The Professional
Services and  Service-Oriented sectors rank second and third, with a^.'s
of  .55  and  .37 respectively.  One would expect the service sectors to
have  larger household coefficients, since they are labor-intensive in-
dustries.   Grocery  and Agricultural Services have the smallest household
coefficients, .0995 and .1049 respectively.

As noticed  earlier, the large quantity of imports (Row 20  of Tables 8
and 9)  purchased by the various sectors in the model indicates that the
economy is  highly dependent upon the "rest of the world" as a source of
goods and services.  The large importers are those sectors that deal in
products that cannot be produced within the economy.  They include Agri-
cultural Services (64 percent of purchases are imports), Product-Oriented
(57 percent), Service Stations (62 percent), Grocery (56 percent), and
Automotive  (67 percent).  Agriculture imports less, in absolute and rela-
tive  terms, than any other sector.

A final characteristic of the Klamath County economy can be seen by
studying the  Export column (20) in Table 8.  The fact that all of the
sectors (except Local Government) export some goods and services illus-
trates  that the economy serves as a trading center for other communities
outside the county.  This is especially true at the wholesale trade level.
The largest exporters are the basic industries of the economy.  Agricul-
ture  exports  more than $12 million, Lumber $66.8 million,  and Manufacturing


                                  81

-------
and Processing about $11.6 million.  A large portion of the exports of
the latter sector are agricultural commodities that have been processed
by local firms.

The foregoing discussion illustrates that, like most small regional econo-
mies, the Klamath County economy is highly specialized.  Lumber and Agri-
culture are the basic industries of the economy.  Many of the goods and
services used in the economy must be imported, while the primary indus-
tries export large quantities of products.
                                 82

-------
                             SECTION  IX

               APPLICATION OF  THE  INPUT-OUTPUT  MODEL

            The Direct  and Indirect  Coefficients Matrix

The Xl-A)   or direct and indirect coefficients matrix is  presented  in
Table 11.  It contains  17 rows and 17  columns,  one  for each  of  the 17
sectors of the economy.  Because of  the  direct  and  indirect  effects  ex-
plained in Section VII,  the  coefficients in  the (I-A)"1 matrix  are larger
than those in the corresponding cells  of the direct coefficients matrix.
That is, the coefficients in the  (I-A)    matrix indicate the total in-
crease in output of  a sector resulting from  a change  in demand  for the
output of a sector.  For example,  assume that there is a one-dollar  in-
crease in final demand  for the products  of Agriculture.  This sets into
motion a series of changes in  the  output of  all the sectors  of  the econ-
omy.  When the change has worked itself  out, Agriculture's output will
have increased $1.29;   Agricultural  Services, 20 cents;  Lumber, one-half
cent;  and Manufacturing and Processing  one  cent, and so on  (reading down
the Agriculture column) .

                         Output Multiplier

The eighteenth row of the  (I-A)   matrix is  the sum of the first 17  rows
of each column.  The figures in the  row  represent the change in total
output of the economy resulting from a one-dollar change in  final demand
of the sector listed at  the  top of the column.  For example, a  one-dollar
increase in final demand of  Agriculture  will cause  the output of the en-
tire economy to increase $1.82. This  is called the output multiplier of
the sector.  The magnitude of  the  output multiplier of each  sector depends
upon the quantity of goods and services  the  sector  purchases from the
local economy.  Earlier it was pointed out that Communications  and Trans-
portation purchased  only 5.6 percent of  its  purchases from the  local econ-
omy.  The output multiplier  of that  sector is small (1.07),  as  would be
expected.  Conversely,  Manufacturing and Processing purchases more of its
goods and services from the  l^cal  economy than  other  sectors , and the out-
put multiplier of the sector is the  largest  (1.96)  in the  economy.   The
output multiplier for each sector  is listed  in  Column 2 of Table 12.

                     Income-Output  Coefficients

Another useful tool  is  the "income-output" coefficient (1^)  listed in
Column 3 of Table 12.   These coefficients are computed as  follows:
                                   a    r
                                         ik
                                  83

-------
Table 11.   Direct and Indirect  Coefficients Matrix,  Klamath  County,  1968

1
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13
14.
15.
16
17.

18.



Agric. Services....
Manufacturing &

Cafes & Taverns ....
Service Stations...
Professional
Product-Oriented. . .
Service-Oriented. . .
Communications &
Transportation. . .
Financial

Resorts &
Marinas 	



Summation^- 	 	

1
1.2924189
.2034219
.0050732
0101022
.0000440
.0000316
.0505540
0143484
.0071093
.0561251
.0260502
.0319157
0334278
.0014715
.0002046
0158705
.0653672

1.8155362

2
0715281
1.0299960
0008269
0010847
0000166
.0000550
.0126468
0022169
0019811
.0400509
.0082094
.0738215
0141072
0001595
. 0000116
0094640
0103294

1.2765056

3
0017655
.0010798
1 1289852
00321Q4
0000156
.0002715
.0137084
0063512
0009765
.0329186
.0043800
.1854346
0095785
0009295
.0000012
0169545
0232081

1.4297781

4
5303589
.0925499
0098114

0000303
.0000441
.0302422
0183536
0042238
.0547837
.0140249
.1288916
0206359
0007187
.0000843
0110564
0351263

1.9557740

5
0114250
.0030697
0031961

1 0000328
.0000198
.0160474
0155995
0036325
.1904989
.0433469
.0899685
0237899
0002092
.0000030
0041341
0305327

1.4569958

P U R C H
6
0705473
.0128891
0060240

0000180
1.0000333
.0166331
0286043
0021509
. 3094692
.0378958
.0391008
0063267
0283062
.0000118
0058081
0188689

1.7024360

A S I N G
7
0002174
.0001018
0006957

0000063
.0000049
1.1151830
0043592
0011723
.0255200
.0080439
.0117901
0039979
0000286
. 0000002
0240042
0046023

1 2001299

S E C T 0
8
0100951
.0026726
0303902

0004081
.0011121
.0316320
1 1378045
0026343
.0657223
.0207097
.0309500
0103770
0001196
0000019
0148427
0081818

1.3866376

R
9
0022119
.0006590
0003767

0000355
.0000039
.0046137
0031860
1 0007314
.0197731
.0102899
.0173916
0043564
0000397
0000006
0048693
0066366

1.0773725

10
0048521
.0010560
0146981

0000123
.0000288
.0071525
0242193
0007632
1.0259884
.0073289
.0453898
0026005
0003072
0000017
0116814
0234121

1 1786482

11
0016276
.0034530
0015116

0000224
.0000209
.0084350
0192585
0052421
.0567059
1.0560206
.0311355
0037810
0001061
0000011
0104146
0214906

1 2218764

12

.0006905
0002068

0000026
.0000038
.0140276
0009295
.0106190
.0015324
1.0044428
0007713
0007512
0000012
0044683
0275880

1 0694795

13
0004275
.0003060
0007493
0007721
0010962
.0000720
.0015410
0104380
0021429
.0338895
.0541728
.0215342
0000550
0000004
0020837
0094686

1 1499592

14
0988687
.0171587
0032411
1703329
0000079
.0000115
.0080539
0063725
0013759
.0880268
.0083282
.0444174
0047084
1 0439570
0000159
0034427
0112854

1 5096049

15
0217108
.0130502
0020454
0398947
0000340
.0000152
.1496507
0089977
0035552
.1049830
.0397294
.0137089
0264103
0002099
1 0021440
0248025
0474299

1.4983721

16
0001674
.0002722
0002621
0002817
0000340
.0000041
.0028894
0024012
0012092
.0142181.
.0157132
.0113082
0233338
0000249
.0000002
1 0580259
004 3110

1.1344570

17
0142783
.0089656
0024848
0259868
.0000574
.0001174
.0167171
0513798
.0092312
.0679513
.0275269
.0168817
0019919
0034470
.0000428
0299144
1.0093432

1.2863176

                                                                                                                                                                                                                                  CO
—  May not sum, due to rounding error.

-------
Table 12.  Output and Income Multipliers  and Income-Output
           Coefficients  for Each Sector of the  Klamath  County
           Economy
(1)


Sector

2. Agricultural Services 	 ,


5. Lodging 	

7. Service Stations 	
8. Construction 	

10 . Product-Oriented 	
11. Service-Oriented 	
12. Communications & Transportation
13. Financial 	

15. Resorts & Marinas 	


(2)

Output
multiplier
,. 1.82
. . 1.28
.. 1.43
.. 1.96
,. 1.47
. . 1.70
.. 1.20
.. 1.39
. . 1.08
.. 1.18
.. 1.22
.. 1.07
.. 1.15
.. 1.51
.. 1.50
.. 1.13
.. 1.29
(3)
Income-
Output
coefficients
.55
.18
.41
.43
.49
.50
.19
.35
.57
.23
.44
.31
.30
.21
.28
.20
.71
(4)

Income
multiplier
1.66
1.71
1.43
2.83
1.32
1.45
1.25
1.38
1.04
1.30
1.18
1.09
1.19
2.11
1.78
1.19
1.11
 where
the value of purchased labor services from the
Household sector by the J*  sector, per dollar
of total output in the j   sector;  that is, it
is the a. row 
-------
For example, the income-output coefficient for the Agriculture sector
(H1 ) is calculated as follows :


Hl = a19x  + a!92 (r21} + a!93 (r31> +  — + a!917  
-------
of a sector must increase to obtain the desired one-dollar change in
income for that sector.

              Estimation of Recreational Expenditures

To estimate the local  benefits  attributable to recreational activities
at Klamath Lake as  it  exists at the present time,  the quantity of recrea-
tional expenditures that were made in the county during 1968 had to  be
estimated.  The data collected  from re creationists at Klamath Lake,  to
estimate the  recreational demand model for the lake,  were also used  to
determine the type  and magnitude of recreational expenditures made in
Klamath County.  Only  those questionnaires completed  by recreationists
who stayed at Klamath  Lake less than 20 days were used.  There are 43
such questionnaires.

Travel Cost

The relevant  travel cost for this analysis is all purchases made in  Klam-
ath County by recreationists while traveling to and from Klamath Lake
during 1968.  Travel cost is composed of six categories:  Automobile,
Cafes and Taverns,  Grocery, Lodging, Camping, and "Other" costs.

To estimate the travel costs incurred in Klamath County, the 43 question-
naires were put into one of two groups.  The first contains those recrea-
tionists who  were residents of  the county, while the  second included the
non-Klamath County  residents.  The first group contains 29 of the 43 ob-
servations.   It is  assumed that all travel costs incurred by the first
group were spent in the  county, since the origin and  destination were
both in the same county.  Automobile expenses for this group were com-
puted by multiplying the number of miles traveled to  and from the site
by an average cost  per mile of  five cents.  All other components of  travel
cost were estimated from data contained in the questionnaire.  All com-
ponents of travel cost for the  resident group were divided by the number
of people in  the party to obtain the desired estimate of average travel
cost per person.

A different procedure  had to be developed for the non-resident group.  It
could not be  assumed that all expenditures associated with traveling to
and from the  site would  be made in Klamath County. However, since the
questionnaire indicated  where the recreationists came from, it was pos-
sible to estimate the  total miles traveled in the county.  The number of
miles traveled inside  the county was then multiplied  by 5 cents to esti-
mate automobile travel expenses incurred within the county.

An additional step  was required to estimate the other five components of
travel cost for the non-resident group.  The number of miles traveled in
Klamath County was  divided by the total miles traveled during the trip
to obtain the percent  of the total distance that was  traveled within the
county.  These percentages were then applied to the other components of
travel cost to estimate  the amount that was spent in  the local economy.
That is, if the distance traveled inside Klamath County was 10 percent of
                                  87

-------
the total distance, it was assumed that 10 percent of the total  expendi-
tures for food, lodging, and so forth was made inside Klamath  County.
It was again necessary to divide these estimates by the number of  people
in the party, to obtain the average travel cost per person.  Table 13,
Column 2, shows the average amount per person spent inside the county  for
each of the components of travel cost.
Table 13.  Average and Total Travel Cost Incurred in
           Klamath County, by Component, in 1968
(1)
Component





Other 	
TOTAL 	

(2)
Average
travel cost
, ... $ .6371
, ... .1088
, ... .6360
, ... .2477
. .. .0458
.1242
, ... $1.7996

(3)
Total travel
cost
$ 93,329
15,938
93,168
36,286
6,709
18,194
$263,624

It should be emphasized that the average travel cost computed here  is not
the same as that computed earlier.  The previous estimate  ($6.84) is
larger than the one computed here ($1.80).  The primary reason  for  the
difference is that the estimate used here represents only  those travel
costs incurred in Klamath County, while the other figure represents total
travel expenditures, without geographical reference.

The estimated average travel cost incurred in Klamath County was multi-
plied by the number of people visiting the site in 1968, to estimate  the
total amount of travel expenditures made in the county during the year.
The U.S. Forest Service estimates that 146,491 people visited Klamath
Lake in 1968.  Therefore, recreationists spent an estimated $263,624
(146,491 x $1.7996) in Klamath County while traveling to and from Klamath
Lake.  A breakdown of this cost into its components is given in Column  3
of Table 13.

On-Site Costs

An estimate of the average on-site cost per person per day was  computed
in a manner similar to the travel costs discus: ed in the previous section.


                                 88

-------
Since it can be assumed  that  all on-site purchases  were made  in  the county,
it was not necessary  to  separate the observations into the local-  and
non-local-resident  groups.

There is one difference  between estimating travel costs and estimating
on-site costs.  Travel costs  are estimated on a per-person basis,  while
on-site costs are estimated on a per-person per-day basis. That is, the
various components  of on-site cost are divided by the number  of  visitor-
days the party stayed at the  lake.  On-site costs are computed in  this
manner because it is  hypothesized that the average  number of  days  spent
at the site during  each  visit will increase as the  water quality of the
lake improves.  This  point  will be expanded upon later.  Column  2  of Table
14 shows the average  daily  cost per person for each component of on-site
costs.
Table 14.  Average and Total On-Site Cost, by Component,
           for Klamath Lake in 1968

           (1)                 (2)                     (3)

                          Average on-site         Total on-site
Component
Cafes 	 	





Bait 	

TOTAL 	

cost /day
$ .2132
.2985
.3360
. 0209
.0517
1.0473
.0169
.0266
$2.0111

cost
$ 61,527
86,143
96,965
6,032
14,920
302,237
4,877
7,676
$580,377

The number of visitor-days enjoyed at Klamath Lake in 1968 is estimated
by multiplying the number of visits (146,491) by the average length of
stay per  visit (1.97), as estimated in the demand model.  This gives
288,587 visitor-days.   An estimate of total on-site costs for Klamath
Lake in 1968  is obtained by multiplying the number of visitor-days by the
on-site cost  per visitor-day.  Total on-site costs are estimated to be
$580,377.   Total on-site cost is allocated among its components in Column
3 of Table 14.  The average on-site cost per day in Table 14 is greater
                                  89

-------
than the value vised in the demand model, because  an adjustment was not
made for food costs the recreationists would have incurred  if they had
stayed at home.

On-site costs and travel costs are aggregated  to  obtain  the total expendi-
tures in Klamath County associated with recreation at Klamath Lake in
1968.  Thus, $844,001 represents the total annual cost incurred in Klamath
County by recreationists at Klamath Lake in 1968,  given  the present water
quality of the lake.  Table 15 shows how much  of  the total  figure was
spent in the various sectors of the economy.   This was determined by assign-
ing each component of travel cost and on-site  cost to the appropriate sec-
tor of the model.
Table 15.  Total Recreational Expenditures and Percentages,
           in Klamath County, by Sector, Associated with
           Recreation at Klamath Lake in 1968

         (1)                     (2)                     (3)
                                                Percent of total
                                                  recreational
      Sector               Expenditures           expenditures






TOTAL .....

... $403,658
75,859
187,302
132,292
12,811
25,370
. .. $837,292

48.21
9.06
22.37
15.80
1.53
3.03
100.00

It should be noted that the total estimated in Table 15  ($837,292)  does
not equal the sum of the totals in Tables 13 and 14  ($844,001).   The
difference ($6,709) is due to camping fees incurred  in Klamath County
while traveling to and from Klamath Lake.  It is assumed that these camp-
ing fees were paid to the state or federal government, and not to any of
the sectors of the local economy.  However, camping  fees incurred while
at the site were allocated to the Resorts and Marinas sector, since all
camping facilities at the site are privately owned.

Note that only six sectors of the economy are directly affected  by  recrea-
tional expenditures.  Column 3 of Table 15 shows the percentage  of  the
recreational expenditures that each of the six sectors received.  The
                                 90

-------
Service Stations sector  received almost half of the total recreational
expenditures made in  the county.

          The Impact  of  Recreational Expenditures  in 1968

Input-output projections are based on Equation (68).  The first  step  is  to
project a new level of final demand that is expected to occur due to  an
exogenous force acting upon the economy.  The new  "projected" level final
demand is post-multiplied with the R matrix to obtain the new projected
total output of each  sector:

                           X = (R) (Y)
                  s\
where            Y =  projected final demand vector,
                  /\
                 X =  projected total output vector,
                  R =  (I-A)"1 matrix.

The expenditures  that recreationists at Klamath Lake made in 1968 are
viewed as a change in final demand in the input-output model. The in-
crease in recreational expenditures are multiplied through the model  to
estimate the total effect of recreational expenditures associated with
Klamath Lake in  1968.  The projected increase in final demand is listed
in Column 2 of Table  16, while the projected total output is listed in
Column 3 of Table  16.

Although only six sectors are directly affected by recreational  expendi-
tures, all  sectors in the economy are indirectly affected.  This provides
an illustration  of the importance of the from-to model in the study.   If
the relationships between the various sectors of the economy had not  been
specified,  the total  effect of the recreational expenditures would have
been underestimated.

The increase in  household income resulting from recreational expenditures
at Klamath  Lake  is also estimated.  Earlier it was noted that the a^  's
of the direct coefficients matrix specify the amount that household in-
comes of that sector  will increase if the output of the sector increases
by one dollar.   Therefore, the increase in household income is determined
by multiplying the estimated increase in output of a sector by the a^ of
that sector.  The  results are listed in Column 4 of Table 16.  It is  esti-
mated that  county household income in 1968 was $227,000 higher as a result
of the recreational  activities at Klamath Lake.
                                  91

-------
Table 16.  Projected Increases in Final Demand, Total Output,
           and County Household Income, by Economic Sector,
           Associated with Recreation at Klamath Lake in 1968
        (1)
      Sector
     (2)
  Proj ected
final demand
      /\
     (Y)
                                          (3)
(4)
                                       Projected
                                     total output  Projected increase
                                          (X)      in household income
 1.  Agriculture	 $      0
 2.  Agric. Services...        0
 3.  Lumber	        0
 4.  Manufacturing &
       Processing	        0
 5.  Lodging	  132,292
 6.  Cafes & Taverns...   75,859
 7.  Service Stations..  403,658
 8.  Construction	        0
 9.  Professional
       Services	        0
10.  Product-Oriented..   25,370
11.  Service-Oriented..        0
12.  Communications &
       Transportation.        0
13.  Financial	        0
14.  Grocery	  187,302
15.  Resorts &
       Marinas	   12,811
16.  Automotive...	        0
17.  Local Gov't		0_
           TOTAL	 $837,292
                                      $   25,870
                                           4,832
                                           2,167

                                          44,736
                                         132,302
                                          75,869
                                         457,144
                                           7,916

                                           1,440
                                         102,841
                                          14,111

                                          29,275
                                           6,527
                                         197,732

                                          12,843
                                          11,936
                                          10.644
                                      $1,138,185
                                 $  8,633
                                      507
                                      618

                                    6,921
                                   48,629
                                   23,974
                                   69,520
                                    2,024

                                      796
                                   18,370
                                    5,226

                                    8,437
                                    1,637
                                   19,674

                                    1,992
                                    1,998
                                    6.811
                                 $227,767
                                 92

-------
                              SECTION X
           •j
           ECONOMIC  BENEFITS OF WATER QUALITY IMPROVEMENT


In Section VI  the demand model for Klamath Lake was  estimated  for  the
existing water quality regime.  The net economic value  of the  lake in
1968 was estimated to be about $82,000.  The regional benefits associated
with recreation at Klamath Lake in 1968 were also estimated by using the
input-output model of Klamath County.  Household income in the county  in-
creased an estimated $227,000 as a result of recreation at Klamath Lake
in 1968.   Now  consider the effects of improving the  characteristics of
Klamath Lake.

                   An Improvement in Water Quality

A possible improvement in the water quality of Klamath  Lake might  consist
of two steps.   The first step would be the removal of the blue-green algae,
while the  second would be decreasing the water temperature and improving
the beaches around the lake.

It is hypothesized that the proposed improvements will  affect  the  use-
intensity  ratings of the lake, since the lake would  be  better  suited for
the various recreational activities.  For example, the  removal of  algae
(Step 1) would increase the use of the lake for boating and water-skiing,
while lower water temperatures and improved beaches  would increase swim-
ming and fishing at  the lake.  Table 17 shows the hypothesized use-intens-
ity ratings associated with Steps 1 and 2 described  above. The use-intens-
ity ratings for Steps 1 and 2 were estimated by personnel at the E.P.A.
Pacific Northwest Water Laboratory in Corvallis, Oregon.
Table 17.   Hypothesized Use-Intensity Ratings  for Klamath Lake
            at  the Present Time,  and after Steps 1 and 2

                                   Klamath Lake Use-Intensity Ratings
       Activity                    Present       Step 1        Step  2





	 0
	 1
	 1
	 1

1
3
3
2

3
3
3
3

One other possibility  has  to be considered:   Will improvements in water
quality affect any of  the  other variables in the demand model?  That is,
                                  93

-------
will the average on-site cost, travel cost, or  income of  recreationists
change as the characteristics of the site improve?  To  determine  this,  the
average values of p. , k, and y for each lake were computed  and  regressed
against the site characteristics.  The site characteristics were  represented
by summing the swimming, boating, water skiing, and fishing use-intensities
for each lake.  The  three equations are:

              p. = 1.638 +  .066Q                 R2 =  .507    (71)
               1            (.046)

              k  - 12.075 +  .475Q                R2 -  .519    (72)
                            (.323)

              y  - 5,031.295 +  529.637Q          R2 -  .743    (73)
                               (220.150)

where Q represents the sum of the use-intensities of each lake.

The sample data suggests that the average values of p, , k,  and  y  increase
as the sum of the use-intensities of a lake increase.   However, the  R  for
each equation, and the coefficients of Q in each of the three equations,
are not significant  at t'-e 10 percent level.  Therefore,  no relationships
between Q and the p., k, and y variables are assumed to exist.  Any  im-
provements of water  quality will enter the demand model only through the
W and F variables.

              Demand Model for Klamath Lake (Step 1)

After removal of the algae from Klamath Lake, the estimated demand model
is:

     k* = 44.151 + .002y + 10.435?                             (74)
     p* - 10.060 +  ,269k - .004k2 +  .000000017y2               (75)

        m e.759 - .0064k* +  .0064k + .0637p* -  ,0637p.         (?6)


The estimated average critical travel cost and  average  critical  on-site
costs are $90.38 and $12.07, respectively.  Introducing these  averages
into the q^ relation, and holding k  at the average value estimated  prior
to Step 1, the demand function is estimated to  be:
                             .993 -  .0637?..                    ,__.
                            e            Kl                    (77)
The difference between this demand function and one previously  derived
for Klamath Lake is the larger constant term.  Thus the  demand  curve has
shifted to the right, indicating that the length of stay per  visit has in-
creased.  The demand curve is presented graphically in Figure 13.   The
average length of stay per visit has increased from 1.97 to 2.41  days.
Note that the average on-site cost is the same value  that was used in the
previous demand model for Klamath Lake.
                                 94

-------
   $12.07
  $1.84
                            e'993 -  '°637 PI
              1.25   2.41          x


Figure 13.  The Average Recreationist's Demand Curve, Per
            Visit, for Klamath Lake (Step 1).
                        95

-------
The value per visit is estimated to be:

                    12.07

Value per visit =   j      (e'"3 "  '0637Pi) d?1 -  [(10.23) (1.25) ]    (78)

                 1.84
               = $18.29 -  $12.79 -  $5.50.

Before the net economic value of Klamath Lake, after  Step  1,  can be de-
termined it is necessary to estimate the number of visits  to  Klamath Lake
after the algae has been removed.   Substituting the larger values  of W
and F into the visits equation gives an estimate of 234,386 visits.  This
represents a 60 percent increase in the number of visits.

The estimated net economic value of Klamath Lake, after  removing the algae,
is:
                     ($5.50) (234, 386) = $1,289,123.

The estimated value of Klamath Lake as it now exists  is  $82,035.   The value
of removing the algae is represented by the difference in  the two  estimated
values, or $1,207,088.  That is, based on 1968 data,  $1.2  million  worth of
recreational benefits could be obtained per year by removing  the algae
from Klamath Lake.

              Demand Model for Klamath Lake (Step 2)

It is hypothesized that lowering the water temperature of  the lake and im-
proving the beaches would  cause the swimming use-intensity to rise from low
to high, and the fishing use- intensity to increase from  medium to  high.
Again, it is assumed that  all other variables are unaffected  by the pro-
posed improvements.  Substituting the higher use-intensity ratings into
the model yields the following demand model for Klamath  Lake  after Step 2:

     k* = 60.426 + .002y + IQ.MSp                             (79)
     P1 - 14.250 +  .269k -  .004k2 +  .000000017y2               (80)

        m e.759 - .0064k* +  .0064k +  .0637p* -  .0637p1         (81)

     *      *
The k  and p  functions have again shifted, representing  an  increase in
the average k* and p. for recreationists at Klamath Lake.  The revised
averages are $106.65 and $16.26 respectively.   The proposed  revised de-
mand function is :

                         ,x - e1'156 -  -0"7"!                 (82)

The demand function has again shifted to the right, since  the constant
term is larger.
                                 96

-------
The value per visit  to  recreationists at Klamath Lake after Step 2 is seen
more clearly with  the use of Figure 14.   The per-visit value is:
                      16.26
Value per visit =     J     (e1'156 ' -0637Pi) dp[ _ [(14.42)(1.13)]    (83)

                    1.84

                -  $26.72 - $16.30 = $10.42.

When the higher use-intensity ratings are substituted into the  visits
equation, the number of visits to Klamath Lake after Step 2 is  estimated
to be 377,947, a 158 percent increase above  the number of visits  estimated
for Klamath Lake in 1968.  The estimated net economic value of  Klamath
Lake after the completion of Step 2 is:

                        ($10.42)(377,947) = $3,938,208

The total increase in economic value of Klamath Lake after completion  of
Steps 1 and 2 is $3,856,173.  That is, the recreational benefits  available
to society would be worth $3.86 million annually if the two-step  improve-
ment in Klamath Lake were undertaken.  $1.2  million worth of benefits
would be associated with the first step.  An additional $2.66 million  are
related to the second step.

               Increase in Recreational Expenditures

The demand estimates of the number of visits and the length of  stay  per
visit are used to  estimate the net increase  in expenditures associated
with recreation at Klamath Lake as water quality improves.  The net  in-
crease in travel costs is estimated by multiplying the averages in Table
13 by the net increase in the number of visits to the site resulting from
water quality improvements.   That is, upon completion of Step 1,  the esti-
mated number of visits to the lake would increase from 146,491  to 234,947.
Therefore, the net increase is 87,895 visits.  This figure is multiplied
by the averages in Table 13 to obtain the estimates in Column 2 of Table
18.  The same procedure is used to estimate  the net increase in expendi-
tures associated with Step 2 (Column 3, Table 18).  (See page 90  for an
explanation of the deletion of camping fees  in Table 18.)

The net increase in on-site  costs was estimated in the same manner.  The
net increase in visitor-days associated with Step 1 is the difference  be-
tween the number of visitor-days estimated after the completion of Step 1
and the number of  visitor-days estimated for 1968:

Net visitor-days (Step 1) =  (234, 386) (2. 41)  - (146, 491) (1.97)     (84)

                           -  564,870 - 288,587

                           •  276,283 visitor-days.
                                  97

-------
  $16.26
    $1.84
        0
                                 1.156 - .0637
              1.13    2.83
Figure 14.  The Average Recreationist's Demand Curve, Per
            Visit, for Klamath Lake (Step 2).
                               98

-------
Table 18.  Net  Increase  in Expenditures,  for Each Component
           of Travel  Cost, Associated with Improvements of
           Water Quality at Klamath Lake

                                          (2)                 (3)
                                       Net Increase in Expenditures
Component

Cafes 	 	


Other, 	
TOTAL 	

Step 1




	 10,964
	 $154.197

Step 2
g1A7 A61
2S 1 ft^
-uj j XO.J
1A7 706
57 T?l
28,823
$406.004

 Completion  of Step 2 would result in an estimated net increase of 781,003
 visitor-days.  The net increases in visitor-days are multiplied by the
 average  values of each component of on-site cost listed in Table 14,  to
 obtain the  net increase in expenditures for each component of on-site
 cost.  The  estimates for Steps 1 and 2 are listed in Table 19.
 Table  19.   Net Increase in Expenditures for Each Component of
            On-Site Cost Associated with Improvements of Water
            Quality at Klamath Lake
(1)
Component








TOTAL 	
(2)
Net Increase
Step 1
	 $ 58,904
	 82,470
	 92,831
	 5,774
	 14,284
	 289,351
	 4,669
	 7,349
	 $555,632
(3)
in Expenditures
Step 2
$ 166,510
233,129
262,417
16,323
40,378
817,944
13,199
20,775
$1,570,675
                                  99

-------
            Impact of Recreational Expenditures (Step 1)

The increase in expenditures of recreationists represents an exogenous
change in the local economy.  That is, the increase in expenditures is
caused by a postulated change in the recreation experience at Klamath
Lake.  This change does not alter the structure of the local economy.
Therefore, the appropriate way to view the increase in recreational ex-
penditures is as a change in the final demand of the model.

Column 2 of Table 20 is the projected final demand associated with the
removal of the algae from Klamath Lake.  The projected final demand is
post-multiplied by the (I-A)~  matrix to determine the direct and indirect
increase in sales resulting from the completion of Step 1.  The projected
output of each sector is listed in Column 3 of Table 20.

The total increase in output associated with Step 1 is $252,286 greater
than the original value of recreational expenditures.  That is, the esti-
mated $709,829 spent in the economy by recreationists would generate an
additional increase in sales of $252,286 in the economy.

As the output of the local economy increases, household incomes also rise.
The exact magnitude of the increase is estimated by multiplying the change
in output of each sector by the household coefficient (a,q  of the direct
                                                          J
coefficients matrix) of the sector.  The results are listed in the last
column of Table 20.  It is estimated that household income in the county
will increase $194,000 if the algae are removed from Klamath Lake.

            Impact of Recreational Expenditures (Step 2)

The procedures described in the previous section are also used to esti-
mate the regional Impact associated with lowering the water temperature
and improving the beaches at Klamath Lake.  The results are contained in
Table 21.  Projected total output has increased to almost $2 million,
while county household income is estimated to increase more than a half-
million dollars per year if Steps 1 and 2 are completed at Klamath Lake.

The net impact associated with Step 2 can also be estimated.  The change
in total output of the economy would increase $1,716,439.  Household in-
come would increase an estimated $347,820 if Step 2 were undertaken after
the completion of Step 1.

                       A Limitation Restated

While Section X has developed some very specific numerical estimates of
several variables, sight must not be lost of the many limitations under-
lying these estimates.  Many shortcomings, both in theory and empirical
techniques, were discussed in the earlier sections.  We hasten to re-
emphasize that the final estimates which have just been presented are
subject to all of these shortcomings.  Consequently, they are difficult
to interpret.
                                 100

-------
Table 20.  Projected Increases  in  Final Demand, Total Output,
           and County Household Income, by Economic Sector,
           Associated with  Klamath Lake after  the Completion
           of Step 1

1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
(1)
Sector

- j ^ •.

Manufacturing & Processing...






Communications &







(2) (3)
Klamath Lake (Step
Projected
final
demand
/v
(Y)
$ o
0
0
0
114,602
68,467
359,633
0
0
15,633
0
0
0
138,371
13,123
0
0
$709,829
Projected
total
output
(X)
$ 20,259
3,832
1,734
35,041
114,611
68,476
407,225
6,692
1,234
81,795
12,243
24,264
5,636
146,433
13,155
10,489
8,996
$962,115
(4)
1)
Projected
increase
in house-
hold income
$ 6,760
402
495
5,420
42,120
23,443
61,927
1,711
682
14,609
4,538
6,993
1,415
14,570
2,040
1,756
5,757
$194,638
 The major limitation to the reliability of our estimates is probably re-
 lated  to  the manner in which the predictions of changes in recreational
 use of Klamath Lake were made, given the postulated two-step improvement
 in water  quality.   It is recalled that "use-intensities" are employed in
 the prediction model to relate recreational demand to changes in water
                                  101

-------
Table 21.  Projected Increases in Final Demand, Total Output,
           and County Household Income, by Economic Sector,
           Associated with Klamath Lake after the Completion
           of Step 2

1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
(1)
Sector



Manufacturing & Processing...




Product-Oriented 	

Communications &




Local Government 	
TOTAL 	

(2) (3)
Klamath Lake (Step
Projected
final
demand
(Y)
$ o
0
0
0
319,748
191,693
1,005,783
0
0
42,022
0
0
0
380,335
37,098
0
0
$1,976,679
Projected
total
output
/\
(X)
$ 56,008
10,606
4,803
96,877
319,773
191,717
1,138,867
18,631
3,440
226,391
34,164
67,432
15,720
402,596
37,187
29,299
25,044
$2,678,554
(4)
2)
Projected
increase
in house-
hold income
$ 18,690
1,112
1,370
14,987
117,548
65,644
173,222
4,764
1,901
40,433
12,651
19,434
3,946
40,058
5,767
4,905
16,026
$542,458
quality.  A logically superior model can be developed for relating
changes in the physical characteristics of the water resource  to  responses
in human behavior.  However, the empirical specification of  these rela-
tionships could not be undertaken within the scope of this study.   The
need for multidisciplinary research is clearly indicated in  this  area.
Failure to focus on the logical relationships between the characteristics
                                 102

-------
of the physical environment  and  recreational behavior by going directly
to the "use-intensity" variables led  to  an inadequate specification of
the prediction model.  One might interpret these variables as proxies for
the dependent variable itself.   If so, one would expect the use-intensi-
ties to "explain"  a  considerable proportion of the  variation in the de-
pendent variable.  This  could lead to an overstatement of the estimated
effects associated with  a water  quality  improvement.
                                   103

-------
                             SECTION XI

                          ACKNOWLEDGMENTS
The authors wish  to  acknowledge  the  valuable  assistance provided by the
staff of the Pacific Northwest Water Laboratory of  the Environmental
Protection Agency.   Their  input  helped  in  the formulation stages of the
project, as well  as  during the actual analysis.

Appreciation is also extended to Dr.  John  A.  Edwards, Department of Agri-
cultural Economics,  Oregon State University,  for helping with  the formu-
lation of the  theoretical  demand model.

The Klamath County  Chamber of Commerce  also provided help by contacting
the businesses in the  county, and soliciting  their  support  for the study.

The authors also  thank Dr. Russell Youmans, Department of Agricultural
Economics, Oregon State University,  and Arnold Hoffman, Economist, Sys-
tems Analysis  and Economics Branch,  Division  of Planning and Interagency
Programs, for  their review of the manuscript.

The financial  support  of the Environmental Protection Agency,  and the
assistance provided by Dr. Roger Don Shull, Project Officer, is also
acknowledged with sincere  thanks.
                                  105

-------
                             SECTION XII

                             REFERENCES


 1.  Bartsch, Alfred F.   1968.   Director,  Pacific Northwest Water Labora-
     tory, Federal Water  Pollution Control Administration.  Eutropica-
     tion Problems in  Reservoirs.  Seminar conducted by Water Resources
     Research Institute.   Oregon State University,  Corvallis, Oregon,
     October 17.

 2.  Beattie, Bruce Robert.  1970.  "Economic Efficiency and Distributive
     Consequences of Interbasin  Water Transfers:  A Framework for Analy-
     sis."  Unpublished Ph.D.  thesis.  Corvallis, Oregon State University.
     146 numb, leaves.

 3.  Board of Governors of the Federal Reserve  System.  1968.  "Economic
     Expansion in 1968."   Federal Reserve  Bulletin  54:941-952.

 4.  Bromley, Daniel Wood.   1967.  "An Interindustry Analysis of the Im-
     portance of Grazing  on  Federal Lands  to the Economy of Grant County,
     Oregon."  Unpublished M.S.  thesis.  Corvallis, Oregon State Univer-
     sity.  135 numb,  leaves.

 5.  Brown, William G., Ajmer  Singh and Emery N. Castle.  1964.  "An Eco-
     nomic Evaluation  of  the Oregon Salmon and  Steelhead Fishery."  Cor-
     vallis.  47 p.  (Oregon Agricultural  Experiment Station.  Technical
     Bulletin 78.)

 6.  Cameron, Burgess.  1952/53.  "The Production Function in Leontief
     Models."  The Review of Economic Studies 20:62-69.

 7.  Chenery, Hollis B. and  Paul G. Clark.  1959-   Interindustry Economics.
     New York, Wiley.  336 p.

 8.  Clawson, Marion.  1959.   "Methods of  Measuring the Demand for and
     Value of Outdoor  Recreation."  Washington, D.C., Resources for the
     Future, Inc.  36  p.   (Reprint No. 10.)

 9.  Draper, N. R. and H.  Smith.  1966.  Applied Regression Analysis.
     New York, Wiley.  407 p.

10.  Gibbs, Kenneth Charles.  1969.  "The  Estimation of Recreational Bene-
     fits Resulting from  an  Improvement of Water Quality in Upper Klamath
     Lake:  An Application of a  Method for Evaluating the Demand for Out-
     door Recreation."  Unpublished Ph.D.  thesis.   Corvallis, Oregon State
     University.  156 numb,  leaves.
                                 107

-------
11.  Guedry, Leo Joseph Jr.  1970.  "The Role of Selected Population  and
     Site Characteristics in the Demand for Forest Recreation."  Unpub-
     lished Ph.D. thesis.  Corvallis, Oregon State University.  378 numb.
     leaves.

12.  Hogg, Robert V. and Allen T. Craig.  1965.  Introduction to Mathe-
     matical Statistics.  New York, Macmillan.  383 p.

13.  Hotelling, Harold.  1947.  Letter cited in U.S. National Park Ser-
     vice.  1949.  "The Economics of Public Recreation:  An Economic
     Study of the Monetary Evaluation of Recreation in the National Parks."
     Washington, D.C.  36 p.

14.  Input-Output Bibliography, 1960-1963.  1964.  New York, United Na-
     tions.  159 p.  (St. STAT/Ser. m/39.)

15.  Johnston, W. E. and V. S. Pankey.  1968.  "Use of Prediction Models
     for Corps of Engineers Reservoirs in California."  In:  An Economic
     Study of the Demand for Outdoor Recreation:  Conference Proceedings
     of the Cooperative Regional Research Technical Committee, San Fran-
     cisco.  1968.  p. 15-47.  (Report 1.)

16.  Knetsch, Jack L. and Robert K. Davis.  1966.  "Comparisons of Methods
     for Recreation Evaluation."  In:  Water Research, edited by Allen V.
     Kneese and Stephen C. Smith.  Baltimore, Johns Hopkins, p. 125-142.

17.  Leontief, Wassily W.  1936.  "Quantitative Input-Output Relations in
     the Economic System of the United States."  The Review of Economic
     Statistics 18:105-125.

18.  Leven, Charles.  1961.  "Regional Income and Product Accounts:   Con-
     struction and Application." In:  Design of Regional Accounts, edited
     by W. Hochwald.  Baltimore, Johns Hopkins,  p. 148-195.

19.  Pearse, Peter H.  1968.  "A New Approach to the Evaluation of Non-
     Priced Recreational Resources."  Land Economics 44:87-99.
                       *

20.  Reiling, Stephen D.  1971.  "The Estimation of Regional Secondary
     Benefits Resulting from an Improvement in Water Quality of Upper
     Klamath Lake, Oregon:  An Interindustry Approach."  Unpublished  M.S.
     thesis.  Corvallis, Oregon State University.  120 numb, leaves.

21.  Riley, V. and R. J. Allen.  1955.  Interindustry Economic Studies.
     Operations Research Office, bibliographic reference No. 4.  Balti-
     more, Johns Hopkins.  280 p.

22.  Stevens, Joe B.  1966.  "Recreation Benefits from Water Pollution
     Control."  Water Resources Research 2:167-182.

23.  Stigler, George J.  1966.  The Theory of Price.  3rd ed.  New York,
     Macmillan.  355 p.
                                 108

-------
24.  Stoevener, Herbert  H.   1964.   "Water Use Relationships As Affected
     By Water Quality  on the Yaquina Bay."  In:   Western  Resources  Con-
     ference's New  Horizons  for Resources Research;   Issues and Method-
     ology.  Boulder,  University of Colorado, p. 87-99.

25.  Stoevener, Herbert  H.  and E.  N. Castle.   1965.   "Input-Output  Models
     and Benefit-Cost  Analysis in Water Resources Research."  Journal  of
     Farm Economics 47:1572-1579.

26.  Taskier, C.  E.  1961.   Input-Output Bibliography 1955-1960.  New
     York, United Nations.   222 p.  (St. STAT/7.)

27.  U.S. Bureau  of the  Census.  1967.  Census of Agriculture 1964; Sta-
     tistics for  the State and Counties.  Oregon.  Washington, D.C. 367 p.

28.  U.S. Bureau  of the  Census.  1968.  Population Estimates;  Estimates
     of the  Population of Counties;  July 1, 1966.  Washington,  D.C.
     September  27,  1968.  (Ser. p-25, Nos. 401,  404, 407.)

29.  U.S. Department of Agriculture.  Agricultural Research  Service.
     1968.   Food Consumption of Households in the U.S.;  Household  Food
     Consumption  Survey. 1965-1966, Washington,  D. C.  212 p.  (Report  No.
      1.)
                                   109

-------
                           SECTION XIII
                     PUBLICATIONS AND PATENTS
Edwards, J. A., K. C. Gibbs, L. J. Guedry, and H. H. Stoevener.  "The
Demand for Non-Unique Outdoor  Recreational Services:  Methodological
Issues.1'  Pending publication  as  an Agricultural Experiment Station
Bulletin.  Corvallis, Oregon State University.
                                   Ill

-------
                             SECTION XIV

                              GLOSSARY


Consumer's Surplus  -  The difference between what a consumer actually
pays for a commodity  and what he would be willing to pay  rather  than  do
without it.

Critical On-Site  Costs  (PI)  - The level of on-site costs  at which  the

recreationist  is  indifferent between recreating and not recreating, given
the utility  function  of the  individual, his income, travel  costs,  and the
price of other goods.  A change in any of these variables will result in
a change in  the value of p*.

Critical Travel Costs (k ) - The value of travel costs at which  the recre-
ationist is  indifferent between recreating and not recreating, given  his
utility function, his income, on-site costs, and the price  of other goods.
A change in  any of  these variables will result in a change  in the  value
of k*.

Demand Curve - The  locus of  points representing the maximum amount of a
commodity the  consumer  will  purchase at different prices  in a given time
period, other  things  being equal.

Direct Coefficients (a   's)  - The value of goods and services Sector  j

must purchase  from  Sector i  to produce one dollar's worth of output.

Income Multiplier (M, )  - The total change in household income of the

region, resulting from  a one-dollar change in income of households in
the kth sector.

Income-Output  Coefficients  (H,) - The total change in income paid  to  all

households in  the region as  a result of a one-dollar change in the out-
put of the k^ sector.

Marginal Utility  -  The  change in utility or satisfaction  resulting from
a one-unit change in  the level of consumption of a good or  service.

Output Multiplier - The total change in the output of the entire economy
resulting from a  one-dollar  change in the output of one sector in  the
economy.

Recreation Visit  -  The  entry of any person upon a site or area of  land
or water generally  recognized as an element in the recreation population,
except those which  are  part  of, or incidental to, the pursuit of a gain-
ful occupation.


                                  113

-------
Recfreation Visitor-Day ^ Twelve visitor-hours, which may be accumulated
continuously, intermittently, or simultaneously by one or more persons
whd are not in pursuit of a gainful occupation at the recreation site.
                                114

-------
                            SECTION XV






                            APPENDICES




                                                   Page No.




A.   Outdoor Recreation Questionnaire	   117




B.   Input-Output Questionnaire	   123
                                  115

-------
                                         APPENDIX  A

                               FIELD SURVEY QUESTIONNAIRE


                                 OREGON STATE UNIVERSITY
                                          July,  1968
                               .   I'm working on a recreation survey for Oregon State University and
Hello, I'm	
would like to ask you a few interesting questions if you don't mind!

1-  1 Visit  lake  (continue)             Was the main purpose of your trip to visit this particular
     2 Other purpose (DISCONTINUE)    lake,  or are you taking your trip for some other purpose?
2-

(AM or PM)
3-

(AM or PM)
4-



Date May I ask when yo
Time date and the appro

Date Now, when do you
Time again the date and

City/Town
Where do you live
County
county auu slate ?
State r

iu arrived at this particular site- -the
ximate time?
plan to leave this particular site--
approximate time of day or night?

at the present time--the city or town.
4a- 1  In city/town (skip to 5)
    2  Suburban area (skip to 5)
    3  Rural outside (ask 4b)
                                        Do you live right in the city (town), a suburban area, or
                                        a rural area outside of the city (town)?
4b-
    1  Nearer site
    2 Away from site
5-
                            Miles       How many miles do you live out of the city (town)?
                                        (INT:  Mark whether nearer or farther away from site)
                               Number  Including yourself, how many persons are there in your
                                        party which is stopping at this particular place?
6-1 Immediate family
    2 Other relatives
    3 Unrelated individuals
    4 Other (explain below)
                                        Does your party consist mainly of your immediate family,
                                        mainly of other relatives, or mainly of unrelated
                                        individuals, such as neighbors and friends?
7-
                           Number      Including yourself, how many persons are there in your
                                        immediate family?

To help the University figure out how valuable recreation is to the state,  I'd like to ask you about
your party's expenditures from your home to this, area.
g _ $                   Enroute         Approximately how much did your party spend for food and
                                        liquor in cafes, restaurants or taverns while you were en-
                                        route to this particular site?  (just your best estimate)
                                            117

-------
                             Here     About how much will your party probably spend in
                                       restaurants, cafes or taverns while you are stopping at this
                                       particular site? (just your best estimate)
9- $
10- $
                            Home     Approximately how much did your party spend for this trip
                                       in grocery or liquor stores before you left home? (just
                                       your best estimate?)
                            Enroute    About how much did your party spend in grocery or liquor
                                       stores while you were enroute to this particular site?
                                       (just your best estimate?)
                            Here       What do you think your party will spend in grocery and
                                       liquor stores while you are stopping at this particular
                                       place?  (just your best estimate?)

                            Enroute    While you were enroute to this site,  about how much did
                                       your party spend for lodging in motels,  hotels, or trailer
                                       parks?  (just your best estimate)
                            Here       What do you think your party will spend for lodging in
                                       motels, hotels  or trailer parks- while you are stopping at
                                       this site? (just your best estimate)
 11- $
                            Enroute    How about camping fees--how much,  if any, did your
                                       party spend for camping fees while you were enroute to  7
                                       this site?  (just your best estimate)
                            Here       What do you think your party will probably pay for camping
                                       fees while you  are here at this site? (just your best esti-
                                       mate)
12  -
     	Miles      How many miles,  if any, did your party drive yesterday
                                       while at this site?
   — — — — — — — — — ___ — ______„ — _.__ — ____.~_._________._.»____..— _ — — ______.._____—,^ — — — _ — — ____________
12a -	Purpose    What was  the purpose of your drive yesterday?

12b -	Miles      (If not at site yesterday) About how many miles,  if any,
                                       will your party probably drive today while at this site?
12c-
                            Purpose    For what purpose will today's drive be for?
Now, please think of the gasoline and oil that will be purchased for your party's car and boat for this
entire trip.
13 -
                            % before   First,  about what percent of gas and oil for the car and
                            leaving    boat was purchased for the  trip before you left home ?
                                       (just your best estimate)
                            % both     Now,  think of all the gas and oil that will be purchased
                            ways       between home and here and between here and back home.
                                       Approximately what percentage of the gas and oil will
                                       be purchased between home and the time you get back
                                       home from here,  that is, both ways? (Estimate)
                            % at this   What percent of all gas and oil purchased for the car and
                            site        boat will you probably make while you are stopping at this
                                       site?  (just your best estimate)
                                             118

-------
14  -  1  Yes(askl4a)                  Did your party bring a boat with you to this site?
       2  No (skip to no. 15)

14a ~ 	,	 Gallons      About how many gallons of gasoline does your boat use a
                            Gas         day at this particular site ?
      	 Quarts       Hov/ many quarts of oil does your boat use  in a day while
                            Oil         here?
(INTERVIEWER:  Refer to question no. 6.  Ask question 15 series only if code 2, 3,  or 4 is
                 circled in no. 6)

IS  - 1  Mine (ask 15a)                Whose car did you bring on the trip--yours or someone
      2  Someone else (skip to 15b)     else's in your party?
    ————— — — •-—••— — — ——————"•————•"— ——•-••——"-—•-—-•———••—-• — — ————— — — — ————— —————....-. ———-.———— — ..————....
15a - $	       How much, if any, did other members of your party
                                       contribute for gas, oil and automotive expenses thus far
                                       on the trip? .
i
iSb - $	       How much, if any, have you contributed thus far to the
                                       owner of the car for gas, oil and automotive expenses?

16  -  _$	       How much money, if any, has your party spent on boat
                                        launching fees while on this trip?
 17 - Thus far,  we have talked about expenses for the automobile, boat,  food and liquor, and for
     lodging and camping fees.  Can you think of any other types of expenses you have had coming
     here, such as camera supplies, souvenirs,  etc.  (If YES)  What type?

     1  No
     2  Yes    Type _     	,	
               Total cost of these expenses                 $  ,	
 17a - What other types of expenses will you have while, stopping at this site?
      1  No
      2  Yes   Type	—	
               Estimated cost of these expenses             $     •	
                                            119

-------
18 - (HAND CARD TO RESPONDENT)  Here is a list of items which either you or other members
     of your party may own, which you have brought with you to this site. Looking over the list,
     will you please tell me which owned items were  brought with you? Do not inclue rented
     items. (INT: Mark X for each item.  Then ask  remaining questions on your card for each X'd
     ltem)                              Amount Paid
                                           for Item
Items
  Year
Purchased
  Type & Location of
Store Where Piirchased
                                      Maintenance
Boat 	
o , , .
^xU uj^jcLni jii vj CvX
Boat trailer 	
Fishing tackle (rod,


Camper (van, truck,


T 4- +~ "1

T
i CHL










Boat equipment not in-
cluded in price of boat
(preserver, fire extinguisher,
BH.» j
Water skiis, ropes, etc. -
Special clothing (such as
rubber boots, coats, rainwear,
avv j-in.iriiii£ a tllLo j CLC. j
Any other items ?
(If YES) What?








































































•


-


,
f











'.















„


—































120

-------
23-1  Male   <                               1  Under 21 years of age
      2  Female                                2  21-29 years
                                               3  30-39
                                               4  40- 49
                                               3  50 - 59
                                               6  60 or over       Age and sex of respondent

24 -  		•  , ..-„_		  Site Where interview wa's taken

25 -  	;_	a	  Telephone number of respondent.
      Area Code                                        {For verification purposes only)

      X I hereby certify this interview was actually taken with the person described above, and
         represents a true and accurate account of the interview.

                                                                                        . 1968
                        (Interviewer's Signature)                          (Date)
 COMMENTS ON INTERVIEW (if any):
                                              121

-------
 19 - (HAND RENTAL CARD)  Looking at this list of items, will you please tell me which, if any,
     of these items you or other members of your party have rented for this particular trip?  (INT:
     Mark X for each item.  Then, for each X'd item, ask the remaining questions on your card
     for each X'd item)
                                 Rental Rate
                                (Daily, Hourly,
      Item                        Weekly)
                                                  Type G Location of
                                                  Store Where Rented
 reel,
     Boat	
     Outboard motor	
     Boat trailer	
     Fishing tackle (rod,
      tackle box, etc, )	
	Camper (van,  truck,
trailer camper, etc. )	
	Tent trailer	
	Tent	
	Back pack	
	Sleeping bag	
	Water skiis	
	Life vests	
	Other equipment for
 boats-
	Any other items?
 (If YES) What?  	
                                                               Total Rent Expected
                                                                  to Pay for Item
20- $
     0 None
                                     About how much will you spend at this site for various
                                     baits—just that amount that will be used at this particular
                                     site?
21 - (HAND RESPONDENT INCOME CARD)  Would you please look at this card and tell me which
     one of these groups best fits your total family income before taxes for last year?  Just call
     your answer by letter, please.
                                              8  (h) $11,000- $11,999
      1
      2
      3
      4
      5
      6
      7
(a)  Less than $3,500
(b)
(c)
(d)
(e)
(f)
(g)
$3, 500 -
$5, 000 -
$7,000-
$8,000-
$9.000-
$10, 000
$4, 999
$6,999
$7,999
$8, 999
$9,999
- $10,999
                                              9 (i) $12,000- $12,999
                                             _0 (j) $13,000- $14,999
                                              1 (k) $15,000 - $16,999
                                              2 (1) $17,000- $19,999
                                              3 (m) $20,000- $24,999
                                              4 (n) $25,000 or over
                                              (INT:  If $25,000 or over, get range from
                                                    respondent)
22 - INTERVIEWER: Mark below the type of activity the respondent was doing when you first
     approached (him) (her), or the type of activity the respondent just finished doing.

     	Activity
                                         122

-------
                             APPENDIX B
                       OREGON STATE UNIVERSITY
                                  ** **«_*._T_fc«AMtJ JL .j. A.
                                  conducting a  survey for Oregon State
                                  you a few questions about your business
               ™UJ.U  XO.KE  co  ask you a few questions  about your busines
             mind.  Everything you say is  confidential,  and the results
               tor  the area as a whole - not for  any one  person or busi-
if you  	
are tabulated
ness
                          First, may I ask what your total sales of mer-
                          chandise and services were during 1968.  This
                          can be either calendar year or fiscal year,
                          whichever is easier for you.
     0 None
                          What was the approximate amount of  your sales
                          to private individuals  during 1968?  Do not  in-
                          clude businesses or government - just private
                          consumers or individuals.
3 - $
     0 None
    $
              Fed.
             State
     0 None
     0 None
            _City/County
During 1968, did you sell any merchandise or
services to government units outside Klamath
County?  (If YES) What was the total amount of
these sales to government units outside Klamath
County?

During 1968, did you sell any merchandise or
services to government units inside Klamath
County?  (INT: If NO, circle all three O's)

(If YES) What was the amount of your sales to
government units in Klamath County?  What was
the amount of your sales to agencies of the
Oregon State government in Klamath County?  And,
what was the amount of your sales to local city
and county governments?
4 - $	 Outside     What was the total amount of your sales to busi-
     0 None               nesses outside Klamath County in 1968?  Again,
                          either the calendar or fiscal year?
5 - $
     0 None
             Inside       What was the total amount of your sales to busi-
                          nesses inside Klamath County in 1968?
    Now, would you please think, of the sales you made to businesses within
    Klamath County during 1968.  (HAND RESPONDENT CARD)  On this card are
    some types of  businesses.   As I read off each type, will you please
    tell me the  amount  or percentage of your sales, if any, you made to
    that type of business in Klamath County?  (INT: Go through list one
    at a time.   There must be an answer recorded for each type of busi-
    ness.  If answer is "None", write in "0" on appropriate line)
6 -
                                  123

-------
                    (a) Agriculture
                    (b) Agricultural services
                    (c) Lumber
                    (d) Manufacturing
                    (e) Lodging
                    (f) Cafes and Taverns
                    (g) Service stations
                    (h) Construction
                    (i) Professional services
                    (j) Product-Oriented wholesale & retail (unless
                        listed elsewhere)
                    (k) Service-Oriented wholesale & retail
                    (1) Communications & Transportation
                    (m) Financial institutions
                    (n) Grocery wholesale or retail
                    (o) Resorts and Marinas
                    (p) Auto and Trailer Sales
     $_	  (q) Other (Specify	)
7-1. Corporation (Ask 8 & 9)  Is this business a corporation, or some
    2. Other (Skip to #10)      other kind of ownership?
8 - $__	        About how much in compensation was paid to cor-
     0 None              poration officers during 1968?  Please include
                         all compensation including bonuses, profit-shar-
                         ing, and firm contributions to retirement.
8a -$	        How much of this compensation, if  any, was paid
     0 None              to officers outside Klamath County?
9  -$	-           About how much in wages were paid  to employees
     0 None              of the corporation during 1968?  Please include
                         all wages including bonuses, profit-sharing, and
                         firm contributions to retirement.
9a -$	        How much of these wages, if any, were paid to
     0 None              employees outside of Klamath County?
     (INTERVIEWER:  If you asked #8 and #9, skip now to #11.  Questions
     #10 and lOa are to be asked only of businesses which are not cor-
     porations)
                                 124

-------
10  - $
lOa - $
       0 None
          Including yourself, how much was paid  in wages
          to all employees of the firm during  1968?
          •••••••^•»,»,.».«,»„» j^^,.,,,^^^_^—_	._in .___t [LI	

          How much of these wages,  if any, were  paid to
          employees outside of Klamath County?
     ASK OF  EVERYONE

11  - $	
       0 No  purchases
    (If None, skip to
    #18)
          Did this business buy  any new  equipment, machin-
          ery, buildings, or other capital  items  during
          1968?

          (If YES) What was the  total  amount  of  these
          capital item purchases during  1968?
12  -  $
       0 None
          Of the capital  item purchases you made  in  1968,
          how much, if  any, were purchased from individu-
          als?
13  -  $
        0  None
          Were any of  these  1968  capital  items purchased
          from government  units outside Klamath  County?
14  -  $
        0  None
        0 None
Fed.
                State
       ?	City/
       0  None  County
Were any of these 1968 capital items purchased
from government units inside Klamath County?
(INT: If NO, circle all three O's)

(If YES) What was the amount of your 1968 capi-
tal items purchased from  federal government
units in Klamath County?

What was the amount of your capital item pur-
chases from agencies of the State of Oregon
in Klamath County?

What was the amount of your capital item pur-
chases from local city and county govern-
ments in 1968?
15  -  $
        0  None
          Of these  capital  item purchases you made  in
          1968, how much, if  any, were  purchased  from
          firms or  businesses outside of Klamath  County?
16  -  $
        0  None
          Of the  capital  item purchases  you made  in  1968,
          what amount was bought  from firms or businesses
          inside  Klamath  County?
17  - Now, will you please think of the purchases of capital items which
      you made from businesses within Klamath County during 1968.  On
      this  card are the same types of businesses which you read before.
      As I  call off each type, will you please  tell me the amount or
      dollar percentage, if any, which was purchased from that business
                                  125

-------
     group in Klamath County?  (INT:  Go through list one at a time.
     There must be an answer recorded for each business type.  (If None,
     write in "0")
          $
          (a) Agriculture
          (b) Agricultural services
          (c) Lumber
          (d) Manufacturing
          (e) Lodging
          (f) Cafes & Taverns
          (g) Service stations
          (h) Construction
          (i) Professional services
          (j) Product-Oriented wholesale & retail
          (k) Service-Oriented wholesale & retail
          (1) Communications & Transportation
          (m) Financial institutions
          (n) Grocery wholesale & retail
          (o) Resorts & Marinas
          (p) Auto and Trailer Sales
          (q) Other (Specify	)
     ASK OF EVERYONE
18  - $
Fed.
               State
               City/
               County
(HAND RESPONDENT LIST OF POSSIBLE TAXES)
What is the approximate amount of taxes which
your firm paid to the federal government  in
1968?
How much in taxes did your  firm pay  the State
of Oregon in 1968?
What was the approximate amount of taxes  which
your firm paid to this county or to  cities
within the county in 1968?
19  - $
       0 No or None
          Did your firm pay any taxes to states outside
          of Oregon in 1968?  (If YES) About how much?
19a - $
       0 No or None
          Did ypur firm pay any taxes to city and  county
          governments outside Klamath County in  1968?
          (If YES) About how much?
                                 126

-------
20 - 1. Yes  (Continue with #21
     2. No  (Skip  to #28)
                            Did your firm receive any interest,
                            rent, royalties, or dividends during
                            1968?
21 - $
                    What was your firm's total receipts from inter-
                    est, rent, royalties, or dividends during 1968?
22 - $
      0 None
                    How much of these receipts, if any, were paid
                    to you by private individuals?
23 -  $	            During 1968, did you receive any interest, rent,
                          royalties, or dividends from government units
                          outside Klamath County?  (If YES) What was the
                          total amount?
24  -  $
         Fed.
       0 None
               State
       0 None
      $
    	 City/
 0 None  County
During 1968, did you receive any interest, rent,
royalties, or dividends from government units
inside Klamath County?  (If NO, circle all three
O's)

(If YES) How much was received from federal
government?  From the Oregon State government?
From this county or cities within the county?
 25  - $
       0 None
                    During  1968,  did you  receive any interest, rent,
                    royalties,  or dividends  from businesses outside
                    Klamath County?   (If  YES) What was  the total
                    amount?
 26 - $
       0 None
                    During  1968,  did you  receive any interest, rent,
                    royalties,  or dividends  from businesses inside
                    Klamath County?

                    (If YES) What was  the total amount?
                    (If NO, skip  to #28)
 27 - Again, here is a list of types of businesses in Klamath County.  As
      I read off each one, will you please  tell me the amount or percent-
           if any, which came from interest,  rent, royalties, or dividends
                                              -------     (INT:
age.
from any of these  types  of  businesses within  Klamath  County,
There must be an answer  recorded on each  line)
      $_
      $
                     (a) Agriculture
                     (b) Agricultural services

                     (c) Lumber
                     (d) Manufacturing

                     (e) Lodging'
                                   127

-------
       $		•._	            (f) Cafes & Taverns
       $ 	            (g) Service stations

       $	            (h) Construction
       $	              (i) Professional  services
       $	            (j) Product-Oriented  wholesale & retail

       $	            (k) Service-Oriented  wholesale & retail
       $                   (1) Communications  &  Transportation

       $	            (m) Financial  institutions
       $	            (n) Grocery wholesale & retail

       $                   (o) Resorts &  Marinas
       $	            (p) Auto and Trailer  Sales

       $	            (q) Other (Specify	)

     ASK OF EVERYONE—
  28  - $	           What was the total  amount of depreciation taken
                           by your firm in 1968?

  29- 1.  Higher (Ask 29a)Was your  inventory higher or lower at
        2.  Lower (Ask 29a)         the end of  1968 than it was at the
        3.  Same (Skip to #30)      beginning of  the year?
        4.  D.K. or no inventory

  29a - $_^	           About how much (higher) (lower) was your inven-
                           tory at the end of  1968?

  30 - X  I hereby certify this interview was  actually taken with the per-
          son listed below, and represents a true and accurate account of
          the interview.
            (Respondent)                  (Firm)                (Date)
           (Phone Number)             (Interviewer's signature)

       FOR OFFICE USE ONLY**

                           Interview  verified by	

                           Date  of  verification
«U.S. GOVERNMENT PRINTING OFFICE:1973 514-152/1771-3        128

-------
    Accession Number
  W
                            ;c( Field & Group


                            06 B
                                               SELECTED WATER RESOURCES  ABSTRACTS
                                                      INPUT TRANSACTION FORM
         Department of Agricultural Economics, Oregon State University
         Economic Benefits  From an Improvement in Water Quality
 J Q 1 Authors)

          Reiling, S. D.
          Gibbs, K. C.
          Stoevener, H. H.
                                 21
                                      EPA,  OEM,  Project 16110 FPZ
                                    Note
 22
     Citation
            Environmental Protection Agency report
            number, EPA-R5-73-008, January 1973.
 23
Descriptors (Starred First)                         ~~~                    ~~—

  Benefits*,  recreation*, water quality*, economics*,  lakes,  camping,
  sport  fishing
 25
Identifiers (Starred First)

  Recreation demand, travel costs, on-site  costs
 27
Abstract
This  report  introduces and empirically tests a new methodology  for estimating  the
economic benefits accruing to society from  an improved  recreational  facility.  The
specific facility under consideration is Upper Klamath  Lake,  Oregon, which presently
has low water quality.  The methodology draws upon previous work  done  in  the evalua-
tion  of recreational demand; however, it focuses upon the  individual recreationist
and separates the traditional price variable into on-site  costs and  travel costs.
The model  is used to estimate the number of days per visit the  recreationist will
stay  at the  site as the water quality improves.

Data  collected at three other lakes with varied characteristics are  used  to derive
a relationship between the number of visits to a site and  the characteristics  of
the site.  This relationship is then used to estimate the  increase in  visits to
Klamath Lake that would be forthcoming with an improvement in water  quality.

The impact of expanded recreational use of  Klamath Lake upon  the  local economy is
also  estimated through the use of an input-output model of the  Klamath County
economy.             .
Abstractor
                              Institution
WR;102 (REV JULY 1969)
WRSIC
                        SEND. W,TH COPV OF DOCUMENT
                                                   TO: WAT.* RESOURCES

                                                      WASHINGTON. D. C. 20240
                                                                               * GPO! 1970-389-930

-------