Understanding Changes in Demographics, Preferences, and Markets for Public Transportation (2018)

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.

35 The characteristics of the residential setting, in terms both of physical characteristics and of transportation options offered, have a profound effect on the frequency of transit use. Chapter 3 includes five sections: 1. Trends in location of residence, 2. Patterns needed to support transit, 3. Trends in the location of employment, 4. Telecommuting and working at home, and 5. Defining new ways to understand neighborhood characteristics. Trends in Location of Residence This section reviews how changes in settlement patterns at the end of the previous century are associated with the decrease in transit use during those decades. In the last half of the twentieth century, populations in the central cities stagnated while all growth occurred in sub urban areas. Over a 60-year period, the nonmetropolitan areas lost population. Changes in Density Patterns Over Several Decades Trends in development since the mid-20th century may be broadly characterized as a move- ment from rural to metropolitan areas. These patterns also suggest a dispersion of both popula- tion and employment outward from historical central cities to the suburban areas surrounding those cities. This pattern of growth, substantially enabled by the popularity of the automobile and programs of major roadbuilding, produced patterns of development most suited to travel by automobile and unfavorable to transit. These trends challenge the design, performance, and utilization of transit service moving forward. Interestingly, early 21st century developments are upending the narrative of the past 50 years or more. Urbanized living has found a new attractiveness among various population segments: those of the millennial and generation X (gen X) cohorts who appear to be attracted to the vital- ity of cities and the opportunity to live without or with less dependence on a car, single profes- sionals or those who are married without children, and, increasingly, empty-nesters and retirees who no longer want the maintenance burden of a large single-family house in the suburbs. To serve these evolving markets, cities have been undergoing major revitalization and infill, drawing a new class of developers to the marketplace; and since the working version of this often highly educated population is regarded as the more highly skilled creative class, employers are also trending back into the center city and surrounding city-like urban centers. These trends, whose effects remain to be quantified, offer a promising lift to transit in overcoming the challenges of the past half century. C H A P T E R 3 Variation in Transit Use by Neighborhood Type and Urban Form

36 Understanding Changes in Demographics, Preferences, and Markets for Public Transportation Residential Settlement Patterns Were Not Supportive of Transit Transit is at its most effective and attractive as a transportation mode when it can provide frequent on-time service, have few restrictions on its travel speed, serve a finite number of stops (to not overly diminish average speed), and be readily accessed on foot at both the origin and destination of a trip. When these criteria are met, transit can provide service that is competitive with—or even superior to—that of the private automobile, and thus become a preferred choice for travelers. The setting in which these performance criteria are maximized is an urbanized region with moderate to high densities along the route and service to one or more nodes with significant attractions. These attributes allow transit to realize a sufficient mass of potential trip flows to justify a high quality of transit service with frequent headways and adequate speeds. Clearly, this description does not fit the growth patterns seen in the United States over the past half century. As a result, it is unsurprising to observe the corresponding decline in the share of person trips made by transit over that period (even though total transit trips have increased with overall growth in travel). Although the land area of the United States is still substantially rural (only about 5% has been developed), the proportion of the population living in metro- politan areas grew from 46% in 1910 to more than 80% in 2010. Significant differences exist in where that metropolitan growth occurred and in the form it has taken. Most metropolitan growth over the past 60 years has gone into the areas outside of central cities. The area at sub- urban or exurban densities (one dwelling unit per 1 to 40 acres) covers 15 times the land area developed at urban densities (National Land Cover Database 2016). Figure 19, taken from Commuting in America 2013, vividly shows not only the growth nationally in metropolitan areas that occurred between 1950 and 2010, but the ever-expanding proportion of that growth that occurred in the suburbs outside of the central cities (Pisarski and Polzin 2013). From a share of only about 20% of all population in 1950, the suburbs grew to be the location of more than 60% of all the U.S. population by 2010. Where the Growth Occurred Over Several Decades Commuting in America 2013 also points out that where this growth occurred is also related to the size of the metropolitan area and the region of the country (Pisarski and Polzin 2013). Source: U.S. Census Bureau (as cited in Pisarski and Polzin 2013, Brief 4, Fig. 4-10). Figure 19. Population growth within metropolitan areas, 1950–2010.

Variation in Transit Use by Neighborhood Type and Urban Form 37 Most of the growth between 1990 and 2010 appears to have occurred in large metropolitan areas with populations of 5 million or more. Figure 20 illustrates how the greatest population growth since 1990 occurred in the Southeast and Southwest regions of the United States. These areas have historically had the lowest rates of transit use attributable to having grown under the shaping force of the automobile. With some important exceptions (e.g., Chicago, Minneapolis, and Milwaukee), the Midwest shares these land use characteristics and the low rates of transit use; however, population growth in the Midwest has occurred at even lower rates than in the Northeast. Transit-Supportive Conditions Analysts have studied key factors to try to categorize activity level thresholds necessary to support types of transit service. Pushkarev and Zupan developed a set of screening criteria back in the early 1980s, suggesting the appropriate levels of residential density and destina- tion activity levels to support several types of transit, from simple bus to heavy-rail and sub- way service (Pushkarev and Zupan 1982). Table 3, taken from TCRP Report 95, Chapter 15 (Kuzmyak et al. 2003), presents a more recent effort sponsored by the Institute of Transportation • Highest growth rates in Southeast and Southwest. • Newer areas developed after the automobile. • Much lower rates of transit use. Source: U.S. Census Bureau, American Community Survey; Pisarski and Polzin 2013, Brief 4, Table 4-10. Figure 20. Population growth trends and transit use rates by region for 40 largest metropolitan areas, 1990–2010.

38 Understanding Changes in Demographics, Preferences, and Markets for Public Transportation Engineers (ITE) that provides a similar and slightly simpler classification. It suggests that for even a basic transit service consisting of a local bus with 1-hour headways, a density of four to six dwelling units per acre is needed, and the bus route should be serving an activity center of at least 5 to 8 million square feet of attractions (office and commercial space). To justify a bus with 30-minute headways requires a minimum of 7 to 8 dwelling units per acre and a destina- tion activity level of 8 to 20 million square feet. Of course, this is minimalist transit service; to support rail transit service with 5-minute headways requires a minimum of 9 dwelling units per acre and more than 35 million square feet of commercial/office space at the destination. Most of suburban America does not achieve such densities, but neither do most activity centers outside of the regional central business district (CBD). Trends in Employment Location The Move to the Suburbs Patterns in residential development and densities are only part of the challenge for transit. Employment also pushed out to the suburbs: first, retail and service activity to serve early resi- dents in the 1950s and 1960s, and later—in the 1970s and 1980s—office and other professional employers in pursuit of lower rents and (seemingly) greater access to employees. The retail and service activities found it most effective to locate in shopping centers or commercial strips adja- cent to major highways to serve visitors arriving by auto. Meanwhile, office and other professional activities gravitated to office parks and isolated campus locations, presuming that employees would drive to their jobs and be provided with free parking. Sometime in the late 1980s, the trends in outward movement of employment resulted in most metropolitan area employment nationwide being located outside central cities. Table 4 indicates that as of 2010, about 30% of all jobs (35% of metropolitan area jobs) were in central cities, while about 42% (50% of metropolitan area jobs) were in the suburbs (Pisarski and Polzin 2013). The 10% of jobs (11.5% of those in metropolitan areas) located in principal cities were a toss-up for transit: while these areas may possess urban characteristics, they are still generally located in the suburbs with surrounding lower densities and suburban road networks and are not necessarily conducive to transit use. Transit routes connecting inner suburbs to major employment concentrations in regional CBDs could offer reasonable levels of service and be competitive with auto travel. However, the dispersion of both population and employment has made it increasingly difficult to connect the dots and compete efficiently with transit in this many-to-many marketplace. A New Pattern for Relocation of Jobs Since 2005? Data with which to examine employment location trends in detail after 2005 are not readily available. Table 5 shows that as of 1996, the share of metropolitan area employment located Table 3. ITE-recommended minimum densities to support transit service. Mode and Service Level Residential Density (dwelling units/residential acre) Employment Center Size (million ft2 commercial or office space) 1 bus/hour 4–6 1 bus/30 minutes 7–8 5–8 8–20 Light rail or feeder buses 9 35–50 Source: Holzclaw 1994, as cited in Kuzmyak et al. 2003.

Variation in Transit Use by Neighborhood Type and Urban Form 39 in the regional core (defined as within 3 miles of the city center) had dropped to 23.3% from about two-thirds in 1950 and one-half in 1963 (Cortright 2015). A study by Kneebone (2009) suggested that employment decentralization continued through at least 2006. This work found that the percentage of metropolitan employment located in the core reached a low of 21.3% in 2006. Suburban job growth averaged 1.8% annually over the period from 1996 to 2006 as compared with only 0.1% in the city centers (of the largest 98 metropolitan areas). The general sense among researchers and representatives of the development community is that a turning point in metropolitan employment trends from suburban to urban occurred somewhere in conjunction with the Great Recession. Between 2007 and 2011, job growth in city centers grew at an average annual rate of 0.5%, while the suburbs lost jobs, shrinking by 0.1% annually (Smart Growth America 2015). Following the recession, jobs based outside cities, such as construction and manufacturing, were hit much harder than urban jobs such as business services. Jobs disappeared everywhere, but the reversal occurred more rapidly outside cities (Miller 2015). Assorted reasons are given for the presumed reversal. A study in 2011 by the Urban Land Institute provides its members in the international real estate industry with a prospective view of key trends likely to shape the industry in the coming decades, particularly in the wake of the recent recession (Urban Land Institute 2012). The study cites several economic and demographic Source: Summary of American Community Survey as cited in Pisarski and Polzin 2013, Brief 5, Table 5-1. *For purposes of analysis, total U.S. jobs set to equal workers. Table 4. Geographic distribution of population, workers, and jobs, 2010. Employment Year 98-MSA Total 3-Mile Share (%) City Center Periphery 1996 70,159,860 23.3 16,347,247 53,812,613 2006 77,411,492 21.3 16,488,648 60,922,844 Source: Kneebone 2009, Appendix A, page 17, as presented in Cortright 2015. Note: Annual growth rate, 1996–2006: city center = 0.1%; periphery = 1.8%; MSA = metropolitan statistical area. Table 5. Zip code–based estimates of city center and periphery employment.

40 Understanding Changes in Demographics, Preferences, and Markets for Public Transportation drivers that it suggests will dramatically influence future real estate investment and urban devel- opment decisions. The Urban Land Institute report argues that these trends in the wake of the recession illustrate systematic changes in the perception of and planning requirements for future cities. The report advises its members that the past decade provides an opportunity to rethink and evolve, reinvent, and renew. It asserts that nearly 100% of robust growth is in urban areas and that growth is embracing a new mixture of land uses, new suburban centers, and in-town reconfigurations. The 2015 City Observatory study provides additional detail and insight into the changing shape of urban areas (Cortright 2015). The study explores what appears to be a growing prefer- ence for urban living, in which it observes the following: • Some young adults are showing a clear preference for close-in urban locations. Cortright reports that between 2000 and 2012 the number of 25- to 34-year-old adults with at least a bachelor’s degree who chose to live in city centers outnumbered those who chose the metro- politan area by a rate of 2 to 1. This suggests that highly educated individuals are choosing city centers. The 2016 TCRP survey showed about 37% of individuals between the ages of 25 and 34 preferred the big city, leaving more than 60% reporting a preference for less-dense residential locations. The results of the analysis support the conclusion in Chapter 1 of the present report that preference for city living by millennials increased between 2004 and 2016, but not necessarily for most of them. • In 2010, college-educated young adults were 126% more likely to live within 3 miles of the center of the CBD of a large metropolitan area than other metropolitan residents, up from about 77% more likely in 2000 (Cortright 2015). • This finding is important because this group of people constitutes an important source of labor for fast-growing knowledge-based firms, which seem willing to alter their growth or expansion plans to tap this labor pool. Perhaps the most definitive study of the changes underway in the shaping of metropolitan areas was performed by Smart Growth America in partnership with Cushman–Wakefield and the George Washington University Center for Real Estate and Urban Analysis. The study report, Core Values: Why American Companies Are Moving Downtown, summarizes a research project that identified almost 500 companies that relocated, opened new offices, or expanded in walkable downtowns and investigated their underlying characteristics, motives, and preferences (Smart Growth America 2015). That project identified key aspects of industries that require highly edu- cated, technically competent young people for whom the urban environment is desirable. The project also documented recent patterns of job relocation by these firms to certain downtown areas (see Technical Appendix 3). Telecommuting and Working at Home One major trend affecting the number of commuting trips taken is the propensity to work at home, and thus have no trip to work. The share of workers who report working at home doubled between 1980 and 2013, according to the Census journey-to-work data shown in Fig- ure 21. Figure 22 shows that the propensity to work at home increases directly with age—a pattern quite different from the propensity to have a job and accept the option of telecommuting for some days, a pattern which decreases with age, as discussed in Chapter 7. Employers incentivize nonauto commuting in different ways. The types and amounts of incentives offered to employees vary widely. Young millennials seem to be more likely to have a job with an employer that offers a bike subsidy, a shuttle to the transit station, or a subsidy on either the transit fare or vanpooling charges. As expected, millennials have a much higher pro- pensity to use the bike incentives offered at the workplace than their older colleagues. Figure 23

Variation in Transit Use by Neighborhood Type and Urban Form 41 Figure 21. Growth of work-at-home share over 33 years. Source: McKenzie 2015. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 1980 1990 2000 2013 Year of U.S. Census S ha re o f J ou rn ey to W or k (% ) Figure 22. Effect of age on work-at-home commute share. Source: McKenzie 2015. 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 16–24 25–29 30–34 35–44 45–54 55 Age Group S ha re o f J ou rn ey to W or k (% ) Figure 23. Incentives used to encourage nonauto commuting, by age group. Source: TransitCenter 2014. 0 5 10 15 20 25 30 35 40 Bicycle Flextime Shuttle service Vanpooling Telecommuting Use by Millennials Use by Older Group N on au to C om m ut in g (% )

42 Understanding Changes in Demographics, Preferences, and Markets for Public Transportation shows that millennials are far more likely than their older colleagues to use the subsidized shuttle service and the subsidies for transit and vanpooling. Flextime use occurs more often with those over age 35 than with those under 35. The 2014 TransitCenter survey found that age was not a major factor in using telecommuting. Defining New Ways to Understand Neighborhood Characteristics Going Beyond Density Density is a widely used measure to describe the “urbanicity” of a setting and its suitability for transit service and use. However, research on the interplay between land use and travel behavior has shown that the influence of land use is considerably more nuanced than just density. Pioneering work by noted researchers including Robert Cervero, Reid Ewing, and Kara Kockelman in the 1990s revealed that while density clearly played a role in terms of reflecting compactness and proximity, other contextual factors, such as the mix and balance of activities, the connectivity between those activities facilitated by design, and the level of accessibility to regional opportunities added critical insights in explaining travel behavior such as vehicle ownership rates, mode choice, and VMT generation (Cervero and Kockelman 1997; Ewing and Cervero 2010). These measures of land use have been widely referred to as the “four D’s”: density, diversity, design, and destinations. These advanced measures have been enabled by technological break- throughs in tools such as the geographic information system (GIS). GIS tools can sharpen the spatial resolution of the built environment to create relationships that were previously lost in large-area aggregation of the characteristics. Traditional transportation planning models used by metropolitan planning organizations were not designed to function at a level of detail fine enough to capture the critical interplay between the design of the microenvironment and its effect on walking, biking, and even transit use (given the importance of walk access/egress). GIS methods helped focus the analysis at the level of individual households, employment sites, or any number of trip generators or attractors. Subsequent research has placed greater emphasis on the relationship between accessibility and travel behavior. NCHRP Report 770 uses accessibility as a major strategy for understanding bicycle and pedestrian travel behavior (Kuzmyak et al. 2014). One of the new procedures devel- oped by the study focused on the calculation of accessibility scores for each mode and trip purpose by leveraging geospatial data on employment, households, travel networks, and other contextual data. The procedure for calculating the scores is similar to that of the popular Walk Score on the Internet, although the calculation is considerably more sophisticated. Geospatial layers of employment establishments are overlaid onto layers with the various transportation networks, and then sophisticated path-building methods are used to locate the number of oppor- tunities (e.g., jobs, grocery stores) that can be reached within a travel time budget. The numbers of such opportunities are added to calculate a total for any given location (e.g., a residence address); the value of each opportunity is reduced (decayed) by the amount of time required to reach it. This is done for each mode and for work- and non-work-trip purposes. The U.S. Environmental Protection Agency’s Office of Sustainable Communities’ Smart Location Database, a web-based resource, contains sociodemographic, employment, land use, and transportation information for each census block group in the contiguous states (Ramsey and Bell 2014). Among the more innovative variables brought into this version of the database are various measures of accessibility, including scores for regional access (from each census block group) to jobs by auto and by transit, and, conversely, employer access to workers, also by auto

Variation in Transit Use by Neighborhood Type and Urban Form 43 and by transit. In working with these data, Ramsey and Bell found a strong relationship between the ratio of the transit- and auto-accessibility scores and the transit/auto mode split for work trips (as captured in the American Community Survey). Creating Five Categories of Neighborhood Accessibility The research team accessed and attached these measures to the databases for the 2014 TransitCenter survey and the 2016 TCRP survey to create an objective measure of the quality of transit service at any of the respondent household locations. This study used five neighbor- hood types, ranging from the most transit oriented to the least. Table 6 summarizes eight demographic categories for each neighborhood type. Moving from the highest transit orientation toward the highest highway orientation, the populations become (1) increasingly old, (2) increasingly white, (3) increasingly married, and (4) increasingly unemployed. Along the same scale, populations become decreasingly employed full time, decreasingly students, and decreasingly without driver’s licenses. The research team used this accessibility ratio as another potential explanatory factor when analyzing the range of attitudes and choices captured in the 2016 TCRP survey. In Chapter 6 of this report, the five levels of neighborhood accessibility are applied in tandem with five age groups and four attitude-based market segments to obtain a better understanding of how key preferences vary in relation to demographic variation, geographic variation, and market seg- ment variation. Neighborhood Type Age (years) White (%) Single (%) Income ($) Student (%) Employment (%) Share Without Driver’s License (%) Full Time Unemployed Most transit- oriented 40.9 71.1 43.9 61,783 9.4 52.5 23.8 9.6 Transit-oriented 45.7 77.5 34.4 60,537 6.4 44.9 32.6 8.2 Mid 48.9 78.7 31.9 61,145 5.5 42.3 38.0 3.9 Highway- oriented 51.6 80.8 21.9 69,359 4.1 38.6 40.1 2.9 Most highway- oriented 52.7 85.8 21.7 70,763 3.3 38.4 44.1 2.4 Table 6. Demographic characteristics of the five neighborhood types, 2016.