Via Kevin Lewis, here is a new paper on that question:
Frank Sloan, Lindsey Eldred & Yanzhi Xu
Journal of Health Economics, May 2014, Pages 64–81
This study investigates whether drinker-drivers attributes are associated with imperfect rationality or irrationality. Using data from eight U.S. cities, we determine whether drinker-drivers differ from other drinkers in cognitive ability, ignorance of driving while intoxicated (DWI) laws, have higher rates of time preference, are time inconsistent, and lack self-control on other measures. We find that drinker-drivers are relatively knowledgeable about DWI laws and do not differ on two of three study measures of cognitive ability from other drinkers. Drinker-drivers are less prone to plan events involving drinking, e.g., selecting a designated driver in advance of drinking, and are more impulsive. Furthermore, we find evidence in support of hyperbolic discounting. In particular, relative to non-drinker-drivers, the difference between short- and long-term discount rates is much higher for drinker-drivers than for other drinkers. Implications of our findings for public policy, including incapacitation, treatment, and educational interventions, are discussed.
Here is an ungated version of the paper.