The impact of auction choice on revenue in treasury bill auctions - An empirical evaluation

I evaluate whether uniform price or discriminatory auctions are revenue-superior for selling Treasury bills. To this end, I apply two structural econometric models, Hortaçsu and McAdams (2010) and Fevrier et al. (2002), to a dataset on Polish zero-coupon bonds. My secondary aim is to analyze mutual inconsistencies in prediction from these models. I find that both agree on the revenue-superiority of discriminatory auctions, by between 0.01% and 1.5%; the models' predictions are contradictory in only 7% of auctions. The large-scale agreement of two vastly different models gives confidence that the conclusions on Polish data are robust to varying many of the underlying economic and econometric assumptions, and not likely just a modeling artifact.