Contrary to well-publicized claims by the anti-auto Surface Transportation Policy Project, cities that have emphasized transit over highways experienced the greatest increases in congestion over the past two decades. Data from the Texas Transportation Institute reveals that cities that concentrated on highways instead of transit did fairly well at minimizing the burden of congestion on their residents. This makes sense because few transit investments carry as many passenger miles as comparably-priced highway investments. But highway investments alone won't solve congestion: What is also needed is a toll system that charges more during congested periods than other periods.
The Texas Transportation Institute put out its annual mobility report on May 7, 2001, showing that congestion in most U.S. urban areas was worse in 1999 than the year before. The Institute's carefully worded publication does not take sides on the vexatious question of what to do about congestion.
But the study does quantify a lot of the problems related to congestion. In the 68 urban areas in the study, researchers estimated that congestion cost $78 billion and wasted 6.8 billion gallons of fuel and 4.5 billion hours of people's time in 1999. These numbers make for exciting newspaper headlines, but (as described on pp. 387-388 of The Vanishing Automobile) have to be taken with a grain of salt as they are based on on a lot of assumptions about road capacities and designs.
For example, the Institute estimates that congestion cost people in the San Francisco Bay Area (including the San Francisco-Oakland and San Jose urban areas) 227 million hours and $3.9 billion in 1998, which works out to about 750,000 hours and $13 million dollars per working day. By comparison, the California Department of Transportation estimates that congestion cost about 112,000 vehicle hours (which works out to about 150,000 person hours) and $1.25 million per day. Caltrans' numbers are significantly smaller than those of the Texas Transportation Institute.
This is not to say that one number is right and the other wrong. Both are in fact fairly crude estimates. The important point is that we have to be careful using these numbers. In fact, Texas Transportation Institute data are most reliable when comparing the changes in congestion over time, and less reliable when comparing congestion among urban areas for any single year.
For example, the report consistently rates Los Angeles as the nation's most congested urban area, while New York is somewhere around number thirteen. Yet, relative to Los Angeles highways, most New York highways tend to have narrower lanes, shorter merge lanes at on- and off-ramps, and tighter curves, all of which reduce the flow capacities of New York freeways. Since the Institute doesn't take these variations into account, New York congestion may actually be worse than in Los Angeles.
This weakness in the data is not recognized by the Surface Transportation Policy Project (STPP), an anti-auto, pro-transit group that issued a report on the same day as the Texas study claiming that "commuters suffer most" in urban areas "with the fewest transportation choices." "Transportation choice" is a euphemism, of course, for spending transportation dollars on transit, especially rail transit, instead of highways.
STPP used the Texas institute's data to compare congestion in various urban areas in 1999 -- exactly the kind of comparison that the Texas data is least suited for. STPP also added its own factors in a convoluted attempt to measure "congestion burden" rather than just the cost of congestion. STPP's congestion burden index is the Texas institute's 1999 travel rate index times the percentage of commuters who drove to work in 1990.
Never mind that this is mixing 1990 and 1999 data. Even if this were not a problem, STPP's index produces some very specious results. For example, according to the Texas Transportation Institute's data, the average Bostonian spends twice as many hours stuck in traffic each day as the average Las Vegan. Yet because Boston has more transportation choices than Las Vegas, STPP ranks Boston one of the least congestion-burdened urban areas and Las Vegas as the second-worst congested-burdened metro area in the nation.
Just why is sitting in traffic less of a burden because there is rail transit system that doesn't go where you need to go? STPP's index may confuse cause and effect. High congestion may force some people to ride transit, but that doesn't mean they are less burdened, especially since transit trips are generally slower than auto trips even in congested conditions.
If inter-metro area comparisons of Texas institute data are not valid (and made even less valid by STPP's politically contrived manipulations), are the Texas data worthless? No, the data are particularly valuable where they provide a time series from 1982 through 1999. These time series show what most urban residents know: that congestion is getting worse in almost every urban area. Moreover, the data are valid for making inter-urban area comparisons of the changes in congestion over that time period.
For example, it may not say much to say that Los Angeles is ranked number one and New York number thirteen in 1999 congestion because this ranking fails to account for the superior capacities of Los Angeles freeways. But it is more meaningful if the data show that New York congestion increased 50 percent faster from 1982 to 1999 than Los Angeles congestion. This suggests that whatever Los Angeles is doing may be working better at solving congestion than what New York is doing.
The Texas study uses several different measures of congestion, including:
There is a strong correlation (r-squared) between most of these indices. For example, there is a 98 percent correlation between the Travel Rate and Travel Time indices, and a 93 percent correlation between the Travel Rate Index and hours of delay per capita. (There is less than a 60 percent correlation between any of these measures and STPP's congestion burden index.)
Looking at the changes in the Travel Time Index from 1982 to 1999, it turns out that New York congestion did increase 50 percent faster than Los Angeles congestion. Hours of delay per capita in New York increased 2.5 times more than in Los Angeles. Thus, having all those rail lines in New York doesn't seem to have eased the burden of congestion all that much.
The table below shows the percentage change in the travel time index, travel rate index, roadway congestion index, and hours of delay per capita (1999 hours minus 1982 hours) for each of the urban areas in the Texas study. The delay per capita is shown as a difference rather than a percentage because percentage changes can be misleading. An urban area which moved from nearly no congestion in 1982 to some congestion in 1999 can have a huge percentage increase in the hours of delay per capita even though the 1999 hours of delay remain relatively small.
A lot of factors caused congestion to increase in some regions faster than in others, including population growth, urban layout, and the location and distribution of employment centers. But the one important factor under the control of transportation planners is how transportation funds are spent. Over the past decade or so, the leaders of many urban areas, including Portland, the Twin Cities, and San Diego, have openly given up on highways to relieve congestion and funneled transportation dollars into transit instead. Other regions, such as Houston, Phoenix, and Kansas City, decided to stick with highways.
Table one shows a clear result. Urban areas that focused on transit had the greatest increases in congestion. My own home town of Portland suffered the greatest increase in the Travel Time Index, the second-greatest increase in the Travel Rate Index (scoring a fraction of a percent less than San Diego), and large increases by the other measures as well.
All of the top seven urban areas (as ranked by change in Travel Time Index) became fixated on rail rather than roads sometime in the 1980s or 1990s. Number eight on the list, Las Vegas, has been growing at nearly 6 percent per year -- far faster than any other urban area -- which is far faster than road agencies can build roads even if they have the political support.
Meanwhile, urban areas that focused on highways had relatively low increases in congestion. Thanks to aggressive construction of toll roads, Houston has barely recorded an increase in congestion. Phoenix and Kansas City also did well.
Urban area TTI TRI RCI Hours/capita Portland OR 151% 130% 153% 30 Twin Cities MN 149% 126% 182% 35 San Diego CA 145% 130% 158% 29 Boston MA 145% 125% 145% 30 New York NY 144% 120% 149% 27 Seattle WA 144% 126% 121% 34 Atlanta GA 143% 125% 165% 42 Las Vegas NV 143% 127% 171% 16 Sacramento CA 140% 122% 158% 25 San Bernardino CA 139% 125% 159% 32 Denver CO 138% 123% 146% 32 Cincinnati OH 137% 121% 160% 29 Tacoma WA 136% 122% 159% 22 Indianapolis IN 136% 121% 173% 34 Albuquerque NM 136% 120% 182% 30 Dallas TX 135% 120% 135% 38 Detroit MI 135% 119% 135% 29 Washington DC 132% 121% 135% 28 Austin TX 131% 117% 145% 36 Los Angeles CA 131% 119% 122% 26 SFO-Oakland CA 131% 119% 131% 22 Columbus OH 130% 117% 167% 42 Chicago IL IN 130% 120% 138% 22 Milwaukee WI 130% 118% 148% 24 Baltimore MD 129% 117% 143% 23 St. Louis MO 129% 117% 118% 33 Charlotte NC 129% 119% 133% 27 Ft. Lauderdale FL 127% 119% 170% 22 Salt Lake City UT 126% 116% 152% 16 Miami FL 126% 116% 129% 25 Cleveland OH 126% 116% 146% 18 Fort Worth TX 125% 116% 132% 28 Phoenix AZ 125% 116% 127% 19 Louisville KY 125% 114% 140% 29 San Antonio TX 125% 118% 148% 19 Providence RI 124% 113% 136% 25 Philadelphia PA 124% 112% 129% 17 Colorado Springs CO 123% 113% 170% 19 Tucson AZ 123% 113% 131% 19 Memphis TN 123% 112% 138% 19 Orlando FL 122% 113% 128% 32 Jacksonville FL 122% 113% 133% 25 Fresno CA 122% 113% 149% 14 Kansas City MO 118% 109% 158% 22 San Jose CA 117% 109% 111% 22 Nashville TN 117% 109% 122% 29 Honolulu HI 117% 112% 134% 12 El Paso TX 116% 111% 152% 12 Omaha NE 116% 110% 145% 17 Oklahoma City OK 115% 108% 135% 13 Norfolk VA 115% 108% 109% 16 Salem OR 114% 107% 152% 12 Eugene OR 113% 106% 172% 9 Tampa FL 112% 107% 121% 22 Houston TX 112% 108% 107% 23 Hartford CT 109% 105% 154% 12 Rochester NY 109% 105% 153% 7 Spokane WA 109% 104% 126% 7 Albany NY 107% 104% 167% 9 Boulder CO 107% 104% 151% 4 Bakersfield CA 107% 104% 143% 5 Brownsville TX 107% 104% 139% 3 Buffalo NY 107% 104% 136% 6 Laredo TX 106% 103% 111% 5 Pittsburgh PA 105% 103% 111% 8 New Orleans LA 104% 104% 108% 8 Beaumont TX 104% 102% 126% 5 Corpus Christi TX 103% 102% 125% 4
(The names of some urban areas were shortened for convenience; e.g. Portland-Vancouver to Portland.)
Thus, STPP's claim that huge investments in transit ease the burden of congestion is almost certainly disproven by Texas Transportation Institute's data. In fact, the leaders of most urban areas that have elected to focus on transit see congestion as a tool to get people to ride transit. As noted in The Vanishing Automobile (p. 112), Portland planners say that any relief from stop-and-go traffic (level of service F) "would eliminate transit ridership." A 1996 transportation plan for the Twin Cities places a moratorium on new roads in the hope that, "as traffic congestion builds, alternative travel modes will become more attractive" (Vanishing Auto, p. 465).
Thus, well-publicized efforts to increase "transportation choices" are almost certain to be accompanied by less-publicized decisions to increase highway congestion to force people to ride transit. This is exactly the opposite of the impression that STPP tries to convey.
Historically, congestion has increased because the states elected to pay for roads using highway user fees based mainly on a cents-per-gallon tax. The inflation of the 1970s drove up highway costs without concurrently increasing highway revenues, while the fuel-efficient cars of the 1980s drove down revenues per mile driven. This left highway builders financially unable to keep up with the growth in driving. As cars become even more fuel efficient, the long-term solution will be toll authorities, such as the one serving Houston, along with rush-hour premiums (value pricing) on those tolls.
Before that financial solution can be adopted, however, we have to overcome the political problem, which is that the anti-auto lobby has convinced many political leaders that transit, not roads, is the solution to congestion. Those who want a real solution to congestion must prove that this is not true before they can press for toll roads and value pricing.
As a side note, a recent Brookings Institution review of census data concludes that transportation policy is an important factor in urban growth. "Cities built for cars grew" between 1990 and 2000, say Brookings/Harvard analysts Edward Glaeser and Jesse Shapiro, "but cities designed for mass transit and pedestrians tended to shrink."
Their report, City Growth and the 2000 Census: Which Places Grew, and Why, notes that "Cities with less than 65 percent of their commuters driving alone grew by less than 2 percent on average, while other cities grew by an average of more than 12 percent." Those who oppose automobiles might applaud this result since many of them also oppose growth. But those who believe that growth is an important part of their regional economies should pause before endorsing transit-intensive and road-starved transportation plans.