ArchWeek - On Travel and the Built Environment

[Kevin Matthews]

<i>Discussion related to this <a href="http://www.ArchitectureWeek.com/">ArchitectureWeek</a> story, <a href="http://www.ArchitectureWeek.com/2010/0818/environment_1-1.html"><b>On Travel and the Built Environment</b></a>:</i><br>
<p>
<SCRIPT LANGUAGE=JavaScript src="http://www.ArchitectureWeek.com/2010/0818/environment_1-1.float.js">
</SCRIPT>

[Kevin]

This new report from 'CEOs for Cities' debunks a whole series of misleading ideas around congestion, commuting, and sprawl, mythology that has been spread over some 30 years using the often-referred to Urban Mobility Index created by the Texas Transportation Institute:

Driven Apart: How Sprawl Is Lengthening Our Commutes and Why Misleading Mobility Measures Are Making Things Worse
http://www.ceosforcities.org/pagefiles/ ... -29-10.pdf

And because one of the key flaws in the methodology used to establish the Urban Mobility Index for cities around the US is that the calculations left out the vital factor of driving distance, these findings are related to and supportive of the findings we reported on in Travel and the Built Environment.

Driven Apart - Executive Summary
http://www.ceosforcities.org/pagefiles/ ... SFINAL.pdf

While peak hour travel is a perennial headache for many Americans — peak hour travel times average 200 hours a year in large metropolitan areas — some cities have managed to achieve shorter travel times and actually reduce the peak hour travel times. The key is that some metropolitan areas have land use patterns and transportation systems that enable their residents to take shorter trips and minimize the burden of peak hour travel.

That’s not the conclusion promoted by years of highway-oriented transportation research. The Urban Mobility Report (UMR) produced annually by the Texas Transportation Institute and widely used to gauge metropolitan traffic problems has overlooked the role that variations in travel distances play in driving urban transportation problems.

This report offers a new view of urban transportation performance. It explores the key role that land use and variations in travel distances play in determining how long Americans spend in peak hour travel.

• Travelers in some cities - those with more compact development patterns - tend to spend less time in peak hour traffic because they don’t have to travel as far.

• If every one of the top 50 metro areas achieved the same level of peak hour travel distances as the best performing cities, their residents would drive about 40 billion fewer miles per year and use two billion fewer gallons of fuel, at a savings of $31 billion annually.

• In the best performing cities the typical traveler spends 40 fewer hours per year in peak hour travel than the average American because of the shorter distances they have to travel.

In the best performing cities—those that have achieved the shortest peak hour travel distances - such as Chicago, Portland and Sacramento, the typical traveler spends 40 fewer hours per year in peak hour travel than the average American. In contrast, in the most sprawling metropolitan areas, such as Nashville, Indianapolis and Raleigh, the average resident spends as much as 240 hours per year in peak period travel because travel distances are so much greater. These data suggest that reducing average trip lengths is a key to reducing the burden of peak period travel.

RANKING METROPOLITAN AREAS ON PEAK PERIOD TRAVEL TIMES

The additional travel time associated with longer average trip distances is the chief determinant of which metropolitan areas have the longest travel times. Longer trip distances add 80 hours a year or more to peak travel times in Nashville, Oklahoma City, Richmond, and Nashville. Areas with the shortest average travel distances, including Chicago, New Orleans, Sacramento and New York, have among the lowest total hours of peak period travel.

These results are a stark contrast to the picture of urban transportation painted by the UMR, which has long been used to measure traffic problems and compare cities. A close examination shows that the UMR has a number of key flaws that misstate and exaggerate the effects of congestion, and it ignores the critical role that sprawl and travel distances play in aggravating peak period travel.

THE TRAVEL TIME INDEX: A FLAWED TOOL FOR DIAGNOSING TRANSPORTATION PROBLEMS

The central analytical tool in the Urban Mobility Report is the Travel Time Index (TTI), which is the ratio of average peak hour travel times to average free flow travel times.

On its face, the Travel Time Index seems like a reasonable way to compare city transportation systems. And if all cities had similar land use patterns and densities and had the same average trip lengths, then the TTI would be a fair measure. But city land use patterns vary substantially, and as a result the Travel Time Index conceals major differences in urban transportation between different cities.

According to the UMR, the worst traffic was in Los Angeles, Washington and Atlanta. But a re-analysis of the data shows that residents in at least ten other metropolitan areas, including Richmond, Raleigh-Durham, Detroit and Kansas City, spent the most time traveling in peak hours. Again, the key reason for the difference is the much longer-than- average peak period travel distances in those cities.

LIMITATIONS OF THE URBAN MOBILITY REPORT’S METHODOLOGY

Our detailed analysis of the methodology of the Urban Mobility Report suggests that it is an unreliable guide to understanding the nature and extent of transportation problems in the nation’s metropolitan areas.

The Urban Mobility Report’s key measure - the Travel Time Index—is a poor guide to policy, and its speed and fuel economy estimates are flawed. In the aggregate, the analysis appears to overstate the costs of traffic congestion three-fold and ignores the larger transportation costs associated with sprawl. Specifically:

• The Travel Time Index used in the UMR is based on a questionable model of how traffic volumes affect traffic speeds, and it uses an unrealistic and unattainable baseline of zero delay in computing congestion costs. The structure of the Travel Time Index conceals the effect of sprawl and travel distance on travel time.

• The key statistic underpinning the UMR’s findings is based on the difference in travel times between peak and non-peak periods, but the study’s travel time estimates are based on volume data, not on actually observed travel speeds.

• The model used to convert volume data to estimated speeds was calibrated by “visual inspection” of the data, and the line chosen to reflect the data isn’t based on statistical analysis; a line fit with a simple quadratic equation would produce much higher estimates of peak hour speeds and consequently lower levels of peak hour delay.

• The UMR speed/volume model relies on daily, rather than hourly (or minute-by-minute) traffic volumes, meaning that the authors must make strong assumptions about the distribution of traffic between peak and non-peak hours.

• The claims the UMR makes about trends in travel times over time and across cities do not correlate with other independent measures of travel times. Survey data on observed speeds from Inrix, a private aggregator of travel time data gathered from commercial vehicles, and self-reported travel times from the Census and National Travel Survey are not consistent with the conclusions of the Urban Mobility Report. Neither the total change in travel time, measured nationally, nor the pattern of changes in travel time across metropolitan areas is consistent with the estimates of increased delay presented in the Urban Mobility Report.

• The UMR claim that travel times have increased is a product not of direct observations but is an artifact of the structure of the UMR’s speed/volume equations, for which there is no independent confirmation. As long as volume increases more than capacity, the UMR model mechanically predicts slower speeds and travel times.

• There are strong reasons to doubt the UMR claim that slower speeds associated with congestion wastes billions of gallons of fuel.

• Data from speed measurements monitored by Inrix suggest that the UMR methodology overstates the Travel Time Index by about 70 percent.

• Data from the National Household Travel Survey show that nearly all of the increase in peak commuting times was due to longer trips rather than slower travel speeds.

• The pattern of changes in commuting times between 1990 and 2000 shows that there is no correlation between changes in peak delays estimated in the UMR and changes in commute times reported in the Census.

• The UMR estimates of fuel consumption are based on a 29 year-old study of low-speed driving using 1970s era General Motors cars, which is of questionable applicability to today’s vehicles and to highway speeds.

• The UMR extrapolates these data outside of the speeds for which they were intended and changes the functional form used in the original study in a way that exaggerates fuel consumption associated with speed changes.

• The UMR fuel consumption results are not consistent with other, more recent estimates of fuel economy patterns and ignore the savings in fuel consumption associated with modest reductions in travel speeds.

• The UMR ignores the fuel consumption associated with longer trips in sprawling metropolitan areas.

Adjusting the UMR estimates to account for each of these issues produces a significantly lower estimate of the cost of congestion. Adopting a more reasonable baseline for congestion-related delays, using the Inrix Travel Time Index, adopting a lower value of travel time, and adjusting fuel consumption estimates would imply that the cost of congestion in monetary terms is perhaps less than 70 percent lower than the figure claimed in the UMR. For the 51 metropolitan areas analyzed here, this means that the UMR overstates the cost of congestion by about $49 billion.

A re-analysis of the data in the UMR paints a very different picture of transport problems. Trip distances grew rapidly in the 1980s and 1990s, but have stopped growing since then. Between 1982 and 2001, average commute trips nationally got three miles longer. Our calculations, based on data from the UMR, suggest that average travel distances increased in three-quarters of the 50 largest metropolitan areas over this time period. Since 2001, however, peak period travel distances have been shrinking in most metropolitan areas, and the average travel distance has declined about 1.0 percent.[/b]