Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
introduction and Overview THE PUBLIC POLICY CONTEXT There are strong indications, from the decennial Census of Population journey-to- work statistics and from the periodic Nationwide Personal Transportation Surveys (NPTS), that the average occupancy of private vehicles used for commute trips has been falling since the 1960s. For example, for 31 of the nation's most populous metropolitan areas, the private vehicle occupancy for trips to work averaged 1.16 persons per vehicle in 1970; by 1990, this statistic had fallen to i.09, a drop of 6.6%.~ More markedly, for the country as a whole the proportion of work trips made by driving alone in a private vehicle increased from 64.4% of all commuters in 1980 to 73.2% by 1990. This growth - implying over 22 million extra vehicles on the nation's roads, many of them during peak periods - came primarily at the expense of ridesharing in private vehicles, and to a lesser degree from transit riders. The switch from ridesharing to drive alone was more marked during the 19SOs than it had been in the 1970s. By 1990, the proportion of private vehicle commuters who were riding in vehicles with four or more travelers was roughly I% nationally, down from about i.7% in 1980. What has happened to carpooling? Given that the large and continued increases in the proportion of commuters driving alone during the 1980's are mostly attributable to the decline in carpooling, the question would seem to be, "Why doesn't anyone carpoo} as much as they used to do?" Perhaps the higher occupancies we observed in two previous decades reflected a passing "fad," spurred on by a sudden awareness of limited energy resources, high gasoline pnces, and the growing interest in the environment. But on closer inspection, there appear to be some important structural changes taking place in the ways we live and work that may go a lot further in explaining the decline. Of the 31 metropolitan areas, only ten had constant geographical boundaries over the twenty years. The comparisons above make some adjustment for boundary changes, but if attention is restricted solely to the ten unchanged metropolitan areas the decline in average occupancy is somewhat less, about 3.9%.
Workplace and Residential Geography Overall population growth in the United States has been quite slow in recent years, averaging less than i% per year since 1980. in fact, with the exception of the great depression of the 1930s, the rate of population growth during the 1980s was the slowest in US history. But this national figure masks some other very important trends. While the total growth in population is at a historic low, it is at the same time becoming increasingly aggregated around large urban centers. Nearly all of the increase in population during the 1980s, about 22 million people, occurred in metropolitan areas.2 The percent of the total population living in metropolitan areas has increased from less than 70% in 1970 to 80% in 1994 (the most recent year for which data are available), and over half of all Americans now live in metropolitan areas with a population of one million or more.3 The trend within metropolitan areas is even more instructive. Figure 2 shows that nearly all of the population growth that has occurred in the post-war penod has been in the suburbs, and that trend continued through the 1980s such that now nearly half of the US population (47%) lives in suburbs, up from 43% in 1980 and more than double the 1950 share of 23%.4 Moreover, this growth has also been fueled by internal migration of metropolitan area residents-central cities lost over two million residents per year to the suburbs during the 1980s.5 250 - ~n o 200 - ._ - o A_ o 150 100 50 a Figure 1. US Population by Major Geographic Area 1950-1990 300 : _ . ::::::::::::::::::::::::: ...... ,..~...................................... W ..... ......... ~ ; .. ..................................... ...... .................. Suburban population Central city population 1950 1960 1970 1980 1990 Source: Pisarski, p. 19, using data from the 1990 Census. 2
Not surprisingly, the pattern of job growth has paralleled that of population, as illustrated in Figure 2. Nearly two thirds of the job growth during the 19SOs occurred in the suburbs, such that 42% of all jobs are now located there.6 Figure 2. Job Distribution by Major Geographic Area Growth in Jobs by Geographic Area 1980-1990 Source: Pisarski, p. 26, using data from 1990 Census. Suburbs 42% \....'.~.2...:...'.'''"'"''""""'"'''"'"""""''""."'''''"'' - , Nonmetro ~ : _ 12O/^ ~ ~ ~ i, ~ . ,, . ~ V..'..~....'....2..._ ... ............ _ ............ , _ ...... I_ 1~ , On metro 20% Distribution of Jobs by Geographic Area 1990 As people have increasingly chosen to live in metropolitan areas, the pressure for these areas to expand outward has intensified. At the same time, migration within metropolitan areas of population and jobs away from central cities and into the suburbs has further exacerbated this trend. The result has been the continued dispersion of homes and workplaces and the rise of the suburb as the dominant pattern of development in the American geography. More disparate residential and workplace locations make carpooling more difficult, and may be a key driver in the sharp decline in ridesharing observed during the 19SOs, and the consequent rise in SOV share. The Role of Demographics Surely the concentration of population on one end of the work trip and jobs on the other make the opportunities for carpooling more numerous, other things being equal. But there may be other trends at work at an even finer level of geography that may also influence these opportunities. The most recent data from the NETS reveal that carpooling is most likely to occur among people living in the same household. For example, over 62% of two person carpools were formed with other household members in 1990, and this proportion has increased from 50% in 1983 (perhaps another result of the 3
dispersion of residences and jobs described above ).7 It would follow logically that the number of persons per household, or (more precisely) the number of workers per household, would be a factor in the rates of carpooling and SOV use we presently observe. Many have observed the changes in household and family structures that have been occurring in recent decades. Indicative of this trend is the fact that the proportion of the population that is either divorced or has never married, when adjusted for changes in the age distribution, has risen from 17.5% in 1970 to 31.~% in 1995.8 This and other socioeconomic and demographic changes have combined to shrink average household sizes markedly since 1960, as shown in Table 1. With fewer people per household, opportunities for carpooling are further reduced. Table 1. US Household Composition 1960-1990 ~ ~ ~ - ~ ~ ~ ~ . ~ ~ At. ~ ~ ~ . ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ . ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ . ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ . . . ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ !? ~ . ~ ~ Year I- ~ ~ ~ Percent Oine ~ ~ ~ ~ ~ Mean ~ ~- Mean : : ~ ~ ~ ~ ~ ~ Person ~ ~ ~ ~ Persons per ~ ~ Workers per Households I- Householcl ~ ~ Household ~ 1960 1 13.1% 3.33 1 1.22 1970 17.1% 3.14 1.21 1980 22.7% 2.76 1.20 1990 1 24.6% 2.63 | 1.25 Source: US Bureau of the Census, 1996. Socioeconomic Factors According to the Census Bureau, disposable personal income per capita rose almost 20% in real terms between 19SO and 1990,9 while real average household incomes increased 12%.~° At the same time, Figure 3 shows that the cost of new cars has actually been declining slightly when adjusted for quality. Even more striking though, is the drop in real gasoline prices that has occurred since 1980, and the fact that they are now at an all time low, as shown in Figure 4. These developments have made automobile ownership and use easier than ever, and it has continued to increase accordingly as shown in Table 2.~ Increased dispersion should make it more difficult to find a neighbor or co-worker to carpool with. To the extent that this may reduce the absolute number of. non-household carpools, other things being equal this would cause household-based carpools to increase as a percent of the total. 4
Figure 3. Measures of New Car Prices 1968 -1993 (thousands of dollars) 25 20 15 10 O ~ 68 70 75 80 85 90 - r Year Average market price per vehicle Market price of constant-quality vehicle Source: DRI/McGraw Hill, 1994. CPI-adjusted price of constant-quality vehicle Figure 4. Real Average Price of Gasoline in the US 1920-1993 160 ~ 150 O 140 cn 1 30 Cal ~ 120 - O 110 Cal C) 100 90 C: 80 70 ~ :~ : ~ ~ 1920 1930 1940 1950 1960 1970 1980 1990 Source: Oil & Gas Journal Energy Database. 5
Table 2. Vehicle Availability 1960-1990 :~ Year ~ ~ ~: ~Vehicles per ~Vehic es : Capital I: ~Household 1960 0.31 1.03 1970 0.39 1 .25 1980 0.57 1 .61 1990 0.61 1 .66 . .. . . .. . . . .. . .. Source: Federal Highway Administration. While the dispersion of workplaces and residences, combined with shrinking household sizes, will reduce the opportunities to carpool, more available vehicles that are also cheaper to own and use will tend to reduce the need to carpool, even among those who have good ridesharing alternatives to driving alone. The Decline in Transit Shares As we have seen, much of the increase in SOV shares observed in the last twenty years came at the expense of carpooling, but the remainder came largely from eroding public transit shares. It would appear that many of the same factors at work in the decline of carpooling may also be involved in the reduction of public transit use in the US. Surely having more household vehicles available might lessen the need to take transit, or at least provide an alternative. Moreover, this alternative has become all the more competitive with the drastic fall in real gasoline prices since 1980, and the fact that cars are now better made and more comfortable than ever, while at the same time cheaper to own. Likewise, even absent these developments n sing incomes have made them easier to afford. But perhaps the factor that the drop in carpooling and the loss of transit ndership have most in common is geography. Because transit services must be provided over a fixed network, they rely heavily on a concentrated base of patrons in order to provide services that are competitive while efficient. It is reasonable to assume that people who can afford to use a private vehicle will want to do so. Private vehicles have certain service qualities that no form of public transit can match and that people value very highly, such as privacy, route flexibility, and infinite departure schedule variability. 6
Transit use and the density of development We can certainly observe this apparent truism in that the larger and most densely developed metropolitan areas tend to be the ones with the highest transit shares. Table 3, taken from our report for TCRP project H-4A,~2 summarizes the use of transit for travel to work in the 39 metropolitan areas whose 1990 populations exceeded one million, and shows that between them, these metropolitan areas account for seven out of every eight of the nation's transit tnps. Table 3. Transit share of work trips in major metropolitan areas, 1990 Population density (persons/sq. ml.) Transit share to work (onto) New York City2,446 26.9 4 MSAs with population 3+ mn., density 1,000+ per sq. ml.1,238 10.5 6 MSAs with population 3+ mn., density < 1,000 per sq. ml.573 5.8 10 MSAs with population 2 to 3 mn.522 4.6 18 MSAs with population 1 to 2 mn.470 3.3 A1139 MSAs with population 1+ mn.664 8.7 Source: Charles River Associates (1997~. New York City is always an extreme outlier in any discussion of the national transit picture. While completely atypical of the situation throughout the rest of the country, it is nevertheless an extremely important outlier. New York alone .. , . .. . . as. accounts for over one-third of the nation's transit trips. The population density in the New York metropolitan area is over 70% greater than that in the next most dense metropolis (Chicago), and transit's share of trips to work in New York (27%) is about twice as high as for the nearest contenders (Chicago and Washington DC). Four other metropolitan areas had 1990 populations of over three million and residential densities exceeding :1,000 people per square mile: Chicago, Boston, Philadelphia, and Miami. With the exception of Miami, transit's share of work trips exceeded 10% in each of these cities. 7
The importance of core area t rave! We can extend this argument further by noting that the primary market for transit services, in cities of any size, tends to be the residents of the central city of the metropolitan area, traveling to and from jobs also located in the central city. This is a logical result, given that other things being equal, it is there that we expect to find the highest concentrations of people and jobs. The importance of central city trips to transit is still very apparent, although the story is necessarily a more complicated one. The statistics in Table 4 show that in 1990 about three-quarters of all transit trips made to work by metropolitan area residents were made to central city workplaces, and that three out of every five transit work trips were ascribable to central city residents commuting to central city jobs. It can also be seen from the exhibit that the two geographical markets of greatest importance to transit-intra-central city commutes and suburb-to-central-city commutes are the ones showing the least proportional growth over the 198Os, 6.2% and 9.6% respectively. Moreover, transit's market share fell most heavily (by almost a third of the 1980 level) for the suburb-to-central-city work trips, and the Toss in market share for the intra-central city commutes, while proportionately much less, was still quite marked. As a result, the number of intra-central city transit work trips fell by almost 13% (or 475,000 trips per day), and from the suburbs to the central city the decline was over a quarter of all 1980 trips (or a loss of about 300,000 trips per day). These results suggest quite strongly that the dispersion of population and jobs away from the central city described above may be a primary factor in the decline in transit use. Given that transit depends heavily on the concentration of people and jobs made possible by the dense core of an urban region, its fixed service networks are ill-equipped to serve the increasingly diffuse geography we currently observe. 8
Table 4. Transit market share for work trips by metropolitan area residents, 1990 and 1980 1990 1 980 Change, 1980 to 1990 Commute trips by all modes (millions) central citytocentral city24.322.9+6.2% suburbsto central city15.313.9+9.6% suburbs to outside MSA6.83.9+72.3% suburbs to suburbs35.427.7+27.5% centralcityto suburbs6.04.6+29.8% centralcityto outside MSA1.91.3+48.1% all commute trips by MSA residents89.674.4+20.5% Transit share of commute trips (DO) central city to central city13.216.1-18.0% suburbs to central city5.38.0-33.2% suburbs to outside MSA6.47.6-15.4% suburbs to suburbs1.21.6-28.0% central city to suburbs5.15.6-8.5% central city to outside MSA8.37.3+13.7% al/ commute trips by MSA residents6.07.9-24.9% Distribution of transit commute trips by geography (TO) central city to central city60.262.5 suburbs to central city15.018.9 suburbs to outside MSA3.15.1 suburbs to suburbs8.07.5 central city to suburbs5.74.4 central city to outside MSA3.01.6 all commute trips by MSA residents100.0100.0 Source: Charles River Associates (1997~.
Implications of These Trencis The trend in recent years has been for population and employment to become more concentrated in metropolitan areas, but at the same time it has aggregated in diffuse suburbs rather than dense central city areas. The vastly larger array of origins and destinations that must now be served for commuting trips makes both carpooling and transit use far less feasible, and the statistics marking their decline reflect this. A continued decline in ridesharing and transit use, particularly for peak period commute trips in major metropolitan areas, would obviously place increased strain on the nation's urban transportation system. Per traveler or per passenger- m~le, single-occupant vehicles consume more road space and fuel, and result in greater noxious emissions, than any other mode of local transportation. As the above discussion makes clear, modal shares and private vehicle occupancy levels are influenced, over time and across localities, by a wide variety of primary and intermediate factors: demographic/socioeconomic variations, private vehicle availability, development patterns, transportation facilities and services, pricing, and so on. Many of these factors are largely outside the direct control of public agencies, but others may be influenced - directly or indirectly - by public investment, financing, pricing, regulatory, or promotional policies. PROJECT OBJECTIVES The primary goal of this work is to gain a stronger quantitative understanding of the factors that have been affecting the dynamics of commuter mode choices and vehicle occupancy levels in recent years, and to use this improved understanding to comment on the potential efficacy of policies designed to increase transit ridership and/or to increase average vehicle occupancy levels. The potential impacts of continuation of the trends in single occupant vehicle (SOY) use are numerous. They include both relatively direct, near-term effects, as well as consequences that would be realized only over a period of many years. The short-term effects include such relatively easy-to-observe factors as energy consumption, moving source emissions levels, and the patterns of use of transportation facilities. The longer-term effects, however, may include such macro-level trends as physical development patterns and metropolitan area dispersion, regional and national economic clevelopment, air quality, and so on. 10
While such phenomena are plausibly linked with travel conditions and behavior, the causal relationships are not well understood. It's just not clear, for example, how far transportation (either private or public) drives dispersion of homes and jobs, or how far the causality flows in the opposite direction. The marketplace evidence appears to indicate that, at least under the public regulatory and financial ground rules that have prevailed throughout the twentieth century, the general public has aspired very strongly to both lower density living and increased automobility. Their decisions about where to locate residences and businesses are influenced by a myriad of factors. Transportation considerations are traded off simultaneously against development densities, property values, the quality of educational and other public services, social linkages, macroeconomic factors, and so on. Disentangling credible causal relationships about the minutiae of land use and transportation interactions is a formidable undertaking, and perhaps an ultimately impossible one. Similarly, the causal links between local air quality and vehicular use patterns is an almost equally difficult analytical problem. It is true, of course, that there are extant quantitative lance use and air quality models that try to capture at least the key ways in which urban form and air pollution are correlated with transportation supply and demand characteristics, given existing (or past) public financing and regulatory postures. The deficiencies of these models are well-recognized, at least by the analysts if not fully by the public policy process that desperately needs some quantitative basis, however deficient, for setting and enforcing standards, promoting policies, forecasting outcomes, and evaluating program performance. So, identifying the more direct, near-term potential impacts of continued SOV growth is inherently more tractable than trying to predict such lonaer-run · ~ ~ . ~ te ~ ~ . ~ ~ ~ ~ ({ · . ~ implications as ~ developmenVurban 0,77 ~ environmental quality," "economic productivity," or "social effects," all of which are influenced by very many forces as well as transportation. In deciding how best to focus the project resources, therefore, tractability was the chief criterion. In addition, we also considered two other important factors: · Overlap with other, more specialize`] TCRP studies. The relationship between transportation considerations and metropolitan area development was considered in much more detail in Project H-l, An Evaluation of the Relationships Between Transit and Urban Form, and is also the subject of current Project H-iO, The Costs of Sprawl - Revisited. Some of the social impacts of transportation policy are being explored in Project H-S, 11
Using Public Transportation to Reduce the Economic, Social, and Human Costs of Personal Immobility. Relevance and complementarily to the previous Project H-4A work. Project M-4A, to which this effort was an adjunct, was concerned with choice of urban travel mode, and in particular, with the influence of various public policies on travel choices. it was therefore intended that this new work add to the understanding of the mode choice process developed in the previous project. For reasons like these, we chose to concentrate on the most direct, near-term implications of the trends in mode choice and private vehicle occupancy levels. in order to be most relevant to the transit community, we also wanted to focus attention on those market situations in which transit has traditionally been regarded as a viable, cost-effective alternative to lower-occupancy modes, primarily commute trips within central areas or between the suburbs and central areas. Most of the existing discussion of trends in commuting mode shares has been primarily descriptive in nature, as distinct from analytical. Therefore, the specific objectives of this project are: . LITERATURE REVIEW To explore credible, quantitative causal relationships concerning the variations and trends in mode shares and in vehicle occupancy levels, particularly in those market situations in which transit has traditionally been considered most competitive with private transportation, using multivariate analysis techniques; To use the identified quantitative relationships to assess the extent to which these observed trends can be explained by factors within the control of transportation policymakers; and To comment on the implications of our findings for potential public policies designed to favor higher occupancy modes (such as road pricing proposals). The societal and technical implications of increasing reliance on private transportation has been examined before, in much larger studies than this one. One such effort that comes to mind was cauied out in the late 1970s by the Congressional Of flee of Technology Assessment. ~ 3 While some of the 12
assumptions and discussion in that report now inevitably appear dated, the study did a strong, even-handec! job in addressing both the negative and positive impacts of continued growth in automotive travel, and the many of the major conclusions still appear to be valid. The study accurately predicted, for example, that auto travel would increase markedly through the 19SOs and 199Os despite stricter fuel economy standards and incentives to reduce pollution, and that as a consequence, air quality standards would be increasingly violated and urban mobility would be hampered by increased congestion levels. There have been several other large studies on the effects of auto ownership and urban form on commuting mode choices. A 1976 study by Kain and Fauth of Harvard University developed a large number of models of automobile ownership and use based on a household sample from the 125 largest metropolitan areas in 1970.~4 These models used as explanatory variables measures such as highway capacity, transit service levels, housing density, and socioeconomic factors such as household income. In a report of the same year, Ben-Akiva and Lerman developed disaggregate models of auto ownership and mode choice from Washington DC data for 1968.~s This study tested the hypothesis that auto ownership and mode choice decisions are interdependent, and are therefore made jointly. The resulting models confirmed this hypothesis, and related auto ownership and use to auto and transit service levels and costs, as well as socioeconomic land use/urban structure variables. A 1981 study by Zahavi, Beckmann, and Golob examined the simultaneous process of housing location choice and mode choice through the development of "travel probability fields", mode] parameters which allow travel patterns to be related to urban structure in a dynamic feedback process. Some of the best-known work in this area, and perhaps the work most closely resembling the present study, is that by the Australian authors Newman and Kenworthy. Their 1989 book Cities and Automobile Dependence: A Sourcebook, analyzes the relationship between auto use and variety of land use and other factors across a series of major cities throughout the worId.~6 This analysis is used to develop a city typology with respect to "automobile dependence", which is then used to classify each of the cities considered in the study. The book also provides a thorough compendium of statistics for each city, as well as an excellent bibliography on the subject. A little more recently, Prevedouros en c! Schofer analyzed auto ownership and use in two low-density, outer ring growing suburbs and two high-density, inner-ring stable suburbs.~7 This 1992 paper found the stage in the lifecycle to be a major 13
determinant of automobile ownership, and that the number of workers, people of driving age, and suburban residence locations are also positive factors. Transit use was found to influence auto ownership negatively, as was the location of workplaces in the central city. The availability of company cars was found to be a positive influence on automobile use. A 1996 paper by Schimek reexamined the issue of residential density and its effect on automobile use. This analysis estimates a mode} of vehicle travel that includes vehicle ownership as an intermediate factor and assumes travel and density are determined simultaneously. The author finds that density has a negative effect on auto travel, but that this effect is quite small, and smaller than the influence of income. On the subject of journey to work trends in mode choice with which this study is primarily concerned, there have been several important national-level analyses. The best known of these is Alan Pisarski's Commuting in America survey of commuting patterns, cited earlier. The study was originally carried out in 1987 using data from the 1980 Census, and has recently been updated with new data from the Census of 1990. These studies examine national trends in commuting flow patterns, mode choice, demographic and socioeconomic measures. Another recent synthesis of national commuting trends using data from both the Census and NETS and also offering some insights into the future of commuting alternatives was published in 1994 by National Urban Transit Institute at the University of South Florida.~9 This study provides recommendations for the future success of public transportation, ridesharing, and working at home. At a slightly more disaggregate [level, there have been two reports commissioned by the Federal Highway Administration that examine these trends for the nation's largest metropolitan areas. The original report, titled Journey-to-Work Trends Based on the 1960, 1970, and 1980 Decennial Censuses, was also co-authored by Pisarski and published in 1986.2° The more recent version by Rosetti and Eversole2~ is an update using data from the 1990 Census, and contains a large amount of additional data and tabulations. A series of special reports have also been prepared based on data from the 1990 NPTS. They include volumes on mode of travel,22 demographics,23 and trip and vehicle attributes.24 Each volume contains several papers on specific topics of interest, including travel by women and the elderly, the decline in carpooling, and trip chaining. 14
Work has recently been cattier out by the Hickling Corporation under funding from the FTA's Office of Policy, to collect and analyze data from some 17 prototypical, diverse congested corridors in US metropolitan areas.25 This work attempts to quantify the magnitude of economic inefficiencies represented in current traffic patterns in those corndors, following a Mogridge/Glaister type of mode! used to analyze transit infrastructure investments in London and other cities. In addition, it explores appropriate road pricing policies that would redress those inefficiencies, and (as a second best solution in the absence of marginal cost pricing of road space) identifies feasible transit investment and operational policies that might also lead to net societal benefits in the corridors. Finally, the results of TCRP Project H-4B, Transit Markets of the Future The Challenge of Change, provide a detailed analysis of future economic, demographic, social, land use, and transportation policy trends and the implications of these trends for the future of public transit in the US.26 THE STRUCTURE OF THIS REPORT The remainder of this report describes the details of the research and analysis that was carried out for the project. Chapter 2 provides details on the development of the databases used in the analyses. The discussion includes a review of relevant data sources that have been used or are potentially useful in studies of this nature, and the subsequent compilation of both cross-sectional and time-series data for use in the analysis. Chapter 3 presents the main findings from the study, including the results of the mode} estimation using the datasets descnbed in Chapter 2. Several models are presented, for several market segments. These results include both cross-sectional and time-genes analyses. Chapter 4 presents our main conclusions based on these findings and an analysis of the implications of the results. The conclusion of the chapter also provides recommendations for further research. Finally, a series of appendices provide more details about the data used in the study. NOTES: Calculations by CRA from data presented in Rosetti, M.A. and Eversole, B.S., Journey-To-Work Trends in the United States and its Major Metropolitan Areas, 1960-1990. Washington, DC, US Department of Transportation, Federal Highway Administration, (1993~. 2 Pisarski, A. E., Commuting in America 11. Lansdowne, VA, Eno Foundation for Transportation, Inc. (1996), p. 18. 15
3 US Bureau of the Census, Statistical Abstract of the United States 1996. Washington, DC, (1996), Tables 40 and 41. 4 Pisarski, p. 19. s Ibid. 6 Pisarski, p. 26. 7 Vincent, M.J., Keyes, M.A., et. al., Nationwide Personal Transportation Survey: 1990 NPTS Urban Travel Patterns. Washington, DC, US Department of Transportation, Federal Highway Administration, ( 1994), Table 5-7. ~ US Bureau of the Census, op. cit., Table 58. 9 US Bureau of the Census, op. cit., Table 700. a US Bureau of the Census, Statistical Abstract of the United States 1992, Washington, DC, (1992), Table 698 and US Bureau of the Census, Statistical Abstract of the United States 1982-83, Washington, DC, (1983), Table 711, with adjustment to constant dollars using CPI from US Bureau of the Census, Statistical Abstract of the United States 1996, Washington, DC, (1996), Table 745. i~ Rosetti, M.A.and Eversole, B.S., op. cit. i2 Charles River Associates, "Building Transit Ridership An Exploration of Transit's Market Share and the Public Policies that Influence It." to be published as TCRP Report H-4A (1997~. |3 US Congress, Office of Technology Assessment, Technology Assessment of Changes in the Future Use and Characteristics of the Automobile Transportation System (3 volumes). Washington, DC, US Government Printing Office (1979~. {4 Kain, J.F. and Fauth, G.R., The Elects of Urban Structure on Household Auto Ownership Decisions and Journey to Work Mode Choices. Prepared for the US Department of Transportation, Cambridge, MA, Harvard University Department of City and Regional Planning (1976). Is Ben Akiva, M.E. and Lerman, S.R., A Behavioral Analysis of Automobile Ownership and Mode of Travel (3 volumes). Washington, DC, US Department of Transportation, (1976~. i6 Newman, P.W.G. and Kenworthy, J.R., Cities and Automobile Dependence: A Sourcebook. Sydney, Australia, Gower Technical Press (1989~. |7 Prevedourous, P.D. and Schofer, J.L., "Factors Affecting Automobile Ownership and Use." Transportation Research Record 1364, pp. 152-160. ~8 Schimek, P., "Household Motor Vehicle Ownership and Use: How Much Does Residential Density Matter?" Transportation Research Record 1552, pp. 120-125. ~9 Ball, W.L., Commuting Trends in the United States: Recent Trends and a Look to the Future. Prepared for the US Department of Transportation, Tampa, FL, Center for Urban Transportation Research, (1994~. 20 Briggs, D., Pisarski, A., and McDonnel, J., Journey-to-Work Trends Based on the 1960, 1970, and 1980 Decennial Censuses. Washington, DC, Federal Highway Administration, (1986~. 2} Rosetti, M.A.and Eversole, B.S., op. cit. 22 Nationwide Personal Transportation Survey: 1990 NPTS Travel Mode Special Reports. Washington, DC, US Department of Transportation, Federal Highway Administration, (1994~. 23 Nationwide Personal Transportation Survey: 1990 NPTS Demographic Special Reports. Washington, DC, US Department of Transportation, Federal Highway Administration, (1995~. 24 Nationwide Personal Transportation Survey: 1990 NPTS Special Reports on Trip and Vehicle Attributes. Washington, DC, US Department of Transportation, Federal Highway Administration, (19951. 25 The Role of Transit in Congestion Management. Hickling Lewis Brod Inc., prepared for the Federal Transit Administration, ( 19961. 16
26 The Drachman Institute for Land and Regional Development Studies, "Transit Markets of the Future The Challenge of Change." To be published as TCRP Report H-4B (1997), Washington, DC, Transportation Research Board. 17
