Girls can't bid?

I just finished reading "Selection Bias, Demographic Effects, and Ability Effects in Common Value Auction Experiments" by Marco Casari, John C. Ham and John H. Kagel published in the September 2007 edition of the American Economic Review. It was a very interesting and quite accessible paper, I was even able to follow the math.

They set up an experiment to observe behavior is repeated auctions where winner's curse from overbidding could cause bankruptcy. Some of their more expected conclusions were that ability (measured by ACT/SAT scores) and experience through successful previous bids lead to higher profits. Interestingly, though, they find that women perform significantly worse than men when they're inexperienced bidders and all other factors are controlled for, but they learn much more quickly and eventually catch up and might even surpass men as experienced bidders. Also, while lower than average ACT/SAT scores are strongly correlated with poor performance, very high (top 5th percentile) are only weakly correlated with better performance. Finally, and perhaps most interesting, they were able to detect selection bias through their experiment design but were unable to do so using advanced statistical methods.




While the result of women's bidding behavior is interesting, I think the last point, that even advanced statistical methods were unable to detect a selection bias when there was one, is the most important. We need to remain skeptical of results, both controlled experiments and natural experiments, and carefully look over the methodology before accepting all conclusions. While statistical analysis can tell us much about what's going on that isn't immediately apparent from observation, careful experiment design and analysis can catch some things that don't necessarily show up statistically.

The paper was also an interesting and entertaining piece of auction theory. Now I need to find and read some papers by John Kagel and Dan Levin as well as John Kagel and Jacques-Francois Richard to find out more about the risk neutral Nash equilibrium bidding function. Following citations in papers is almost as addictive and probably even more time consuming than following Wikipedia links.