Evidence on the Long-Run Elasticity of Labor Supply

Taxi drivers appear to have worked just a little bit less in response to an increase in the fare structure.

Many public policies regarding taxation, social safety nets, and the redistribution of income are designed based on assumptions about the long-run effect of after-tax wage rates on labor supply. For men, many of the estimates of this elasticity of labor supply suggest values near zero, implying that permanent wage increases have relatively small effects on labor supplied. However, all of these studies have faced the problem that most workers cannot alter their hours of work without changing jobs, and that it is difficult to measure the actual changes in net-of-tax wages that workers face.

In A Shred of Credible Evidence on the Long-Run Elasticity of Labor Supply (NBER Working Paper No. 15746), co-authors Orley Ashenfelter, Kirk Doran, and Bruce Schaller introduce a simple, natural experiment to deal with these problems. They rely on a dataset of New York City taxi drivers who choose their own hours, and who experienced two permanent fare increases instituted by the New York City Taxi and Limousine Commission in March 1996 and May 2004. While this approach has the obvious advantage of transparency, the authors note that it may not be appropriate to generalize the findings here to other workers.

The data indicate that the effect of increases in the fare structure on the number of hours worked is small, and negative: taxi drivers appear to have worked just a little bit less in response to an increase in the fare structure. Indeed, the fare increases analyzed here resulted in higher total revenue per hour -- the number of passengers hailing cabs did not drop off enough to offset the additional revenue from higher fares per trip. Miles driven did decline, however, by 4.2 percent on average following a fare increase.

The wages of the cab drivers, on the other hand, were strongly affected by the increases in the fare structure. The fare increases appear to be associated with an average 19 percent increase in revenue per mile.

Taken together, the evidence implies that the long-run uncompensated elasticity of labor supply is -0.23, and that the income effects of a fare increase dominate the substitution effects. This result is consistent with a broad variety of historical evidence that suggests that the massive increases in real wages seen in the United States and Europe since 1879 have been accompanied by significant declines in annual hours worked per worker.

-- Claire Brunel