We present the …rst equilibrium search model with learning about an important characteristic of the aggregate market, tightness. In a steady state dynamic matching and bargaining game, the inflow of new buyers ("Demand") can be either high or low, depending on the state of the world. Buyers participate in sealed bid second price auctions. Bidders learn about the state through unsuccessful bids. There is no other source of information, i.e., all learning is endogenous. Beliefs depend on the bidding history, inducing endogenous heterogeneity into a population of ex ante identical agents. We show that there exists a unique equilibrium in symmetric, increasing bidding strategies.  We show that a seller has no incentive to reveal information about bids after the auction, since this extra information increases continuation values and thus depresses bids in the current auction. This result di¤ers markedly from static second price auctions where the linkage principle implies that revealing any information the seller has (or obtains from the bidders) increases revenues.