Millennials are driving more, but only those making the least money

By Chris McCahill
The new 2017 National Household Travel Survey gives us our first look at changing travel habits since the recession. From what we can tell, the average American drives less in 2017 than eight years earlier. Driving also seems to have increased considerably among Millennials—but mostly among those with the lowest incomes—bucking expectations.
Unfortunately, changes in the way that trip mileage was calculated in the new NHTS have made it difficult to compare the survey to previous ones. Instead of using reported trip distances, the 2017 survey used Google Maps API to estimate shortest trip distances, which could be about 10 percent lower than self-reported trip distances used in previous surveys. Average per capita VMT dropped by 12 percent in the 2017 survey or by 2 percent if that adjustment factor is right.
No matter how much average VMT changed, however, we can dig into the data to understand who’s driving more and by how much, for each survey year, and compare those trends over time. Our analyses show per capita VMT for different groups of individuals, relative to the average.¹
One striking finding (Figure 1) is that driving among those aged 26 to 33 dropped considerably compared to the population average in 2009—at the height of the recession—and then increased again in 2017 (today’s Millennials). This pattern suggests that young adults were highly affected by the recession, having just recently entered the job market.
In general, though, what we see in comparing age groups in the three periods is that younger people are driving less compared to the population average, while older people are driving more. As an example of the latter, in 2001 those aged 74 to 81 drove 36 percent less than the population average, while in 2017 they drove just 18 percent less.

Figure 1. Relative VMT per capita by age, 2001 to 2017. Source: FHWA/NHTS

Like age, income has always been a useful predictor of travel demand, and the recession widened that gap (Figure 2). In 2009, people in high income households (those making $100,000 or more) drove 27 percent more than average, compared to those in households making under $50,000, who drove 23 percent less than average. In 2017, VMT dropped among those in high- and medium-income groups, relative to the average. A rising average could help explain this, but since average VMT most likely held constant or decreased, this indicates that those with means are possibly choosing to live in places that facilitate less driving and use of alternative modes.
Figure 2. Relative VMT per capita by income, 2001 to 2017. Source: FHWA/NHTS

Among Millennials, it appears that those in the lowest income group are mainly behind the post-recession driving increase, compared to population averages. In 2009, driving dropped considerably among all income groups in the 26 to 33 age range. However, in the most recent survey, driving among those in low-income households returned to pre-recession levels—36 percent above average. In high-income households, driving continued dropping to 31 percent above average, down from 73 percent above average in 2001. Those in medium income households held steady around 44 percent above average.
Figure 3. Relative VMT per capita by income for those age 26 to 33, 2001 to 2017. Source: FHWA/NHTS

This makes sense, given a recent analysis by City Observatory showing, “the number of 25-34 year olds with four-year degrees living in large cities is growing almost five times faster than the overall rate of population growth in [large] cities.” Many young people with the means to move into urban areas are gaining better access to transit, shorter travel distances, and more opportunities for walking and biking, which lessens their need to drive.
Meanwhile, Millennials without the incomes, skill sets, or preferences to live in more urban places are forced to drive more, costing them upward of $10,000 per year, according to new estimates from Inrix, which include vehicle ownership, lost time, and parking.
¹VMT per capita is based on weighted total estimates of VMT and persons. Weights annualize the travel-day trip mileage while adjusting for MSA type, race, ethnicity, sex, travel month, and day of week. None of the values depicted here is based on fewer than 1,000 survey responses.
Chris McCahill is an Associate Researcher at SSTI.