On the Predictability of the Decline of Transit Ridership in the US

Noting that transit ridership is generally down in the past few years, Jarrett Walker tweeted:

“Would somebody please do a quick but defensible study on why this is happening in so many US cities? I’m tired of having only anecdotes.”



Boom times

Income since 2009 has risen. At the trough of the recession, per capita income in nominal terms was just under $40,000, as of 2015 it is about $48,000. It is up in real terms as well. Elasticity of Transit demand with respect to income suggests that as incomes rise, transit demand falls. In other words, for most people transit is an inferior good (demand falls as income rises), with an income elasticity of about -0.98 (a 1% increase in income resulted in about a 0.98% decrease in ridership) found in Hawaii (McLeod et al 1991). In another example, the Mexico City Metro is an inferior good for middle and upper income Mexicans (Crotte et al. 2009). There are many other references, and while rail is sometimes disputed as to whether it is inferior or normal, bus is generally considered inferior.


Cheap Gas

The Price of gasoline peaked in 2008 at just above $4/gallon. It is presently $2.24. Estimates of the  Elasticity of Transit demand with respect to the price of gas vary considerably, but all suggest when the price of gasoline falls, transit demand falls with it.

The price of fuel increased sharply in the run-up to the Great Recession, as shown in Figure 3.8; this certainly discouraged car travel. Interestingly, it also reduced car crashes by more than the reduction in distance traveled, which the research team attributed to worse than average drivers (especially the young) being more likely to be priced off the road. How much less travel is there because of increases in the price of gasoline? For every 100% increase in the price of gas, there is a 5% decrease in gasoline consumption (which correlates to driving in the short run, in the long run there is also a shift in vehicle fuel economy, and the elasticity is higher). From Levinson and Krizek (2015) The End of Traffic and the Future of Transport.  Figure 3.8 Source: US Energy Information Administration http://www.eia.gov/petroleum/gasdiesel/.
  • “the elasticities of monthly transit ridership with respect to the real gasoline price are positive and inelastic, ranging from 0.08 to 0.80. ” (Wang and Skinner 1984)
  • ” Research in 2007 established that for every 10% increase in gas prices, U.S. transit demand has increased by 1.2%, a cross-elasticity of demand to gas prices (e) of 0.12. Results also showed much higher effects on U.S. light rail systems (e = 0.27 to 0.38) for heavy rail (e = 0.17) but low sensitivity for bus (e = 0.04). In Australia, the impact of gas prices on ridership has been larger (e = 0.22) because of higher gas prices (20% to 30% higher than U.S. prices).” (Currie and Phung 2008)
  • ” the cross elasticities when gas prices were less than $3 a gallon were small, with a magnitude of less than 0.05. When prices exceeded $3 a gallon, the elasticity was larger, in the range of 0.12–0.14, for the rail modes. In the summer of 2008 when prices exceeded $4 a gallon, there was considerable responsiveness with elasticities of 0.28–0.30 for city and suburban bus, and 0.37 for commuter rail. These values are similar to, or even larger than, those found during the oil crises of the 1970s and early 1980s.” (Nowak and Savage 2013)



Obviously there are other effects, in terms of service and fares, that effect ridership, but the macro-economic condition is too large an effect to dismiss out of hand. While it is hard to attribute the particular amount of the fall of transit ridership to a particular factor (econometrics can help, but in the end at best there will be a range of results, not one specific number that people will be able to rely on), it is clear that that a 50% fall in the price of gas could easily explain more than a 5% fall in transit ridership. Similarly a 10% increase in real incomes (a 20% increase in nominal incomes) could explain perhaps a 10%  fall of ridership. Elasticities are measured at the margin, and the changes in the price of fuel are far from marginal, so one can’t be too enamored of a particular elasticity effect, but the directions are robust and clear and not surprising.