But what if driving has been falling most rapidly on rural roads – roads that tend, all else being equal, not to get congested very often – and less rapidly on gridlocked urban highways and streets?
A cursory glance at the Federal Highway Administration’s (FHWA) flagship data publication, Highway Statistics, suggests this possibility. Rural driving, according to the FHWA, fell off a cliff right around 2002 – declining by 13 percent since then – while urban driving remains higher than it was a decade ago and appears to be down only slightly from its 2007 peak.
So, is falling rural travel responsible for the recent decline in driving in the U.S.?
What appears to be a sharp decline in rural driving in the 2000s is likely an artifact of the inconsistent, opaque and confusing way that travel is assigned to “urban” and “rural” categories by the FHWA and state Departments of Transportation. Rural driving may have fallen as a share of the whole over the last decade or two – as one might expect in an urbanizing country – but the decline has been much more gradual than the FHWA data suggest. On the other hand, the decline in urban driving over the last decade is likely sharper – and therefore, more significant – than it appears.
If you’re looking for the answer to why driving is declining in the United States, therefore, you’re unlikely to find it on dusty country roads and lonely rural Interstates. The remainder of this post explains why.
“Rural” versus “Urban”
In getting to the bottom of why the data on recent rural driving trends can’t be trusted, you need to understand two important things.
The first is that “urban” driving is defined by the FHWA as that which takes place on roads within an “urbanized area.” “Rural” driving is everything else.
What is an urbanized area? It is a geographic entity that’s generally bigger than a city but smaller than a metropolitan area – essentially a city and most of its suburbs.
The second important thing to know is that the boundaries of urbanized areas change over time. Here, for example, is a map of fast-growing Austin, Texas, showing the U.S. Census-defined boundaries of the Austin urban area in 1990, 2000 and 2010.
(Credit: U.S. Census Bureau, ESRI)
Put yourself, for a moment, in the place of someone who moved into a newly developed neighborhood in one of the green areas on the Austin map during the 1990s – say, 1992. Because the Census redraws urban area boundaries only once a decade, you would have been considered a “rural” resident from the day you moved into your suburban house, condo or apartment up until the 2000 Census. Then you, your neighbors, and all the other residents of the green zone would have been subtracted from the “rural” total and added to the “urban” column in one fell swoop.
How fell a swoop are we talking about? The Census Bureau just released its first population estimates based on urban area boundaries from the 2010 Census. Between 2011 and 2012, among urban areas for which a valid comparison could be made, the U.S. urban population was shown to have grown by about 8 percent. The Phoenix urban area alone was shown to have gained 641,000 people in the single year between 2011 and 2012. Phoenix may be growing fast, but not that fast. Similarly, if you were looking at changes in the rural population of Arizona during that same time period, you would expect to see a major drop in the number of rural residents corresponding with the spike in urban residents.
Missing: 84,000 Miles of Rural Roads
The same principle applies to highways and vehicle travel. The roads in the green zone (and the driving that occurs on those roads) would also have been reclassified from “rural” to “urban” at some point along the line, leading one to expect a dramatic drop in rural vehicle travel and corresponding spike in urban travel in the year that the boundaries were redrawn.
And if you look closely at Highway Statistics, you can see this happening. Between 2002 and 2006 (the years when most states redrew their urbanized area boundaries following the 2000 Census), rural vehicle-miles traveled (VMT) declined by 8 percent – this during a period when overall VMT nationwide was still rising. During that same period, America “lost” 84,000 miles of rural roads, including 7 percent of its rural Interstates and 4 percent of its rural principal arterial roads. Travel on rural Interstates plummeted by 8 percent and travel on principal arterials by 10 percent during that short, four-year span of time.
Needless to say, those rural roads – and the driving that occurred on them – weren’t “lost” any more than the rural population of Arizona was decimated between 2011 and 2012. They were simply shifted into the “urban” category.
Hoosiers and Gophers and Texans, Oh My!
If the FHWA and state DOTs revised their urbanized area boundaries regularly and all at once like the Census Bureau does, it would be easy for analysts to recognize which comparisons were valid and which weren’t. Unfortunately, some states take much longer to update their boundaries than others. During the research for our recent report, Transportation in Transition, for example, my colleague Ben Davis found that several states were still using urbanized area boundaries from the 1990 Census as late as 2006!
Let’s go back to Austin. According to the Census Bureau, you might have been a “rural” resident from the time you moved into your home in the green zone in 1992 until 2000 or so – about eight years – at which point you became “urban.” But according to the data in Highway Statistics, the road outside your house might have been considered “a little bit country” for 14 years or more!
There are three states that seem to embody this problem: Indiana, Minnesota and Texas. Between 2007 and 2008, vehicle-miles traveled (VMT) on rural roads nationwide declined by 44.6 billion miles. Roughly 45 percent of that decline took place in just those three states. Indiana, Minnesota and Texas saw rural VMT decline by an astounding 14 percent in that single year, while the rest of the country saw only a 2.8 percent decline. And, sure enough, that same year, those three states “lost” more than 12,000 miles of rural roads – about 3 percent of their rural road networks – suggesting that a major revision of urbanized area boundaries may have taken place that year.
Separating Truth from Fiction: How Much Did Rural Driving Really Fall?
How does this play out over a longer time scale? Highway Statistics tells us, for example, that VMT in the United States peaked in 2007 and has since fallen by 2.1 percent (as of 2012). Rural VMT, according to the FHWA, fell by 5.4 percent during that period, compared with a paltry 0.3 percent drop in urban VMT.
In Indiana, Minnesota and Texas, however, rural VMT has fallen by 14 percent since 2007, while urban VMT actually increased by 9 percent. Take those three states out of the national figures and you find that, in the other 47 states, rural VMT fell by 4 percent, while urban VMT fell by a much more notable 1.5 percent.
The examples above show that any comparison of trends in rural and urban driving based on Highway Statistics that fails to account for urbanized area boundary revisions is likely to lead to erroneous conclusions by dramatically overstating the decline in rural driving and understating the decline in urban driving over the last decade.
The steep decline in rural driving that appears in FHWA data in the mid-2000s is largely the result of an accounting issue – not any unusual change in rural driving patterns since the end of the Driving Boom. Rural driving may be falling, but declining rural travel likely does not come anywhere close to explaining the drop in U.S. vehicle travel over the last decade.
The fall in urban driving in the United States since the end of the Driving Boom, on the other hand, is real, it is significant, and it is likely a lot larger than the FHWA data would lead one to believe. It is nearly impossible (at least without a much deeper dive into the data) to tell how steep the decline in urban travel really is. That’s frustrating, and it provides yet another reminder of why we desperately need better data if we’re going to come to intelligent decisions about how to invest America’s transportation resources for the future.
 The 374 urban areas with consistent definitions from the 2000 and 2010 censuses.