This paper proposes a semi-parametric method to uncover the distribution of bidders' private information in the market for highway procurement when unobserved auction heterogeneity is present. I derive sufficient conditions under which the model is identified and show that the estimation procedure produces uniformly consistent estimators of the distributions in question. The estimation procedure is applied to data from Michigan highway procurement auctions. I estimate that 75% of the variation in bidders' costs may be attributed to the factors known to all bidders and only 25% may be generated by private information. My results suggest that failing to account for unobserved auction heterogeneity may lead to overestimating uncertainty that bidders face when submitting their bids. As a result both inefficiency of the auction mechanism and mark-ups over the bidders' costs may be overestimated.