A Report on the
SOCIOECONOMIC IMPACTS of the
PROMULGATED TRANSPORTATION
CONTROL PLAN for SAN DIEGO
Prepared under
BASIC ORDERING AGREEMENT 68-01-1551
TASK ORDER NO. 1
Prepared for
ENVIRONMENTAL PROTECTION AGENCY
FEBRUARY 20, 1974
Prepared by
(PMM&) PEAT, MAKWICK, MITCHELL & CO.
""
-------
PEAT, MARWICK, MITCHELL & Co.
15446
1O25 CONNECTICUT AVE., N. Vf.
WASHINGTON. D. C. 2OO3G
February 20, 1974
Mr. Robert Borlick
Office of Planning and Evaluation
Environmental Protection Agency
1023 Waterside Mall
Washington, B.C. 20460
Mr. David Calkins
Region IX
Environmental Protection Agency
100 California Street
San Francisco, California 94111
Gentlemen:
This is a draft report of Phase I of the study of the socioeconomic impact
of the Air Quality Control Plan for San Diego promulgated by the Environ-
mental Protection Agency on November 12, 1974.
Authority for completing this work was received under Task Order #1 of
Contract No. 68-01-1551 issued to Peat, Marwick, Mitchell & Co. (PMM&Co.)
and subcontractor, the San Diego Comprehensive Planning Organization (CPO)
on September 21, 1973.
Primary objectives of this phase of the study are to assess the economic
and social impacts which might result from implementing the EPA promulgated
plan, and to assess the impact of selected alternatives to certain measures
found in Phase II of the study. Phase I has been similar to a pilot study
in the sense that it has developed methodologies for evaluating costs,
estimating measures, and assessing allocation techniques and socioeconomic
Devaluation data. The project team feels that these data will be useful
to EPA and other urban areas as a basis for the development of policy
analysis and management tools.
Initially, several cities were under consideration for pilot socioeconomic
studies. The major reasons San Diego was selected include:
its detailed transportation planning networks and working data
base that are suitable to support a modal split analysis that
could be accomplished in a minimum amount of time;
-------
R M. M.&CO.
Mr. Robert Borlick
Mr. David Calkins
February 20, 1974
2
its location in an air quality control region where substantial
adjustment to travel habits or equipment is indicated on the
basis of present standards, monitoring results, and projected
growth patterns; and
its substantial sources of data, i.e., the air pollution control
studies previously completed in San Diego by EPA and the local
authorities, and the capability to provide assistance as a result
of prior modeling efforts.
The study has proceeded generally as planned, except that major events
have occurred since the study was authorized which have already influenced
the future of the program. On the one hand, a number of study findings
(e.g., the estimated impact of the full surcharge program) will probably
be of less concern to the San Diego Central Business District as a result
of the new EPA position removing the mandatory requirement for that measure
from the proposed plans. On the other hand, the continuing discussions in
Congress about increased relaxation of new automobile manufacturing require-
ments and energy issues should heighten the interest and increase the value
of findings relevant to these items.
The transportation control measures were analyzed in considerable detail,
and an evaluation of the impact of the EPA promulgated plan was completed
utilizing projections of the San Diego regional transportation networks and
modifications to the appropriate models for projecting travel changes
resulting from the surcharge program and car pooling. Trips are projected
by type, time, and geographic locations relative to the income of the
travelers in the San Diego region.
Cost projections for all of the program measures were prepared for startup
through 1966 and for continuing annual operating costs beginning 1977
(assumed to be the first year of the implemented program). San Diego costs
were used where actual localized values could be found. Cost related
information was requested of EPA, all state and local organizations, and
consultants known and obtainable. PMM&Co.'s research revealed that a
similar program management oriented cost estimate of the overall program had
H£t__previously been attempted. Substantial portions of the cost data used
in our estimate were furnished by San Diego County and regional staffs.
While every program cost has not been estimated and could not be developed
within the scope of this effort, the project staff of PMM&Co. believes
that the economic elements of major substance have been estimated, and
that the elements involving as much as 80 to 85 percent of program costs
of any socioeconomic significance have been identified.
-------
P. M. M.a CO.
Mr. Robert Borlick
Mr. David Calkins
February 20, 1974
3
An initial attempt at developing a revenue flow, representing only the first
wave of the economic ripple effect, reveals an interesting group impact in
terms of net money flow as a result of the EPA promulgated plan. This
effect is identified to federal, state and local government groups and
also to the industry and consumer/citizen group.
Phase I of the study reveals clearly the past difficulties in reaching an
agreement or even a general consensus on the fundamental issues relative
to health effects, rollback calculations, and other technical difficulties
not directly within the scope of this effort. There are, however, sig-
nificant indications pointing both to a consensus on the degree of technical
disagreement and to sufficient agreement on many other issues. In San Diego,
there is also substantial indication that an institutional transition
phase is being reached by which program evaluation activities are beginning
to evolve into an organized program planning and implementation phase.
The project staff believes that future tasks should be designed to rein-
force the further development of the San Diego program planning and
implementation capabilities. Section V outlinej3_briefly the major
challenges to the program management stafl
Information from many sources has been utilized in the preparation of this
report. Among the sources employed were reports published by the Environ-
mental Protection Agency and other governmental agencies, consulting firms
employed by EPA, the County of San Diego, and others. Rand Corporation data
was used in some cases as baseline information. Although the project staff
believes these sources to be reliable, we have not verified their data and
offer no opinion on the assumptions contained in their work. To the extent
that the information provided by such sources is not accurate, or the
assumptions they utilized are not realized, the projections may differ
from future events. In addition, changes in economic conditions, shifts in x
population, natural catastrophies and many other uncertainties could
adversely affect the projections. Therefore, we cannot represent the
projections herein as future results that actually will be achieved.
The terms of the contract are such that we have an obligation to submit a
final report, but we have no obligation or intention to update or to revise
that report, unless we are subsequently engaged to do so, because of events
occurring after the date of this draft report.
Substantial assistance was provided to the PMM&Co. study team by members of
the San Diego Task Force and by the EPA Project Directors, Mr. D. Calkins
of Region IX, and Mr. Robert Borlick of the Plans and Programs Office,
Washington, D.C. - -
Very truly yours,
PEAT, MARWICK, MITCHELL & CO.
fcJLJ'
John A. VfaiVder, Principal
-------
A Report on the
SOCIOECONOMIC IMPACTS of the
PROMULGATED TRANSPORTATION
CONTROL PLAN for SAN DIEGO
Prepared under
BASIC ORDERING AGREEMENT 68-01-1551
TASK ORDER NO. 1
Prepared for
ENVIRONMENTAL PROTECTION AGENCY
FEBRUARY 20, 1974
Prepared by
PEAT, MAKWICK, MITCHELL & CO.
-------
TABLE OF CONTENTS
SECTION PAGE
I OBJECTIVES OF THE SAN DIEGO STUDY 1
II SUMMARY DISCUSSION OF FINDINGS AND CONCLUSIONS 8
III PRIMARY TASKS 14
Impacts of the EPA Promulgated Transportation
Control Plan Upon Travel Behavior 14
Development of Program Costs 63
Allocation Rationale 86
Revenue Analysis 98
IV IMPACTS OF STRATEGIES 106
V PROGRAM IMPLEMENTATION 109
Bibliography HI
Appendix A A-l
Appendix B B-l
Appendix C C-l
-------
I. OBJECTIVES OF THE SAN DIEGO STUDY
BACKGROUND
On March 20, 1973, by publication in the Federal Register, the
Administrator of the Environmental Protection Agency (EPA) acting in
response to a court order, requested transportation control plans for
certain air quality control regions (AQCR). By April 15, 1973,
California had not submitted a San Diego control strategy. As a con-
sequence, EPA proposed substitute regulations for San Diego on July 16,
1973. Considerable public discussion followed and the state submitted
Revision 3 of the California State Implementation Plan (SIP) on July 25,
1973.
EPA held public hearings on its plans in each affected AQCR. Many
witnesses in the San Diego AQCR expressed the belief that economic in-
centives and disincentives might reduce the vehicle miles traveled
(VMT) more effectively than other measures, and that, as a bonus,
revenues could be produced for mass transit. Among possible strategies
presented were to:
. increase surcharges on gasoline sales;
. establish surcharges on excessive automobile ownership
involving related features such as weight, displacement; and
. place surcharges on facilities that house or service auto-
mobiles.
It was emphasized that any scheme incorporating surcharges for parking
must extend beyond the urban central business district (CBD) in order
to keep core areas competitive with outlying employment and shopping
centers.
Significant controversy has existed in San Diego over two particular
requirements under the Clean Air Act:
. Is the National Ambient Air Quality Standards (NAAQS) for
photochemical oxidents (0.08 parts per million for an hour)
the appropriate standard; and
. Are socially disruptive measures, e.g., extensive gasoline
rationing necessary by 1977 or might they be delayed?
Both of these points of controversy were discussed during the hearings.
-------
While the November 12, 1973, plan promulgated for San Diego by EPA was
different from its earlier plan and plans generated by the local task
force and the California Air Resources Board, it was hoped that many of the
measures would be acceptable in the AQCR and that alternate plans
developed by state agencies and the local task force could be approved
by EPA.
STUDY OBJECTIVES
A complete analysis of the social and economic effects of the Trans-
portation Control Plans (TCP's) had not been made due to time and
money constraints and the complexity of the problem. On September 21,
1973, the Environmental Protection Agency contracted with PMM&Co. to
conduct an initial socioeconomic impact study of the EPA-promulgated
regulations for the San Diego Air Quality Control Region. In partic-
ular, this study was to assess the several transportation control
measures initially promulgated and then to project the socioeconomic
effects of such controls. In addition to providing information specific
to the San Diego plan, it was hoped that a systematic methodology for
assessing the socioeconomic impacts in other metropolitan areas of the
country could be developed and that the experiences of the San Diego
effort might be helpful to other urban AQCR efforts.
Overview of the Program Structure
Figure 1 identifies the major components of the program, indicates the
scope of institutional alignments which necessarily have been included
in the development of the study structure, and illustrates the basis
for the cost allocation rationale. The major components of the socio-
economic program .structure include:
program basis;
, program strategies and measures;
direct program cost groups; and
cost allocations.
Program Basis
The initial component of the total study overview shown in Figure 1 is
the program basis. In the context of this study, the program basis in-
cludes those technical and environmental health issues which are the
basic underpinning of the entire air quality control program. Some
of the major issues indicated are:
-------
PROGRAM BASIS
• HEALTH
• MEASUREMENT
• CALCULATIONS
•TECHNICAL
PROJECTIONS'
STRATEGIES
STATIONARY
MOBILE
VMT
DIRECT PROGRAM COSTS
1
LOCAL
GOVERNMENT
I
STATE
GOVERNMENT
I
FEDERAL
GOVERNMENT
I
I
INDUSTRY
1
CITIZEN
I
• TAX RATES
ALLOCATIONS
• INCOME LEVEL
AUTO OWNERSHIP
T
T
$ BY HOUSEHOLD
BY INCOME
% OF INCOME
o/
7o
0 10 20 30 40
0 10 20 30 40
FIGURE 1: AIR POLLUTION PROGRAM COST DISTRIBUTION METHODOLOGIES
-------
. the amount, types, concentrations, and time of exposure to
those air pollutants which are considered detrimental to human
health and other related objectives based on the values to our
society which have supported the development of legislation,
guidance, and standards;
capabilities of measuring pollutant sources, concentrations,
and dangers to the required degree of accuracy over time con-
sistent with the prescribed standards; and
the bases for projecting short and longer range pollution
sources and quantities generated, as well as the projections
of the cost and effectiveness of the technical and administra-
tive measures proposed for corrective actions.
Program Strategies and Measures
As indicated on Figure 1, there are specific action areas identified
as those most likely and least costly in socially disruptive areas
which, if implemented effectively, would contribute to the solution
of reducing specific pollutants.
Past studies have been necessarily concerned with the medical, engi-
neering and legal facets of the program. As progress is made and
additional facts are developed, increasing attention is now being
focused on the socio-economic and institutional impacts of the pro-
gram, and the practical feasibility and procedures involved in com-
pleting implementation and monitoring the effectiveness of the
programs.
These measures are categorized loosely into subgroupings identified
as:
. stationary sources—which include issues not directly related
to travel-caused pollution, primarily industrial emissions;
. mobile sources—which generally refer to those measures and
devices associated with mobile equipment;
. VMT reduction—which are measures generally related directly
to reducing travel, or to measures which enhance efficiency of
travel, or are now perceived to be measures which might be
compatible with the overall objectives of the air pollution
control program;
-------
. other measures—which include those not categorized above,
including areas which might not be subject to local regulation,
such as aircraft operations, both ground and local air, rail
and ship movements, and local mass transit operations.
Direct Program Cost Groups
Important components defined during the study are the cost groups. /
These groups reflect general institutional and societal arrangements
at a level where manageable and traceable socioeconomic discussions
of the relationship between the air quality control program measures
and the various segments of institutional groupings can be held. It
was determined that cost data relevant to the directly identifiable
program costs could best be collected under five broad "cost group-
ings." These groupings are:
. local government—for the San Diego study, this title includes
the City of San Diego, San Diego County within the AQCR and
component organizations;
. state government--includes California and all state organiza-
tions, procedures, and fiscal policies utilized in this study;
federal government--includes the impact of funding and taxing
programs plus granting programs and federal policies;
industry—includes all AQCR industrial operations and is in-
tended to incorporate other businesses directly affected by
costs or revenues even though located outside the AQCR or the
State of California; and
. citizen—includes all individuals directly affected by the
air quality control program flow of funds or by eventual
implementation of policies and regulations.
Cost Allocation Alternatives
No previous socioeconomic studies of these impacts on San Diego could
be found during literature reviews or subsequently through discussions
with EPA personnel or members of the San Diego Task Force. It was con-
cluded that a cost allocation methodology would need to be developed
specifically for San Diego, but that the allocation rationale and
techniques would be adopted to reflect existing federal, state, and
local fiscal policies. In addition, industrially related costs would
be applied on the basis of present general business practice for those
segments affected. It is believed that this methodology and alloca-
tion decision will provide maximum benefits for standard application
of socioeconomic evaluations to other programs and will provide the
best basis for subsequent program monitoring and evaluation efforts.
-------
In consonance with this decision, costs for implementing the measures
were considered to be one-time and nonrecurring. Operating costs
during full 1977 application of the program were developed for each
measure. The estimated recurring costs are shown separately for each
of the five groups:
Allocation Procedures
Several study tasks were directed toward the development of method-
ologies and the conduct of the cost allocations of this component of
the study objectives. Program costs were allocated from the defined
groups to determine impacts on the citizens on the basis of:
. household incomes;
. consumption patterns and travel habits by income groups; and
. present financial policies and practices of all levels of
government and industry.
On a group basis, a revenue analysis, at the same level as the program
cost analysis, was completed to indicate the net revenue flow among
groups. This analysis also provided some additional insight into the
proportions of program funding expected to be assigned to various
governmental components. This subject is discussed in Section IV.
A nonquantified procedure for analyses and sensitivity evaluations was
developed to estimate the significance of strategies on selected user
categories. As an example, an analysis by the study team resulted in
a judgment that for San Diego eleven of fifteen proposed measures
would result in the reduction of gasoline consumption. This proce-
dure applied to various groups or mixtures of groups could serve to
measure reaction sensitivities to a series of issues. The approach
is illustrated later in the report.
The recent actions by the oil-producing and exporting countries and
the resulting sudden public interest in petroleum and the general
energy issue have generated immediate and emotional responses by both
government and industry. The Revisions to Surcharge and Management of
Parking Supply Regulations, released by EPA on January 15, 1974, elim-
inates the mandatory requirement for parking surcharges and modifies
parking supply management. It also recommends that the "employer
incentive" regulations are to be more generally worded. Anticipated
energy legislation is also concerned about these measures.
-------
This report is based on the impact of the EPA plan of November 12,
1973, for San Diego. The strategies utilized are included in our
report as is the primary cost information obtained from the Environ-
mental Protection Agency, and the cooperating state and local organi-
zations .
It is expected that Phase II will emphasize the relative socio-
economic sensitivities, i.e., changes in strategies and related items,
such as increases in petroleum prices and extension of certain program
milestones.
A substantial amount of data utilized in this study had been developed
by the Office of Air Programs and others. The designation of
Mr. Robert L, Borlick as Project Director for the Office of Policy
and Planning and Mr. David Calkins for EPA's Region IX ensured utiliza-
tion of prior studies in Washington and improved access to Region IX
and San Diego data sources.
The San Diego County Comprehensive Planning Organization supported one
study of transportation measures as a subcontractor to PMM&Co. and the
San Diego County Air Quality Control District Office and the County of
San Diego Environmental Development Agency Office of Environmental
Management provided substantial data to the study team.
Additional assistance was provided by members of the County of San
Diego Air Pollution Task Force including the City of San Diego, the
San Diego Transit Corporation, the U. S. Navy, the State of California
Air Resources Board, and the State of California Department of Trans-
portation.
-------
II. SUMMARY DISCUSSION OF FINDINGS AND CONCLUSIONS
This section includes a brief discussion of both the general
and the specific findings that are covered in the following sections
of this report and detailed in the appendices. We have included
issues and needs pertinent to recent events and future implementation.
The findings and issues are summarized in three parts:
0 general conclusions relevant to applicability of methodologies
used in the study to other AQCRs;
0 specific findings relative to the projected impacts of the
Environment Protection Agency Promulgated Transportation
Control Plan; and
e future issues.
GENERAL CONCLUSIONS RELATIVE TO APPLICABILITY OF
METHODOLOGIES USED IN THIS STUDY TO OTHER AQCRs.
At the outset of the study, one of the primary interests was
the potential development of methodologies and structures which might
be utilized as a semi-standardized procedure in other Air Quality
Control Regions (ACQRs). In view of that interest, considerable
attention was given to the documentation of certain methods and
structures applied during the work.
For example, an extensive bibliography of documents collected
during this study has been reviewed and an attempt has been made
to incorporate the advice and counsel obtained during our discussions
with organizations extending from EPA through all levels of state
and local government agencies to technical consultants and selected
industry representatives.
It was found that substantial cost information had been
developed during prior studies and that with considerable restructuring,
portions of those estimates could be utilized as a basis for this
report. However, the general application of national average
information was not acceptable, and, as a result, locally developed
values were selected for a number of the cost and socioeconomic
elements. Price estimates of program costs failed to segregate
start-up and operating costs and tended toward theoretical fiscal
and economic assumptions. Consequently, a consistent program
framework and ground rules on cost allocations and projections were
required. A program structure (Figure 1) and defined ground rules
for cost groups were developed.
The program components, cost and revenue groups, and basic
allocation techniques applied in this study can be expected to be
-------
uniformly applicable to all AQGRs, and could be readily applied by
local and regional task forces developing AQCR program plans.
However, actual values and ratios would be expected to vary.
Similar standardized formats will need to be generally described and
accepted if state, regional, or federal program evaluation and
monitoring procedures are to be meaningful.
A decision was made to allocate costs in accordance with existing
financial policies affecting the citizens of San Diego. Application
of existing policies will be required if meaningful programs are
to be developed for maintaining a national and regional evaluation
and regulatory policy compatible with actual state and local program
planning and management priorities.
At the start of the study, particular interest existed in the
potential usefulness of existing models that associated control
measures with existing transportation data. While an appraisal of
existing and new models was not a part of the study, several current
efforts to evaluate or develop air pollution monitoring models and
related tools were reviewed during the work. Specific discussions
were held with representatives of the National Science Foundation,
the Lawrence Livermore Laboratory of the University of California, and
San Diego County. The MOVEC model developed by the Rand Corporation
was utilized to provide specific data items in support of this
study effort.
Discussions of transportation models are detailed in later
portions of this report. From the results of this study, it
appears that a potential for application of such transportation
models exists in those specific AQCRs where substantial transportation
planning networks have already been developed. It might eventually
be desirable to categorize specific AQCRs possessing such transpor-
tation information into groups having similar travel characteristics.
A selective study program may produce sufficient information for
cities so that extrapolations of certain travel behavior might be
accomplished for critical issues with some additional degree of
confidence. Potential cities where such studies might be conducted
are Boston, Denver, Baltimore, and Atlanta.
One issue that should be considered prior to determining the
need for these additional studies is the existing degree of accuracy
and relative impact of other program components (e.g., costs and
mechanisms of major car-pooling programs, potentials of massive
public supported urban transit systems, or new socioeconomic incentives
or penalties such as the four-day work week). _
-------
SPECIFIC FINDINGS RELATIVE TO THE PROJECTED IMPACTS OF
THE TRANSPORTATION CONTROL PLAN
A summary of the most significant transportation conclusions
discussed in Section III follows.
0 Application of the surcharge would result in a reduction of
vehicle miles traveled (VMT) of approximately 12 percent.
EPA calculations included in the promulgated plan for San
Diego indicate that a greater reduction, of approximately
30 to 33 percent, achieved through other measures will be
required to meet the standards.
0 The surcharge program would result in a greater travel
mobility penalty being imposed on people living within the
San Diego central zone than on people living in more
outlying zones.
0 The factors contributing to the reduction in VMT in order
of decreasing importance are:
0 increase in car pooling,
0 foregone trips (excluding work trips which would not be
reduced), and
0 diversion from automobile to mass transit.
0 The average travel time for individual trips in each of the
six San Diego major statistical areas (MSA) are as follows:
MSA Transit System Automobile Travel
(Minutes) (Minutes)
0
1
2
3
4
5
6
Base
29
39
28
45
19
0
31
Plan
31
37
29
41
21
0
32
Base
13
14
14
16
17
32
14
Plan
14
16
16
17
18
34
16
10
-------
The total travel time for people traveling within the San
Diego region increases marginally, 3 percent, even though
person-trips and person-miles of travel decreases.
There is a direct recurring cost of $351 million expressed in current
dollars by the parking surcharge andmass transit _ingfntive
measures. If one half this amount were dedicated to improving
local mass transit, a substantial expense/revenue system
imbalance would probably result.
Numerous additional findings and conclusions are contained in the
transportation discussions and in the appendices to the report.
Selected findings discussed in Section III relative to the develop-
ment and allocation of costs are:
0 Considering all measures and groups selected, a total of
132 cost elements were identified where some cost impact
might be incurred. Of these, 116 cost items were developed
with the assistance of the groups identified in the study. On
the basis of information furnished, it is expected that these
items represent the substantial majority of the program
costs. However, data for sixteen items was not available
or was not sufficiently supportable to be included.
0 The economic impact on individual citizens would be severe
if the surcharge program were to be implemented as promulgated.
The total program costs (in current dollars) to citizens
by household income accumulated through 1977 would be as
follows:
Implementation
Household Income Costs Recurring Costs Total Costs
$0 - 5,000 $22 $599 $621
$5,000 - 10,000 29 803 832
$10,000 - 15,000 40 1,012 1,052
$15,000 and over 71 1,276 1,347
0 Stationary measures represent the most cost effective strategy.
These could be implemented with virtually no net cost to
industry or the consumer. In addition,_an annual savings
of petroleum will be achieve^ through conservation equipment.
0 The mobile source program would cost approximately $54 million
for equipment, installation, and construction and would have approximately
$31 million in recurring costs. There is a manpower question
11
-------
regarding the feasibility of training sufficient inspection
and particularly maintenance personnel by 1977 due to the
high degree of proficiency required to service the complete
range of retrofitted and new vehicles equipped with increasingly
complex devices. Examination of longer range (beyond 1977)
considerations of retrofitted vehicle life reveals that this
measure appears to represent one of the least cost effective
approaches. In addition, an annual increase in petroleum
consumption of approximately $6 million per year is anticipated.
0 Without the surcharge, an increase of the mass transit system
to 550 buses would result in a recurring deficit of approximately
$23 million.
Specific findings discussed in Section IV relate to the social
impact of strategies on unique interest groups.
0 No effective procedures have been found for reviewing with
the citizens Complex socioeconomic issues of the air quality
control program.
0 Little in-depth socioeconomic evaluation of a four-day work
week has been referenced in prior EPA or other studies of
alternate feasible strategies or measures.
0 Substantial additional work will be required to refine and
develop the appropriate content and procedures for
evaluating and accurately summarizing the majority opinions
and optimum socioeconomic measures preferred by other AQCRs
where such a tool might productively be employed. A similar
structure would also assist in communications and presentations
for federal, state, and local legislative bodies.
0 The timing of issuance of evaluation procedures, to be of
benefit to other AQCRs, should be expedited. The procedures
and formats should be prepared in a clear, concise, untechnical
manner to be easily comprehended and followed. Standardization
will aid evaluators and program management to compare
attitudes and reactions in different AQCRs and to facilitate
implementation.
0 Program implementation, cost, and socioeconomic data are
developmental in nature. However, the study team believes
that this study has established that uniform program evalua-
tion and planning criteria and procedures can be accomplished.
The selection of next-step AQCRs should involve more than
our example and should include consideration of several
characteristics such as:
12
-------
0 the variety and availability of transit options,
0 the size of the metropolitan area,
0 different government structured interfaces, and
0 different energy supply situations.
It is recommended that EPA offer to provide financial and
technical assistance to develop implementation programs in AQCRs in
several states.
FUTURE ISSUES
Energy conservation and air pollution control can be viewed as
compatible, not competing programs. Seventy-three percent of the
measures considered in the San Diego air pollution control program
were estimated to promote gasoline savings. An assessment of
Energy impacts relate3~To~envirbnmental consequences could include
considerations such as:
0 Positive impacts including increased transit usage, increased
car pooling, and decreased numbers and lengths of trips;
0 Indeterminate impacts including increased usage of smaller ^
vehicles and revised scheduling and timing of trips; and
0 Negative impacts including increased idling of vehicles
waiting in line for gasoline, increased driving in search
of gasoline, and increased gasoline consumption by retrofitted
vehicles.
Thus, new and innovative joint energy conservation—air pollution
control programs should be designed and implemented.
13
-------
III. PRIMARY TASKS
IMPACTS OF THE EPA PROMULGATED TRANSPORTATION CONTROL PLAN UPON
TRAVEL BEHAVIOR
Introduction
An analysis of the impacts of the EPA promulgated transportation
control plan upon travel behavior and the transportation system in the
San Diego region is presented in this part of Section III. In this
analysis, "impact" is defined as the difference between the transporta-
tion system and travel patterns which would have existed in the San Diego
metropolitan area on July 1, 1977, had the transportation control plan
not been promulgated (i.e., the base system); and the transportation
system and travel patterns which would exist in the San Diego region at
this same date were the transportation control plan promulgated by the
EPA on November 12, 1973, to be implemented (i.e., the EPA plan). The
analysis is structured to analyze differential impacts upon various
strata of residents of San Diego — as defined in terms of household in-
come and geographic location within the region.
Distinguishing between the base and the EPA plan, which are both
conceptually straight forward, was particularly difficult with respect
to locally implemented measures proposed to meet EPA's air quality
standards.
Various highway, public transit, and bike system improvements were
planned for the San Diego region prior to promulgation of the transporta-
tion control plan. In several cases, it was difficult to assess whether
transporation facilities and services were planned due to the EPA program,
or whether implementation schedules of previously planned projects were
accelerated to satisfy EPA's air quality standards. To the greatest
possible extent, the reasonableness of the treatment of base and EPA plan
definitions was checked through conversations with appropriate local
officials and EPA representatives. These assumptions are documented in
subsequent sections of this chapter.
The principal criteria used to measure the travel behavior impacts
of the promulgated plan are the passenger miles of travel and number of
daily trips made by automobile drivers, automobile passengers, and transit
passengers. The principal criteria used to assess the changes in the
service provided by the transportation system are travel time (accessibility)
and travel costs for the study area travelers.
Description of the Base System
The regional transportation planning data base which had previously
been assembled for the San Diego region by CPO was utilized as a basis for
analysis of the transportation measures. In conducting its long-range
planning activities, CPO had developed an extensive data base for the 1975
time frame which has been used to evaluate the implications of a number
of different land-use transportation development programs for the region
in the period between the 1975 base case and the 1995 forecast year.
14
-------
The following data files were obtained from CPO:
1975 higlway network;
1975 transit network;
the trip tables with forecasts of interzonal travel for
1975 for five trip purposes;
total employment and the number of employees in each zone
working for employers with more than one hundred employees
(the number of employees working for employers with less than
100 employees was derived from these employment estimates);
the distribution of traveler's household incomes by zone of
residence; and
the most recent version of a computer program for applying the
modal split and automobile occupancy models currently used by CPO.
These data sets were reviewed in detail by the project team for con-
sistency with the requirements of this study. Updates and adjustments
made to the CPO's data base, modal split, and automobile occupancy models
to create and analyze the base and plan transportation systems in the
region are presented below.
Base System
Highway Network
The highway network incorporated all of the highway improvements
anticipated to be opened to traffic in 1975 at the time that CPO prepared
the network. The following modifications were implemented to develop a
network that corresponded to the highway system as it is currently ex-
pected to be in operation on a ten-mile section of Westbound State High-
ways 125-94 between 1-8 and 1-5. Autombile traffic entering the freeway
will be delayed to the extent necessary to prevent the volume on the main-
line lanes of the freeway from becoming large enough to slow the flow of
traffic below 50 mph. Based on discussions with California Department of
Transportation personnel, it was assumed that each automobile would be
delayed an average of three minutes prior to entering the freeway. After
entering the freeway, vehicles would travel at an average speed of 50 mph.
It is anticipated that preferential treatment would be given to buses
utilizing the affected freeway on-ramps and that buses would therefore not
encounter the three-minute delays at the on-ramps.
15
-------
Specific projects and their relevance to our project are listed
below:
Ramp metering projects are expected to be implemented at
four westbound on-ramps along Interstate 8 between Interstate 15
and State Route 125. A three-minute delay was also assumed
for each automobile using these on-ramps.
State Route 252 — a proposed freeway connecting Interstate
805 and Interstate 5 between National Avenue and Division
Street — was not included in the base network because this
route will probably not be available to traffic by July 1, 1977.
Mera Mesa Boulevard, which is proposed to be constructed west
of Interstate 15 north of the Miramar Naval Station, was not
included in the network because it is not anticipated to be
completed by July 1, 1977.
The proposed interchange at Johnson Street and Interstate 8
was included in the 1977 highway network, but the direct
ramp from eastbound Interstate 8 to northbound State Highway
67 was not.
The remaining highway improvements which have been planned for
the San Diego region had already been incorporated~into the
base 1975 highway network prepared by CPO.l
Highway Operating Policies
Estimates of all-day parking costs for each zone in the San Diego
region for 1975 and 1995 were provided by CPO. An estimate of parking
costs for each zone for 1977 was developed by linearly interpolating between
the 1975 and 1995 estimates. For those zones in which CPO estimated
that there would be a parking fee in 1995 but that parking would be free
in 1975, it was assumed that parking would be free in 1977.
Considerable uncertainty exists with respect to both the supply and
price of gasoline in the immediate future, and there is even greater un-
certainty with regard to these factors in the analysis year of 1977. For
the purposes of this analysis, it was assumed that sufficient gasoline
supplies would be available in 1977 to satisfy the demands for gasoline
resulting from autombile driving — that is, that no gasoline rationing
schemes would be in effect because of limited supplies.
1. Cleaning of the Air, Local Transportation Control Plan for the San
Diego Air Basin; San Diego Air Pollution Control Task Force; San Diego,
California, October 1973.
16
-------
The price of gasoline in 1977 was estimated by projecting the
November 1973 price of 41.9 cents per gallon for regular gasoline to
1977 assuming a 4.5 percent annual increase in the cost of gasoline.
An additional 15 cents was added to the 1977 cost per gallon of gasoline
to account for the possibility of significantly higher gasoline prices
in 1977. The resulting cost of gasoline was approximately 65 cents per
gallon. The cost per gallon was discounted to 1966 dollars for use in
the modal split model and converted into a cost per mile using an average
gas consumption rate of 13.22 miles per gallon. A price surcharge was
also added to the per mile cost of oil to account for possible price
increases through 1977. It was further assumed that out-of-pocket costs
for automobile maintenance and tires would increase at the same rate as
consumer prices. The resulting 1977 cost of operating an automobile ex-
pressed in 1966 dollars is 5.5 cents per mile. This figure was used in
both the base and plan system modal split analyses.
Transit Network
The 1975 CPO transit network included the existing bus system as
well as a number of improvements to the 1973 system. In particular,
it incorporated the remaining six routes of a nine route express bus
system which is being developed by San Diego Transit Corporation (SDTC).
The express routes that it is assumed would be implemented by 1975 are:
express route 1-La Jolla-Pacific Beach;
. express route 4-Lakeside El Cajon-La Mesa;
express route 6-Casa de Oro-Spring Valley-Lemon Grove;
. express route 7-San Ysidro-Chula Vista;
express route S-I^perial Beach-Chula Vista-National City;
and
. express route 9-Santee-El Cajon.
The size of SDTC's bus fleet was expected to increase from its
current size of 250 buses to 350 buses by fiscal year 1977-78 to serve
the planned local and express bus system. Bus service within the San
Diego area is also operated by the Oceanside Transportation System,
Chula Vista City Lines, San Diego Economy Lines, and Weston Greyhound
Lines, although the San Diego Transit Corporation is the major carrier
within the region. Based on a review of the 1975 transit network and
discussions with transit officials and planners in the San Diego region,
it was concluded that the 1975 CPO transit network represented the base
transit system as defined in this study (i.e. , the system which would
be in operation in 1977 were the EPA plan not to be implemented).
1. Transit Development Plan and Program, Fiscal 1972-1973 Update;
San Diego Comprehensive Planning Organization and California;
May 1973.
17
-------
Transit Operating Policies
On September 1, 1972, SDTC established a single fare of 25c
with free transfers for travel to all destinations within the region.
Transfer privileges were improved in that the time interval of one hour
between transfers was raised to three hours with stop-over privileges
on the same route. Based on discussions with personnel at SDTC, it was
assumed that the 25
-------
A system to provide preferential treatment to buses and
high-occupancy automobiles is to be implemented at the
entrance ramps of selected freeway sections. These were
the same sections of freeway for which the California
Department of Transportation was planning to implement
the ramp metering system and included:
California State Highway 125 from Interstate 8 to
its junction with California State Highway 94; and
California State Highway 94 from the junction with
California State Highway 125 to the junction with
Interstate 5.
Car-Pool Matching
The plan highway and transit networks were appropriately modified
to reflect these projects. The travel time for buses on Broadway was
reduced and this street was removed from the highway network. It was
assumed that buses and high-occupancy automobiles would encounter no
delays upon entering freeways whereas low-occupancy automobiles would
encounter an average of a three-minute delay, as noted above.
A computerized car-pool matching system that is conveniently-
available to the general public and to all employees of businesses
having more than 100 employees is to be implemented in the San Diego
region. This system is designed to provide information to home-to-work
travelers regarding other travelers who have similar origins, destinations,
and travel schedules thereby enabling the travelers to form car pools.
The availability of a high-quality car-pool matching system is an essential
element of any program to improve automobile occupancy for home-to-work
travel.
Parking Supply Management
The EPA plan includes a program to require EPA approval of any
significant additions to the supply of parking facilities in the San Diego
region. It would be incumbent upon the developer of a parking facility
to demonstrate that the proposed facility would not prevent or interfere
with the attainment or maintenance of any national air quality standard
at any time within ten years from the date of application. The level of
detail required in the developer's submittal is dependent upon the size
of the proposed facility with one level of detail being required for
projects which would increase capacity by 50 to 249 vehicles and a greater
level of detail required for projects which would increase capacity by
250 or more vehicles.
19
-------
The availability of parking facilities could have an important
impact upon the travel choices of urban residents — particularly if
the supply of parking facilities is insufficient to serve the demand
at particular locations or at particular time periods. Thus, this
measure could have an important impact upon traveler decisions if the
number of automobile trips were to significantly increase. Inasmuch
as the measure only controls increases in parking capacity and no
rollbacks in the supply of parking are required elsewhere in the EPA
plan, the impact of the parking supply management measures is more
limited if the demand for parking does not increase. Since the EPA
plan is designed to decrease significantly the volume of automobile
travel in the San Diego region, it is assumed that this measure would
not be a constraint upon automobile travel~or mode choice in the time
"Trame of this study.
Mass Transit Incentives for Employees
The provisions of this measure impact primarily upon "large"
employers — where a large employer is defined as maintaining more
than 70 parking spaces. Employers in this category are required to
implement two programs — a surcharge program and a reimbursement
program.
The parking surcharge program has differential rates for varying
levels of automobile occupancy as follows:
one-person automobiles - daily parking surcharge of
$2.50 plus a commercial parking rate, if applicable;
two-person automobiles - fifty percent of the cost of
one-person automobiles;
three-person automobiles - no parking cost (i.e., either
commercial or surcharge) and the allotment of those parking
spaces located closest to the employment facility.
The large employer, in addition to charging the parking surcharges,
is required by the EPA plan to charge the average daily commercial rate
charged by the three operators of parking facilities each containing
100 or more commercial spaces within two miles of the employer. If
the large employer is more than two miles from such commercial facilities,
only the parking surcharge is to be charged. This effectively means that
large employers within two miles of the San Diego CBD would be required
to charge their employees who drive to work a commercial parking rate plus
the applicable parking surcharges. The average commercial parking rate
used to estimate total daily parking cost was the average daily parking
charge in zones on the periphery of the CBD. The differences in parking
costs by automobile occupancy class and by location of the large employer
relative to the San Diego CBD have been incorporated in the inputs to the
plan modal split analysis.
20
-------
Reimbursement was considered to all employees using public transit
service to and/or from work of the first $200 per year of their transit
fares. Since transit fare was assumed to be 25
-------
The EPA surcharge measure relating to free parking facilities
imposes a regulatory fee of $450.00 per year per space upon each owner
or operator of more than five free parking spaces including free on-
street parking spaces outside of residential areas. The amount of
this fee was determined by calculating the annual proceeds of a 25c per
hour surcharge — the same as imposed on commercial parking spaces —
and assuming a 250-day year and a 60 percent average occupancy for each
space during the days and hours for which the fee is collected. This
measure impacts the majority of workers in San Diego inasmuch as about
75 percent of the workers in San Diego are employed by employers with
less than 100 employees and 90 percent of the employees in the region
are provided with free parking facilities. (Commercial parking spaces
in the San Diego region are located largely in the CBD, and less than 10
percent of the workers in the region are employed there.) Similarly,
this is the parking surcharge measure that would have the greatest impact
upon nonwork travel in the San Diego region inasmuch as well over 90 per-
cent of the nonwork travel in the region is not oriented to the CBD. In
effect, this measure would impose a parking surcharge upon the large
number of free parking spaces provided by commercial, recreational, and
even religious organizations. Any organization providing free parking
facilities for use for other than parking by residents of a structure
would be subject to this surcharge.
The surcharge on free parking spaces is the only surcharge which is
imposed upon the owner and operator of the facility rather than upon the
parker. In assessing the impact of this measure upon travel behavior,
it is necessary to consider whether this surcharge will be imposed
directly or indirectly upon the traveler. For example, if a shopping
center pays the surcharge from its operating budget and correspondingly
increases the price of the goods sold to the travelers — the traveler
would not directly perceive the cost of parking in making his travel
decisions. On the other hand, if a parking meter to collect a surcharge
of 25
-------
travel analysis should assess the combined effects of all of the EPA
measures — save the limitations upon gasoline consumption. In this
manner, the analysis provides an initial assessment of whether such
limitations upon gasoline consumption are required. If such limitations
were required, the impact of these upon travel patterns in the San Diego
region will be assessed in subsequent phases of this project.
Locally Implemented Measures
The EPA plan takes note of a series of locally implemented strategies
as being applicable toward attainment of the oxidant standard and encourages
their implementation, although the reductions are not credited as part of
the EPA plan. Inasmuch as EPA recognized that the locally implemented
measures will be considered in assessing progress toward attainment of
the oxidant standard, these locally implemented measures were incorporated
in this analysis. Only those locally implemented measures which would
result in incremental changes to the base 1977 transportation systems
described above are considered in this section.
Bus System Improvements
The following bus system improvements are proposed as part of the
locally implemented improvements:
. Purchase of an additional 200 buses to improve the level
of service and expand geographic coverage of the public
transit system in the San Diego region. These buses would
be in addition to the 350 buses which were planned for operation
in the San Diego region in 1977 prior to the advent of the EPA
transportation control plan.
. An exclusive bus by-pass lane would be constructed to allow
buses traveling through Balboa Park onto State Highway 163
northbound to by-pass existing traffic congestion. The
estimated two to five minutes savings in travel time for
buses using the by-pass lane was appropriately incorporated
into the analysis.
A set of express transfer facilities has been proposed at
freeway ramps on reconstructed Interstate 15 — the Escondido
Freeway. It is proposed that each transfer facility be equipped
with a parking lot for approximately 200 cars and secure facilities
for bicycle storage. Provision is being made for the construction
for three such facilities — at Mira Mesa Boulevard (which would
be constructed first as a demonstration project), at Poway Road,
and Rancho Vernardo. The locally implemented plan notes that
20 such lots strategically located throughout the region could
have the potential for reducing daily vehicle miles of travel but
that the likelihood of implementing such a system by the year 1977
seems remote. Based on these considerations, only the fringe
parking facility at Mira Mesa Boulevard was incorporated into the
analysis of travel patterns.
23
-------
Two non-capital intensive programs are proposed:
a comprehensive marketing program — incorporating
scheduling, schedule information, system maps, and
information — designed to increase public awareness
of the public transportation system; and
allowing employers to purchase transit tickets at
reduced cost and provide them to their employees to
encourage the employees to utilize transit travel to
work.
An assessment of the impact of the 200 additional buses upon
travel patterns within the San Diego region requires that the services
provided by these buses—that is the routes and schedules of the busses—
be described in detail. Discussions with planning personnel of SDTC did
not at this time have specific recommendations regarding the utilization
of these additional 200 buses. The SDTC planning personnel indicated
that priority in the allocation of the additional buses would be given
to providing additional service on (1) those existing bus routes which
were currently experiencing over-crowding of buses and (2) reducing
headways on those existing routes with excessive headways. A listing of
the peak hour and midday period headway changes to the existing routes
and the proposed express routes which would result from implementation
of these guidelines is presented in Table 1.
In addition, the following new bus routes were included in the plan
transit network:
feeder bus service connecting Scripps Ranch to Route 6;
feeder bus service connecting the Tierra Santa area to
proposed express route 9; and
shuttle bus service from San Diego Stadium to the San Diego
Central Business District running on ten minute headways
during the peak hours of travel.
24
-------
TABLE 1
1977 BASE AND PLAN CONDITION
BUS ROUTE HEADWAYS
Bus Route
1
2
3
4
5
5X
6
7
9
11
11X
12
Express Rt. 1
Express Rt. 4
Express Rt. 6
Express Rt. 7
Express Rt. 8
Express Rt. 9
B
C
E
_F
G
H
J
K
L
0
0-1
R
T
S. D. Economy Line
Bus Route Headways (Minutes)
Peak Hour
Base*
25.0
35.0
25.0
30.0
20.0
15.0
50.0
25.0
35.0
30.0
15.0
30.0
20.0
20.0
20.0
20.0
20.0
20.0
30.0
50.0
30.0
30.0
30.0
30.0
30.0
30.0
30.0
15.0
60.0
15.0
30.0
99.9
Plang
12.5
17.5
12.5
15.0
10.0
7.5
25.0
12.5
17.5
15.0
7.5
30.0
8.0
6.0
5.0
4.0
10.0
7.0
15.0
50.0
15.0
30.0
30.0
30.0
30.0
30.0
30.0
7.5
60.0
7.5
15.0
99.9
Midday
Base*
25.0
25.0
25.0
30.0
40.0
30.0
40.0
20.0
45.0
30.0
30.0
50.0
40.0
40.0
40.0
40.0
40.0
40.0
60.0
50.0
50.0
40.0
60.0
30.0
30.0
40.0
30.0
60.0
60.0
30.0
50.0
99.9
Plan@
12.5
12.5
12.5
15.0
20.0
15.0
20.0
10.0
22.5
15.0
15.0
50.0
15.0
15.0
15.0
15.0
15.0
15.0
30.0
50.0
25.0
40.0
60.0
30.0
30.0
40.0
30.0
30.0
60.0
15.0
25.0
99.5
*Base condition headways were obtained from 1975 transit system
computer listing provided by CPO.
@Plan condition headways were derived by FMM&Co. and discussed
with SDTC.
Note: (1) Routes E-l, E-2, and S, which are relatively short
routes, are onitted from the above listing.
(2) Headways shown are generally the lowest headways
on each bus route as coded in the transit networks.
25
-------
Implementation of all of the schedule and route changes suggested
by the SDTC personnel would require full utiliEation of the 200 additional
buses to be purchased. If it were desired to add many new bus routes in
the San Diego region—particularly to non-CBD destinations—it would appear
that more than 200 buses would have to be added to the bus fleet.
Methodology for Analyzing Impact of the EPA Promulgated Plan
The methodology utilized in this study was based on regional
transportation data obtained from the San Diego Comprehensive Planning
Organization. Certain refinements were required to more fully consider
the impacts of the EPA promulgated plan. These refinements included:
(1) an analysis of car-pooling impacts, and (2) a more thorough consid-
eration of foregone trips (i.e., those trips which will not be made because
of increased travel cost). Significant revision of the regional modal
split model computer program was required in order to implement these
refinements of the original transportation planning data base.
Utilization of the Regional Transportation Planning Process
An overview of the transportation planning process of the San Diego
region is presented in Exhibit 1. This exhibit illustrates each of
the six major phases of a strategic planning process, and the major
intermediate products which result from each phase of the analysis.
The analyses described in this report were based on the fourth
step in the regional transportation planning process — the modal
split analysis. Essentially, the modal split analysis estimates the
share of the urban travel market which is captured by each of the
major modes — where the modes are defined as transit passenger, auto-
mobile driver, and automobile passenger. The modal split model is
predicated on the assumption that the market shares held by each mode
are a function of the level of service provided by each of the modes
(e.g. , price, travel time for different trip components including
walking, waiting, line-haul, and transfer) and the socioeconomic char-
acteristics of the traveler (e.g., income). The major inputs to the modal
split analysis are:
. matrices of person trips by purpose between all
zones of the San Diego region for the forecast year;
a set of matrices describing the highway and transit
travel times for each component of a trip between all
zones in the San Diego region for the forecast year; and
the household income distribution of the urban travelers
for each traffic analysis zone.
The major data inputs for the modal split process were provided by
CPO and modified as necessary for use in this analysis. CPO also pro-
vided trip tables for total trips for a 1975 forecast year; as described
below, these were modified to develop trip tables for a 1977 forecast
year.
26
-------
EXHIBIT 1.
OVERVIEW OF TRANSPORTATION PLANNING PROCESS
FOR SAN DIEGO REGION
LAND USE FORECASTING
TRIP GENERATION
POPULATION
AND
EMPLOYMENT
SUMMARIES BY
ZONE
TRIP DISTRIBUTION
MODAL SPLIT
HIGHWAY AND TRANSIT
NETWORK ASSIGNMENT
TRIP TABLES
FOR
AUTO DRIVER,
AUTO PASSENGER,
AND
TRANSIT
PASSENGER
SYSTEM EVALUATION
HIGHWAY
AND
TRANSIT NETWORK
LINK VOLUMES
27
-------
The inputs made to the 1975 highway and transit systems in order
to describe (1) the base system which would have existed in the San
Diego region prior to the advent of the EPA transporation control plan
and (2) the transportation system which would result from implementation
of the EPA promulgated transportation control plan, have been described
above.
Description of the Modal Split Model
The modal split model is briefly described below. A further
description of the structure, estimated parameters, and testing and
validation of the modal split model is presented in Appendix A. •*•
For purposes of the modal split analysis, travel in the San Diego
region is stratified into two sets: travel to CBD destinations, and
travel to non-CBD destinations. This stratification was selected because
investigations indicated that the propensity to use transit for CBD travel
was greater for equivalent service alternatives than it was for non-CBD
travel.
A detailed analysis was undertaken to select that set of independent
variables, and to estimate the best parameters for each of the independent
variables that provide the highest quality modal split model for each travel
stratification for the San Diego region. Table 2 presents the recommended
CBD and non-CBD models. Three transportation variables, the differences in
line-haul time, excess time, and modal costs, and one socioeconomic variable
(i.e., household income) are used in the models. All of the coefficients in
the model are significant at the 95 percent level.
The major contrast between the CBD and non-CBD models is to be
found in the time and cost coefficients: C, D, and E. For the CBD
medel, the effect of excess time as compared to line-haul time (i.e. ,
ratio of £) is about 1.63. The derived value of time (i.e., the ratio
of E to C or D) is $3.18 per hour for line-haul time and $5.17 per hour
for excess time. These estimates confirm results observed elsewhere.
Excess time becomes a critical determinate of modal choice for non-CBD
trips as shown by the value of the coefficient C and the weight of excess
time (relative to line-haul time), both of which are substantially higher
than in the CBD model. The increase in the weight of excess time from
1.6 to 6.8 affects both the major differences in CBD and non-CBD transit
service and the differences in travel behavior resulting therefrom.
Implementation of the n-Dimensional Logit Model, Final Report
Prepared by Peat, Marwick, Mitchell & Co., Washington, D.C., for the Compre-
hensive Planning Organization, San Diego County; California, May 1972.
28
-------
TABLE 2
SAN DIEGO MODAL SPLIT MODELS
Model Form; Parameters of the Models;
Parameter CBD model Non-CBD model
P = 1.0 + exp(aTI35 + b) + exp(cDX3 + dDL3 + eDCH + f) ° 0.0295 0.0268
, T-nc 4. KN b -1.4809 -0.5441
p _ exp (aT 13 5 + b)
D 1.0 + exp(aTI35 + b) + exp(cDX3 + dDL3 + eDCH + f) c 0.0916 0.1314
exp (cDX3 + dDL3 + eDCH + f d 0.0563 0.0192
PT 1.0 + exp(aT135 + b) + exp(cDX3 + dDL3 + eDCH + £)
e 0.0106 0.0184
f 1.1635 1.6600
N>
VO
Nomenclature;
Pp> PD» PT = auto passenger, auto driver, and transit passenger modal choice probabilities
exp = exponential operator
a, b, d, = calibrated coefficients and constants for the model
d, e, f
TI35 = transformed household ijicome
= 100/L-exp (-0.035*INC1/
INC = = household income in hundreds of dollars
DX3 = difference in excess time
= walk to/from auto time - (walk to transit time + first wait for transit + walk from transit)
DL3 = difference in line haul time
- auto driving time - (transit in vehicle time + transit transfer time)
DCH = difference in modal cost
auto operating cost (5
-------
The models were tested to see if they could reproduce the total
number of automobile passenger, automobile driver, and transit passen-
gers trips in the calibration sample. The CBD and non-CBD models
reproduced almost exactly the total number of trips by each of the three
modes considered. Calculated trips were within a (negligible) fraction
of one percent of the observed trips.
The models also reproduce the observed trip distributions by trip
length. Cumulative distributions of transit trips as a function of trip
length (in minutes) as observed in the calibration data set and estimated
by model are presented in Figure A-3 of Appendix A. In all trip length
ranges, the calculated distribution of transit passengers is within
about 15 percent or less of the observed distributions. The accuracy
of these results is indicative of the overall quality of the models.
Modifications to Develop Analysis for 1977 Forecast Year
The data inputs provided by CPO were predicated on an analysis year
of 1975. Inasmuch as the transportation control plan promulgated by EPA
was to achieve its full effect by July 1, 1977, it was necessary to develop
a 1977 analysis capability. The modifications of the 1975 highway and
transit networks to develop networks which represent (1) the base system
which would exist in San Diego in 1977 prior to the advent of the EPA pro-
mulgated transportation control plan, and (2) the transportation system
which would exist as a consequence of the EPA promulgated transportation
control plan, have been described above. Modifications of the 1975 trip
tables of total person trips to produce 1977 trip tables are described
below.
Over the last .ten years, the best available estimates are that the
vehicle-miles of travel (VMT) by automobile in the San Diego region have
increased at a compounded annual growth rate of 5.5 percent. This increase
in automobile travel has resulted from (1) increases in the total number
of trips caused by increased population and increased income, and improve-
ments to the highway system, and (2) increases in the average trip length
resulting from increases in income and improvements to the highway system.
The growth in annual VMT is a reasonable indicator of travel growth,
because over 98 percent of all trips are made by automobile in the San Diego
area. Hence, each entry in the 1975 trip table was multiplied by a factor
of 1.12, thereby, ensuring that the total VMT increased at a rate of 5.5
percent per year under the base condition.
In view of the relatively short two-year time frame for which this
extrapolation was accomplished, and the scope of this study, this approach
was considered acceptable. If the analysis were to be accomplished for a
more extended time frame—beyond five to 10 years—it would be desireable
to proceed through the various phases of the transportation planning process
to ensure that the forecasts of total trips appropriately reflected differ-
ential changes in land use, trip generation, distribution, and the
major variables.
30
-------
The EPA promulgated transportation control plan provides major
incentives for car-pooling and significantly increases the out-of-pocket
cost of an automobile trip (as a result of the parking surcharge) which
may result in certain trips no longer being undertaken. These aspects
of the EPA promulgated plan are of such significance that modifications
of the previously developed modal split process were undertaken to ensure
that these impacts were appropriately assessed; these modifications are
described below.
Analysis of Car-Pooling Impacts and Surcharges
Because of the major pricing incentives created by the EPA plan to
encourage car-pooling, it was necessary to develop a more explicit mechan-
ism for assessing the impacts upon automboile occupancy. An assessment of
the total cost per person of using an automobile for home-work travel in
1977 under the EPA promulgated transportation plan for various levels of
automobile occupancy is presented in Table 3. In considering the costs
for automobile travel, it was useful to distinguish between non-CBD oriented
travel—which largely involves employer provision for free parking facilities—
and CBD oriented travel—which largely requires employees to utilize commer-
cial parking facilities. In considering the non-CBD oriented travel, it is
further useful to distinguish betwen employees working for employers with
more than 100 employees and employees working for employers with fewer than
100 employees inasmuch as the surcharge program differs for these two groups
of employees. This distinction is not relevant in the case of CBD oriented
travel, since the surcharge for commercial parking facilities is imposed
upon employees irrespective of the size of the organization by which they
are employed.
The total one-way driving cost and the average one-way cost per person
for each of these travel groups, assuming automobile occupancies of 1, 2,
and 3 persons, are presented in Table A. The average travel distances for
home-to-work travelers with CBD and non-CBD destinations were those observed
in the 1966 home interview survey. While the average travel distances for
home-to-work travel in 1977 are greater than they were in 1966, the increases
in average trip distance do not significantly affect the results.
In conducting the analysis, a major effort was made to identify data
pertaining to car-pooling arrangements in San Diego or elsewhere in the
country to assess the increases in average travel distance and average travel
time associated with car pools of two and three individuals. Unfortunately,
it was not possible to locate any survey data pertaining to the incremental
travel times and travel distances associated with car pooling. The impedances
associated with car pooling were therefore based on the following assumptions:
for trip interchanges involving 100 or more trips, the
incremental travel time associated with one added passenger
in the car is five minutes for the driver, and the incre-
mental travel time associated with two passengers in the car
is eight minutes for the driver and three minutes for the
second passenger—this is the total incremental time for the
trip and considers the added travel time for both the origin
and the destination;
-------
TABLE 3
IMPACTS OF CAR POOLING UPON COSTS OF
HOME-WORK AUTOMOBILE TRAVEL IN 1977 UNDER EPA
PROMULGATED TRANSPORTATION CONTROL PLAN
Non-CBD Oriented Travel -
Free Parking Facilities
Employees working for Employers with
greater than 100 employees
Average one-way travel distance (miles)
Out-of-pocket distance cost (cents)*
Parking cost (cents)*
Total one-way trip cost (cents)*
Average one-way trip cost per person (cents)*
Incremental charge in one-way trip cost
per person (cents)*
Incremental travel time (minutes)
Employees working for employers with
less than 100 employees
Average one-way travel distance (miles)
Out-of-pocket distance cost (cents)*
Parking cost (cents )*
Total one-way trip cost (cents)*
Average one-way trip cost per person (cents)*
Incremental charge in one-way trip cost
per person (cents)*
Incremental travel time (minutes)
CDB Oriented Travel - Commercial
Parking Facilities
All Employees
Average one-way travel distance (miles)
Out-of-pocket distance cost (cents)*
Parking cost (cents)*
Total one-way trip cost (cents)*
Average one-way trip cost per person (cents)*
Incremental charge in one-way trip cost
per person (cents)*
Incremental travel time (minutes)
Number of Persons in Car Pool
1
7.1
63.4
125.0
188.4
188.4
-
-
7.1
63.4
125.0
188.4
188.4
-
-
7.7
68.8
163.1
231.9
231.9
-
-
2
9.3
83.0
62.5
145.5
77.8
115.6
6.5
9.3
83.0
125.0
208.0
104.0
84.4
6.5
9.9
88.4
163.1
251.5
125.8
106.1
6.5
3
10.5
93.8
0.0
93.8
31.3
41.5
3.9
10.5
93.8
125.0
218.8
72.9
31.1
3.9
11.1
99.1
163.1
262.2
87.4
38.4
3.9
*1977 prices, 1972-3 base.
32
-------
TABLE 4
CAR POOLING PARTICIPATION RATES
Income Level
Income between 0
and $5,000 per year
Proportion of car
pools with this
occupancy
Proportion of persons
with this occupany
Income between $5,000
and $10,000 per year
Proportion of car
pools with this
occupancy
Proportion of persons
with this occupancy
Income greater than
$10,000 per year
Proportion of car
pools with this
occupancy
Proportion of persons
with this occupancy
Number of Passengers in Automobile
123
.09
.24
.3
.26
.5
.59
.43
.4
.57
.5
.65
.1
.17
.0
.0
Average Auto-
mobile Occupancy
2.3
1.7
1.4
33
-------
the automobile driver will achieve an average speed of
20 mph in picking up and dropping off passengers at each
end of the trip; and
corresponding incremental travel times and travel dis-
tances for interchanges involving fewer than 100 trips
are 130 percent of the corresponding values for inter-
changes involving more than 100 trips.
About 85 percent of the home-to-work travel in San Diego involves inter-
changes with fewer than 100 trips; thus, most of the home-to-work trav-
elers are assumed to have incremental times of 6.5 and 10.4 minutes and
incremental distances of 2.2 and 3.4 miles for automobile occupancies
of two and three respectively. The incremental times and distances for
each number of persons in a car pool are summarized in Table 5.
A conclusion that can be drawn is that travelers achieve substantial
savings—on the order of $.84 to $1.50 per one-way trip—by increasing the
automobile occupancy from one to two persons. The incremental savings,
which is achieved by increasing the automobile occupancy from two persons
to three persons is relatively speaking smaller—ranging from about $.31
to $.42 for each one-way trip.
These resutls suggest two findings regarding traveler behavior in
response to the pricing incentives to car pool. First, those travelers
with a lower income will have a greater incentive to car pool than those
travelers with a higher income—all other factors being equal. Thus, the
rate of automobile occupancy might be expected to decreases as the trav-
eler's income increase. Second, the incremental savings which are to
be derived by increasing automobile occupancy from one to two persons is
about 2.7 times greater than the incremental savings which are derived by
increasing the automobile occupancy from two to three persons. In this
context, two-person car pools might be expected to offer significant
attractions to travelers in spite of the fact that ever greater savings
are provided by three-person car pools.
The relatively high average automobile occurancy rates of 2.3, 1.7,
and 1.4 persons for the low, medium5 and high income ranges may be con-
trasted with the base period average automobile occupancy for home-to-
work travel of 1.1 to 1.2 persons per automobile. Given these car-pooling
participation rates, over half of the persons with low incomes would travel
to work in three-passenger car pools, whereas over half of the persons with
medium and high incomes would travel to work in two-passenger car pools.
In contrast, about 82 percent of the persons in the base condition traveled
to work in automobiles in which they were the only occupant.
The car-pooling participation rates for different income groups as
well as the incremental impedances associated with car-pooling were utilized
to analyze the car-pooling impacts of the transportation control plan. Inas-
much as the base plan involved incremental improvements of the existing system
with no major pricing incentive to encourage car-pooling, the automobile
occupancy analysis incorporated into the calibrated modal split model for the
San Diego region (as described above.) was utilized in this analysis.
34
-------
TABLE 5
INCREMENTAL TIMES AND DISTANCES
Interchange
involves 100
or more trips
Interchange
involves fewer
than 100 trips
Travel Time
No. of Persons
in Automobile
2 3
(minutes) (minutes)
6.5
10.4
Travel Distance
No. of Persons
in Automobile
2 3
(miles) (miles)
1.7
2.2
2.6
3.4
Proportion of
Total Work Trips
in this Category
Percent
15.0
85.0
35
-------
Consideration of Foregone Trips
It was recognized at the outset of this study that the combination
of an enhanced level of public transportation in the San Diego region as
well as restrictions imposed upon automobile usage might affect the number
of trips undertaken in the region as well as their destinations. The
difficulty of reliably estimating induced or foregone trips was also
recognized; aside from the nodal split model, the existing travel demand
analysis for the San Diego region is not sensitive to the costs of auto-
mobile travel nor to the level of service provided by the transit system.
Thus, while the importance of carefully assessing induced and foregone
trips was recognized, it was also understood that the results of such an
analysis are more open to question than the results of the modal split
analysis.
The parking surcharge measure promulgated in the transportation
control plan significantly increased the out-of-pocket cost of automobile
travel for essentially all work and non-work trips undertaken in the
region. These increases in the cost of automobile travel are so great that
they would undoubtedly motivate changes in the frequency and destinations
of automobile travel, in addition to encouraging automobile users to divert
to transit. If the impacts of the increases in automobile costs upon the
frequency and destinations of trips were not consdiered, the results of this
analysis vrould be highly unrealistic. It was, therefore, imperative that
some mechanism be incorporated into the analysis for assessing the extent
to which trips would be foregone as a result of the automobile price in-
creases.
The increase in the bus fleet from the planned 350 buses under the
base condition to a total of 550 buses represents a most important 57 per-
cent increase in the size of the bus fleet relative to the base condition
for 1977 and is about 220 percent greater than the existing 1973 fleet of
250 buses. While this represents a most significant improvement to the
regional bus service, it should be recognized that the parking surcharge
represents a greater potential impact on the total regional transportation
service than does this increase in the number of buses. In this context,
it was evident that the analysis of induced and foregone trips needed to
focus primarily upon the impacts of the parking surcharge and secondarily
upon the impacts of the bus service improvements. While an assessment of
the induced travel demands resulting from the recent transit fare reductions
in San Diego and Atlanta provides some insight into the induced demands
resulting from improved bus service, these results do not provide sufficient
information to assess the impacts of increases in the out-of-pocket cost of
automobile travel upon automobile users. A literature review was therefore
undertaken to identify and assess the existing models for assessing the
travel consumption affects of the automobile price increases.
The literature review identified five approaches for achieving this
objective:
. Demand Generator and Modal Split Model—Rand Corp.;
36
-------
. Direct Demand Estimation Model—Kraft;
. Attitudinal Model—Wallace;
Shopping Trip Frequency Model—Charles River Associates;
and
Econometric Model of Gasoline Consumption—Data Resources, Inc.
A description and assessment of each of these approaches is presented in
Appendix B.
It was concluded that the Econometric Model of Gasoline Consumption
provides, at this time, the most reliable available method for assessing
the impacts of the increases in the price of automobile usage upon the
consumption of person-miles of travel. The approach taken by Rand—while
conceptually interesting—was rejected because the relationship postulated
in the report did not appear to have been empirically calibrated or valid-
ated to a sufficient extent. The Direct Demand Estimation Model suggested
by Kraft and a number of other authors is also conceptually attrative but
it has proven extremely difficult to calibrate and validate reliable dir-
ect demand estimation models using aggregate data. The Attitudinal Models,
such as those suggested by Wallace, differ from the other approaches in
that they are based upon ellicited attitudes rather than actual travel
behavior. While the approach appears to offer promise, fully developed
models of this type have yet to be implemented in a form such that they
can be used in this study. The Shopping Trip Frequency Model developed
by Charles River Associates is a most useful development—particularly
because it is feasible to calibrate this model using a disaggregate
approach. The work is insufficiently developed, however, for use at this
time in this sutdy for two reasons: (1) corresponding models have not
been developed for the other trip purposes, and (2) while the model
addresses the issue of trip frequency, it does not allow for consideration
of changes in trip destination.
The Econometric Model of Gasoline Consumption developed by Data
Resources, Inc. has been empirically validated and directly addresses the
issue of changes in the vehicle-miles of travel; for these reasons, it
was selected for use in this study.
The basic assumption underlying the use of the results of this econ-
ometric model of the quarterly demand for gasoline as a function of gasoline
price and income is that the price elasticities estimated for gasoline con-
sumption could also be used as estimates of the price elasticities for
consumption of vehicle-miles of travel (and thus, person-miles of travel
assuming constant automobile occupance and relatively minor diversion to
transit). Factors to be considered in assessing the reasonableness of
this approach include:
37
-------
. the price of gasoline represented essentially a constant
proportion of the total out-of-pocket costs of automobile
travel during the time period represented by the calibration
data;
changes in gasoline consumption may have resulted from
changes other than vehicle-miles of travel (e.g., substi-
tution of automobiles which provide a greater number of
miles per gallon); and
. changes in travel patterns are "appropriately" related to
the changes in gasoline consumption.
A review of available information regarding the per-mile costs of
operating a car suggests that during the period covered by the cali-
bration data set for the model, the cost of gasoline has constituted
essentially a constant proportion of the costs of operating a car.
Admittedly, gasoline consumption is being used in this analysis as
a surrogate for vehicle-miles of travel. If the average miles per gallon
achieved by automobiles increase or decrease, the vehicle miles of travel
would correspondingly increase or decrease with no change in gasoline con-
sumption. Changing one's consumption of vehicle-miles of travel is,
however, an immediately feasible action which consumers can take in res-
ponse to changes in gasoline prices. Other changes—such as the substi-
tution of automobiles which achieve a higher gas mileage—require a longer
time to achieve. The Econometric Model of Gasoline Demand provides both
a short-run price elasticity (which is defined as the elasticity for the
immediate three months following the price change), and a long-run price
elasticity (which is defined as the elasticity pertaining to a period
approximately 10 quarters following a price change). In this context,
the short-run elasticity may be viewed as a conservative estimate of the
change in vehicle miles of travel. For this reason, the estimate of short-
run elasticity of gasoline demands was utilized in this study to estimate
foregone trips as a function of increased travel costs.
The analysis of the demand for gasoline was conducted with four
different data sets, each of which measured a slightly different version
of the aggregate sales of gasoline. Of these four data sets for gasoline
consumption, the "Federal Highway Administration's Highway Motor Fuel
less Highway Special Fuels" estimate of gasoline consumption pertained
most closely to an estimate of the gasoline consumed for highway driving.
Thus, the short-run price elasticity for gasoline consumption based on
this data which has a value of -.16 using an error component estimation
technique with a linear model — was utilized in this study. The short-
run price elasticities for gasoline consumption which were estimated using
the other three data sources were comparable, although somewhat lower,
than this estimate. -
The Cost of Operating An Automobile, Federal Highway Administration,
Washington, B.C., published periodically.
38
-------
This estimate of -.16 for the short-run price elasticity was
developed using consumption data for all states. Prior to utilizing
such an estimate, it is necessary to consider whether it is applicable
to the case study city — San Diego. While it was not possible to
validate this model utilizing data from San Diego, the model has been
validated using California data. Examination of the actual versus the
estimated values for California suggested that the utilization of this
parameter for California did not cause significant errors in prediction
of gasoline consumption. The simple correlation coefficient when com-
paring predicted and actual values of gasoline consumption for California
using the "Federal Highway Administration Highway Motor Fuel less Special
Fuel" data base was .996.
One further modification of this elasticity was required in order
that it be specified on a basis comparable to its use in this study. It
appears reasonable to assume that within the range of price changes caused
by the transportation control plan, travelers would not change their con-
sumption of home-to-work travel — except in the most marginal and stat-
istically insignificant employment situations. The work trip will be one
of the last trips to be foregone, since this would imply that the house-
hold is also foregoing its income. Since the surcharge program would be
applied essentially uniformly throughout the San Diego region, there
would also be no incentive for workers to shift employment sites in order
to alleviate the economic effect of the parking surcharge, except in those
instances where transit service would serve one job location but not another.
Thus, within the price range of the increase in automobile costs resulting
from the parking surcharge, it appears reasonable as a first order of
approximation to assume that all of the changes in travel consumption
would be for non-work travel. Inasmuch as the price changes for gasoline
during the period represented by the calibration data set for the Econo-
metric Model were also relatively minor, a comparable assumption could be
made that the changes pertain solely to the changes in travel consumption
for non-work purposes. Inasmuch as the short-run elasticity estimated by
the model considers gasoline consumption for all trip purposes — both work
and non-work — one could argue that this average elasticity overestimates
the short-run elasticity of travel consumption for work travel and under-
estimates the short-run elasticity of consumption for non-work travel. The
travel analyses indicate that in the 1975-1977 time frame, about 52.9 percent
of the travel in the region is non-work related and the remaining 47.1
percent is work related. Given the previous assumptions that consumption
of work travel is essentially inelastic and the estimate of the proportion
of travel which is non-work, it can be seen that the short-run price elas-
ticity of non-work travel is -.16 divided by .5294 or -.302. This estimate
of the short-run price elasticity of travel was utilized in this analysis.
Finally, it is necessary to consider the range of validity for this
elasticity estimate. In discussions with the model developers, it was
suggested that the elasticity estimates were valid for gasoline prices
up to 60 cents per gallon — a 50 percent price increase as compared to
the early 1973 price or 40 cents per gallon. Consequently, it was assumed
that the elasticity estimate was valid for up to a 50 percent increase in
39
-------
the out-of-pocket costs of automobile travel. Inasmuch as the parking
costs (including the surcharge) are constant irrespective of the length
of the trip and the other out-of-pocket costs increase with the average
trip length, the relative increase in the total out-of-pocket cost repre-
sented by the parking surcharge decreases with increasing trip distance.
Assuming a parking duration of 1.6 hours or less for all non-work travel,
it was found that the relative increase in the out-of-pocket cost of
automobile travel exceeded 50 percent for trip lengths less than 3.5
miles. The anlaysis was therefore predicated on the conservative
assumption that the relative increase in price for trips less than 3.5
miles was no more than 50 percent; and that for trip lengths greater
than 3.5 miles, the relative increase in the out-of-pocket costs was
correspondingly small.
Implementation of the Analysis Refinements
The computer code previously developed to apply the regional modal
split model was modified to incorporate the following:
stratify home-to-work trips by income category and according
to whether they work for employers with more than 100 em-
ployees or employers with fewer than 100 employees;
apply the appropriate car-pool participation rate to each
stratification;
compute the appropriate automobile and transit impedances for
the workers in each category (these impedances vary according
to car-pool size, whether the organization employed more than
or fewer than 100 employees, and whether commercial parking
facilities were available in the zone);
apply the modal split model to each segment of home-to-work
travel to estimate transit usage and then appropriately
compute the number of automobile users;
estimate the number of foregone trips using the price
elasticity of travel applied to non-work trips;
estimate the modal split of non-work trips using the non-
work modal split factors after reducing total non-work
trips by the number of trips foregone;
estimate the modal split of foregone non-work trips
using the* non-work modal split factors added to other
non-work transit trips; and
output the work and non-work travel patterns, trips
foregone, and transportation system characteristics data
in a form suitable for use in the evaluation process.
40
-------
RESULTS OF THE TRANSPORTATION ANALYSIS
A discussions of the findings of this analysis with respect to the
impacts of the EPA promulgated transportation control plan upon travel
behavior and the transportation system in the San Diego region is pre-
sented in this section. This discussion focuses upon the total impacts
at a regional scale, on the incidence of the impacts with respect to
household incomes, and on the geographic incidence of the impacts. A
detailed description of the results for both the base and the plan sys-
tems—stratified with respect to trip purpose and geographic and house-
hold income incidence—is presented in Appendix C.
Impacts of the transportation control plan are measured in terms of
the following criteria:
. person trips—the average number of trips which residents
of the San Diego region would undertake on an average work
day in mid-1977;
. person miles—the total miles of travel which residents of
the San Diego region would undertake on an average work day
in mid-1977, the person miles of travel for automobile
drivers is the vehicle-miles of travel (VMT);
. travel costs—the total daily cost of travel expressed in
1973 dollars which residents of the San Diego region would
pay on an average work day in mid-1977; and
. travel time—the total daily hours associated with the
travel of residents of the San Diego region in mid-1977.
Distribution of the out-of-pocket costs and travel time associated with
individual automobile and transit trips under the base and plan conditions
are also presented.
Impacts at a Regional Scale
The impacts of the transportation control plan upon the San Diego
region—as measured by changes in the above identified four criteria—
are presented in Table 6. The major findings which can be developed from
this exhibit are:
. whereas the volume of automobile driver trips decreases by
about 18 percent, the VMT decreases by only about 12 percent;
. the factors contributing to the reduction.in VMT in order of
decreasing importance;
41 ..
-------
TABLE 6
Regional Impacts of the Transportation Control Plan
Dally Person Trips
Daily Person Miles
Daily Travel Cost
(100' a of 1973
dollars)
Daily Travel Time
•(hours) ,
Transit
Base
130,000
459,000
200
68,000
Plan
204,000
678,000
400
110,000
Percent of
Difference
57.1
47.8
61.9
62.5
Automobile Driver
; Base
3,563,000
24,786,000
Plan
2,912,000
21,716,000,
Percent of
Difference
-18.3
-12.4
Automobile Passenger
Base
1,288,000
8,866,000
Plan
1,524,000
11,380,000
Percent of
Difference
18.1
28.4
Total For Auto Person
21,000
1,170,000
48,000
1,165,000
132.8
-.4
-
Total
Base
4,980,000
34,111,000
21,000
1,238,000
Plan
4,637,000
33,773,000.
48,000
1,276,000
Percent of
Difference
-6.9
-1.0
131.9
3.1
N>
-------
• increases in carpooling transportation rates
r trips being foregone
. diversion of automobile travel to transit travel
. the cost of daily travel in the San Diego region increases
by about 132 percent at the same time as the number of trips
and person-miles of travel are decreasing by about 7 and
1 percent respectively; and
. the total travel time for the San Diego region marginally
increases—about 3 percent—although person trips and person-
miles of travel are decreasing.
An understanding of why a reduction of about 18 percent in the
number of trips results in a reduction of only about 12 percent in the
VMT can be developed from a consideration of the average trip lengths
for automobile driver trips (see Table 7.) Because of the additional
travel distance required to pick-up and drop-off the carpool passengers,
the average trip length for an automobile driver work trip is 11 percent
greater under the plan condition than under the base condition. The
average trip length for non-work trips also increases under the plan
condition but for a different reason. The relative increase in the cost
of an automobile trip resulting from the parking surcharge decreases as
the length of the trip increases and for this reason, it might be
expected that the relative proportion of shorter trips which would be
foregone would be greater than the relative proportion of longer trips
which would be foregone. Thus, it is not surprising that the average
trip length for the foregone trips is about 6.3 miles as compared to an
average trip length under the base condition for all non-work trips of
6.8 miles. Thus, the average trip length for those non-work automobile
trips which are not foregone is longer than the corresponding trip
length under the base condition. Inasmuch as the average trip length
for those trips which continue to be made increases, the VMT reduction
resulting from the 1870 reduction in the volume of trips is smaller than
might have otherwise been expected.
An assessment of the factors resulting in the reductions in auto-
mobile travel is presented in Table 8. The diversion of the automobile
driver trip to a carpool represents the single largest cause of the re-
duction in automobile travel—as defined in terms of both person trips
and VMI. Because of the increase in automobile travel distance asso-
ciated with carpooling, this factor has a greater influence upon the
reduction in person trips than it does on the relative reduction of
VMT. The foregoing of the trip entirely is the second largest factor
influencing the reduction in automobile travel. Whereas only about
35 percent of the reduction in person trips were -foregone trips, about
46 percent of the reduction in VMT were foregone trips. Finally,
transit accounted for about 7 percent of the reduction in automobile
43
-------
TABLE 7
AVERAGE TRIP LENGTH FOR
AUTOMOBILE TRIPS
Work Trips
Non Work Trips
Total Trips
Foregone Trips
Base
(Miles)
7.14
6.81
6.96
Plan
(Miles)
7.93
7.13
7.46
6.27
Percent
Difference
11.1
4.7
7.2
-
44
-------
TABLE 8
REDUCTION IN AUTOMOBILE TRAVEL
FACTORS
CAUSING
REDUCTION
IN AUTOMOBILE
TRAVEL
Trip Foregone
Trip diverted to
transit
Trip diverted to
car pool
Total
AUTOMOBILE
DRIVER
TRIPS
NUMBER
227,000
48,700
-374,900
650,600
PERCENT
OF TOTAL
REDUCTION
34.9
7.5
57.6
100.0
VEHICLE
MILES
NUMBER
1,423,000
144,000
1,504,000
3,071,000
PERCENT
OF TOTAL
REDUCTION
46.3
4.7
49.0
100.0
45'
-------
driver trips and about 5 percent of the reduction in VMT. The average
transit trip length is only about 3.3 miles—less than half of the
average trip length for automobile trips.
Although the volume of trips decreases by about 7 percent and the
person miles of travel decreases by about 1 percent, the total travel
costs for residents of the San Diego Region increases by about 132 percent
under the plan condition. This cost increase is largely a result of
the parking surcharge. As might be expected, the relative increase in
the transit cost is essentially the same as the relative increase in
the number of transit trips. It should be recognized that the cost
increase for automobile occurs in spite of the fact that the number of
trips decreases by about 7 percent, and the number of person-miles
decreases by about 1 percent.
The total travel time for residents of the San Diego region
increases by about 3 percent under the plan condition. The total travel
time for automobile passengers essentially remains constant, although
it should be recognized that the total number of automobile trips will
decrease by approximately 417,000. The total travel time for transit
travelers increases in approximately the same relative proportion as
the increase in the total number of transit trips.
Incidence of the Impacts Upon Various Income Groups
It was recognized at the outset of this study that although the
overall effects of the transportation control plan might be acceptable,
the plan might have unacceptable impacts upon specific subgroups of
residents of the San Diego region—as defined in terms of their incomes
and their geographical residence within the region. The incidence of
the impacts with respect to residents with different incomes are
summarized in Tables 9A through 9D. Three income ranges were util-
ized in this analysis representing low, medium, and high incomes, re-
spectively. It should be recognized that these income ranges are ex-
pressed in 1966 dollars and that they would be about 26.5 percent
greater were they expressed in 1973 dollars.
The relative decrease in the number of automobile driver trips de-
creases as the household income increases with a maximum decrease of
26 percent for the low income range and about 15 percent for the high
income range. This result is as might be expected inasmuch as house-
holds with greater incomes would have a greater capacity to pay the
significantly increased costs of automobile usage. Similarly, the
relative increase in the number of automobile passengers decreases as
the household income increases from a high of about 35 percent for the
low income range to a low of about 10 percent for the high income
range. In contrast, however, the relative increase in the number of
transit trips increases as the household income increases from a low
of about 14 percent for the lowest .income range to a high of about
46
-------
A •
TABLE 9A
INCOME INCIDENCE OF THE IMPACTS
PERSON TRIPS
Household
Income
(1966 dollars)
0,000-4,999
5,000-9,999
10,000 +
Total
Transit
Base
22,000
62,000
46,000
130,000
Plan
25,000
97,000
82,000
204,000
Percent
Difference
13.6
56.5
78.4
57.1
Automobile Driver
Base
216,000
1,496,000
1,850,000
3,563,000
Plan
159,000
1,176,000
1,576,000
2,912,000
Percent
Difference
-26.3
• -21.4
-14.8
-18.3
Automobile Passenger
Base
83,000
547,000
658,000
1,288,000
Plan
112,000
683,000
726,000
1,521,000
Percent
Difference
34.9
24.9
10.4
18.1
Total
Base
321,000
2,105,000
2,554,000
4,980,000
Plan
296,000
1,956,000
2,385,000
4,637,000
Percent
Difference
-7.8
-7.1
-6.7
-6.9
-------
TABLE 9B
PERSON MILES
Household
Income
(1966 dollars)
0,000-4,999
5,000-9,999
10,000 +
Total
Transit
Base
67,000
211,000
181,000
459,000
Plan
67,000
309,000
301,000
678,000
Percent
Difference
1.0
46.8
66.2
47.8
Automobile Driver
Base
1,339,000
10,282,000
13,166,000
24,786,000
Plan
1,061,000
8,704,000
11,951,000
21,716,000
Percent
Difference
-20.8
-15.4
-9.2
-12.4
Automobile Pasoenger
Base
508,000
3,700,000
4,659,000'
8,866,000
• Plan
784,000
5,111,000
5,485,000
11,380,000
Percent
Difference
54.4
38.1
17.7
28.4
Total
Base
1,913,000
14,192,000
18,006,000
34,111,000
Plan
1,912,000
14,124,000
17,737,000
13,773,000
Percent
Hfference
-.1
-.5
-1.5
-1.0
00
-------
TABLE 9C
TRAVEL COST
($000, 1973)
Household
Income
(1966 dollars)
0,000-4,999
5,000-9,999
10,000 +
Total
Transit Automobile
Base
42
127
94
265
Plan
52
205
172
429
Percent
Difference
23.8
61.4
"83.0
61.9
Base
1,137
8,534
10,876
20,550
Plan
2,532
19,247
26,072
47,850
Percent
Difference
122.7
125.5
139.7
132.9
Total
Base
1,179
8,661
10,970
20,815
Plan
2,584
19,452
26,244
48,279
Percent
Difference
119.2
124.6
139.2
131.9
VD
-------
TABLE 9D
TRAVEL TIME
(Hours)
Household
Income
(1966 Dollars)
0,000-4,999
5,000-9,999
10,000 +
Total
Transit
Base
10,000
31,000
26,000
68,000
Plan
12,000
51,000
47,000
110,000
Percent
Difference
15.9
61.6
82.1
62.5
Automobile
Base
66,000
488,000
616,000
1,170,000
Plan
68,000
490,000
608,000
1,165,000
Percent
Difference
2.1
.5
-1.3
- .4
Total
Base
77,000
519,000
642,000
1,238,000
Plan
80,000
541,000
655,000
1,276,000
Percent
Difference
4.0
4.2
2.0
3.1
1/1
o
-------
78 percent for the highest income range. The relationship between auto-
mobile car pooling and transit usage is evident from these results;
essentially, the lower income travelers are tending to carpool relatively
more and use transit relatively less, whereas the higher income travlers
are carpooling relatively less and using transit relatively more.
The relative changes in the person-miles of travel essentially
correspond to the relative changes in the number of person trips. For
each income group, the relative decrease in the VMT is less than the
corresponding relative decrease in the number of trips for that income
group. Similarly, the increase in the person-miles of travel for auto-
mobile passengers is relatively greater for each income group than the
corresponding increase in the number of automobile passengers—this is
a result of the increase in trip lengths which are required to suc-
cessfully accomplish carpooling. For each income group, the relative
increase in the person-miles of transit travel is less than the corres-
ponding increase in the number of transit trips; this suggests that the
trips which are diverted to transit have a relatively short trip length.
The relative increase in the travel costs increases with increasing
household income from a low of about 119 percent for the low income
group to a high of 139 percent for the high income group. In a relative
sense, then, those residents with the greatest incomes have the greatest
relative increases in their travel costs. On the other hand, the in-
creases in travel costs are so great that their absolute impact on
residents in all income ranges cannot be ignored.
The relative increase in the total travel time for the low and
medium income groups is about 4 percent, whereas the relative increase
for the high income group is only about 2 percent. This is probably a
consequence of the lower rate of carpooling which occurs among the high
income group. Though the increase in total travel time is relatively
small—about 3 percent—it should be recognized that the total number
of trips has decreased by approximately 7 percent.
Geographic Incidence of the Impacts
The San Diego region was subdivided into six major statistical
areas (MSAs) to assess the geographical incidence of the impacts. A
map describing the MSAs is presented in Exhibit 2. MSA 0 essential re-
presents the City of San Diego and certain close-in suburbs. MSAs 1,
2, and 3 contain essentially all of the remaining suburbs of the City
of San Diego. Portions of the MSAs 1 and 3 are located over 30 miles
from the central business district of the City of San Diego. MSA 4
contains several urbanized areas which are in the North County and are
generally beyond commuting range from the CBD of San Diego. Neither
San Diego Transit Corporation nor the Oceanside Transit System operates
bus service from MSA 4 to MSA 0. MSA 5 contains essentially undeveloped
desert and no transit service is available.
51
-------
EXHIBIT 2
SAN DIEGO MAJOR STATISTICAL AREAS
52
-------
The impacts of the transportation control plan with respect to
each of the MSAs is presented in Tables 10A through 10D. Essentially
all of the MSAs experience the same relative change in the number of
automobile driver and automobile passenger trips. Two of the three
suburban MSAs—MSAs 1 and 3—experience a rate of increase of transit
trips approximately twice that experienced by the other MSAs. Overall,
however, residents of MSA 0 have the greatest number of transit trips
of any of the MSAs under both the base and the plan conditions.
All of the MSAs experience approximately the same relative decrease
in the VMT and the same relative increase in automobile passenger miles
of travel. MSA 4 has the greatest relative^Jn the person-miles or" $
transit travel reflecting longer average transit trip lengths. Con- \
sidering the total person-miles of travel as in auto transit passenger,
in automobile driver, and in automobile passenger, it is interesting to
note residents of MSAs 0 and 1 experience decreases in their person-
miles of about 19 and 13 percent, respectively, whereas the decreases
in the person-miles of travel in the other MSAs is essentially negli-
gible. Thus, the decrease in the person-miles of travel of residents of
the most densely settled portion of the San Diego region is significantly
greater than the decrease for the residents of the less densely settled
portions.
Considering the relative changes in the travel costs, it is evident
that the greater the density of the area, the greater the relative in-
crease in the travel costs of residents. The central area—MSA 0—
experiences a relative increase in travel cost of about 156 percent,
whereas the San Diego suburbs—MSA's 1, 2, and 3—experience relative
increases of about"130 percent and the other MSA's experience even
smaller relative increases. Thus, residents of MSA 0—essentially the
City of San Diego and National City—experience the greatest relative
increase in the travel costs at the same time as they are experiencing
the greatest relative decrease in the person-miles of travel.
The relative change in the average travel time of trip of each MSA
is presented in Table 10E. The average travel time for transit trips
is approximately twice the travel time for the automobile trips under
both the base and plan conditions. The average travel times for
residents of MSAs 0 and 2 increase about twice as much as the times for
residents of the other MSAs.
Travel Times and Travel Costs Per Trip
The preceding discussion focused upon the total time and the total
costs for all of the trips in a category. Thus, if the cost per trip
increased by 50 percent but the number of trips in that category
decreased by 33 1/3 percent, the total cost of travel would not change.
This section focuses upon the average travel time and the average travel
cost per trip under the base and plan conditions.
53
-------
TABLE IDA
PERSON TRIPS
MSA
0
1
2
3
;A ,
'5 .
Total
Transit
Base
70,000
21,000
23,000
11,000
5,000
1 0
130,000
Plan
99,000
40,000
35,000
22,000
8,000
0
204,000
Percent
Difference
41.2
92.0
52.3
108.7
48.9
-
57.1
Automobile Driver
Base
1,051,000
889,000
353,000
645,000
624,000
200
3,563,000
Plan
847,000
734,000
285,000
532,000
514,000
200
2,912,000
Percent
Difference
-19.4
-17.4
-19.4
-17.5
-17.6
-9.2
-18.3
Automobile Passenger
Base
390,000
323,000
126,000
229,000
221,000
100
1.288.000
Plan
451,000
375,000
150,000
273,000
272,000
100
1.521,000
Percent
Difference
15.8
16.1
19.6
19.2
23.2
4".3
18.1
Total
Base
1,510,000
1,233,000
502,000
885,000
850,000
300
4.980,000
Plan
1,397,000
1,149,000
470,000
827,000
794,000
300
4,637,000
Percent
Difference
-7.5 •
-6.8
-6.4
-6.5
-6.6
-7.6
-6.9
Ul
-------
TABLE 10B
PERSON MILES
MSA
0
1
2
3
4
5
Total
Transit
Base
197,000
104,000
83,000
69,000
6,000
0
489,000
Plan
260,000
168,000
124,000
115,000
10,000
0
678,000
Percent
Difference
31.8
61.9
50.1
67.4
73.0
-
47.8
Automobile Driver
Base
5,723,000
6,213,000
2,452,000
4,788,000
5,607,000
4,000
24,786,000
Plan
4,962,000
5,495,000
2,116,000
4,231,000
4,932,000
4,000
21,716,000
Percent
Difference
-13.3
-11.6
-13.7
-11.6
-12.0
- 2.7
-12.4
Automobile Passenger
Base
2,120,000
2,268,000
855,000
1,698,000
1,921,000
2,000
8,866.000
Plan
2,662,000
2,813,000
1,123,000
2,176,000
2,613,000
2,000
11,380,000
Percent
Difference
25.5
24.0
31.2
28.1
36.0
.2
28.4
Total
Base
8,040,000
8,585,000
3,390,000
6,555,000
7,534,000
6,000
34.111,000
Plan
7,884,000
8,476,000
3,364,000
6,522,000
7,556,000
6,000
33,773,000
Percent
Difference
-19.4
-12.8
- .9
- .5
.3
- 1.8
- 1.0
\J\
Ui
-------
TABLE IOC
TRAVEL COST
(100's of 1973 dollars)
Transit Automobile Total
MSA
0
1
2
3
4
5
Total
Base
147
41
46
20
10
0
265
Percent
Plan
210
83
74
45
15
0
429
Difference
42.9
102.4
60.9
125.0
50.0
-
61.9
Base
5,035
5,248
1,983
3,886
4,393
3
20,550
Percent
Plan
13,044
12,174
4,629
8,889
9,129
4
47,850
Difference
159.1
132.0
133.4
128.7
107.8
33.3
132.9
Base
5,182
5,289
2,029
3,906
4,403
3
20,815
Plan
13,254
12,257
4,703
8,934
9,144
4
48,279
Percent
Difference
155.8
131.8
131.8
128.7
107.7
33.3
131.9
-------
TABLE 10D
TRAVEL TIME
(Hours)
Transit Automobile Total
MSA
0
1
2
3
4
5
Total
Base
34,157
13,483
10,696
7,893
1,598
0
67,795
Percent
Plan
50,367
24,868
16,998
15,218
2,742
0
110,194
Difference
47.5
84.4
58.9
92.8
71.6
-
62.5
Base
304,459
292,530
113,651
226,359
232,908
152
L, 170, 060
Percent
Plan
299,211
288,943
112,483
225,011
235,416
149
1,165,434
Difference
-1.7
-1.2
1.0
.6
1.1
-2.0
-.4
Base
338,616
306,013
124,347
234,252
234,506
152
1,237,855
Plan
349,578
313,811
129,481
240,229
238,158
149
1,275,628
Percent
Difference
3.2
2.6
4.1
2.6
1.6
-2.0
3.1
-------
TABLE 10E
AVERAGE TRAVEL TIME
PER TRIP (MINUTES)
MSA
0
1
2
3
, .4
5
Total
Transit
Base
29
39
28
45
19
0
31
Plan
131
37
29
41
21
0
32
Percent
Difference
6.9
-5.1
3.6
-8.9
10.5
-
3.2
Automobile
Base
13
14
14
16
17
32
14
Plan
14
16
16
17
18
34
16
Percent
Difference
7.7
14.3
14.3
6.3
5.9
6.3
14.3
Total
Base
13
15
15
16
17
32
15
Plan
15
16
17
17
18
34
17
Percent
Difference
15.4
6.7
13.3
6.3
5.9
6.3
13.3
oo
-------
The distribution of. average travel times per trip for automobile
and transit trips under the base and plan conditions is presented in
Exhibit 3. The distributions of travel time for the base and plan
conditions for each of the modes are relatively similar. Generally,
the proportions of trips under the plan conditions is smaller for
lower ranges of travel time (up to about 15 to 20 minutes) and greater
for travel times of 20 to 25 minutes or more. With automobile trips,
this reflects an increase in the travel times for the average trip
resulting from an increase in the travel distance of the average trip
(the reasons for which were noted above). This increase in travel
times is more difficult to explain for transit trips inasmuch as the
average trip length for transit trips in the plan condition (about
3.3 miles) is smaller than the average trip length for transit trips
under the base condition (3.5 miles). These results would suggest
that those passengers who were provided with the best transit service
from a travel time perspective were already using transit in the base
condition, and that the travelers who were diverted to transit had a
relatively longer travel time (in spite of their relatively shorter
transit travel distance). The diversion of the travelers to transit
occurred, of course, as a result of the significant increase in the
cost of automobile travel.
The corresponding travel cost distribution for automobile vehicle
trips under the base and plan conditions is presented in Exhibit 4.
Inasmuch as the same transit fare was used under the base and the plan
conditions, and this fare provides a constant 25c fare for transit
travel throughout the San Diego region, it was not necessary to present
the travel cost distribution for transit. The significant increase in
the cost of an automobile trip that occurs under the transportation
control plan is evident. Under the base condition, relatively few
automobile travelers in the San Diego region paid parking charges for
either their work on their non-work trips. Hence, the travel cost
distribution for the base period essentially reflects the distance
related component of the trip, and the travel cost distributions for
the base period decays approximately as the travel distance distribution
would decay. Under the plan condition, there are no automobile trips
with an out-of-pocket cost of 50
-------
EXHIBIT 3
TRAVEL TIME DISTRIBUTION FOR TOTAL AUTO AND TRANSIT
PERSON TRIPS FOR BASE AND PLAN CONDITIONS
0-5 5-10 10-15 15-20 20-25 25-30 30-35 35-40 40-45 45-50 50-55 55-60 60+
TRAVEL TIME, (MINUTES)
60
-------
EXHIBIT 4
TRAVEL COST DISTRIBUTION FOR TOTAL AUTO
VEHICLE TRIPS BASE AND PLAN CONDITIONS
100.00
50
1.00
1.50 2.00 2.50 3.00 3.50 4.00 450
TRAVEL COST (1966 DOLLARS)
550 6.00 6.50+
61
-------
Thus, the most effective approaches for reducing VMT were increases in
carpooling or foregoing the trip entirely, and these factors had a
relatively equal impact upon the VMT reduction. While transit travel
increased by 60 percent between the base and the plan conditions, the
diversion for transit accounted for only a relatively small proportion
of the VMT reduction.
In the short-run, there are other modifications of travel behavior
which a traveler might make in response to a transportation control
plan. In particular, travelers might choose alternative destinations
for both their work and their non-work trips which were closer to their
residences, thereby reducing the average travel distance per trip. The
EPA promulgated transportation control plan contains little if any
incentive to achieve this objective, however, inasmuch as the parking
surcharge is independent of the distance of the trip.
62
-------
DEVELOPMENT OF PROGRAM COSTS
As outlined in Section I of this report, direct program costs were
obtained for program strategies and measures and distributed among the
five groups:
. local government (including San Diego County AQCR, and
city of San Diego organizations);
. state government (including California state financing
policies and organizations);
. federal government (including federal granting programs,
financing policies, and federal regional activities);
. industry (relating primarily to local San Diego and
California industry, all inclusive of business and
commerce); and
. citizen (includes citizens and consumers in this group,
and is generally referred to in the case of all employ-
ment impact discussions).
This portion of the report discusses the nature of and basis for
cost data obtained for each of the measures, and the general methodolo-
gies utilized to assess, categorize, and apply these costs to each of
the groups outlined above.
Technical Basis for Costs
Prior technical studies had resulted in a basis for the development
of substantial cost data by EPA and state and local organizations. The
Region IX studies had concluded that in the case of the San Diego AQCR:
. The National Ambient Air Quality Standard for photochemical
oxidants had been exceeded.
. The nitrogen dioxide ambient air quality standard had not
been exceeded. Furthermore, it was anticipated that this
standard would not be exceeded in the future.
. The carbon monoxide standard had been exceeded. However,
control measures could be taken to reduce photochemical
oxidants through reactive hydrocarbon (RHC) reductions to
achieve a carbon monoxide level below the standard.
The State of California's Implementation Plan, Revision 3, submitted
July 28, 1973, for the San Diego air basin concluded that:
63
-------
. The photochemical oxidant standard could not be met by 1975
and would require additional control strategies to meet the
standard by 1977.
. The nitrogen dioxide standard had not been exceeded in 1970
and was not projected to be exceeded through 1980.
. The carbon monoxide standard would be exceeded in 1975, but
could be met by 1977.
EPA's conclusions were that implementation of the state plan would
result in much of the required reduction of oxidants, but would not be
sufficient to meet the standards. The Administrator of EPA granted the
allowable extension of the deadline under Section 110(e) of the Clean
Air Act Amendments of 1970 from 1975 to 1977. The EPA plan requires
compliance for most emission sources other than motor vehicles by
June 1975.
An initial observation made on the distribution of socioeconomic
impacts throughout the AQCR was that the majority of the population
that would bear the costs would benefit directly from the improvements
in air quality. This fact is contrasted to other regions where air
pollutants are carried across regional borders. San Diego is affected,
however, by the inflow of uncontrolled vehicles from Mexico.
For this study, costs were identified and separated into implemen-
tation costs and recurring costs estimated for 1977. Attempts were made
to obtain longer range costs to achieve the objectives of assessing the
duration, the cost-effectiveness, and the degree of impact of each of
the various strategies. It was not possible to develop specific post-
1977 cost projections from available sources. For consistency, 1973
dollars have been used throughout the study.
Implementation costs were not annualized because the majority of
the costs would be one-time costs. A series of annualized approaches
utilizing various assumptions have been used in past studies. However,
policies for determining direct or annualized bases for taxing the
population and methods of payment await action by the State of California
legislature and may take entirely different formats than any of the
study approaches that have been suggested.
The totals for implementation costs to 1977 are shown by impacted
organizational segments in Table 11. Similarly recurring costs during
1977 are totalled in Table 12. Each of the component strategy-cost
areas are detailed in the following discussions.
64
-------
TABLE 11
TOTAL IMPLEMENTATION COSTS BY 1977
($000)
Cn
Source Impact
Stationary
Mobile
VTM
Other
Total Costs
Local
Gov't.
79
-
4,925
-
5,004
State
Gov't.
78
10,000
337
2,000
12,415
Federal
Gov ' t .
77
-
338
8,000
8,415
Industry
10,610
-
-
-
10,610
Citizen
-
44,250
-
-
44,250
Total
10,844
54,250
5,600
10,000
80,694
-------
TABLE 12
TOTAL RECURRING COSTS IN 1977
($000)
Source Impact
Stationary
Mobile
VMT
Other
Total Costs
Local
Gov ' t .
40
678
5,440
4,886
11,044
State
Gov ' t .
40
1,695
69
18,379
20,183
Federal
Gov't.
40
678
69
-
787
Industry
424
-
-
-
424
Citizen
-
33,892
351,019
-
384,911
Total
544
36,943
356,597
23,265
417,349
-------
^Implementation Costs
Stationary Sources
The San Diego Air Pollution Control District (APCD) provided data
and cost estimates associated with reactive hydrocarbon reductions. A
study* by the Rand Corp. was utilized to evaluate data resulting from
in-depth investigations. The Rand projections to 1975 are generally con-
sidered comparable to 1977 since the majority of fixed installations
are required to be corrected by 1975 in order to comply with the EPA
promulgated plan and the APCD Rules and Regulations.
In its July 16, 1973, proposed plan, EPA included gasoline vapor
control. However the gasoline strategy was deleted from the final
promulgated plan in view of the advanced status of the program being
implemented by the San Diego APCD. This study incorporates this area
in its scope of work. Estimates of RHC reductions utilized by EPA were
drawn from work performed by the Rand Corp., discussions with the APCD,
and modifications that reflect the EPA reactivity factors for petroleum
plus growth projections. It should also be noted that the Rand 1975
projections had not incorporated EPA proposals in either RHC reduction
projections or cost estimates. Cost figures were provided by the APCD
in the areas of local, state, and Federal Government administration and
for industrial equipment and its installation. The APCD industry imple-
mentation costs per ton of RHC reductions per day is $966 and compares
closely with Rand figures of $950.
Other stationary sources covered in the promulgated plan included:
. degreasing operations;
. metal surface coating thinners and reducers;
. organic solvent usage; and
. dry cleaning solvent vapor losses.
The promulgated plan reduced the requirements previously proposed.
Estimated RHC reductions were provided by the APCD. The Rand study
analyzed the processes and costs in these four areas under a series of
projected requirements. Their nominal strategy conformed to the proj-
ected requirements of the APCD for 1975. Rand's series of reference
strategies (i.e., those technically feasible by 1975) were reviewed for
cost significance.
*DeHaven, W.C., and B. M. Woodfill. Cost and Effectiveness of
Strategies for Reducing Emissions from Fixed Sources: The San Diego
Air Basin, 1975. Rand Corporation WN-8141-SD.
67
-------
The study team's conclusions—that implementation costs in these
four areas appeared to be insignificant when compared to the magnitude
of costs associated with implementing other strategies—follow:
Degreasing, the major area of RHC reduction, has been projected
to require little change in equipment by substituting nonreac-
tive solvents.
. Metal coatings, the second major reduction, also involves
shifting to nonreactive solvents and water-based, powdered,
or high solids coatings. No implementation costs were iden-
tified.
. Organic solvents implementation is to be accomplished by sub-
stitution. Some vapor recovery equipment may be required.
Dry cleaning establishments are difficult to control, and
Rand conluded that solvent substitution did not appear to
be technically or economically feasible. As an alternative,
Rand recommended that activated carbon adsorbers be installed
at an estimated equipment cost of $425,000 for 166 establish-
ments. Solvent recovery values were estimated to permit
equipment amortization within two years. EPA's promulgated
plan addressed apparent technical difficulties, extended com-
pliance time by one year (i.e., to 1975), allowed four percent
RHC content in solvents, and dropped the required emission
reduction from 95 percent to 90 percent. Costs were not
estimated for implementation of this area because of the
difficulties cited and the lack of cost data to comply with
the promulgated plan.
A summary of RHC reductions projected by our sources is shown in
Table 13. The associated direct implementation costs provided by the
APCD are listed in Table 14.
Mobile Sources
Catalytic converter retrofit implementation costs plus costs for
devices required under state programs were developed and supplied by
the Office of Environmental Management, County of San Diego, utilizing
the Motor Vehicle Emission and Cost (MOVEC) model developed by the Rand
Corporation. In addition, projected reductions in RHC were furnished.
Inspection and maintenance implementation costs were developed
by PMM&Co. based upon the projected vehicle population and the need
for loaded dynamometer inspection stations together with estimates
of associated installation costs.
68
-------
TABLE 13
STATIONARY SOURCE REACTIVE HYDROCARBON REDUCTION BY 1977
(TONS/DAY)
Stationary Source
Gasoline Vapor
Degreasing Solvents
Metal Coating
Solvents
Organic Solvents
Dry Cleaning Solvents -
Total
>
EPA
31.9
15.8
47.7
APCD
30.1*
6.9
4.2
1.3
0.6.
43.1
>13.0@
Rand
13.9*
14.6
3.7
0.7.
32.9
>19.0@
*7-day week
@5-day week
69
-------
TABLE 14
•vj
o
IMPLEMENTATION COSTS FOR STATIONARY SOURCE RHC REDUCTION BY 1977
($000)
Stationary
Sources
Gasoline Vapor
Degreasing Solvents
Metal Coating
Solvents
Organic Solvents
Dry Cleaning Solvents
Total Costs
Local
Gov ' t .
51
2
17
2
7
79
State
Gov ' t .
51
2
16
2
7
78
Federal
Gov't.
51
2
16
2
6
77
Industry Citizen
10,610
0
0
0
0
10,610
Total
10,763
6
49
6
20
10,844
-------
Catalytic converter retrofits apply to the 1966-1974 model years,
involving 556,300 vehicles of a projected 1977 total of 902,790. Two
sets of costs were developed, based upon alternate retrofit eligibi-
lities. Universal Oil Products (UOP) recently advised the San Diego
Office of Environmental Management that their analysis indicated that
significantly more vehicles utilizing the same device could run on 91
RON gasoline. The basic costs were more conservative, based upon EPA
estimates of eligibility.
Both sets of costs are illustrated here, but the study team has
elected to use those cost figures based upon the EPA promulgated plan.
Table 15 shows a distribution of vehicles by model year, the percen-
tage of eligible users of devices, and the projected RHC reductions.
Table 16 covers implementation costs for the same alternatives. Costs
for equipment and installation under the state program apply to
657,090 vehicles representing model years 1955-1974.
Inspection and maintenance programs to be conducted under the
EPA promulgated plan involving loaded testing will require inspection
stations equipped to perform test programs of significantly more com-
plexity than those performed under the random sampling program cur-
rently in effect.
Since no cost data have been developed, other than an estimated
range for dynamometers, a set of calculations were made by PMM&Co.
to establish inspection station costs. On the basis of a state-
conducted annual inspection program including in addition to pollution
control equipment, allowance for the potential incorporation of
vehicle safety inspection, the following rationale was established:
. for 900,000 vehicles = 3 600 vehicles per day;
250 days y
. with each vehicle requiring 30 minutes, including
loaded dynamometer test, safety tests, and certi-
fication or rejection:
. during an 8-hour day, 16 vehicles per lane (each
station to be built with 6 lanes, handling 96 vehicles -
rounded to 100);
. yielding 3.600 vehicles/day = 36 stations. And
100
. for each inspection station, 6 lanes wide by 60 feet long
(to accommodate portions of inspection plus other required
facilities) = 3,600 square feet;
71
-------
TABLE 15
PROJECTED MOBILE SOURCE CONTROLS APPLICABILITY AND RHC REDUCTIONS BY 1977
Numbers of Vehicles Affected by Strategies
Model
Year
1975-77
1972-74
t
1971
1966-70
1961-65
1958-60
Total
Number
245,700
239,400
73,800
243,000
74,700
26,190
902,790
Catalyst Eligibility
UOP
95%1
757.J
60%
EPA
75% of
313,200
20%
Crankcase
Blowby
100%
NOX Control
and/or
VSAD
1 100% of
[556,200
J
Pre-1966 Exhaust
Retrofit;
NOX Control
and /or
VSAD
}100% of
100,890
Inspection
Maintenance
L 100% Of
' 902, 790
Reactive Hydrocarbon Reductions - Tons/Day
Rand plus UOP Eligibility 18.6
Rand plus EPA Eligibility
EPA Promulgated Plan
12.0
1
1 19, <; 1
1 I?. ^ I
19.0 1
RHC
Reduction
Totals
31.1
24.5
12.9
-------
TABLE 16
A.
IMPLEMENTATION COSTS TO
RETROFIT VEHICLES BY 1977
($000)
Source
Equipment
Installation
Total
EPA*
24,850
7,360
32,210
ARB@
8,600
3,440
12,040
Total
33,450
10,800
44,250
B.
Source
Equipment
Installation
Total
UOP#
37,150
11,380
48,530
ARB@
8,600
3,440
12,040
Total
45,750
14,820
60,570
^Catalytic retrofit eligibility--75% of 1971-74, 207o of 1966-70.
©California Air Resources Board equipment.
. crankcase blowby--100% of 1955-60;
. pre-1966 exhaust retrofit—NOX control and/or VSAD, 100% of
1955-65; and
. NOX control and/or VSAD, 100% of 1966-70.
^Catalytic retrofit eligibility—95% of 1972-74, 75% of 1971,
60% of 1966-70.
73
-------
. with building at $20 per square foot = $72,000
six dynamometers at $20,000 = 120,000
other equipment = 50,000
land at 15% of above = 36.000
•
TOTAL $278,000
. then 36 stations at $278,000 $10,008,000
Adequate geographic dispersion to accommodate the population of the
Air Quality Control Region both to avoid concentration and to allow
for growth, will require detailed study.
In view of the limiting regulations on motorcycle operations,
which will be modified if a national regulation on RHC emissions is
promulgated, no implementation costs were developed. Hardware costs
incorporated in 1976 and later year models are envisioned along with
associated administrative and inspection/maintenance costs. Table 17
shows the mobile source implementation costs allocated to the impacted
groups.
VMT Reduction
Exclusive Bus/Carpool Lanes. The Comprehensive Planning Organi-
zations (CPO) of the County of San Diego is initiating a pilot study
covering 22 ramp and meter locations including bypasses. This cost
has been shown at $450,000, split between state and federal funds.
Bus/Carpool Matching. A computer matching and promotion system
was estimated by the state/local task force to cost $175,000 for imple-
mentation. The promulgated plan modified the total area approach to
begin with an initial pilot evaluation utilizing selected U. S. Navy
facilities.
Parking Supply. Implementation of the program for the review
of new commercial parking facilities involves an initial review and
approval by Region IX with potential transfer to state and local
authorities. PMM&Co. has included an estimated figure of $50,000 to
cover the development of operational procedures. The implementation
date of this program has been deferred to January 1, 1975. (Federal
Register, January 15, 1974.)
Mass Transit Incentives. Implementation cost assignment for
this employer incentive program did not appear applicable prior to
the revision contained in the Federal Register (January 15, 1974)
which advised that more general regulations will be issued.
74
-------
TABLE 17
Ul
IMPLEMENTATION COSTS FOR MOBILE SOURCES BY 1977
($000)
Mobile Source
Catalyst Retrofit Equipment
ARB Equipment
Inspection/Maintenance Stations
Total Costs
Local
Gov't.
~
-
-
-
State
Gov ' t .
-
-
10,000
10,000
Federal
Govlt.
-
-
-
-
Indus try
-
-
-
-
Citizen
32,210
12,040
-
44,250
Total
32,210
12,040
10,000
54,250
-------
Parking Surcharge. Those parking surcharge implementation costs
that could be identified were supplied by the City of San Diego. Al-
ternatives were evaluated for the estimated 76,000 free non-residen-
tial parking spaces, including parking meters. The capital outlay,
including installation, for parking meters was estimated at $4.8 mil-
lion. Extensive surveys would be required to identify and inventory
all on-street non-residential and all off-street parking areas.
Ajr Quality Monitoring. VMT air quality monitoring was assessed
by the APCD to require six permanent stations plus a number of mobile
installations. Continuous monitoring is to be incorporated in future
plans. No implementation cost data were identified.
Bicycle Lanes. The cost of implementing bicycle lanes was esti-
mated by the San Diego Task Force at a preliminary estimate of $5 mil-
lion which could reduce VMT approximately one percent. The City of
San Diego is evaluating a plan covering 30 bike path proposals at
$125,000. This later figure has been incorporated in our implementa-
tion costs.
Four-Day Week. The four-day week concept has been initiated in
San Diego County for county employees. The program permits adjustment
of working hours so that the required 80 hours are worked during each
two weeks and so that rescheduling does not require additional person-
nel. No implementation costs appeared applicable.
Table 18 is a listing of VMT reduction implementation costs
accrued by 1977.
Other Measures. Aircraft operational impacts have been studied
extensively by Rand and the California Air Resources Board. EPA's pro-
mulgated plan notes that the Agency will continue to assess the feasi-
bility of proposed controls working with the Federal Aviation Adminis-
tration. It appears premature to assign implementation costs at this
stage of the program.
The expansion of mass transit involves increasing the number of
buses in the fleet in addition to allowing for the replacement of re-
tired vehicles. The San Diego Transit Corporation plans to expand its
present fleet of 250 buses to 350 by 1977. It has been estimated by
the Comprehensive Planning Organization (CPO) that to respond to the
EPA promulgated plan, an additional 200 buses will be required in 1977.
San Diego Transit's estimate of $50,000 per bus translates into an
implementation cost of $10 million. Under current law, the Urban Mass
Transit Agency (UMTA) is permitted to fund 80 percent of the capital
cost. The remaining 20% has been allocated to the state for funding
under its S.B. 325 program which is accomplished, through state sales
taxes. In 1972, the sales tax structure was increased 0.25 percent,
and the base was extended to include gasoline taxes. Additional funds
generated by this measure were to be allocated to transportation devel-
opment. In counties with populations greater than 500,000, the funds
76
-------
TABLE 18
VMT REDUCTION IMPLEMENTATION COSTS BY 1977
($000)
VMT Source
Exclusive Bus/
Carpool Lanes
Bus/Carpool
Matching
Parking Supply
Mass Transit
Incentives
Parking Surcharge
VMT Air Quality
Monitoring
Bicycle Lanes
4-Day Week
Total
Local
Gov't.
-
-
-
-
4,800
-
125
-
4,925
State
Gov ' t .
225
87
25
•*
-
-
-
-
337
Federal
Gov't.
225
88
25
-
-
-
-
-
338
Industry
-
-
-
-
-
-
-
-
Citizen
-
-
-
-
-
-
-
-
Total
450
175
50
-
4,800
-
125
-
5,600
-------
were to be spent for the development, maintenance, and operation of
public transportation systems. Thus, $8 million implementation cost
has been allocated to the Federal Government and $2 million to the
state. Cost allocations are shown in Table 19.
Recurring Costs
Recurring cost data, expressed in 1973 dollars, represent those
identified costs projected to occur during 1977 and are based upon
completion of the implementation portion of the strategies prior to
1977.
Stationary source costs were supplied by APCD. A total of $80,000
was assigned to administration and field operations involved in enforce-
ment of APCD Rules and Regulations in effect at that time which would
incorporate state requirements combined with the EPA promulgated plan.
An estimated $424,000 was assigned to the petroleum industry for
permits, equipment maintenance, and operations. It was estimated that
the use of vapor control equipment would recover gasoline vapor equiv-
alent to $1.1 million annually. This savings to the industry at $0.40
per gallon would be equivalent to an additional 2,750,000 gallons of
gasoline.
The significance of projected costs and savings resulting from
vapor recovery and materials substitution in the other four stationary
areas were addressed in the Rand study. In the Rand nominal strategy.
for 1975, available costs and savings data were minor and offsetting,
resulting a net figure approximating zero. Table 20 indicates recurring
costs for stationary sources.
Mobile source recurring costs include conducting an annual inspec-
tion and maintenance program to assure the operational efficiency of
vehicles equipped with the required RHC reduction devices. Estimated
increases in operating costs resulting from additional gasoline con-
sumption were also provided. The costs, provided by the Office of
Environmental Management of the County of San Diego, were developed
utilizing the MOVEC model, incorporating the same series of inputs used
in generating the implementation costs for hardware and installation.
Table 21 compares alternate costs utilizing EPA and UOP eligibilities
for catalytic converters. Costs for California Air Resources Board
equipment for maintenance/operating cost increases are included. No
local estimates were available for loaded dynamometer testing. A figure
of $4.50 per vehicle was obtained from the national average figure
contained in an EPA white paper.*
*Holtnes, J., J. Horowitz, R, Reid, and P. Stolpman. The Clean Air Act
and Transportation Controls, an EPA White Paper. Office of Air and
Water Programs, Environmental Protection Agency. Washington, D.C. :
August 1973.
78
-------
TABLE 19
OTHER REDUCTION IMPLEMENTATION COSTS BY 1977
($000)
Other Sources
Aircraft Ground Operations
Bus Requirements
Total Costs
Local
Gov ' t .
-
-
-
State
Gov ' t .
-
2,000
2,000
Federal
Gov ' t .
-
8,000
8,000
Industry
-
-
-
Citizens
-
-
-
Total
-
10,000
10,000
-------
TABLE 20
RECURRING COSTS FOR STATIONARY SOURCES DURING 1977
($000)
CO
o
Stationary
Source
Gasoline Vapor
Degreasing
Solvents
Metal
Coating
Organic Solvents
Dry Cleaning
Solvents
Total Costs
Local
Gov ' t .
26
1
8
1
4
40
State
Gov't.
26
1
8
1
4
40
Federal
Gov't.
26
1
8
1
4
40
Industry
424
-
-
-
-
424
Citizen
-
-
-
-
-
-
Total
502
3
24
3
12
544
-------
TABLE 21
RECURRING COSTS FOR INSPECTION/MAINTENANCE
PROGRAM AND INCREASED OPERATING COSTS DURING 1977
($000)
A. EPA Eligibility Assumptions
Mobile Source
Inspection
Loaded Mode
Maintenance!?
Operating
Total*
EPA .
4,930
3,930
8,860
ARE
18,750
22,220
20,970
Total
4,062*
23,680
6,150
33,892
B. UOP Eligibility Assumptions
Mobile Source
Inspection
Loaded Mode
Maintenance^
Operating
Total*
EPA
7,260
5,700
12,960
ARE
18,750
2,220
20,970
Total
4,062*
26,010
7,920
37,992
* Based upon 902,790 vehicles at $4.50 per vehicle.
@ OEM, San Diego.
# OEM San Diego at fuel cost of $0.40 per gallon.
81
-------
Administrative costs for government involvement in the program were
estimated by PMM&Co. Two percent of the inspection and maintenance
poritions of the combined total of $33,892,000 was allocated to local
and federal governments and five percent was alloted to the state govern-
ment. All of the above costs are allocated in Table 22.
VMT reduction recurring costs are based upon the EPA promulgated
plan. The parking surcharge cost to local government was supplied by
the City of San Diego and represents an estimate covering the administra-
tion, maintenance, and collection of fees for the additional parking
meters.
The balance of the costs associated with this series of strategies
were developed by PMM&Co. Figures covering exclusive bus/car-pool lanes were
estimated at $2,300 per ramp to cover maintenance and operations based on
data developed from Los Angeles. Bus/car-pool matching costs were based
upon manpower requirements to operate and update the system at $58,000
per year. The parking supply strategy was estimated to incur $30,000
per year in administrative costs.
Mass transit incentive and parking surcharge costs to industry and
citizens were .developed utilizing the PMM&Co. n-dimensional logit model.
This area is discussed in detail in Section II. VMT air quality moni-
toring, bicycle lane maintenance, and four-day work week program impacts
were not assigned recurring costs because of non-availability of quanti-
tative data. Table 23 portrays the allocation of costs among the groups.
Aircraft ground operations were studied by Rand and the California
Air Resources Board. No recurring costs were incorporated for this
potential measure.
The recurring costs for operating buses—reflecting an increase in
operating losses over revenues for a 1977 bus fleet of 550 vehicles,
were developed by PMM&Co. based upon data furnished by the San Diego
Transit Corporation.
Operating costs are estimated to run $70,000 per bus in 1977. This
figure adjusted to 1973 dollars becomes $58,874 or a total for 550 buses
of $32,353,000. Operating revenues, based upon 1973 history of 19.3£
per rider realized from a fare rate of 25c, with deductions for students,
elderly, and monthly rates, projected at 19C per rider to 1977, but
adjusted in 1973 dollars to 16.2c, combined with a projection of 204,000
daily transit trips in 1977,and utilizing a 275 day year as the transit
equivalent of workdays per year, results in a revenue figure of $9,088,000.
The resulting operating deficit becomes $23,265,000. This deficit was
allocated 21 percent to local government and 79 percent to state support
under S.B. 325 funding as shown in Table 24.
82
-------
TABLE 22
RECURRING COSTS FOR MOBILE SOURCES DURING 1977
($000)
Mobile Source
Inspection
Maintenance
Operating
Administrative
Total
Local
Gov ' t .
-
-
-
678
678
State
Gov ' t .
-
-
-
1,695
1,695
Federal
Gov't.
-
-
-
678
678
Industry
-
-
-
*
Citizen
4,062
23,680
6,150
-
33,982
Total
4,062
23,680
6,150
3,051
36,943
O3
U)
-------
TABLE 23
VMT REDUCTION STRATEGY RECURRING COSTS IN 1977
($000)
VMT Strategies
Exclusive Bus/
Carpool Lanes
Bus/Carpool
Matching
Parking Supply
Mass Transit
Incentives
Parking Surcharge
VMT Air Quality
Monitoring
.Bicycle Lanes
4-Day Week
Total
Local
Gov ' t .
-
-
-
-
5,440
-
-
-
5,440
State
Gov't.
25
29
15
-
-
-
-
-
69
Federal
Gov't.
25
29
15
-
-
-
-
-
69
Industry
-
-
-
-
-
-
-
-
-
Citizen
-
-
-
56,626
294,393
-
-
-
351,019
Total
50
58
30
56,626
299,833
-
-
-
356,597
09
•P-
-------
TABLE 24
OTHER RECURRING COSTS DURING 1977
Mobile Sources
Aircraft Ground
Operations
Bus Requirements
Total
Local
Gov ' t .
-
4,882
4,882
State
Gov ' t .
-
18,365
18,365
Federal
Gov't.
-
•
Industry
-
-
Citizen
-
-
Total
\
23.247
23,247
00
Ul
-------
ALLOCATION RATIONALE
Allocation of Costs to Citizens
The previous chapters described the methodologies for obtaining
the estimated costs for each of the measures, and then identifying the
direct cost impact on the five major groups.
This portion of the report further develops the rationale and
describes the methodologies for further allocation of the program costs
from four of the groups to the citizens in such a manner that program
costs can be determined by the various income stratifications.
In addition to describing the techniques applied for distribution
of each group, the program costs for one-time implementation, and the
estimated annual operating costs are shown as a percent of income of
the citizens of San Diego. Four income strata were utilized for
determining the effect on households in the San Diego AQCR. These are:
. under $5,000;
. 5,000 to 10,000;
. 10,000 to 15,000; and
. over $15,000.
Allocation Methodologies
For allocation purposes, it has been assumed that certain costs
will be incurred by local government, state government, and Federal
Government as well as industry in controlling stationary and mobile
pollution sources. Additionally, costs will be incurred in enforcing
regulations to be employed in reducing vehicle miles traveled and
diminishing other air pollution sources.
Ultimately, the taxpayer in the case of government and the user in
the case of industry are called upon to pay the costs of improved living
conditions or advances in technology. In the past, government has gone
to great lengths to soften the economic impact of taxation. The real
property tax is the major source of local government funding while state
and Federal Government rely primarily on income taxes. Industry, for its
part, attempts to pass government imposed taxes directly onto the
consumer in the form of direct and indirect price increases.
There are numerous potential ways of allocating the burden of
government taxation by imposing special taxes on households, automobile
owners, gasoline, energy/utilities, employee payrolls, parking, admis-
sions, sales, corporate incomes, and employers. This report of the
86
-------
impact of direct implementation and recurring costs to San Diego area
residents has utilized present policy and is generally based on real
property taxes in the case of local government costs, personal income
taxes of households (married taxpayers filing a joint return) for both
state and Federal government costs, and increased gas prices at the
pump in the case of industry.
Costs were allocated to households by income group and in some
instances, on the basis of vehicles owned and estimated miles driven
per household. Utilizing the strategies described under Program Costs,
a methodology for allocating source costs at the local, state, Federal,
industry, and citizen levels was developed.
Base line data on population and number of households, vehicle
ownership, and average miles driven per vehicle by various income
groups in the San Diego area was drawn from previously referenced Rand
estimates for 1975 and from the County Office of Environmental Manage-
ment data and cost estimating model (MOVEC). The estimated number of
households by income group in 1975 was:
Income Group Number of Households
Under $5,000 142,500 (287,)
$5,000 - 10,000 154,800 (30%)
$10,000 - 15,000 119,900 (23%)
Over $15,000 97.200 (19%)
Total 514,400 (100%)
On an average annual basis between 1970 and 1975, the San Diego
area is expected to grow at a rate of approximately 9,300 households
containing 2.9 persons. Between 1975 and 1980, approximately 10,000
additional households per year are projected for the study area. The
assumption was made that the percentage distribution of households by
income group would be the same throughout the six years.
Personal consumption expenditure data developed by the U.S. Depart-
ment of Labor/Bureau of Labor Statistics was used as the basis for
establishing estimated taxable incomes per household within each of the
income ranges of $2,600, $7,700, $13,800, and $30,100.
Implementation costs for federal and state sources were then
allocated to each income group using the estimated taxable income per
household as the basis for determining economic impact. Local impact
was estimated based on property taxes and industry impact on miles
driven per household unit with one or more vehicles. Rand data on
estimated user costs was employed in estimating citizen costs.
87
-------
The basis for allocations by major groups follow:
Local Government
The assumption was made, based on statistical experience in other
communities, that 45 percent of total 1977 San Diego households would
own their place of residence according to income group as follows:
Income Group % Owning Home
Under $5,000 - 0 -
$5,000 - 10,000 5%
$10,000 - 15,000 35%
Over $15,000 60%
A further assumption was made that the market value of an individual
home would be approximately 2-1/2 times the estimated taxable income of
a household within each income group as follows:
Average Market Average Property Tax
Income Group Value of Dwelling Per Dwelling
Under $5,000 -0- -0-
$5,000 - 10,000 $19,250 $436
$10,000 - 15,000 34,500 $870
Over 15,000 75,250 $1,900
San Diego property taxes in 1973 were $10.10 per $100 of assessed value,
which represented 25 percent of market value and amounted to the esti-
mated property taxes shown above.
On the assumption that the local portion of stationary implementa-
tion costs are to be allocated on a real property basis, these taxes
per household unit were multiplied by total households within each of
the four income groups to produce an estimated 1977 property tax. Dis-
tribution of the estimated local portion of air pollution stationary
costs was completed on a percentage basis so that total group contribu-
tion would impose the same proportionate burden as real property taxes
by income group. Household costs by income group are expected to be
0.0, 9.2 cents, 18.3 cents, and 40.1 cents.
88
-------
State Government:
Estimated state personal income taxes as filed by married taxpayers
filing a joint return has been selected as the basis for measuring the
impact of these costs on San Diego households. Utilizing average esti-
mated taxable incomes per household multiplied by the estimated number
of 1977 household units within each income group would generate approxi-
mately $340.3 million in total state personal income taxes. Allocating
costs on the same percentage basis as each income group contributes to
state personal income taxes produces the impact per household by income
group.
Federal Government
In this case, estimated federal personal taxable incomes for
married taxpayers filing a joint return were used as the basis for
determining the impact of these costs on San Diego households. Average
estimated federal taxable income by household within each income group
was multiplied by the expected 1977 households to generate a tax of
$1,407.1 million. Following a percentage distribution approach, the
the Federal Government pollution control costs plus individual house-
hold impacts were developed.
Industry
For purposes of determining the impact of an estimated $10.6 mil-
lion in industry stationary implementation costs, PMM&Co. utilized
Office of Environmental Management MOVEC computed estimates that 8,320.2
million miles will be driven in the San Diego area during 1977. Of
this figure, 837o or 6905.8 million miles will be driven by vehicles
owned by households.
Under this assumption, the estimated miles to be driven in 1977 by
households with vehicles in each income group were divided by an average
of 13.22 miles per gallon to produce the 522.4 million gallons of gas
consumed by household vehicles in 1977. These estimated gallons distri-
buted by income group for households with vehicles were then multiplied
by 1.7
-------
Est. Miles Driven
Income Per Household With
Group Vehicles 1977
Under $5,000 13,045
$5,000 - 10,000 17,132
$10,000 - 15,000 20,975
Over $15,000 23,869
Average Gallons
of Gas Required
Per Household
With Vehicle
986
1295
1586
1806
Average Estimated
Cost Per Household
With Vehicle
$13.91
18.27
22.38
25.48
The rationale behind these adjustments rests on the fact that
statistical data, taken from the previously referenced Rand studies,
indicates that only 750,000 (83%) of the area's estimated total of
900,000 vehicles were allocated to San Diego area households by income
group.
Therefore, in the case of estimated industry stationary implementa-
tion costs of $10.6 million, only 83 percent ($8.8 million) was distri-
buted among households by income group. The remaining $1.8 million
would have to be allocated to the remaining 150,000 non-household vehi-
cles (e.g., commercial) at an average cost of $12.02. Similarly,
industry stationary recurring costs of $424,000 were distributed on the
same 83 percent basis, and the remaining $72,000 are allocated to the
150,000 non-household vehicles (e.g., commercial) at an average cost of
48 cents.
Citizen/Consumer
In distributing estimated implementation and recurring costs for
mobile strategies the Rand statistical data reflecting estimated "user"
costs was employed.
Again, the costs utilized represented 83 percent of the total costs
involved, and therefore, the remaining $7.4 million in mobile implementa-
tion cost must be allocated to the 150,000 non-household vehicles at an
average cost of $49.87 per vehicle. Similarly, mobile recurring costs of
approximately $5.0 million would have to be allocated to 150,000 non-
household vehicles at an average cost of $33.61 per vehicle.
Implementation Costs
Table 25 portrays the allocation of estimated implementation costs.
Expected implementation costs by 1977 of $80.7 million were identified.
Of this amount, local government would pay $5.0 million, state govern-
ment $12.4 million, Federal Government $8.4 million, industry $10.6 mil-
lion, and the citizen owning vehicles $44.3 million.
90
-------
TABLE 25
ESTIMATED IMPLEMENTATION COSTS FOR AIR
POLLUTION CONTROL IN SAN DIEGO AREA BY 1977
INCOME GROUP
Under $5,000
I - Stationary
II - Mobile
III - VMT
IV - Other
$5,000-510,000
I - Stationary
II - Mobile
III - VMT
IV - Other
$10,000-$15,000
I - Stationary
II - Mobile
III - VMT
IV - Other
OVer $15,000
i I - Stationary
II - Mobile
III - VMT
IV - Other
LOCAL
$ 0
0
.09
5.78
.18
11.55
—
.40
25.30
STATE
$ 0
.74
.03
.15
.04
5.16
.17
1.03
.13
17.10
.58
3.42
.54
68.61
2.30
13.72
FEDERAL
$ .02
.09
.59
.07
.32
4.12
.15
.65
13.68
.44
1.89
54.88
INDUSTRY (1)
(Household Only)
$13.91
—
18.27
—
22.38
—
—
25.48
CITIZEN (2)
(Household Only)
$ ~
70.83
—
80.62
__
89.11
—
—
90.37
TOTAL
$ 13.93
71.57
.12
.74
$ 86.36
18.47
85.78
6.27
5.15
$ 115.67
22.84
106.21
12.78
17.10
$ 158.93
26.86
158.98
29.49
68.60
$ 283.93
(1) Non-household industry costs of $1.8 million would be allocated to 150,000 non-household vehicles at an average cost of $12.02 per vehicle.
(2) Non-household citizen costs of $7.4 million would be allocated to 150,000 non-household vehicles at an average cost of $49.87 per vehicle.
-------
Table 25 reflects the expected economic impact of implementation on
the average household in each of four income categories. In the case of
households earning less than $5,000, air pollution implementation costs
have been estimated at 86.36 by 1977 with major cost being the $70.83
charge for mobile costs assigned to households. Figure 2 depicts the
estimated impact of implementation on households. Implementation costs
then range upward for the average household as income increases as
follows:
Household Total Implementation Cost °L of Net Taxable Income
Under $5,000 $ 86.30 3.3
$5,000 - 10,000 115.61 1.5
$10,000 - 15,000 158.93 1.1
Over $15,000 283.93 0.009
Implementation costs for local, state, and federal government were
estimated at $79,000, $78,000, and $78,000 respectively as shown in
Table 11 and are expected to involve very minor charges to households in
income groups below $15,000. For household earning over $15,000, the
allocation of mobile costs for required state inspection stations, and
the federal charge of $8.0 million for new buses under Other Costs
appear substantial. Industry costs of $10.6 million by 1977 for vapor
control equipment were distributed on the basis of estimated miles
driven, as described earlier in this section, and result in a more equal
allocation of these costs against households with vehicles within each
income group. Citizen costs of $44.2 million for retrofit hardware and
installation are expected to exert the major implementation cost impact
on households with the most number of vehicles, because the cost differ-
ences by income group in the Citizen column Table 25 reflect the slight
difference in average number of vehicles owned per household. For pur-
poses of this impact analysis, only 83 percent of the estimated $10.6
and $44.3 million in industry and citizen implementation costs were
distributed among households. The remaining $1.8 and $7.4 million in
these categories were allocated to 150,000 non-household vehicles at
an average cost of $12.02 and $49.87 per vehicle.
In summary, the $44.3 million cost of retrofit hardware and instal-
lation is expected to be the single largest outlay of any implementation
cost. It accounts for 82 percent of the total cost for the average
households earning less than $5,000, but only 32 percent of the much
92
-------
VO
OJ
DOLLAR
OUTLAY
$300
250
200
150
100
1 50
0
(IN 1973 DOLLARS)
DOLLAR
OUTLAY
PERCENT DISTRIBUTION
I
I
I
I
$5,000
$10,000 $15,000 $20,000
ANNUAL INCOME PER HOUSEHOLD
$25,000
$30,000
PERCENT
OF
INCOME
5%
4
3
2
1
0
FIGURE 2: ESTIMATED IMPACT OF IMPLEMENTATION COSTS ON SAN DIEGO
HOUSEHOLDS AT TAXABLE INCOMES ACCUMULATED THROUGH 1976
-------
larger total implementation cost allocated to household earning over
$15,000 annually. To the extent that households in income categories
below $10,000 and particularly those under $5,000, could be attracted
to mass transit and thereby reduce the current average of 1.3 and 1.6
vehicles per household, the remaining dollar impact of implementation
costs on these households of $15.53 and 35.05 respectively would be
minimized.
Recurring Costs^
As previously indicated, recurring costs are those of a non-
capital nature that are projected to recur beginning with 1977. They
have been estimated at $417.3 million, with local and state government
each incurring over $11 and $20 million respectively,Federal Government
approximately $800,000, industry $424,000, and San Diego citizens the
major share $385 million. Table 26 indicates the anticipated impact of
1977 recurring costs on an average household in each of four income
categories:
Households Total Recurring Costs % of Net Taxable Income
Under $5,000 $ 598.59 23.0
$5,000 - 10,000 802.74 10.4
$10,000 - 15,000 1,011.98 7.4
Over $15,000 1,276.56 4.2
As shown in Figure 3, the most significant economic impact of
recurring costs during 1977 would be on the Citizen group as a result
of the November 12 promulgated surcharge program.
Including the surcharge program,annual recurring costs of between
$598 (under $5,000), and $1,276, (over $15,000) would result. For
example, within the first income category, taxable income per household
was estimated at $2,600. For the average household earning under
$5,000 recurring Citizen/VMT costs of $552 alone would amount to 21 per-
cent of total taxable income. For households in the over $15,000 income
range in which average taxable income is expected to approach $30,100 by
1977, this same recurring cost would amount to between three percent and
four percent of average total taxable income. While it is conceivable
that many households earning over $15,000 annually would continue to
consider an automobile essential, it seems likely that the second car
might be eliminated by all but the very high income households. For
94
-------
TABLE 26
ESTIMATED RECURRING COSTS FOR AIR POLLUTION
CONTROL IN SAN DIEGO AREA DURING 1977
vo
(M
INCOME GROUP
Under $5,000
I - Stationary
II - Mobile
III - VMT
IV - Other
$5, 000- $10, 000
I - Stationary
II - Mobile
III - VMT
IV - Other
$10,000-$15,000
I - Stationary
II - Mobile
III - VMT
IV - Other
Over $15,000
I - Stationary
Hi- Mobile
III - VMT
IV - Other
LOCAL
0
0
0
0
.05
.80
6.39
5.74
.09
1.59
12.76
11.46
.21
3.48
27.95
25.11
STATE
0
.13
0
1.36
.02
.87
.04
9.48
.07
2.90
.12 •
31.44
.27
11.63
.47
126.10
FEDERAL
.01
.19
.02
—
.04
.64
.06
—
.08
1.31
.13
—
.22
3.81
.39
—
INDUSTRY (1)
(Household Only)
.52
__
--
—
.69
—
--
—
.85
—
—
—
.98
«
..
__
CITIZEN (2)
(Household Only)
—
44.20
552.16
—
52.96
725.20
—
—
61.02
888.16
—
—
65.14
1,010.80
—
TOTAL
.53
44.52
552.18
1.36
598.59
,80
55.27
731.69
802! 98
1.09
66.82
901.17
_42A90_
1,011.98
1.68
84.06
1,039.61
151.21
1,276.56
(1) Non-household industry costs of $72,000 would be allocated to 150,000 non-household vehicles at an average cost of 48 cents per vehicle.
(2) Non-household citizen costs of $5.0 million would be allocated to 150,000 non-household vehicles at an average cost $33.61 per vehicle.
-------
vo
DOLLAR
OUTLAY
$1,400
1,300
1,200
1,100
1,000
900
800
700
400
500
400
(IN 1973 DOLLARS)
DOLLAR
OUTLAY
PERCENT DISTRIBUTION
I
I
I
PERCENT
OF
INCOME
- 40%
35
30
25
20
15
10
5
0;
$5,000 $10,000 $15,000 $20,000 -$25,000
ANNUAL INCOME PER HOUSEHOLD
$30,000
FIGURE 3: ESTIMATED IMPACT OF RECURRING COSTS ON SAN DIEGO
HOUSEHOLDS AT TAXABLE INCOMES: 1977
-------
those households earning between $5,000 and $15,000 and required to pay
recurring costs amounting to more than 10% and 7% of their respective
average taxable incomes of $7,000 and $13,800, continued vehicle owner-
ship could be expected to significantly change the life-style of the
average household within these income groups.
In summary, the $351.0 million of proposed recurring surcharge cost
to be imposed on San Diego households owning vehicles appears substantial.
Some reduction in vehicle ownership among the lower income households
could be attained by the imposition of proposed implementation costs, but
the imposition of the recurring surcharges on citizens who found a vehi-
cle necessary for regular use would result in a heavy financial burden.
The estimated recurring cost of measures excluding the surcharge
program would be as follows:
Hous ehoId s
Under $5,000
$5,000 - 10,000
$10,000 - 15,000
Over $15,000
Total
Recurring Costs
$ 45.90
77.54
143.81
265.77
% of Net
Taxable Income
1.7
1.0
1.0
0.008
97
-------
REVENUE ANALYSIS
Previous portions of this report have described the sources and
allocation methodologies of costs directly associated with program
measures. For each dollar of those costs allocated to the groups identi-
fied, there has also been a dollar of revenue generated for a group.
Some data for evaluating the flow of revenue has been referenced in
prior air quality studies.-^ The need for revenue analysis had recently been
noted in a Council on Environmental Quality Study2, but as yet, no com-
pleted revenue analysis of San Diego or any other AQCR has been found.
Consequently, in order to complete the development of socio-economic
methodologies, PMM&Co. has devised a rudimentary revenue analysis.
The revenue analysis was limited to direct revenue flow at group
levels and should be tempered by the following qualifications.
The potential impact of the flow of revenue to international,
other non-California or non-San Diego locations was not
evaluated.
Potential lags in revenue expenditures and impact on other
economic multiplier effects were not studied.
Probably impacts and the application of local government
generated revenue, new employment (e.g., skill mixes, and
inflation) were not considered except for developing an
estimate of increased direct AQCR employment at an average
wage level.
Possible consumer psychological behavior changes due to
unusual or uncommon influences such as gasoline rationing.
1 Economic Impact Assessment of Alternative Transportation Control
Strategies, Final Report of EPA Contract #08-01-0565, Stephen Sobotka &
Co., Environmental Protection Agency, Washington, D.C., May 3, 1973.
2 The Impact on the Distribution of Income of Financing Federally
Required Pollution Control, N.S. Dorfman and A. Snow, Public Interest
Economics Center, Washington, D. C., August 15, 1973.
98
-------
A significant point concerning possible reluctance of employees to
use mass transit could be that, under present interpretations of the
income tax regulations, funds provided by employers to employees to
encourage transit use could be considered a portion of commuting expense
and therefore would not be normally deductable.
An analysis of the revenue flows generated by the implementation
portion of the program indicated that governmental agencies would not
benefit directly. Industry, covering equipment requirements, and the
group designated as citizen/employees, those whose services would be
required for construction, equipment installation, and administration
of the program, plus those selling needed land, would be the direct
beneficiaries. Approximately 75 percent of the funds would flow to
industry, generally located outside of the local area, and 25 percent
would be directed to new, mostly one-time, employment. The distribution
is shown in Table 27.
Approximately eighty percent of the $19,838,000 of revenue directed
to the citizen/employee group could be expected to be utilized for jobs,
the remainder being absorbed primarily in overhead. Assuming an installa-
tion phase of four years and an average income of $10,000, an annual em-
ployment impact of approximately 397 could be expected during this period.
The majority of these opportunities could be anticipated to develop within
the San Diego AQCR.
Recurring Revenues
PMM&Co. has considered transit operations a part of local government
for purposes of displaying the flow of revenue at ninety percent of funds
collected from surcharge programs. It was assumed that the remaining ten
percent of the funds collected would be utilized as direct income to the
citizen/employee group in the form of wage-income for administrative and
maintenance. Direct revenue accruing also includes new employment require-
ments. A more complete analysis of revenue flow could show flow from
local government revenue directly into industry and also into employment.
Total revenue accruing to the three groups, shown in Table 28, is projected
at $417,349,000. Of this amount, $356,459,000 can be attributed to parking
surcharge and mass transit incentives measures.
Assuming that 80 percent of the citizen/employee and 60 percent of the
industry/employee group revenues would be utilized for employment, the
amount of $74,421,000 would provide 7,442 positions at the average projected
income of $10,000. The majority of these positions would be related to
vehicle inspection and maintenance, fee collection, transit operations,
and administration.
Analysis of Implementation Cost Per Revenue Group
The flow of money on a group-by-group basis as-a result of the im-
plementation of the Clean Air Act is projected in Table 29. This projection
indicates that during the 1972-76 periods, there is a negative flow for all
government agencies and local citizens and a positive flow for industry
99
-------
TABLE 27
IMPLEMENTATION REVENUES
($000)
Measures
Indus try
Citizen/Employee
Stationary Sources
Equipment
Construction
Administrative
Mobile Sources
Retrofit
Equipment
Installation
Inspection Stations
Equipment
Construction
Land
VMT Reduction
Parking Meters
Equipment
Installation
Administration
Bicycle Lane
Construction
Other
Buses
Total
6,366
33,450
7,200
3,840
10,000
60,856
4,244
234
10,800
1,504
1,296
960
675
125
19,838
100
-------
TABLE 28
RECURRING REVENUES
($000)
Measures
Stationary Sources
Maintenance
Adminis tration
Mobile Sources
Inspection
Maintenance
Operation
Administration
VMT Reduction
Mass Transit
Parking Surcharge
Adminis tration
Other
Transit Operations
Total
Local
Gov't.
50,963
264,954
315,917
Industry/
Employee
424
23,680
6,150
3,370
33,624
Citizen/
Employee
120
4,062
3,051
5,663
34,879
138
19,895
67,808
101
-------
TABLE 29
IMPLEMENTATION REVENUE CASH FLOW
($000)
Local
Gov't.
]ost 5,004
levenue
Set Flow -5,004
State
Gov't.
12,415
-
-12,415
Federal
Gov't.
8,415
-
-8,415
Industry
10,610
60,856
+50,246
Citizen
44,250
19,838
-24,412
Total
80,694
80,694
_— .
102
-------
of $50.2 million. A major item causing the negative cash flow from
the citizen group combined with the positive flow to industry is the
potential retrofit converter program. Although it is considered beyond
the scope of this study, comparative economic analyses of the cost-
effectiveness relating projected lives of retrofitted models over an
extended period of time beyond 1977 compared to RHC reductions antici-
pated to be achieved could be useful for policy and regulation inputs.
Analysis of Recurring Cost per Revenue Group
Annual projected net money flows on a recurring basis following
program implementation is projected in Table 30. The projection indicates
a potential negative flow in 1977 from the citizens, the state government
and the federal government compared to positive flows to local government
and industry. The impact of the parking surcharge and mass transit in-
centives are substantial. Their deletion would eliminate the positive
flow of ..money to the local government, specifically to mass transit.
Elimination of these would also reduce to $6.6 million the projected nega-
tive flow of money from the citizen group as shown in Table 31. A pro-
gressive financial policy could result by elimination of these measures,
if mass transit were provided with money through other government financing
policies. The negative impact on costs to the citizens could thus be
reduced while maintaining the direct positive impact on that portion of
local employment opportunities related to mass transit operations.
Fuel Consumption Impacts^
The impacts of the strategies upon a reduction in gasoline consumption
have been compiled and are summarized in Table 32. Their relative signi-
ficance in volume and value are indicated. Vapor control in gasoline
marketing is a positive conservation tool that combined with RHC reduction
remains relatively low in cost. An increase in gasoline consumption result-
int from retrofitting existing vehicles is anticipated due to the lowering
of efficiencies. The most significant area of gasoline usage reduction
results from lower VMT and occurs in direct proportion to the amount achieved.
The basis for determining VMT reduction savings was estimated conservatively
using the daily VMT reduction of 3.07 million miles coupled with the rationale
that this would apply on work days only, since surcharges would not be in
effect on weekends.
103
-------
TABLE 30
RECURRING REVENUE CASH FLOW
($000)
Cost
Revenue
Net Flow
Local
Gov't.
11,044
315,917
+304,873
State
Gov't.
20,183
-
-20,183
Federal
Gov't.
787
-
-787
Industry
424
33,624
+33,200
Citizen
384,911
67,808
-317,103
Total
417,349
417,349
—
TABLE 31
RECURRING REVENUE CASH FLOW MODIFIED BY THE
ELIMINATION OF MASS TRANSIT INCENTIVES AND
PARKING SURCHARGE
($000)
Cost
Revenue
Net Flow
Local
Gov't.
5,604
-
-5,604
State
Gov't.
20,183
-
-20,183
Federal
Gov't.
787
-
-787
Industry
424
33,624
+33,200
Citizen
33,892
27,266
-6,626
Total
60,890
60,890
—
104
-------
TABLE 32
IMPACT OF STRATEGIES UPON FUEL CONSUMPTION
Gasoline ($0.40 per gallon)
Stationary Vapor Control
Vehicle Retrofit
VMT Reduction
Total
Diesel ($0.15 per gallon)
200 Additional Buses
Fuel Saved
(600 gal.)
2,750
(15,375)
58,056
45,431
Fuel Increase
1,508
Savings
($000)
1,100
(6,150)
23,222
18,172
Cost
226
105
-------
IV. IMPACT OF STRATEGIES
Under the Clean Air Act Amendments of 1970, EPA was directed to set
national air quality standards which would protect the public health and
welfare from the known effects of the major air pollutants.
At the regional, state and local levels, institutional arrangements
are being modified in order to deal with those air quality programs which
cross jurisdictional boundaries. New arrangements have been found
necessary for completing the process of planning a program for a given
AQCR. It can be anticipated that even more effective organizational and
budget arrangements will be required when the individual differences,
which will always occur because of the different impacts on certain
interest and user groups within a given community, come into conflict.
The degree of acceptance or rejection of a given measure will eventually
be determined by the cumulative positions of the various individuals
affected. It is extremely difficult to evaluate the overall effects of
measures where divergent views could be expected from what would normally
be considered homogeneous and predictable. For example, certain measures
are expected to be unpopular with the business community. Yet, due to
the creation of a new product, service, or other economic feature, that
same measure could generate strong positive support from selected portions
of the same community. Positive and negative interest groups can be
described in a scenario of every measure.
The lack of predictable reactions from individuals and consumer groups
makes an authoritative summary statement of an impact on society extremely
difficult. The data from studies of the transportation control measures
conducted by consultants and the government and from the overall implementa-
tion plans failed to produce a specific methodology for completing this
portion of the program assessment. While several prior studies completed
in-depth surveys of selected consumer interests (e.g., parking costs, modal
choices) and other studies contained discussions of the relevant social and
socioeconomic topics, no completed evaluation tools or decision matrices
designed to assist in the process of characterizing each measure in relation
to specific interest groups and issues were found. An initial non-quantita-
tive but structured review process appears to be appropriate as a decision
tool that could be utilized by public administrators, task forces and public
interest citizen groups engaged in formulating implementation measures
and strategies acceptable to that particular community. The process might
be most suitable for presentation in a public forum.
For purposes of demonstrating the methodology and for completing a
study consensus, each measure has been evaluated by each study team member
on the basis of expected preference or perception by the majority of
involved San Diego individuals or interest groups.
Figure 4 is a matrix developed to show the results of that attempt to
evaluate the overall program strategies from the point of view of a selected
number of specific consumer and societal interest groups and issues. Where
106
-------
STRATEGIES
STATIONARY SOURCES
VAPOR CONTROL
MOBILE SOURCES
CATALYST RETROFIT
OTHER HARDWARE
INSPECTION AND MAINTENANCE
MOTORCYCLE
V.M.T. REDUCTION
BUS/CAR-POOL LANES
BUS/CAR POOLS
PARKING SUPPLY
MASS TRANSIT
PARKING SURCHARGE
AIR QUALITY MONITORING (V.M.T.)
BICYCLE PATHS
FOUR-DAY WEEK
GASOLINE RATIONING
OTHER
AIRCRAFT
0
0
0
0
0
+
+
—
+
0
0
0
+
—
0
0
—
—
—
0
0
0
—
0
—
0
0
+
—
0
0
0
0
—
0
0
0
0
—
0
0
0
+
—
0
+
—
—
+
—
+
+
+
+
+
0
+
-t-
+
+
+
+
+
0
0
+
+
+
+
—
+
+
+
—
+
0
+
+
0
0
+
+
—
+
—
-f
0
—
—
0
0
/
/
0
0
0
0
0
0
0
0
0
0
/
0
•
0
0
/
/
/
./
/
/
•/
/
/
/
0
/
FIGURE 4: STRATEGIES
107
-------
a favorable reaction was indicated by or for a majority, a "+" symbol
was used; a "-" indicates an unfavorable reaction; and a "0" indicates
no major impact, no clear preference, or an equally divided interest
without a major unfavorable impact on the specific groups affected. For
purposes of this analysis, future strategies and program implementation
were considered on the basis of existing financial, tax, and economic
practices and policies.
For most measures involving specific and direct impacts on individual
groups, a consensus of the majority position could be expected to be
reached through sufficient discussion and sampling. However, in the
event that it is necessary to apply generally unpopular measures to
achieve necessary standards (e.g., the surcharge) an attempt to gain
consensus would be impractical.
An interesting item noted during reporting of this analysis is that
11 of 15 of the measures, or 73 percent, are believed to have a positive
impact on the energy situation and would produce substantial fuel savings.
The results of certain measures, such as the four-day work week, are
difficult to determine using present methods and require further research
of the programs in cities such as Atlanta or in the over 4,000 corporations
and agencies reported to have implemented various test and trial programs.
108
-------
V. PROGRAM IMPLEMENTATION
Based on the urban air quality data collected since 1969, Environ-
mental Protection Agency reports indicate that approximately 29 Air
Quality Control Regions encompassing major population centers will be
unable to achieve adequate emission reductions through the control of
stationary sources alone.
The corrective measures and modifications required to meet projected
standards have been outlined. Recent legislative modifications to imple-
mentation plans have resulted in a reduction of some of the immediate
pressures for compliance. However, regardless of scheduled modifications
it appears that there will be a^ significant long-range socioeconomic /"
impact upon our present life style^including"changes on~the automobile
dependency, industrial development, and land-use patterns.
During PMM&Co.'s study of the San Diego implementation plan, an organiza-
tional modification, resulting in an institutional rearrangement was
announced as the future representative air pollution control task force
for the area. A primary reason for this modification was the need to ensure
that all interest groups and jurisdictions representing the public interest
were fully represented in future program decisions regarding implementation.
A study of the institutional arrangements in the San Diego AQCR revealed that:
a number of such adjustments have been necessitated by the
program;
such adjustments are representative of the environmental
management challenges that exist at the state and regional
levels; and
these adjustments reflect the problems, common concerns, and
individual differences relative to physical boundaries and air
sheds.
When one considers the magnitude of the management problem, it must be
recognized there are approximately 100 AQCRs and that approximately 29 of
these will require substantial air pollution control programs. In addition,
all of the states have been required to submit state implementation plans
(SIP's) to EPA. The agency has the authority and responsibility to modify
those plans determined to be inadequate to comply with the Clean Air Act.
In addition, the agency has the responsibility for outlining the procedures,
monitoring implementation, and determining the progress of the nation's air
pollution control program.
The Environmental Protection Agency has recently initiated some studies
of the methodology and criteria which might be applied in establishing a
mechanism to assist and evaluate regions and states. However, there has been
109
-------
little indication of longer-range planning to overcome the substantial
technical difficulties and to identify financial resources which might be
required for air pollution control program management and implementation
during the next decade.
Technical, legislative, financial, and social issues are involved in
the development of a system which transcends jurisdictional and political
interests and affects the daily living habits of the citizen. At the present
time, certain standards and measures are sufficiently established and little
technical disagreement exist. In some cases, measures have been partially
or fully implemented. Overall, the program appears to be maturing to the
point where major decisions with regard to implementation management must
be made. For example, the methods of sharing costs and of making program
decisions when there are differences among the jurisdictions must be
developed so that an orderly program implementation process can continue
while discussions and solutions of differences are carried out. In addition,
where implementation of certain measures favor one jurisdiction over another,
a management process must be established where consensus can be reached.
It is believed that the San Diego program has progressed to the point
where the active design of program, structures, procedures, and implementa- /
tion methods for sharing both costs and (state and federal) assistance L
should be developed. In addition, the concurrent efforts of the San Diego
AQCR, the state of California, and the region should be supported in dev-
eloping an initial fully implemented air quality control reporting system.
Once a feasible system has been outlined and is operating, an evaluation
can be conducted to assess the suitability of summary information. Such an
assessment will help to determine if monetary requirements can be met. It
is recommended that an additional similar pilot program be identified and
implemented in another independent region and state.
110
-------
BIBLIOGRAPHY
Air Pollution: Alternative Controls for the San Diego Air Basin;
County of San Diego Air Pollution Task Force; November 12, 1973.
Air Quality Implementation Plan Development for Critical California
Regions; Summary Report; TRW Transportation & Environmental Operations;
Redondo Beach, Calif.; Environmental Protection Agency: July 1973.
Alternative Regional Growth Forecasts/Population and Employment -
1975-1995;San Diego Comprehensive Planning Organization; May 1972.
Analysis of the Existing Trends Regional Development Alternative/A
Working Paper; San Diego Comprehensive Planning Organization; June
1973.
Application of the San Diego Transit Corporation for a Mass Transportation
Capital Improvement Grant under the Urban Mass Transportation Act of 1964.
Background Material on Public Hearings for Transportation Control Strategies;
Environmental Protection Agency; Washington, D. C.;1973
Bigelow, J. H., Goeller, B. F., and Petruschell, R. L. ; A Policy-
Oriented Urban Transportation Model; The Los Angeles Version; Rand
Corp. Santa Monica, Calif.; October 1973.
1970 Census Data by Census Tract (Income, Education, Employment,
Spanish-American Population; San Diego Comprehensive Planning Organiza-
tion: January 1973.
Centre City Transportation Study/Circulation Study; Office of the City
Engineer, Transportation Planning Section, City of San Diego; San Diego,
Calif.; February 1971.
Centre City Transportation Study/Satellite Parking; Office of the City
Engineer, Transportation Planning Section; City of San Diego; San
Diego, Calif.; February 1971.
Control Equipment for Hydrocarbon Vapors at Gasoline Service Stations;
San Diego County Air Pollution Control District; San Diego, Calif.;
September 1973.
Cost Effectiveness of Methods to Control Vehicle Refueling Emissions;
Refinery Management Services Co.; American Petroleum Institute;
April 1973
Cumulative Regulatory Effects on the Cost of Automotive Transportation
(RECAT) ; Ad Hoc Committee on the Cumulative Regulatory Effects on the
Cost of Automotive Transportation; Office of Science_and Technology;
February 28, 1972.
Ill
-------
DeHaven, and Woodfill; Cost and Effectiveness of Strategies for Reducing
Emissions from Fixed Sources: The San Diego Air Basin, 1975; Rand Corp.;
Santa Monica, Calif.; October 1973.
Dorfman and Snow; Who Bears the cost of Pollution Control?/The Impact
on the Distribution of Income of Financing Federally Required Pollution
Control; Council on Environmental Quality; Washington, D. C.; August 12,1973.
Economic Impact Assessment of Alternative Transportation Control Strategies/
Part I Final Report; Stephen Sobotka & Co.; May 1973.
Environment Reporter; excerpt pp. 494-507. The Bureau of National Affairs,
Inc.; July 1973.
Evaluating Transportation Controls to Reduce Motor Vehicle Emissions in
Major Metropolitan Areas;Institute of Public Administration, Teknekron,
Inc., & TRW, Inc.; National Technical Information Service; Springfield,
Va.; November 1972.
Evaluation of EPA's Proposed Air Pollution Control Plan with Alternatives;
County of San Diego Air Pollution Task Force; August 1973.
Goeller, B. F., Bigelow, J. H. DeHaven, J.C., Mikolowsky, W. T.,
Petruschell, R. L., Woodfill, B. M., and McFarland, H. W.; San Diego
Clean Air Project; Rand Corp.; Santa Monica, Calif.; July 1973.
Hocker, A. J.; Fifth Progress Report/Project CO/State Fleet Equipped
with UOP Device; Air Resources Board; El Monte, Calif.; August 1973.
Holmes, Horowitz, Reid, and Stolpman, The Clean Air Act and Transportation
Controls/An EPA White Paper; Environmental Protection Agency; Washington,
D. C.; August 1973.
Implementation of the Dimensional Logit Model, Peat, Marwick, Mitchell
ft Co.; (for San Diego County) Comprehensive Planning Organization;
Washington, D. C.; May 1972.
The Implications of Lead Removal from Automotive Fuel;Commerce Technical
Advisory Board Panel on Automotive Fuels and Air Pollution: U. S.
Department of Commerce; Washington, D. C.; June 1970
A Joint Effort for Clean Air in San Diego;Universal Oil Products Company;
Des Plaines, 111.; September 1973.
Kooner and Grant; San Diego Transit Corporation Operations Analysis; Office
of Regional Transportation and Land Use Programs, Environmental Development
Agency, County of San Diego; San Diego, Calif. September 1973.
Lonergan; Financial Report for the Year Ended June 30,. 1972; Office of
City Auditor and Comptroller, San Diego, Calif.; 1972.
Lonergan; Financial Report for the Year Ended June 30, 1973; Office of
City Auditor and Comptroller; San Diego, Calif.; 1973.
112
-------
Merenda and Kuhrtz; Control Strategies for In-Use Vehicles; U. S.
Environmental Protection Agency; Washington, D. C.; November 1972.
Mikolowsky: The Motor Vehicle Emission and Cost Model (MOVEC): Model
Description and Illustrative Applications; Rand Corp.: Santa Monica,
Calif.; October 1973.
Polygon Information Overlay System;San Diego Comprehensive Planning
Organization; September 1971.
Prediction of the Effects of Transportation Controls on Air Quality
in Major Metropolitan Areas; TRW, Inc.; McLean, Virginia; November 1972.
Progress Report on Implementation of the Dimensional Logit Model;
Peat, Marwick, Mitchell & Co.; for San Diego County, Comprehensive
Planning Organization, Washington, D. C.; September 1973.
The Regional Model System and the Planning/Decision Making Process;
A Non-Technical Description; San Diego Comprehensive Planning Organization;
April, 1972.
Rules and Regulations/Air Pollution Control District; Revised Edition;
County of San Diego Department of Public Health; February 1972.
"Rules and Regulations"; Federal Register; Vol. 38, No. 217, Environmental
Protection Agency; Nov. 12, 1973; pp. 31232-31255.
Sage, W. G.; The City of San Diego Annual Financial Report - Fiscal 1973;
Office of City Auditor and Comptroller; San Diego, Calif.; 1973.
San Diego Transit Corporation SB-325 Application FT 1973-74; March 1973.
Sauter, G. D. and Ott, W. R.; "A Computer Program for Projections of
Vehicular Pollutant Emissions in Urban Areas"; Journal of the Air Pollution
Control Association, Vol. 24 No. 1; January 1974, pp. 54-59.
Schwartz, S. I.; "Reducing Air Pollution by Automobile Inspection and
Maintenance: A Program Analysis"; Journal of the Air Pollution Control
Association, Vol. 23, No. 10; October 1973; pp. 845-852.
Semiannual Report; The Committee on Motor Vehicle Emissions of the National
Academy of Sciences; Washington, D. C.; January 1972.
Socio-Economic Impacts of the Proposed State Transportation Control Plans;
An Overview; Transportation & Environmental Operations; Redondo Beach,
Calif.; August 1973.
The State of California Implementation Plan for Achieving and Maintaining
the National Ambient Air Quality Standards; Revision 3; The California
Air Resources Board; June 1973.
113
-------
The State of California Implementation Plan for Achieving and Maintaining
the National Ambient Air Quality Standards; The California Air Resources
Board, January 1972.
System Development for Modeling the Relationship Between Land-Use and
Air Quality in San Diego County; Transportation & Environmental Operations
of,TRW, Inc., Redondo Beach, Calif.; August 1973.
Taylor; Candidate Transportation Systems;San Diego County Comprehensive
Planning Organization; December 1971.
Taylor, Hayden, Sloop, LaRosa; Air Pollution; Alternative Controls for
the San Diego Air Basin;San Diego, Calif.; November 1973.
Technical Support Document for the San Diego Intrastate Air Quality
Control Region; Environmental Protection Agency, Region IX: San
Francisco, Calif. July 31, 1973.
Transit Development Plan & Program Fiscal Year 1972-1973 Update; San
Diego Comprehensive Planning Organization; May 1973.
Transit Development Plan & Program;Alan M. Voorhees & Associates, Inc.;
for San Diego Comprehensive Planning Organization; June 1970.
Transportation Control Strategy Development for the Metropolitan Los
Angeles Angeles Region; TRW Transportation & Environmental Operations;
Redondo Beach, Calif.; Environmental Protection Agency; December 1972.
Watkins, Borden, Kirchner; Highway Research Report/Can Vehicle Travel
be Reduced 20 Percent in the California South Coast Basin?; Sacramento,
Calif.; January, 1974.
Weiner, Edward; DOT-EPA Technical Conference on Air Quality Analysis/
Techniques Needed for Better Air Quality Analysis; Office of The
Secretary of Transportation; Washington, D. C., May 1973
We're Going Your Way I San Diego Transit Corporation; San Diego, Calif.;
July 1973.
Winkler; Transportation Control Strategy Development for the Denver
Metropolitan Area; Springfield, Va.; December 1972.
Winkler: Transportation Control Strategies for the State Implementation
Plan City of Philadelphia; National Technical Information Service;
Springfield, Va.; December 1972.
114
-------
APPENDIX A
DESCRIPTION OF SAN DIEGO
MODAL SPLIT MODEL
-------
Appendix A
DESCRIPTION OF SAN DIEGO MODAL SPLIT MODEL
SELECTION OF THE n-DIMENSIONAL LOGIT MODEL
Two factors governed the selection of the n-Dimensional Logit Model (logit
model) for implementation in San Diego: (1) the sparsity of transit data;
and (2) the desire to simultaneously evaluate traveler choices among three
modes — transit passenger, automobile driver, and automobile passenger.
Development of a modal split model requires data describing travel by each
of the alternative modes. The principal source of this data for the San
Diego region is the 1966 Home Interview Survey. Inasmuch as a uniform sample
of households was surveyed, and because only about four percent of the home-
to-work travel used transit, only a small number of transit users are in the
sample.
Previous efforts to model modal splits have encountered difficulties because
they required an estimate of the proportion of travelers using transit. To
obtain accurate estimates of an observed modal split, analysts were obliged
to aggregate data. This solution resulted, however, in a significant
reduction in the detail — and thus the meaning — of the transportation
service and socioeconomic measures used to predict modal split. These
problems are not encountered by the logit model, since it is developed using
data describing the travel mode choice of individual travelers. In San Diego,
development of the models used all of the valid recorded transit trips in
the region without recourse to aggregation.
It is desirable to simultaneously estimate transit use and automobile
occupancy, inasmuch as these are interrelated phenomena. A key problem with
previously established modal choice models is that they are limited to binary
choices (i.e., a choice between only two alternatives). Analysts have
generally attempted to surmount the limitations imposed by a binary choice
model by using a step-wise or staged approach for estimating splits between
one mode and "all other," and repeating the procedure as necessary. This
step-wise approach obliges the analyst to use aggregate modal service
measures for the initial steps of the process, which often requires
arbitrary decisions regarding, for example, weighted or arithmetic averages.
This artificial, staged analysis inherently implies the risks of error
propagation. For these reasons, San Diego requires a modal split model
capable of simultaneously estimating traveler choices selected from the
three available options — transit passenger, automobile driver, and auto-
mobile passenger.
STRUCTURE OF THE n-DIMENSIONAL LOGIT MODEL
In the selection of a suitable structure for a multi-mode modal split model,
several properties are required:
A-l
-------
the sum of the modal shares should equal one;
a given mode's share (i.e., percent traveling by mode) decreases
with increasing levels of that mode's service measures (e.g.,
time); and
the shares of other modes increase with increases in the given
mode's service measures. -
These theoretical considerations lead to the formulation of the n-dimen-
sional logit model.
In the logit model, the modal split of a given mode m is given by
exp( I «x + «)
w
m 2 exp( 2 a..x.. + a.)
3 J
where:
. x.. is the i impedance attribute of mode j (e.g., transit
in-vehicle time);
ex.. is the calibrated coefficient corresponding to x..; and
or. is a calibrated mode-specific constant.
This formulation does not require that the same set of variables be used to
define the transportation attributes of each of the modes. Further, the
above formulation can be shown to hold when some of the x's, in fact, are
socioeconomic variables, such as income.
IMPLEMENTATION OF THE n-DIMENSIONAL LOGIT MODEL
IN SAN DIEGO
Based on careful review of the CPO's requirements, the capabilities of the
logit model, and the characteristics of the calibration data, it was concluded
that the logit model should be used to predict the modal choice among the
basic alternatives of transit passenger, automobile driver and automobile
passenger. The model defines the choice of mode as a function of several
independent variables which describe the characteristics of the transporta-
tion system, the traveler, and the trip. Selection of the independent
variables to be used in the model was constrained by the availability of
forecasts for these variables. Consequently, it was not possible to
incorporate such socioeconomic variables as the age, sex., or education of
the traveler within the modal split model, since detailed forecasts of these
variables are not available from the Protective Land-Use Model (PLUM) being
developed for CPO. The transportation system variables used to characterize
automobile travel are in- and out-of-vehicle times, and operating and parking
A-2
-------
costs. Similar variables are used to describe transit service; these are
walking times Cat both ends), in-vehicle, waiting, and transfer times,
and transit fare. Socioeconomic variables that were considered included
household income and automobile ownership, variables for which forecasts
are available.
A calibration sample that fulfilled all of the following criteria was
extracted from the 1966 Home Interview Survey. In order to qualify for
the the sample an entry had to:
include travel from home-to-work (work-to-home trips were
not considered);
take place during morning travel period (5:00 a.m. to 10:00 a.m.);
have origins and destinations in the area served by transit;
utilize one of the following modes - automobile driver, automobile
passenger, or transit passenger;
contain valid responses to the income questions; and
involve interzonal travel.
The calibration data set was stratified into two sets: travel to the
Central Business District (CBD) and travel to non-CBD destinations. This
stratification was selected because investigations indicated that the
propensity to use transit for CBD travel was greater for equivalent service
alternatives than it was for non-CBD travel. Characteristics of the CBD
and non-CBD data sets are presented in Figure A-l; these samples contain
809 and 3,109 observations, respectively.
A search process was undertaken to select that set of independent variables
which provide the best modal split models for the San Diego region. The
recommended CBD and non-CBD models are presented in Figure A-2. Three
transportation variables, the differences in line-haul time, excess time,
modal costs, and one socioeconomic variable, household income, are used in
the recommended models. All of the coefficients in the recommended models
are significant at the 95 percent level.
The major contrast between the CBD and non-CBD models is found in the time
and cost coefficients c, d, and e. For the (BD model, the effect of excess
time as compared to line-haul time (i.e., ratio of c/d) is about 1.63. The
derived value of time (i.e., the ratio of e to c or d) is $3.18/hour for
line-haul time and $5.17/hour for excess time. These estimates confirm
results observed elsewhere. Excess time becomes a critical determinant of
modal choice for non-CBD trips as shown by the value of Coefficient c and
the weight of excess time (relative to line-haul time), both of which are
substantially higher than in the CBD model. The increase in the weight
of excess time from 1.6 to 6.8 reflects both the major differences in CBD
and non-CBD transit service and the differences in travel behavior resulting
A-3
-------
Mean Values
Trip Characteristics
Auto Drivers
Auto Passengers
Transit Passengers
Income Level
Auto Network Access Time
Auto Terminal Time
Total Auto Time
Auto Travel Time
Auto Distance
Auto Parking Cost
Walk-to-Transit Time
Transit Vehicle Time
Transit Wait Time
Transit Transfer Time
Total Transit Time
Transit Fare
CBD
570 trips (70.5%)
174 trips (21.5%)
65 trips ( 8.0%)
$9-10,000
0.8 minutes
4.6 minutes
17.8 minutes
12.3 minutes
7.7 miles
47 cents
8.3 minutes
29.8 minutes
13.0 minutes
6.0 minutes
57.1 minutes
44 cents
Non-CBD
2610 trips (83.9%)
406 trips (13.1%)
93 trips ( 3.0%)
$9-10,000
1.7 minutes
2.3 minutes
11.6 minutes
15.7 minutes
7 .1 miles
0.0 cents
10.4 minutes
30.8 minutes
13.7 minutes
11.4 minutes
66.3 minutes
47.4 cents
FIGURE A-l:
CHARACTERISTICS OF CALIBRATION DATA SETS
-------
Model Form:
Po =
1.0
p 1.0 + exp(aT!35 + b) + exp(cDx3 + dD13 + eDCH + f)
PD
expJaT135 + b)
1.0 + exp(aT!35 + b) + exp(cDX3 + dD13 + eDCH + f)
pT = exp(cDX3 + dDL3 + eDCH = f)
1.0 + exp(aT!35 + b) + exp(cDX3 + dDL3 + eDCH + f)
Parameters of the Models:
Parameter
a
b
c
d
e
f
CBD Model
0.0295
-1.4809
0.0916
0.0563
0.0106
1.1635
Non-CBD Model
0.0268
-0.5441
0.1314
0.0192
0.0184
1.6600
Nomenclature :
V PD» 1
exp
a, b, c,
d, e, f
TI35
INC
DX3
DL3
DCH
= auto passenger, auto driver, and transit passenger modal choice probabilities
= exponential operator
= calibrated coefficients and constants for the model
= transformed household income
= 100(l-exp(-0.035*INC))
= household income in hundreds of dollars
= difference in excess time
= walk to/from auto time - (walk to transit time + first wait for transit + walk from transit)
= difference in line haul time
= auto driving time - (transit in vehicle time + transit transfer time)
= difference in modal cost
auto operating cost (5p/mile) + half of parking cost - transit fare
FIGURE A-2:
SAN DIEGO MODAL SPLIT MODELS
-------
therefrom.
The recommended models were tested to see if they could reproduce the
total number of automobile passenger, automobile driver, and transit
passenger trips in the calibration sample. The recommended CBD and non-
CBD models reproduce almost exactly the total number of trips in each of
the three modes considered. The calculated trips were within a (negligible)
fraction of one percent of the observed trips.
A further test of the recommended -model is its ability to reproduce the
observed trip distributions by trip length for each of the three modes under
consideration. The cumulative distributions of transit trips as a function
of trip length (in minutes) as observed in the calibration data set and
estimated by the model are presented in Figure A-3. In all trip length
ranges, the calculated distribution of transit passengers is within about
15 percent or less of the observed distribution. The accuracy of these
results is indicative of the overall quality of the recommended models.
TESTING OF THE SAN DIEGO MODEL IN SAN FRANCISCO
AND BOSTON
The modal split models will be used to evaluate transit systems offering
service which is significantly better than existing transit service. Thus,
there is a concern that the modal split model appropriately reflects consumer
response to significantly improved service, (i.e., extended regions of the
transit service variables). Thus, the model was tested using extended ranges
of transit service from two other metropolitan areas. Validation analyses
were performed for the Golden Gate corridor in the San Francisco metropolitan
area and the Reading corridor in the Boston metropolitan area.
Relative to San Diego, excellent public transportation service is provided
in the Reading corridor by commuter railroad, rapid transit, and bus. The
difference between San Diego's and Boston's transit as compared to auto-
mobile service is evident in the difference in line-haul time; average transit
line-haul time in Boston is six minutes less than automobile time, while it
is 23 minutes longer in San Diego. On the other hand, transit is, relatively
speaking, more expensive in Boston than in San Diego; the average cost
difference (automobile - transit) is 14£ in Boston and 22<: in San Diego.
The San Diego model applied to the Reading corridor data yielded an average
transit modal split of 23.0 percent, compared to an observed share of 27.6
percent. The slight underestimation of transit modal split is ascribed to
travelers in the Boston area having been "conditioned" to transit travel.
The cumulative observed and calculated transit trip distributions were
developed by automobile travel time, as shown in Figure A-4. The distri-
butions are very comparable and lend credence to the assumption that the
under-reporting results from a general difference in behavior between Boston
and San Diego travelers and not to systematic biases of Che San Diego model.
The successful validation of the model with Boston data indicates that the
model can be used to evaluate the ridership and revenue resulting from a
rail transit system in San Diego.
A-6
-------
60
w
H
PS
40
20
Observed
/
^—Calculate d
10
TIME
15
20
FIGURE A-3:
CUMULATIVE TIME DISTRIBUTION OF TRANSIT TRIPS
SAN DIEGO MODEL
-------
100
I
oo
-a
PI
g
0
10
20
50
30 40
TIME
FIGURE A-A:
CUMULATIVE TIME DISTRIBUTION OF TRANSIT TRIPS
SAN DIEGO MODEL APPLIED TO BOSTON DATA
>60
-------
The Golden Gate Bridge connects San Francisco to suburban Marin and Sonoma
Counties. The average automobile travel time is over three times longer in
the Golden Gate corridor than it is in San Diego — 58 versus 18 minutes.
In contrast, the average transit travel time is only 150 percent longer in
the Golden Gate corridor than in San Diego - 85 versus 57 -minutes. Thus,
in terms of the difference in line-haul time, the automobile is 23 minutes
faster in San Diego, whereas transit is two minutes faster in San Francisco.
Because of the higher parking costs in San Francisco Cabout $2.00 versus
$.45), the tolls for the Golden Gate Bridge, and the longer commuting
distances, average automobile costs are much higher in San Francisco than
San Diego. Thus, whereas transit is 22
-------
Transit (%)
Auto
Average CBD Modal Splits (%)
Relative Changes (%)
>
1
o
Excess Line-Haul Parking
Time Time Fare Cost
—
+20
+20 — -- +25
+25
+20 — +25
-10 -20
-10 -20 +25
-10 -20 +25 +25
Auto Auto
Drivers Passengers
70.45
72.01
71.66
70.01
72.38
67.49
68.70
68.15
21.51
22.02
21.91
21.37
22.15
20.54
20.93
20.76
Transit
Passengers
8.04
5.97
6.42
8.62
5.47
11.97
10.37
11.09
Auto
Drivers
—
+2.21
+1.17
-0.62
+2.74
-4.20
-2.48
-3.26
Auto
Passengers
+2.37
+1.86
-0.65
+2.98
-4.51
-2.70
-3.49
Transit
Passengers
-25.75
-20.15
+7.21
-31.97
'•48.88
+28.98
+37 . 94
FIGURE A-5:
SENSITIVITY ANALYSIS OF MODEL
-------
The elasticities of the average modal split estimated for the CBD sample
were determined. Transit modal split is elastic with respect to transit
excess and line-haul times and the difference in access time. In other
words, a one percent relative increase in any of these variables results
in a greater relative decrease in transit modal split. Transit fare,
however, is inelastic: a one percent increase in transit fare results in
a one-third of one percent decrease in transit ridership in the area served
by transit in San Diego. A uniform 10 percent fare increase for each trip
would result in about a net three percent decrease in transit ridership.
CONCLUSION
A graph of the recommended CBD model displaying transit modal split as
a function of the difference of excess times for several levels of cost
and line-haul travel time differences (DC and DL respectively) is presented
in Figure A-6. This graph can be used to rapidly estimate the order of
magnitude transit modal split associated with various combinations of
independent variables. For example, for a $10,000 annual income and equal
service by both modes in terms of times and costs, transit's share would
be 46 percent of the market. This result suggests that the low utilization
of transit in San Diego in 1966 was due to the relatively low quality of the
transit service, not to a systematic automobile bias.
The difficulty of improving the quality of transit service in San Diego
should not be underestimated. In 1966, the average difference in excess
time for all travelers destined to the CBD was -17 minutes (i.e., the
excess time for automobile was 17 minutes less than the excess time for
transit), the average difference in line-haul times was -23 minutes, while
the average difference in cost was -fl8
-------
>
i
100
HOUSEHOL
DIFF
DIFF
) INCOME =
RENCE IN
RENCE IN
$10,000
:OST=
:OST= o
80
TRANSIT
MODAL SPLIT
DL = -30
DL = 15
DL = 0
DL = -15
X
x-
•10 -5 0 5 10
DIFFERENCE IN EXCESS TIME (AUTO-TRANSIT)
FIGURE A-6:
GRAPH OF SAN DIEGO MODEL
-------
Because it specifically addresses multiple choice situations, the logit
model developed for San Diego allows transit usage and "automobile occupancy"
to be simultaneously analyzed. This multi-modal capability offers great
promise for transit planning, inasmuch as the-model correctly incorporates
the competition among the combinations of transit submodes (rail rapid and
all-bus) and access modes (feeder bus, park-and-ride, kiss-and-ride, and
walk) and the automobile submodes (automobile driver and passenger).
The calibrated model very accurately reproduced the total number of observed
automobile passengers, automobile drivers, and transit passengers. In
addition, the model reproduced very well the aggregate modal splits observed
in two other cities for extended ranges of the transit service variables.
This result is most encouraging since transit in Boston and San Francisco
consists of express-bus and rapid rail services not yet available in San
Diego. Overall, these results support the validity of using the developed
logit model as a planning tool for evaluating higher quality transit services
than are currently available in the San Diego region.
A-13
-------
APPENDIX B
REVIEW AND ASSESSMENT OF TRAVEL DEMAND MODELS
-------
APPENDIX B
REVIEW AND ASSESSMENT OF TRAVEL DEMAND MODELS
In order to analyze the travel impacts of the EPA promulgated plan, it was
necessary to incorporate an approach capable of relating the demand for
automobile travel to the price of automobile travel. A literature search
revealed the following travel demand models as approaches for achieving
this objective:
* Demand Generator and Modal Split Model — Rand Corp.;
• Direct Demand Estimation Model — Kraft:
Attitudinal Model — Wallace;
Shopping Trip Frequency Model — Charles River Associates; and
Econometric Model of Gasoline Consumption — Data Resources, Inc.
Appendix B provides a review and assessment of these approaches.
DEMAND GENERATOR AM) MODAL SPLIT MODEL - RAND
A demand generator and modal split model was developed in 1973 by the Rand
Corporation for evaluating various environmental control policies in the
Los Angeles metropolitan area. To that end, its builders attempted to
avoid the conventional Urban Transportation Planning process (UTP) involving
large data requirements, cumbersome zonal systems and networks, excessive
human and computer resources, decomposition of a complex and probably
simultaneous process into a set of consecutive steps — to name only a few
of the shortcomings of the UTP process. To say that they achieved such an
ambitious aim which has eluded both planning researchers and practitioners
in the past 20 years is a Rubicon which a skeptic would not cross without
some hesitation, (whatever the temptation!).
Overview of the Model
The modeling approach has three major features: it does not rely on a
zonal system, it combines demand and modal split in one step, and it takes
account of foregone trips. The model generates aggregate numbers of the
persons trips, vehicle-miles, passenger-miles and vehicle occupancies broken
down by mode, purpose and time of day. To this end, it assumes as given
a "conventional" forecast of total travel by bus and automobile in what the
authors call the "nominal" case. Also assumed as given is the limit of the
total number of trips which would be made regardless of the service level
of the available modes. Foregone trips are then defined as the difference
between the total trips which would be made and the trips actually made.
For any given transportation policy, modal as well as foregone trips would change
while the total number of trips remains constant. Trips are divided into essential
B-l
-------
trips and inessential trips. This distinction applies to inessential trips
inasmuch as there are no essential trips foregone.
Mathematically, the above is expressed as follows:
T* 4- T* + T* = T
A B F
T + T + T = T
A B F
where T^, Tg, Tp, and T are respectively the total number of automobile
trips, bus trips, foregone trips and total limiting or "saturating" trips,
and where the asterick denotes the nominal case.
The transportation system variables are combined into a conductance function
(to be defined later) so that for each mode the change in trips resulting
from a given policy must be in proportion to the ratio of their conductances,
namely
X*" ~ VJ*
A A
B
Tl
where 9 is a proportionality factor, W represents conductances and the
subscipts A, B, and F denote automobile, bus and foregone trips, respectively.
The conductances of foregone trips is set equal to 1.
The conductances of each mode are postulated to be a function of excess
time XT, line-haul time LT and travel cost C (mileage plus parking or fare)
which are combined into a disutility or unattractiveness U. The form of
the function U is derived from the Utilitarian Modal Split model developed
by Alan M. Vorhees and Associates, namely
U = 2.5 X T 4- LT + C/0.25 I
I being the average traveler income. For trips of given length D, line-haul
times by automobile and bus, are derived from speeds and a street mix
usage function which was derived experimentally. As will be discussed later,
automobile occupancies are taken into account as a function of distance
for estimating costs.
B-2
-------
The model is based on two additional assumptions. First, the trip length
distribution function f (D) is assumed known. Second, the conductance
function of each mode follows a logit structure. Mathematically, the
number of essential trips by mode m is given by
/
-
T - T m Tm f (D) dD
m e
•a T*
Tn m
where Te is the total number of essential trips, and where the conductance
WH! of a given mode m is defined by
TT U* (D)
WEI = exp m
exp Um (D)
Similar equations are also presented for estimating passenger-miles and
vehicle-miles. The integrations over D are performed numerically from 0
to 50 miles by one-half-mile increments.
For inessential trips, occupancy is assumed to remain constant for a given
trip purpose. A function r(D) of trip length is postulated to describe
the attractiveness of potential destinations which is assumed to increase,
reach a peak, and decrease. In fact, the function used has a horizontal
asymptote, namely
KQ D
T(D) = ——
where K^ was determined so that the average of T matches the average trip
length and KQ so that/*50 Whence, the conductance function:
/ T(D) = 1.
0 W «= T(D) exp [u*( D)]
m exp [U (D)]
The number of inessential trips by mode m (i.e., bus, automobile or foregone
trips) is given by the same inegration process as for essential trips, namely
m.
Tm '
m = —=-
m
m
where T is the number of "saturating"trips and where «/D for foregone
trips. It should be noted that if T(D) is interpreted as a trip length
distribution function — which in fact it appears to be — then there is
no structural difference in the treatment of essential and inessential
trips except as far as foregone trips are concerned.
The modeling approach lends itself to breaking down trips into peak
and off-peak periods as well as bus-eligible and non-bus-eligible areas.
B-3
-------
As mentioned earlier automobile occupancy was specifically introduced
into the analysis process, in particular for essential trips where it
was supposed to vary as function of trip length instead of remaining
constant as for inessential trips. This is done by first estimating the
delay experienced by a potential carpooler as
ALT = flOD exp (.5 I) )
\~"p~ • \ 10 i
where is a calibrated parameter, P the density of residents living near
the potential carpooler, and D the trip length. Then, for a car pool of
size n the average delay experienced by a member is given by
>
ALT = (n-1) ALT
n
to which corresponds a distance penalty AD derived from ALT by assuming
a speed of 15 mph.
Average occupancy is then determined for a given distance D so that, given
a parametrically expressed proportion!^ of people in cars of occupancy n,
the unattractiveness of essential trips averaged over all carpool sizes
is minimized. The proportion Ha of people in cars of occupancy n is taken
to be
n-1
where % is the positive parameter determined by minimizing the average
unattractiveness for a trip of length D. The resulting average occupancy
is given by
I [-
0(D) » n=l
Comments
According to its authors, the Rand transportation model is a "much needed
flexible policy oriented tool for evaluating alternative transportation
management stratagies." Presently, there is no generally accepted and truly
satisfactory model available to provide information to policy makers. In
reviewing and evaluating Rand's work, two basic and interrelated questions
come to mind. First, what information and what level of detail does it offer?
Second, on what theoretical foundations is the model built and how well does
it reproduce the phenomena it addresses? The model essentially estimates,
from transportation service and cost levels, total travel demand at
a regional level. This travel demand is broken down by mode, including so-
called foregone trips, by vehicle-miles and passenger-miles, as well as by
B-4
-------
peak and off-peak periods. Concurrently, it estimates various other derived
measures, such as, average automobile occupancy and bus loads, average trip
lengths by mode and so forth.
These data are obviously of interest to policy-making but only at a very
"macro" level. There is no information as to how major subareas of the
region or its various population strata are affected. This is so because
of the modeling structure used by Rand.
The authors claim that the modeling framework they developed is simultaneous
and leaves aside the stepwise UTP approach. Unfortunately, this is strictly
speaking not the case: their approach is also stepwise inasmuch as T* and
T, the nominal and saturating trip levels are exogeneous to the model.
Even worse, the inputs are provided by conventional studies. The assertion
that the model is not a forecasting tool is "puzzling", not to say dubious:
the policy-maker seeks information on the consequences of policies which are
not yet implemented, a process which is by definition a forecast. It jnight
be a short-term forecast but it often deals, as is the case of pollution
control policies, with conditions which can be drastically different from
the nominal case.
Rand has imbedded in its approach the notion of conductance which has been
used with little success in intercity demand/modal split modeling without
a formal constraint on the total number of trips (the PML model and the
Composite Analytic model which incidentally was a simultaneous model). It
also relies on previous work in a cumbersome way by combining a logit model
with the "Utilitarian" modal split model in a rather obscure fashion. The
disutility or unattractiveness function used has not been calibrated; it
has been "postulated" from empirical experience and then a "curve was
drawn." Had it been formally estimated, the model would be called a probit
model. One wonders why mention the logit model which is essentially a
disaggregate model within a (very ) aggregate approach. Further, the logit
or the probit model includes a constant in the difference of disutilities
inasmuch as the absence of a constant term would imply that for equal
disutility, modal choice would be equal — an implication which is true
if excess time, line-haul time and out-of-pocket cost fully describe a mode.
The difference in the integration process between essential and inessential
trips is difficult to explain. There is no structural difference between
the treatment of the two types of trips if one considers, as mentioned_
earlier, the function T(D) as a trip length distribution and if essential
trips foregone set equal to zero. Essential trips by automobile (for
example) are calculated as: - I
TAT, » T-, I F [AU(D)] f (D) dD
AE b/_
with *
F[AU(D)] - TAE WA(D)
TAE VD) + TBE VD)
which leads to
jQQ .
TAE WA(D) f(D) dD
TA*TT WA + TTfp WU(D>
AJb A Jj£» D
B-5
-------
However in the case of inessential trips, automobile trips are given by:
T*/* W.(D)
T ' - f J-0 A
XAI XI - ^
IjC
«/jj
instead of
TBIC VD) + TFI
T*W(D)
T* W.(D) + 1
A-L A £>j. D co.
two formulations which are not equivalent.
The modeling structure addresses automobile occupancy or carpooling in a
way which raises several questions. It is understandable that, individually,
"people use their cars so as to make them least unattractive." But one
wonders if this remains true if people in the same trip length behave
similarly regardless of the trip destination, an issue which the model
"ignores" totally. Then for a given distance an average occupancy is
derived for a given distance. How would such an average figure be used
for testing parking policies advocating free parking for groups of three
of more people in a car or, generally speaking, setting parking prices
as a function of automobile occupancy?
The authors propose a formulation of the distribution of people in cars
as follows: _
IIn = x11"1 (1 £ n £ 6)
1 •»• x + x2 4- x3 + x4 + x5
This formulation implies that if x is less than 1, the proportion of cars
of occupancy equal to 1 is larger than that of occupancy equal to 2, and
so forth. If x is greater than one, the reverse is true. There is no
mention in the report on how does x vary when distance varies. One would
doubt that an occupancy of 6 is more likely than an occupancy of, say, 2
or 3. This could happen "inside" the computer program inasmuch as the
search for x(D) is constrained by x being positive.
The estimation of the carpooling time penalty also raises several
questions. This penalty is derived from trip length distributxon for
work trips and calculated as /
ATT = A /10.D
ALT = 9 / exp
The use of the residential density p seems incompatible with a modeling
structure which deliberately casts aside any zonal system and relies only
on a trip length distribution function. If p is considered constant, then
it can be factored out and included in (J) . If residential densities were
B-6
-------
used for one end of the trip, it appears that employment densities
would also be used inasmuch as carpooling opportunities hinge on both
ends of a trip. The description of the calibration method of the
coefficient $ would have been of interest, not to mention its numerical
value. No indication is given as to whether carpooling penalties are
considered as excess time or line-haul time. Finally, the term ALT
represents a delay incurred by the driver to pick up "each of his riders."
The report states that the "average delay experienced by a member" of a
carpool of size n is
This seems wrong as can be seen by considering a two-person carpool: the
driver experiences a delay ALT and the rider no delay (as far as carpooling
is concerned). Hence, an average delay of l/2ilLT instead of ALT as
predicted by the above formula.
The calibration question raised earlier applies to the entire report. Very
little is said about this step which is very important in any modeling
exercise, especially in a case of new framework. In a sens^e, the approach
suggested cannot be calibrated. Many "borrowings" are necessary. This
is very understandable but one would expect to learn more about the sen-
sitivity of the model. Except for the discussion in an appendix to the
report, which is comewhat unconclusive and too general, there is little
said about this critical issue. There are too many coefficients involved—
K!> Ko, <{> trip length distributions—not to address this issue. In par-
ticular, what is the effect of very small levels of transit trips, es-
pecially for nonessential trips, in a metropolitan area such as Los Angeles?
The need for a modeling framework which is sensitive to policy variables,
"easy" to use and requiring little resources, is obvious. The model
attempts to achieve these objectives and the reader would certainly welcome
such an achievement. However, there are too many unanswered questions.
The lack of testing, as evidenced from the report, is unfortunate. The
"transferability" of a model "calibrated" in so specific an area as Los
Angeles is not demonstrated. There are theoretical contradictions which
seem to cast doubt, on the carpooling treatment in particular. Nonetheless,
it is hoped that this approach, which has more merits than would appear
from a deliberately critical standpoint, may help improve the state-of-the-
art.
DIRECT DEMAND ESTIMATION MODELS - KRAFT
The Economic Demand Model developed by G. Kraft is typical of the simultaneous
generation/distribution/modal split models. The model extends to urban trips
the work the author undertook in analyzing intercity trips.
The model derives its origin from economic theory which, according to Kraft,
provides useful guidelines for specifying a transportation demand model for
two broad reasons: it identifies the variables determining demand and
specifies the general nature of the relationship between demand and these
B-7
-------
variables. Briefly speaking, consumer behavior theory indicates that the
variables influencing demand for a given service are the price of this
service, the prices of the competing services and income. Translated in
the context of travel demand, a mode has at least two "prices", travel
times and travel costs. Economic theory indicates that demand for a given
mode should be negatively related to its own prices and positively to the
prices of the competing modes.
Relevant socioeconomic variables include both the individuals and the
measures of output of the activities which attract trips inasmuch as
transportation can be considered as a derived demand commodity. In other
words, it is desired, not for its own sake, but to satisfy another
demand such as work, shopping or recreation. Thus, trips are stratified
by purposes to each of which correspond special activities measured — e.g.,
employment for work trips. It is assumed that generally demand is positively
related to income and activity measures. Finally, as the model is set at
an aggregate leval, it is expected that aggregate demand will be positively
related to the population comprising the market, i.e., the zone of origin.
Finally, demand is expressed in round-trips generated at a given origin.
Among others, Kraft cites two reasons for this approach: the traveler
considers time and cost conditions on both legs of the trip and the outbound
mode strongly influences the inbound mode. The general functional form
adopted for the model is as follows :
,
C(t>3»i\P9,M )] • •
where — ' UJ a
" ' *•* j t i \Ptit MD) - the number of round trips between origin i and destina
tion j for purpose Po by mode Mo,
— I o/ = vector of socioeconomic characteristics appropriate
to purpose Po describing the travelers residing in
zone i,
= vector of socioeconomic and land-use characteristics
describing the level of activity appropriate to pur-
pose Po in destination zone j ,
~ vector of travel time components for the round trip
from origin i to destination j for purpose PO by mode
C(i) Jt i \PO^MQ) = vector of travel cost components for the round trip
cost components for the round trip between origin i
and destination j, for purpose Po by mode MQ,
B-8
-------
_< ljJj i \Po,M^J - vector of travel time components for the round trip
between origin i and destination j, for purpose P0
by each of the alternative modes (a - l,...,n),
C(i-, 3 ji\P * *M J= vector of travel cost components for the round trip
~~ between origin i and destination j for purpose P0 by
each of the alternative modes (a = l,...,n).
Four alternative forms of the general equation were tested:
Model Equation Elasticity Form Estimated
Logarithmic
N = KXa a InN = InK + alnX
Linear
N = K + aX a & N = K + aX
Mixed Log-Linear
N « KXaeBX a + BX InN = InK + alnX + BX
N = K + alnX + BX a + BX N = K + alnX + BX
N
where
N = Dependent Variable,
X = Independent Variable,
K,a,B = Parameters to be estimated.
Although a multi-mode model was available, only the binary choice between
automobile and a general transit category was analyzed. Similarly, the
number of purposes was reduced from six to two, work and shopping, because
the study constraints were relatively strict. As a great number of the
independent variables tested exhibited substantial colinearity, classical
regression (following appropriate log transformations) was unsuccessful in
providing the relationships expected from economic theory. To overcome
this difficulty, the least squares estimation of the coefficient was set
into a quadratic programming format,i.e., to find the set of coefficients
minimizing the sum of the squared residuals subject to a set of constraints
on the said coefficients. Through the use of such constraints on the
signs of the coefficients, the model was fitted with elasticities compatible
with economic theory assumptions. However, because of high residual errors,
Kraft is very cautious about recommending the calibrated models for planning
purposes.
B-9
-------
The Economic Demand Model is a good example of a model, based on apparently
sound theoretical foundations, which encountered serious, if not severe,
estimation problems. The validity of the theoretical approach should not
be dismissed because of these problems, in view of the time and resources
constraints within which the urban application was performed in Boston.
ATTITUDINAL MODEL - WALLACE
In recent years, application of market research techniques has received
increased attention as a potential transportation planning tool. Through
attitudinal surveys, a substantial amount of information has been gathered
concerning the perceived importance of modal attributes as well as the
travelers' level of relative satisfaction with these attributes. Attitudinal
studies have been useful in identifying variables to be incorporated in
transportation models. They have confirmed the importance of such variables
as travel time or cost used in most models. In addition, they have pointed
out the great importance, from the user's standpoint, of such attributes as
reliability, convenience and comfort. On the other hand, if modal split
models are to replicate users behavior, should they be founded on an
attitudinal base? In other words, to what extent should behavior be inferred
from attitudes? Economic studies have relied on the concept of revealed
preferences. As far as modal split analysis is concerned, attitudinal
techniques should not be dismissed, at least at the present time inasmuch
as they offer two very interesting features. First, they allow the use of
other types of variables which cannot be quantified by "engineering" measures
such as pleasantness of the trip or privacy. They offer a promise for
solving the new mode problem for which other models offer only partial
answers, if any.
An example of an attitudinal model based on market research techniques and
developed by Wallace (1968) at General Motors Laboratories is presented
hereafter. Wallace's approach is predicated upon the requirement of defining
modes in terms of attributes which are mode independent. The author identi-
fied three problems to be solved, namely;
determination of a set of attributes which adequately describe
a mode of transportation in the eyes of the travelers for a
given trip purpose;
specifications and calibration of a model to predict modal usage
based on the above set of attributes; and
if a new mode is to be considered, identification of procedures
for estimating the attributes of a new mode of transportation as
perceived by the traveler.
In the first problem, namely the determination of the set of attributes,
Wallace suggests that engineering measures can be understood by the consumer.
However, they do not provide an adequate description of the mode as perceived
B-10
-------
by the user. These shortcomings pertain very often to those qualitative
attributes for which there is no single quantitative measure that is both
unambiguous and adequate to both the engineer and the traveler. Wallace
offers the following list of potential variables:
comfort,
• dependability of on-time arrival,
• protection from weather while waiting,
frequency of vehicle departure times,
pleasantness of trip,
attractiveness of vehicle,
• noise (in vehicle),
chance of accidents,
exposure to undesirable behavior of others,
• traffic,
bodily crowding,
out-of-pocket cost for trip,
total time spent riding,
total time spent walking,
total time spent waiting.
The next question that arises is how to quantify these attributes. Wallace
proposes the use of a technique often used in market research, -namely,
Semantic Differential. Briefly, for a given attribute a seven or eleven-
point rating scale is defined. For example, if noise in the vehicle is
the attribute to be rated, the traveler is asked to rate "very noisy" as
0 and "very quite" as 10. To this end, it might be expected that a traveler
would first determine his perception of the noise level and then use 10, 0
or an intermediate rating if the noise level is respectively very high,
like thunder, or inexistent. However, what most travelers do after deter-
mining their perception of the noise level is to provide a rating which
relates their satisfaction with that level of the noise attribute to the
satisfaction derived from the same attribute of some existing or ideal
mode.
To maintain an element of consistency in the rating process, end points
must be used which will be most likely followed by the travelers. To this
end, the traveler is asked to rate on a scale which best describes his
satisfaction or lack thereof. Thus, to "very noisy" is substituted
"unsatisfactory" or "poor" and to "very quite" "completely acceptable"
or "excellent."
Despite a strong tendency to maintain attribute rating as value judgment
free as possible, Wallace maintains that better results are obtained when
the notion of satisfaction is explicitly introduced, in which case ratings
are called Attribute Satisfaction Ratings. The next step of the analysis
is the determination of the relationship between the attribute satisfaction
ratings and attribute ratings. Wallace admits that this is an area where
considerable research is required.
The second major problem identified by Wallace is the specification and"
calibration of a model. Several approaches have been used for predicting
demand for new products. Wallace proposes a simple utility model in which
B-ll
-------
the utility of a given mode is defined as being equal to the sum of the
utility or satisfaction derived from each attribute. Thus, the total
utility IP- of mode i is expressed as
n .
U1 = S U.(A )
j=l J 3
where Uj is the satisfaction or utility obtained from attribute A^ of mode i
(j - 1.....N). J
It is assumed that the relationship between Aj and Uj (A^) exhibits a
diminishing return property, i.e., there is some maximum level, U ., of
satisfaction attainable with the respect to the jth attribute and^alled
the importance of attribute j. It is then Reasonable to assume that the
relationship between the attribute level A3- and its attribute satisfaction
rating Qj(Al) is linear, i.e., U^ (A^) = U- and 0 correspond respectively
to Qj (Aj ) i 10 and 0 (on an 11 point scale) . Thus , the utility of mode i
expressed in terms of the attribute satisfaction ratings is given by:
.
It is then assumed that product (or mode) i will be preferred to k if
Focusing on those individuals who prefer i to k (denoted by | i) , the follow-
ing inequality should hold for all individuals:
2 U . (Q i'i n kK
j raj 0
A set of Umj can be found such that the inequality holds for each individual
or individuals in a given group. Despite some problems (simplicity of a
linear model, different utility functions for different individuals) Wallace
claims that the method is quite successful in predicting choice among alter-
native products. Specifically, the model's accuracy has ranged from 70 to
95% with an average of 85% in identifying an individual's first choice,
generally among 3 candidate products. On this basis, Wallace believes that,
although never used in transportation, the model would perform satisfactorily
in predicting the modal split of new modes.
The third and last problem mentioned by Wallace is the rating of the attri-
butes of a mode yet to be designed. These ratings are required as inputs
to the model for modal split prediction. Wallace believes that engineering
measures might be helpful in evaluating the concept of a new mode. He
also believes that the questionnaire technique can be useful to a limited
extent because of the difficulty of communicating with printed matters.
He also feels that it would be premature, at this point of the planning of
B-12
-------
a new system, to propose a demonstration program because of the often
high costs of such projects.
Before designing a demonstration project, Wallace suggests, as an inter-
mediate step, to use a technique commonly used in market research and
called a product clinic. The purpose of such an approach is to thoroughly
familiarize a sample of travlers with the new mode under study. Travelers
would be introduced to the new mode as realistically as possible by means
of animated films, vehicle mock-up, charts and so forth, avoiding above
all any "super sell" methods. They would then be asked to rate the modal
attributes which would be used as inputs to the model. Not only could
such an approach provide modal split estimates for a new mode, but also
it could prove to be very helpful in pointing out changes in systems design
or operation which could result in greater user acceptance, thus reducing
the risk of failure of demonstration programs.
In summary, market research models have not been fully tested in transporta-
tion planning applications. There is some doubt that future behavior can be
accurately predicted based on existing attitudes elicited in a survey,
especially if the survey is not carefully designed. However, attitudinal
surveys are helpful in identifying factors influencing modal split. Atti-
tudinal models such as those used in market research offer promise in
taking account of non-quantifiable attributes, treating the new mode
problem, and developing ridership estimates which can be valid for different
geographic areas.
SHOPPING TRIP FREQUENCY MODEL - CHARLES RIVER ASSOCIATES, INC.
This model compares households which took one shopping trip in the 24-hour
survey period with those which took no shopping trip. The hypothesis is
that the more accessible the household is to its relevant shopping areas}
the more frequently it will travel to them. Alternatively, the more costly
or time consuming it is to make a shopping trip, the more carefully the
household will plan its shopping trips and the less frequently trips will
be made.
Although the model is limited to the choice between no trip and one trip
per day, it can readily be extended in a multiple choice framework to daily
frequencies greater than one. The analysis in this study was limited to
zero and one trip because only four households in the sample had shopping
trip frequencies greater than one.
The sample for this model consists of 80 households, 26 of which recorded
no shopping trips and 59 of which recorded one trip in the survey period.
All the households are drawn from the southern suburban corridor. Again
this was done to enable alternate destinations to be identified and
correctly located. The sample is essentially the same one described
earlier for the choice of destination model, with the households recording
more than one trip per day deleted, and a sample of zero trip households
from the same areas added. The sample provides a good cross-section of
downtown and suburban cross-town trips, but is limited to auto travel and
B-13
-------
is further limited in terms of socioeconoraic variety.
Before undertaking this analysis it is necessary to construct the potential
shopping trips for the zero-trip households. That is, if the household were
to take a shopping trip, where would it go, by what mode, and at what time
of day? This is done in a straightforward way, employing the results of
the previous stages of analysis.
First, the relevant destinations are selected by use of the matrix of
shopping destinations developed earlier to analyze choice of destination.
Then the times and costs of travel to each destination are generated from
the network information. Finally inclusive prices of travel to each
destination are computed from the parameters estimated in the modal choice
model. In this study, the inclusive prices of travel for the zero trip
households were based only on the times and costs of auto trips, because
the sample of zero trip households was drawn from a suburban area where all
the observed shopping trips were made by auto. But as described earlier,
this could be generalized by constructing a weighted average of the
inclusive price of travel for each mode to a given destination, using as
weights the computed probabilities from the modal choice function.
To generalize further, the inclusive price could also be averaged over times
of day of travel, using as weights the computed probabilities of time of
day of travel from the time of day model. However, given the quality of
the alternate time of day data or suburban auto trips, there seemed little
point in applying this refinement for this analysis.
The final preparatory step before attempting to model the effect of the
cost or disutility of travel on trip frequency, is to develop an overall
price of travel for each household. Each household has a number of
alternative destinations, for each of which there is an inclusive price
of travel. The natural measure of the overall price of travel to shop for
each household is some weighted average of the inclusive prices of travel
to each of the household's alternative destinations. The obvious weights
to use are the computed probabilities from the choice of destination model.
Thus, we define the overall price of shopping for houshold t as
d
where ht(d) is the estimated probability of household t selecting its dth
destination and pt(d) is tne inclusive price of travel for household t to
its dth destination. This overall price was computed for both the zero
trip households and the one trip households in the sample, using an
earlier equation.
Similarly, a weighted average of the shopping opportunities at each destina
tion was developed by weighting the zonal employment figure by the computed
probabilities from the choice of destination mode. Thus, we define
s'd>
B-14
-------
where ht(d) is as defined above and Et(d) is the percentage of the region's
retail trade employment found in observation t's destination zone d. The
two variables ft and ^t provide measures of the cost of travel to shop for
household t and the shopping opportunities which the traveler t receives
if he takes a trip.
The trip frequency models were estimated using the binary choice logit
model. The best results were as follows:
log Q = -1.72(p-o) + 3.90(E-o)
(3.16) (3.61)
where Qt = the odds that household t will make a daily shopping trip.
/s
-o) = the comparison between taking a trip and incurring price p
and not taking a trip and incurring zerio cost of travel.
(Et-o) - the comparison between taking a trip and obtaining proximity
to £ shopping facilities or opportunities, and not taking a
trip and obtaining proximity to zero facilities.
Both the price and employment variables in the above equation have the cor-
rect sign and are highly significant. The model states that for shopping
the frequency of travel is significant inversely related to the time and
cost of travel. If travel times are reduced through improvements in the
transportation system the frequency of shopping should increase. The
model can be used to predict the expected increase in the number of shop-
ping trips induced by the change in the transportation system.
A number of alternative models were estimated to test the significance of
each of the available socioeconomic variables (including car ownership)
and weather variables. None of them was significant. Interestingly, the
only such variable which resulted in a t-statistic greater than 1.0 was
family income, and this was the only choice model for which income was
even remotely significant. The equation is as follows:
log Q = -2,25p + 2.8SE - .199!
9 (3.30) (2.39) (1.02)
The results suggest that wealthier families shop less frequently, which
seems plausible since they are more likely to have necessary facilities
to carry larger stocks of goods.
The predictive record of the trip frequency model is reasonably good;
60 of the 80 observations were correctly predicted.
B-15
-------
ECONOMETRIC MODEL OF GASOLINE CONSUMPTION
DATA RESOURCES, INC.
An econometric analysis of the demand for gasoline is made with four
different data sets, each of which measures a slightly different version
of the aggregate sales of gasoline. The following listing indicates
that the short term price elasticity (which measures the percentage
decrease in consumption with respect to a one percent change in price)
of demand for gasoline is .12 for the broadest aggregate of gasoline sales
and .16 for the narrowest definition of gasoline consumption (short term
is defined as three months). The short term income elasticity (defined
as the percentage increase in consumption with respect to a one percent
change in income) is approximately.27 with estimates running as low as .23
and as high as .32. The long term income and price elasticities (which
require roughly ten quarters to be fully achieved) are 1.01 and .48. They
range from .42 to .54 for price and .94 to 1.08 for income.
B-16
-------
RESULTS USING THE ERROR COMPONENT
ESTIMATION TECHNIQUE WITH
A LINEAR QUARTERLY MODEL
FOR ALL STATES
Data Source
API Gasoline
SRE
LRE
API Motor Gasoline
SRE
LRE
FHA Motor Fuel
SRE
LRE
FHA Highway Motor
Fuel less Highway
Special Fuels
SRE
LRE
Price
-487.2
(66.1)
-.121
-.417
-467.7
(62.7)
-.120
-.441
-495.9
(64.7)
-.114
-.527
-581.5
(61.7)
-.160
-.540
Income
13658.
(828.)
.275
.944
13288.
(829.)
.276
1.014
12605.
(844.)
.234
1.084
14308.
(813.)
.318
1.062
Lagged
Consumption
.709
(.016)
.727
(.016)
.783
(.014)
.700
(.016)
Standard errors in parentheses
SRE represents short run elasticities
LRE represents long run elasticities
Estimation period is from 63:2 to 72:4
R on
Transformed
Data
.89
,91
.93
.91
B-17
-------
APPENDIX C
STATISTICAL TRANSPORTATION
ANALYSIS TABLES
-------
BA.SE CONDITION
ANALYSIS TABLES
C-l
-------
TRAVEL TIME - WORK
0
1
2
3
4
5
Total
Income of $0,000-4,999
Automobile Transit
Person Passenger
Income of $5.000-9.999
8,688
3,718
2,269
2,452
5,150
0
22,275
2,041
552
413
244
88
0
3.338
Automobile
Person
49,667
38,982
21,429
35,732
45,199
3
191,012
Transit
Passenger
7,062
3,247
2,322
2,172
352
0
15,156
Income of $10,000 and Over
Automobile Transit
Person Passenger
Total For
All Income Groups
53,020
69,429
24,317
55,485
51,880
2
254.133
5,473
4,655
1,996
2,714
232
0
15,069
Automobile
Person
111,373
112,130
48,014
93,669
102,229
5
467.421
Transit
Passenger
14,575
8,455
4,730
5,131
671
0
33,562
TRAVEL TIME - OTHER
0
1
2
3
4
5
Total
Income of $0.000-4,999
Automobile Transit
Person Passenger
Income of $5,000-9,999
20,197
8,133
3,354
4,478
7,915
7
44,082
5,099
673
732
263
170
0
6.937
Automobile
Person
90,334
66,714
29,878
52,344
57,281
85
296,637
Transit
Passenger
9,589
2,032
3,035
1,193
492
0
16,342
Income of $10,000 and Over
Automobile Transit
Person Passenger
Total For
All Income Groups
82,556
105,553
32,405
75,869
65,483
56_
361,921
4,893
2,323
2,198
1,305
235
0
10,955
Automobile
Person
193,086
180,400
65,637
132,690
130,679
147
702,639
Transit
Passenger
19,582
5,028
5,966
2,762
867
0
34,233
-------
TRAVEL TIME - TOTAL
Total For
o
CO
Income of $0,000-4,999
0
1
2
3
4
5
Automobile
Person
28,885
11,851
5,623
6,930
13,065
7
Transit
Passenger
7,140
1,225
1,145
507
258
0
Income of $5,000-9,999
Automobile
Person
140,001
105,696
51,307
88,076
102,480
88
Transit
Passenger
16,651
5,279
'5,357
3,365
844
0
Income of $10,000 and Over
Automobile
Person
135,576
174,982
56,722
131,354
17,363
58
Transit
Passenger
10,366
6,978
4,194
4,019
467
0
All Income Groups
Automobile
Person
304,459
292,530
113,651
226,359
232,908
152
Transit
Passenger
34,157
13,483
10,696
7,893
1,538
0
Total
66,357
10,275
487,649
31,498
616,054
26.024
1,170.060
67,795
-------
TRAVEL COST - WORK
Income of $0,000-4,999
0
1
2
3
4
5
Total
Automobile
Person
124
55
34
34
81
0
328
Transit
Passenger
6
1
1
1
0
0
9
Income of $5,000-9,999 Income of $10,000 and Over
Automobile
Person
706
582
319
512
723
0
2,843
Transit
Passenger
19
8
6
4
1
0
39
Automobile
Person
748
1,042
361
803
827
0
3,781
Transit
Passenger
14
12
5
5
1
0
38
Total For
All Income Groups
Automobile
Person
1,579
1,679
714
1,348
1,631
0
6,952
Transit
Passenger
39
21
12
10
2
0
84
TRAVEL COST - OTHER
0
1
2
3
4
5
Total
Income of $0,000-4,999
Automobile Transit
Person Passenger
224
100
38
49
95
0
505
18
1
2
1
1
0
23
Income of $5,000-9.999
Automobile Transit
Person Passenger
Income of $10,000 and Over
990
799
336
591
691
1
3,409
35
4
11
2
3
0
55
Automobile
Person
895
1,266
365
861
799
1
4,187
Transit
Passenger
17
4
8
2
1
0
32
Total For
All Income Groups
Automobile
Person
2,110
2,166
739
1,499
1,587
2
8,103
Transit
Passenger
69
9
22
5
5
0
110
-------
TRAVEL COST - TOTAL
Total For
o
0
1
2
3
4
5
Income of $0,000-4,999
Automobile
Person
348
155
72
83
176
0
Transit
Passenger
24
2
3
1
1
0
Income of $5,000-9,999
Automobile
Person
1,696
1,381
655
1,103
1,414
1
Transit
Passenger
54
12
17
6
4
0
Income of $10,000 and Over
Automobile
Person
1,643
2,308
726
1,664
1,626
1
Transit
Passenger
31
16
13
7
2
0
All Income Groups
Automobile
Person
3,689
3,845
1,453
2,847
3,218
2
Transit
Passenger
108
30
34
15
7
0
Total
833
31
6,252
93
7,968
69
15,055
194
-------
PERSON TRIPS - TOTAL
MSA
0
1
2
3
4
5
Total
Automobile
Person
139,974
51,834
24,932
27,933
50,646
12
295,331
Income of
Automobile
Driver
101,878
37,809
18,467
20,481
37,461
8
216,104
$0,000-4,999
Automobile
Passenger
40,181
14,769
6,731
7,816
13,742
4
83,243
Income of $5,000-9,999
Transit
Passenger
15,628
1,961
2,382
711
883
0
21,564
Automobile
Person
660,589
436,437
216,724
336,578
370,509
162
2,020,999
Automobile
Driver
486,307
322,247
161,174
250,218
275,982
113
1,496,040
Automobile
Passenger
182,292
119,166
57,639
76,733
97,951
54
546,734
Transit
Passenger
34,706
8,273
11,576
4,580
2,846
0
61,980
o
I
Income of $10,000 and Over
MSA
0
1
2
3
4
5
Total
Automobile
Person
623,224
710,968
232,972
501,681
416,194
106
2,485,142
Automobile
Driver
462,430
528,776
173,787
374,525
310,863
73
1,850,453
Automobile
Passenger
167,147
189,030
61,255
131,514
108,896
35
657,876
Transit
Passenger
19,761
10,620
8,945
5,323
1,427
0
46,076
Total For All Income Groups
Automobile
Person
1,423,787
1,199,239
474,628
866,191
837,349
281
Automobile
Driver
1,050,615
888,831
353,428
645,226
624,305
195
Automobile
Passenger
389,620
322,966
125,623
228,961
220,588
93
Transit
Passenger
70,094
20,854
22,903
10,614
5,156
0
4,801,472
3,562,596
1,287.852
129,620
-------
PERSON TRIPS FOR NON-WORK PURPOSE
MSA
0
1
2
3
4
5
Total
Income of $0,000-4,999
Automobile
Person
99,950
36,442
15,589
18,892
32,026
11
202,912
Automobile
Driver
66,192
24,134
10,324
12,511
21,209
7
134,379
Automobile
Passenger
33,758
12,308
5,265
6,381
10,817
4
68,533
Transit
Passenger
11,923
1,042
1,630
363
632
0
15,591
Automobile
Person
433,584
279,863
130,990
204,047
220,893
154
1,269,530
Income of $5,000-9,999
Automobile
Driver
287,142
185,340
86,748
135,130
146,287
102
840,748
Automobile
Passenger
146,442
94,523
44,242
68,917
74,606
52
428,782
Transit
Passenger
22,493
2,978
7,349
1,557
1,922
0
36,298
Income of $10,000 and Over
MSA
0
1
2
3
4
5
Total
Automobile
Person
383,952
429,434
137,149
293,301
242,242
100
i
1,486,176
Automobile
Driver
254,273
284,393
90,827
194,239
160,425
66
984,222
Automobile
Passenger
129,679
145,041
46,322
99,062
81,817
34
501,954
Transit
Passenger
10,718
3,202
5,309
1,650
866
0
21,745
Total For All Income Groups
Automobile
Person
917,486
745,739
283,727
516,240
495,160
267
2,958,617
Automobile
Driver
607,607
493,867
187,899
341,881
327,921
177
1.959,349
Automobile
Passenger
309,879
251,872
95,828
174,359
167,239
90
999,268
Transit
Passenger
45,133
7,222
14,288
3,571
3,420
0
73,634
-------
PERSON TRIPS FOR WORK
MSA
0
1
2
3
4
5
Total
Income of $0,000-4,999
Automobile
Person
42,109
16,136
9,609
9,405
19,177
1
96,435
Automobile
Driver
35,686
13,675
8,143
7,970
16,252
1
81,725
Automobile
Passenger
6,423
2,461
1,466
1,435
2,925
0
14,710
Transit
Passenger
3,974
930
783
352
265
0
6,304
Automobile
Person
235,015
161,550
87,823
135,804
153,040
13
773,244
Income of $5,000-9,999
Automobile
Driver
199,165
136,907
74,426
115,088
129,695
11
655,292
Automobile
Passenger
35,850
24,643
13,397
20,716
23,345
2
117,952
Transit
Passenger
12,620
5,278
4,355
3,010
961
0
26,224
Income of $10,000 and Over
Total For All Income Groups
MSA
0
1
2
3
4
5
Automobile
Person
245,625
288,372
97,893
212,738
177,517
8
Automobile
Driver
208,157
244,383
82,960
180,286
150,438
7
Automobile
Passenger
37,468
43,989
14,933
32,452
27,079
1
Transit
Passenger
9,174
7,367
3,718
3,649
575
0
Automobile
Person
522,749
466,058
195,324
357,947
349,733
21
Automobile
Driver
443,008
394,964
165,529
303,345
296,384
18
Automobile
Passenger
79,141
71,094
29,795
54,602
53,349
3
Transit
Passenger
25,768
13,574
8,856
7,009
1,801
0
Total 1.022,153
866,231
155.922
24,482
1,891.831
1,603.247
288.584
57.010
-------
PERSON MILES - OTHER
o
i
vo
0
1
2
3
4
5
Total
0
1
2
3
4
5
Income of $0,000-4
Automobile
Driver
342,481
151,927
64,428
82,957
166,075
157
808,027
Income
Automobile
Driver
1,417,537
2,052,127
618,802
1,452,696
1,392,669
1,382
Automobile
Passenger
174,665
77,482
32,858
42,308
84,699
81
412,093
of $10,000 and
Automobile
Passenger
722,944
1,046,585
315,590
740,875
710,261
705
,999
Transit
Passenger
31,202
6,178
6,267
2,601
6,071
0
46,855
Over
Transit
Passenger
30,604
20,754
16,287
13,244
868
0
Income
Automobile
Driver
1,543,592
1,281,096
571,893
998,918
1,205,104
2,114
5,602,716
of $5,000-9,
Automobile
Passenger
787,231
653,358
291,666
509,449
614,603
1,078
2,857,385
Total For All Income
Automobile
Driver
3,303,610
3,485,149
1,255,123
2,534,572
2,763,847
3,655
Automobile
Passenger
1,684,840
1,777,426
640,112
1,292,632
1,409,562
1,864
999
Transit
Passenger
57,145
18,586
23,670
12,343
1,780
0
113,523
Groups
Transit.
Passenger
118,950
45,517
46,225
28,188
3,256
0
Total
6,935,211
3,536,959
81,757
13,345,952
6.806.436
242,136
-------
PERSON MILES - WORK
o
i
0
1
2
3
4
5
Total
0
1
2
3
4
5
Automobile
Driver
186,698
88,777
57,481
56,862
140,799
6
530,620
Income of $0,000-4
Automobile
Passenger
33,605
15,979
10,347
10,235
25,343
1
95,511
Income of $10,000 and
Automobile
Driver
1,154,450
1,691,219
603,422
1,339,878
1,441,801
51
Automobile
Passenger
207,801
304,420
108,616
241,178
259,524
9
,999
Transit
Passenger
10,492
3,817
3,439
1,829
320
0
19,896
Over
Transit
Passenger
30,098
31,785
14,848
21,573
977
0
Income of $5,000-9,
Automobile
Driver
1,078,034
948,071
535,679
857,116
1,260,114
79
4,679,092
Total
Automobile
Driver
2,419,182
2,728,065
1,196,582
2,253,856
2,842,712
135
Automobile
Passenger
194,096
170,6521
96,423
154,281
226,821
14
842,236
For All Income
Automobile
Passenger
435,452
491,052
215,384
405,694
511,688
24
999
Transit
Passenger
37,657
22,509
18,433
17,131
1,457
0
97,187
Groups
Transit
Passenger
78,248
58,111
36,720
40,532
2,755
0
Total
6,230,820
1,121,548
99.283
11,440.531
2,059,295
216,366
-------
PERSON MILES - TOTAL
0
1
2
3
4
5
Total
0
1
2
3
4
5
Automobile
Driver
529,179
240,704
121,909
139,819
306,874
163
1,338,647
Income of $0,000-4
Automobile
Passenger
208,270
93,461
43,205
52,543
110,042
82
507,604
Income of $10,000 and
Automobile
Driver
2,571,987
3,743,346
1,222,224
2,792,574
2,834,470
1,433
Automobile
Passenger
930,745
1,351,005
454,206
982,053
969,785
714
,999
Transit
Passenger
41,694
9,995
9,706
4,430
927
0
66,752
Over
Transit
Passenger
60,702
52,539
31,135
34,817
1,845
0
Income of $5,000-9,
Automobile
Driver
2,621,626
2,229,167
1,107,572
1,856,034
2,465,218
2,193
10,281,808
Total
Automobile
Driver
5,722,792
6,213,214
2,451,705
4,788,428
5,606,559
3,790
Automobile
Passenger
981,327
824,010
388,089
663,730
841,424
1,092
3,699,621
For All Income
Automobile
Passenger
2,120,292
2,268,478
855,496
1,698,326
1,921,250
1,888
999
Transit
Passenger
94,802
41,095
42,103
29,474
3,237
0
210,710
Groups
Transit
Passenger
197,198
103,628
82,945
68,720
6,011
0
Total
13,166,031
4,658,507
181,040
24,786,483
8.865,731
458,502
-------
EPA PROMULGATED PIAN
ANALYSIS TABLES
C-12
-------
PERSON TRIPS - TOTAL
0
1
2
3
4
5
Total
o
>-*
LO
0
1
2
3
4
5
Total
Income of $0,000-4,999
Automobile
Person
128,301
47,426
22,902
25,651
46,141
11
271,545
•
Automobile
Person
570,026
659,150
214,286
267,080
391,763
101
T
2,302,403
Automobile
Driver
76,197
27,968
13,123
14,949
26,249
7
159,231
Income of $10,
Automobile
Driver
389,671
451,022
146,803
320,113
268,646
68
1,576,319
Automobile
Passenger
52,104
19,458
9,778
10,702
19,892
3
112,314
000 and Over
Automobile
Passenger
180,355
208,128
67,483
136,967
123,117
33
726,084
Transit
Passenger
16,620
2,834
2,832
1,145
1,065
0
24,496
Transit
Passenger
32,255
21,733
14,313
11,644
2,261
0
82,208
Automobile
Person
599,567
402,696
197,863
312,184
346,988
155
1,859,453
Automobile
Person
1,297,893
1,109,272
435,051
804,914
786,008
266
4,433,401
Income of $5,000-9,999
Automobile
Driver
380,735
255,215
124,849
197,008
218,542
102
1,176,438
Total For All
Automobile
Driver
846,589
734,205
284,776
532,070
514,176
177
2,911,989
Automobile
Passenger
218,832
147,481
73,014
115,176
128,446
53
683,015
Income Groups
Automobile
Passenger
451,304
375,067
150,275
272,844
271,832
89
1,521,412
Transit
Passenger
50,071
15,464
17,733
9,361
4,351
0
96,979
Transit
Passenger
98,946
40,031
34,879
22,149
7,678
0
203,682
-------
o
I
PERSON TRIPS - OTHER
0
1
2
3
4
5
Total
Income of $0,000-4,999
Automobile
Person
89,744
32,302
13,924
16,689
27,199
10
180,983
Automobile
Driver
59,433
21,392
9,221
11,052
18,013
7
119,856
Automobile
Passenger
30,311
10,910
4,703
5,637
9,186
3
61,127
Transit
Passenger
9,094
892
1,418
350
565
0
12,319
Automobile
Person
378,779
247,709
114,289
180,641
194,981
142
1,116,541
Income of
Automobile
Driver
250,860
164,046
75,688
119,630
129,126
94
739,431
$5,000-9,999
Automobile
Passenger
127,919
83,663
38,601
61,011
65,855
48
377,110
Transit
Passenger
23,224
3,623
9,129
2,090
2,357
0
40,423
Income of $10,000 and Over
Total For All Income Groups
0
1
2
3
4
5
Automobile
Person
336,136
380,507
120,289
259,746
214,94,9
93
Automobile
Driver
222,607
251,991
79,662
172,017
142,350
62
Automobile
Passenger
113,529
128,516
40,627
77,729
72,599
31
Transit
Passenger
11,346
4,637
6,699
2,591
983
0
Automobile
Person
804,658
660,518
248,502
457,075
438,245
244
Automobile
Driver
532,886
437,429
164,571
302,699
290,228
162
Automobile
Pas_senger_
271,772
223,089
83,931
154,376
148,017
82
Transit
Passenger
43,664
9,153
17,247
5,031
3,905
0
Total 1,311,718
868.687
443,031
26,258
2.609,242
1,727,975
881.267
78,999
-------
o
I
PERSON TRIPS - WORK
0
1
2
3
4
5
Total
Income of $0,000-4,999
Automobile
Person
38,557
15,124
8,978
8,962
18,942
1
90,562
Automobile
Driver
16,764
6,576
3,903
3,897
8,236
1
39,375
Automobile
Passenger
21,793
8,548
5,075
5,065
10,706
0
51,187
Transit
Passenger
7,526
1,942
1,414
795
500
0
12,177
Automobile
Person
220,788
154,987
83,574
131,543
152,007
13
742,912
Income of
Automobile
Driver
129,875
91,169
49,161
77,378
89,416
8
437,007
$5,000-9,999
Automobile
Passenger
90,913
63,818
34,413
54,165
62,591
5
305,905
Transit
Passenger
26,847
11,841
8,604
7,271
1,994
0
56,556
Income of $10,000 and Over
0
1
2
3
4
5
Automobile
Person
233,890
278,643
93,997
207,334
176,814
8
Automobile
Driver
167,064
199,031
67,141
148,096
126,296
6
Automobile
Passenger
66,826
79,612
26,856
59,238
50,518
2
Transit
Passenger
20,909
17,096
7,614
9,053
1,278
0
Total For All Income Groups
Automobile
Person
493,235
448,754
186,549
347,839
347,763
22
Automobile
Driver
313,703
296,776
120,205
229,371
223,948
15
Automobile
Passenger
179,532
151,978
66,344
118,468
123,815
7
Transit
Passenger
55,282
30,878
17,632
17,118
3,773
0
Total 990,685
707,632
283,053
55,950
1.824,159
1.184.014
640,145
124.683
-------
o
I
PERSON MILES - TOTAL
0
1
2
3
4
5
Total
Income of $0,000-4,999
Automobile
Driver
427,988
129,531
92,434
110,273
238,298
159
1,060,700
Automobile
Passenger
305,033
138,871
72,832
82,408
180,665
84
783,927
Transit
Passenger
37,764
11,333
11,527
5,462
1,290
0
67,377
Income of $5,000-9,999
Automobile
Driver
2,223,526
1,899,847
924,489
1,577,568
2,087,732
2,132
8,703,895
Automobile
Passenger
1,291,117
1,105,213
547,597
928,173
1,245,333
1,099
5,110,544
Transit
Passenger
128,089
63,924
63,712
48,039
5,679
0
309,412
Income of $10,000 and Over
0
1
2
3
4
5
Automobile
Driver
2,310,966
3,402,363
1,099,296
2,543,425
2,605,590
1,397
Automobile
Passenger
1,065,445
1,569,100
502,244
1,155,406
1,186,911
708
Transit
Passenger
94,157
92,462
49,225
61,535
3,429
0
Total For All Income Groups
Automobile
Driver
4,962,465
5,494,741
2,116,351
4,231,265
4,932,359
3,687
Automobile
Passenger
2,661,607
2,813,183
1,122,741
2,175,986
2,613,285
1,891
Transit
Passenger
259,980
167,719
124,463
115,034
10,398
0
Total
11,951,148
5,485.058
300,807
21,715.743
11.379,527
677,596
-------
PERSON MILES - OTHER
o
i
0
1
2
3
4
5
Total
Income of $0,000-4,999
Automobile
Driver
318,168
141,037
59,914
77,149
155,851
154
753,143
Automobile
Passenger
162,266
71,429
30,556
39,346
79,484
78
384,104
Transit
Passenger
17,503
4,001
5,389
1,869
611
0
29,374
Income of $5.000-9,999
Automobile
Driver
1,396,480
1,182,531
523,924
926,972
1,137,420
2,069
Automobile
Passenger
712,185
603,091
267,201
472,755
580,085
1,055
Transit
Passenger
50,766
17,874
28,456
12,794
2,400
0
5,169,382
2,636,385
112,289
Income of $10,000 and Over
Total For All Income Groups
0
1
2
3
4
5
Total
Automobile
Driver
1,282,351
1,892,317
568,415
1,345,775
1,315,226
1,353
6,405,435
Automobile
Passenger
653,999
965,082
289,891
676,346
670,766
690
3,266,773
Transit
Passenger
30,154
25,102
19,873
16,304
1,137
0
92,569
Automobile
Driver
2,996,984
3,215,885
1,152,385
2,349,895
2,609,236
3,575
12,327,960
Automobile
Passenger
1,528,462
1,640,101
587,716
1,198,446
1,330,711
1,823
6,287.260
Transit
Passenger
98,423
46,977
53,717
30,965
4,148
0
234,232
-------
PERSON MILES - WORK
o
i
CO
0
1
2
3
4
5
Total
Automobile
Driver
109,820
51,494
32,520
33,124
82,447
5
307,557
Income of $0,000-
Automobile
Passenger
142,767
66,942
42,276
43,062
101,181
6
399,823
-4,999
Transit
Passenger
20,261
7,332
6,138
3,593
679
0
38 , 003
Automobile
Driver
827,046
717,316
400,565
650,596
950,312
63
3,534,513
Income of $5,000-9,999
Automobile
Passenger
578,932
502,122
280,396
455,418
665,248
44
2,474,159
Transit
Passenger
77,293
46,050
35,256
35,245
3,279
0
197,123
Income of $10,000 and Over
0
1
2
3
4
5
Automobile
Driver
1,028,615
1,510,046
530,881
1,147,650
1,290,364
44
Automobile
Passenger
411,446
604,018
212,353
479,060
516,145
18
Transit
Passenger
64,003
67,360
29,352
45,231
2,292
0
Total For All Income Groups
Automobile
Driver
1,965,481
2,278,856
963,966
1,881,370
2,323,123
112
Automobile
Passenger
1,133,145
1,173,082
535,025
977,540
1,282,574
68
Transit
Passenger
161,557
120,742
70,746
84,069
6,250
0
Total
5.545.713
2,218.285
208.238
9.387,783
5,092.267
443,364
-------
TRAVEL TIME - WORK
0
1'
2
3
4
5
Total
Income of $0.000-4,999
Automobile Transit
Person Passenger
Income of $5,000-9,999 Income of $10,000 and Over
10,811
4,543
2,697
2,961
6,448
0
27,205
4,077
1,150
776
532
192
0
6,727
Automobile
Person
56,676
44,435
23,961
40,344
51,181
4
220,422
Transit
Passenger
15,205
7,173
4,686
4,931
813
0
32,808
Automobile
Person
58,683
75,484
26,404
60,313
56,740
2
278,283
Transit
Passenger
12,317
10,583
4,155
6,263
560
0
33,878
Total For
All Income Groups
Automobile
Person
126,170
124,462
53,062
103,618
114,369
6
525,910
Transit
Passenger
31,598
18,906
9,617
11,726
1,565
0
73,412
o
I
TRAVEL TIME - OTHER
0
1
2
3
4
5
Total
Income of $0,000-4,999
Automobile Transit
Person Passenger
Income of $5,000-9,999 Income of $10,000 and Over
18,593
'7,453
3,087
4,101
7,321
7
40.562
3,587
524
663
229
179
0
5,183
Automobile
Person
80,657
60,818
26,976
47,916
52,996
83
269,445
Transit
Passenger
9,835
2,273
3,849
1,434
685
0
18,077
Automobile
Person
73,791
96,210
29,357
69,375
60,730
54
329,518
Transit
Passenger
5,347
3,164
2,869
1,829
314
0
13,522
Total For
All Income Groups
Automobile
Person
173,041
164,481
59,421
121,393
121,047
143
639,524
Transit
Passenger
18,769
5,962
7,381
3,492
1,177
0
36,782
-------
TRAVEL TIME - TOTAL
Total For
Income of $0,000-4,999
o
i
ro
o
0
1
2
3
4
5
Automobile
Person
29,404
11,996
5,784
7,062
13,769
7
Transit
Passenger
7,664
1,674
1,439
761
371
0
Income of $5,000-9,999
Automobile
Person
137,333
105,253
50,937
88,260
104,177
87
Transit
Passenger
25,040
9,446
8,535
6,365
1,498
0
Income of $10,000 and Over
Automobile
Person
132,474
171,694
55,761
129,688
117,470
56
Transit
Passenger
17,664
13,747
7,024
8,092
874
0
All Income Groups
Automobile
Person
299,211
288,943
112,483
225,011
235,416
149
Transit
Passenger
50,367
24,868
16,998
15,218
2,742
0
Total
67,767
11.910
489.867
50.885
607.801
47.400
1,165.434
110.194
-------
TRAVEL COST - WORK
Income of $0,000-4,999
0
1
2
3
4
5
Total
Automobile
Person
188
77
47
47
104
0
461
Transit
Passenger
12
3
2
1
1
0
19
Income of $5,000-9,999 Income of $10,000 and Over
Automobile
Person
1,465
1,092
590
942
1,185
0
5,268
Transit
Passenger
41
18
13
11
3
0
87
Automoible
Person
1,890
2,386
809
1,802
1,667
0
8,548
Transit
Passenger
32
26
12
14
2
0
86
Total For
All Income Groups
Automobile
Person
3,543
3,555
1,446
2,791
2,956
0
14,277
Transit
Passenger
85
48
27
26
6
0
192
I
NJ
TRAVEL COST - OTHER
0
1
2
3
4
5
Total
Income of $0,000-4,999
Automobile Transit
Person Passenger
, 666
257
106
131
234
0
1,394
15
1
2
0
1
0
19
Income of $5.000-9,999 Income of $10,000 and Over
Automobile
Person
2,824
2,000
890
1,468
1,647
2
8,832
Transit
Passenger
36
6
15
3
3
0
63
Automobile
Person
2,523
3,107
949
2,122
1,851
1
10,552
Transit
Passenger
18
7
10
4
1
0
40
Total For
All Income Groups
Automobile
Person
6,013
5,364
1,945
3,721
3,732
3
20,778
Transit
Passenger
69
14
27
7
5
0
122
-------
TRAVEL COST - TOTAL
Total For
o
to
to
0
1
2
3
4
5
Income of $0,000-4,999
Automobile
Person
854
334
153
178
338
0
Transit
Passenger
27
4
4
1
2
0
Income of $5,000-9,999
Automobile
Person
4,289
3,092
1,480
2,410
2,832
2
Transit
Passenger
77
24
28
14
6
0
Income of $10,000 and Over
Automobile
Person
4,413
5,493
1,758
3,924
3,518
1
Transit
Passenger
50
33
22
18
3
0
All Income Groups
Automobile
Person
9,556
8,919
3,391
6,512
6,688
3
Transit
Passenger
154
61
54
33
11
0
Total
1,855
38
14,100
150
19.100
126
35.055
314
------- |