Facts and Data Evaluation of Austin’s Proposed Transit System Indicates Continuing Death Spiral
How Vital Is Transit In Your Region? Part 1: Census Data
Transportation Policy Brief Number 4, by Randal O’Toole in The Antiplanner blog, May 21, 2019
Transit ridership is plummeting almost everywhere, yet officials in many cities are still devising hugely expensive plans for transit projects. One such city is Austin, whose leaders are talking about spending between $6 billion and $10.5 billion on new transit lines (and the final cost always ends up being more than the projections).
The need for these plans is contradicted by the rapid decline in transit ridership in Austin. Using Austin as an example, this policy brief will show how people in any urban area can use census data to find out just how important transit is to their region and whether it makes sense to spend a lot more money on transit. This is the first of two briefs on this subject; the next one will look at Department of Transportation data.
Since 2005, the Census Bureau has sent an annual questionnaire to about 3.5 million households a year asking, among other things, how those who have jobs in those households get to work. Known as the American Community Survey, these data can be downloaded for just about any geographic area — state, county, city, metropolitan area, urban area, congressional district, or zip code. Since the results are based on a sample, the Census Bureau does not publish data for small geographic areas because the margin of error is too high.
Data from every year from 2005 to 2017 can be downloaded from the American FactFinder web site. However, starting in July, the agency is transitioning to a new web site called data.census.gov. To avoid having to explain how to use a web site that will disappear in a few months, and to save you time using that site, I’ve already downloaded all of the tables that will be mentioned in this brief and posted them, with some enhancements such as calculations of percentages, for you to use.
The first question is how many people in the Austin urban area commute to work by transit and whether that number is growing or shrinking. This can be answered with table B08103, “means of transportation to work.” I’ve downloaded these data for the nation, states, counties, cities (or, in Census Bureau nomenclature, “places”), and urbanized areas and put them in one file for 2017 and, for comparison, a second file for 2007.
“Urbanized areas,” by the way, include all of the urbanized land in and around cities such as Austin, while “metropolitan areas” include all of the land, both urban and rural, in the counties surrounding such cities. I prefer to use urbanized areas since most people in rural areas aren’t going to have access to transit. However, the Census Bureau remaps urbanized areas with each decennial census, so the data from 2007 and 2017 aren’t based on exactly the same land area.
Transit’s Share of Commuting
The 2007 survey found that the Austin urban area (which is on row 684 of the spreadsheet) had about 560,000
workers in 2007, growing to nearly 890,000 by 2017 (row 736). That much growth shouldn’t be surprising because, on a percentage basis, Austin has been the fastest growing major urban area in America.
Of those employees, the 2007 survey found that about 22,000 of them (4.0 percent) usually took transit to work. Despite the nearly 60 percent growth in the total number of workers in the region, the number com- muting by transit shrank to well under 20,000, or just 2.2 percent, by 2017.
Even that number is probably considerably more than the number of people who actually take transit to work on any given workday. According to a 2017 Department of Transportation survey (see p. 78), people who say they “usually” take transit to work actually take transit only about 71 percent of the time while people who say they usually drive to work in fact drive almost all of the time. Correcting for this would require reducing transit’s numbers by almost 25 percent. I’m going to ignore this for the rest of this brief, as the adjustment factors may vary by state and region, but it’s likely that the American Community Survey probably overstates the number of people who commute by transit on any given workday.
Transit Commuting by Income
The American Community Survey also provides information on who rides transit to work. According to table B08119 for 2017, most Austin-area transit riders have low incomes, but their numbers are declining. Since 2007, the number of transit commuters earning under $35,000 a year declined by nearly a third while the number earning more than $50,000 a year nearly tripled.
Austin-area workers who earned less than $50,000 a year were significantly less likely to ride transit in 2017 than in 2007, while those who earned more than $50,000 a year were more likely to ride transit. People who earned more than $75,000 a year were twice as likely to commute by transit in 2017 as in 2007.
Though the number of high-income transit commuters is small—fewer than 7,500 transit commuters in
Between 2007 and 2017, the number of Austin-area transit commuters declined in every income bracket below $35,000, and grew in every bracket above $35,000.
2017 earned more than $35,000 a year—that is the only growth market for Austin transit. As a result, according to table B08121, the median income of transit riders grew by 85 percent between 2007 and 2017, while the median income of the region as a whole grew by only 49 percent.
Transit’s Share by Race
The American Community Survey also breaks down commute habits by race. According to table B08105B, the share of black workers commuting by transit declined from 7.7 percent in 2007 to 4.9 percent in 2017, while the share of non-Hispanic white workers commuting by transit declined from 2.6 percent in 2007 to 1.8 percent in 2017. The biggest change was among Latino workers, whose transit commute share declined from 5.1 percent in 2007 to 1.8 percent in 2017.
Between 2007 and 2017, the share of commuters who relied on transit declined for blacks and whites, but the decline was particularly large for Latinos. In this chart, “white” refers to non-Hispanic whites.
Latino commuting underwent another startling change: a decline in carpooling from 24.4 percent in 2007 to 14.6 percent in 2017. This contributed to an in- crease in the drive-alone share of Latino commuting from 65.1 percent in 2007 to 80.4 percent in 2017. It seems likely that Latinos significantly increased their motor vehicle ownership rates during this period
The Growth of Three-Car Households
While it isn’t broken down by race, table B08141 indicates that the share of Austin-area workers who live in households with no vehicles declined from 3.2 percent in 2007 to 2.7 percent in 2017, while the share who lived in households with three or more vehicles grew from 22.4 percent in 2007 to 29.2 percent in 2017.
Table B08141 also reveals that, as of 2017, little more than a quarter — 26.3 percent — of the people who live in households with no vehicles commuted by transit. This is down from 41.8 percent in 2007. People without cars were almost twice as likely to commute by automobile than by transit in 2017.
Curiously, more people who live in households without cars — 40.0 percent — commuted by driving alone
The number and share of Austin-area workers who live in households with three or more vehicles significantly grew between 2007 and 2017 while the share with no vehicles declined. This left fewer people than ever dependent on transit to get to work.
to work than by transit. How do they drive alone if they don’t have a car? Probably they use an employer-supplied vehicle.
In sum, American Community Survey data show that transit has become all but irrelevant for commuters in the Austin urban area. Less than 5 percent of black workers and less than 2 percent of both Latino and non-Latino white workers commute by transit. The number of low-income workers who rely on transit is rapidly shrinking, while transit’s only real growth market is among high-income workers who don’t need to have their commutes subsidized.
Growing automobile ownership is a likely explanation for transit’s decline. Yet, Transit no longer even works well for most commuters who don’t own cars.
The next policy brief will show how Department of Transportation data can be used to assess the value of transit in Austin and other urban areas. I’ll then make some recommendations for improving Austin’s transit system without spending $6 billion to $10.5 billion.
The Antiplanner, Randal O’Toole, is a transportation policy analyst and author of Gridlock: Why We’re Stuck in Traffic and What to Do About It as well as a review of Austin’s 2014 light-rail transit plan. The header photo on page 1 shows Austin’s Congress Avenue Bridge.
Summary of Downloadable Tables
B08301: Commute to Work 2007 2017
B08119: Commute by Income 2007 2017
B08121: Median Income by Mode 2007 2017
B08105B:Commute, Blacks 2007 2017
B08105L:Commute, Latinos 2007 2017
B08105H:Commute, Non-H. Whites 2007 2017
B08141: Commute by vehicles in H.H. 2007 2017
Transportation Policy Brief #3 http://ti.org/antiplanner/?p=16036
Transit Death Spiral: 1st Quarter Riders Down 2.6%
By The Antiplanner | May 14, 2019 | Policy brief, Transportation
Nationwide transit ridership in the first quarter of 2019 was 2.6 percent below the same quarter in 2018, according to data released by the Federal Transit Administration (FTA) last week. Transit’s most recent downward spiral began in 2014, and ridership over the twelve months prior to March 31 was 8.6 percent below the same twelve months four years ago.
Ridership is declining for all major forms of transit travel. First quarter bus ridership was 2.1 percent below 2018 while first quarter rail ridership declined by 3.2 percent. Commuter rail, light rail, heavy rail, and streetcars all lost riders.
Since transit agencies depend on fare revenues to cover part of their operating costs, declining ridership can force them to cut service or raise fares, either of which is likely to lose them more riders. This is known in the industry as the “transit death spiral,” and even major agencies such as the Bay Area Rapid Transit District (BART) are worried about it.
The FTA data show that first quarter ridership had fallen in all but twelve of the nation’s fifty largest urban areas. It even fell in Seattle, the one urban area that has, up until 2019, consistently shown ridership growth.
Ridership over the past four years has declined in every state except Washington.
Thanks to Seattle’s previous ridership growth, Washington is the only state that saw more transit riders in the year prior to April 2019 than the same period four years ago. To understand why ridership in Seattle was growing, it is first necessary to look at where ridership has declined the most.
Decentralization of Older Cities
Although the nationwide ridership decline began in 2014, in many places it has been declining for far longer. A recent article in the Cleveland Plain Dealer showed that the number of riders carried by the Greater Cleveland Regional Transit Authority has declined by 73 percent since 1980.
Cleveland has lost nearly three-fourths of its transit riders since 1980.
Cleveland is not the only urban area to have seen such massive declines. Based on 1982 data, the earliest that are available from the FTA, Detroit, St. Louis, Cincinnati, and Milwaukee have all seen declines fo 40 to 70 percent since that year. What all of these urban areas have in common is a massive decentralization of people and jobs from their cores to their suburbs.
At the end of World War II, many of these central cities had dense populations with high levels of multifamily housing. Since 1950, these cities have lost large numbers of people even as most of their urban areas have grown. This represents a preference for single-family housing, but it also was accompanied by a decline of the importance of downtown job centers. Since most transit systems are hub-and-spoke systems focused on downtown, they work for bringing commuters into downtown but not for commuters who work elsewhere.
For example, census data compiled by Wendell Cox shows that, as of 2010, 57 percent of downtown Chicago workers took transit to work. But the area around O’Hare Airport has 210,000 jobs — more than all but seven downtowns in the United States — and only 5.5 percent of those commuters took transit to work.
This shows the change in central city populations from 1980 to 2010 and the change in ridership from 1982 to 2018.
The chart above compares the change in central city populations from 1980 to 2010 with the change in ridership from 1982 to 2018. While the correlation isn’t perfect, it shows that suburbanization has reduced ridership.
Downtown Jobs Key to Ridership
Many people presume that transit ridership has something to do with population densities. But the correlation between urban area densities and transit’s share of commuting is only about 0.4 (where 1 is perfectly correlated and 0 is no correlation). However, the correlation between the number of downtown jobs and transit’s share of commuting is nearly 0.9. As Wendell Cox frequently says, “transit is about downtown.”
New York is not shown on this chart, but it is very close to the trend line with 1.9 million downtown jobs (including midtown Manhattan) and 30 percent transit commute share.
As of 2010, only six downtowns in the United States had more than 240,000 jobs: New York (1.9 million), Chicago (500,000), Washington (380,000), San Francisco (300,000), Boston (242,000), and Philadelphia (240,000). Not coincidentally, those were also the only six urban areas where transit carried more than 10 percent of commuters to work.
Seattle’s ridership has grown because of a huge increase in jobs in its downtown, growing from 216,100 jobs in 2010 to 301,000 jobs in 2018. Seattle may be the only major city in the United States that has more than half its jobs downtown.
At 0.97, the correlation between downtown jobs and transit’s share of commuting in the Seattle urban area is nearly perfect.
As it happens, downtown Seattle reached 240,000 jobs in 2013, the same year transit’s share of Seattle-area commuting reached 10 percent. Over the last decade, the correlation between the number of downtown Seattle jobs and transit’s share of Seattle-area commuting is a remarkable 97 percent.
This doesn’t mean that any urban area can increase transit’s share of commuting to more than 10 percent by attracting more than 240,000 jobs downtown. For one thing, few urban areas are in a position to reach 240,000 downtown jobs. As of 2010, downtown Atlanta and Houston were closest at around 170,000 jobs. Downtown Los Angeles was under 140,000; and downtown Denver 120,000. Other downtowns were under 100,000 jobs.
Even if they could pack more jobs into their downtowns, the share of downtown commuters in those urban areas who are taking transit to work is too low to make much of a difference: around 20 percent in Denver and Los Angeles and 14 percent in Atlanta and Houston, while Seattle’s was closer to 40 percent in 2010.
Seattle’s downtown grew because Amazon and Microsoft decided to move many of their office workers from suburban Bellevue and Redmond into downtown. If any government policy played a role in those decisions, it was the urban-growth boundary that has pushed prices of suburban real estate, making downtown relatively more competitive. But that same policy has increased traffic congestion and made housing far less affordable than it was a few years ago, which in turn has increased homelessness.
Per Capita Ridership Falling
Many urban areas haven’t seen the decentralization experienced by Chicago, Cleveland, and other older cities because they were never very centralized in the first place. Many Sunbelt regions have seen their populations grow primarily after World War II, which high auto ownership rates allowed most people to live in low-density neighborhoods and jobs were similarly decentralized.
Many of these urban areas have seen their long-term ridership grow, but this is often due solely to population growth, while their per capita ridership has often massively declined. In 1985, Atlanta transit carried 83 trips per urban area resident; by 2017, this had fallen to 26. The Miami urban area (including Ft. Lauderdale and West Palm Beach) saw per capita ridership fall from 49 to 22 trips per year.
Even some of the biggest transit regions have seen per capita ridership decline. Chicago dropped from 110 trips per resident in 1985 to 68 in 2017; Washington from 102 to 82; Boston from 106 to 87; Philadelphia from 92 to 62; and San Francisco-Oakland from 121 to 108. Nationally, per capita ridership was 36 trips per urban resident in 2018, the lowest ever recorded, and 2019 is on its way to being lower still.
New York Growth Hid Losses Elsewhere
The one major exception to the per capita ridership trend is the New York urban area, where ridership grew from 201 trips per resident in 1985 to 223 in 2017. In the 1980s, New York City was mismanaged and losing people and jobs. Improvements made by the Giuliani administration in the 1990s reduced crime and made the city and its transit system more attractive.
More recently, New York City’s recovery from the September 11, 2001 terror strike has contributed to transit ridership growth which obscured declines in many other parts of the nation. In 1993, transit in the New York urban area carried just under one-third of all transit rides in the nation. Since then, New York transit ridership has grown by 73 percent, while transit in the rest of the nation has grown by only 7 percent (which, considering population growth, represents a 21 percent decline in per capita ridership outside of New York). As a result, as of 2018, transit in the New York area carried nearly 44 percent of all transit riders in the nation.
Even New York, of course, wasn’t immune to the effects of ride-hailing on transit ridership. New York-area ridership peaked in 2014 and has dropped about 4 percent since then. But even without ride-hailing, New York ridership was not likely to grow at the rates it had been enjoying before 2014.
Last week, the Alliance for Downtown New York reported that job numbers in lower Manhattan have recovered to their pre-9/11 levels. With an 11 percent vacancy rate in lower Manhattan office buildings, there is still room for a little growth, but once that is filled up, job growth in downtown New York will slow.
Thus, the outlook for transit is dimmer than ever. While the transit industry would like people to believe that the most important goal of government land-use, tax, and transportation policies is to get people to ride transit more, there is really no reason why that should be so. The one factor that can increase transit ridership is to significantly increase downtown jobs. Cities have few tools to do that and even if they could do that, the negative side effects — congestion, high real estate prices, and homelessness — outweigh the benefits.
The ridership data posted by the FTA last week shows monthly ridership for every month from January 2002 through March 2019 by transit agency and mode of transit. For those who wish to explore these data further, I’ve posted an enhanced spreadsheet that totals the monthly data into annual data in columns HI through HZ, and provides totals for major modes in rows 2142 through 2149, transit agencies in rows 2153 through 3151, and the nation’s 200 largest urban areas in rows 3153 through 3351.