Value of time: Speeding behavior and gasoline prices

'Value of Time' (VOT) is a key parameter in economics and policy. This paper presents an alternative method to estimate VOT by analyzing an hourly dataset on drivers speeding behavior as a function of the gasoline price. Our identification strategy is novel as it is based on the intensive margin. In comparison, previous studies reveal VOT on the extensive margin, but choice alternatives have multiple attributes thereby potentially confounding estimates. Consistent with the range of the prior literature, we find a VOT of about 50% of the wage rate and analyze sources of bias from accidents and traffic tickets. These bias functions suggest that previous stated preference VOT estimates are likely downward whereas previous revealed preference estimates are likely upward biased.