Design of selling strategies for heterogenous divisible goods auctions with endogenous informational asymmetry is an important policy question. This problem can be analyzed empirically using the distributions of ex- ante valuations of bidders, the value of information and the degree of informational asymmetry. In this paper, I estimate these by a three step procedure from a dynamic auction model with endogenous informational asymmetry. The seller sells multiple goods via a sequence of first price auctions. While bidders are ex-ante symmetric, the first period winner has an informational advantage in the second period bidding game and becomes a strong bidder. This endogenous asymmetry leads to excessive entry and overbidding in the first period relative to a one period game. I characterize the equilibrium in terms of the observed bid distribution and entry behavior. I apply a three step estimation procedure to data on OCS oil tract auctions. I find that the federal government is only ecovering 23% of the 'strong' buyers' willingness to pay in the second period. Bidders perceive the value of information to be at most 12% of their first period's informational rent. A new semiparametric structural test cannot reject the hypothesis of the strong bidder's informational superiority in the second period and sets it at 18% relative to the weak bidder. I use the estimates to design alternate mechanisms and empirically show that government's revenue increases when the asymmetry is taken into account in allocating the goods.