The United States and many other countries are dramatically tightening fuel economy standards for passenger vehicles. Higher fuel economy reduces per-mile driving costs and may increase miles traveled, known as the rebound effect. The magnitude of the elasticity of miles traveled to fuel economy is an important parameter in welfare analysis of fuel economy standards, but all previous estimates from micro data impose at least one of three behavioral assumptions: (a) fuel economy is uncorrelated with vehicle and household attributes; (b) for multi vehicle households, each vehicle can be treated as an independent observation in statistical analysis; and (c) the effect of gasoline prices on vehicle miles traveled is inversely proportional to the effect of fuel economy. Two approaches to relaxing these assumptions yield a large estimate of the rebound effect; a one percent fuel economy increase raises driving 0.2 or 0.4 percent, depending on the approach, but the estimates are not statistically significantly different from one another.