C93 - Field ExperimentsReturn

Learning is a subject of intense research in economics. We present persuasive evidence that learning took place among uninformed heterogeneous agents during a large-scale naturally-occurring set of auctions. Empirical study employs a unique bidding data set of 5000 individual investors that placed their bids in the voucher scheme that in terms of size, incentives, and variation is one of the largest experiments ever conducted. To detect and quantify learning we develop new measures of individual performance during the bidding process on the artificial stock market where prices of goods vary over successive stages of bidding on the basis of supply and demand.