Friday, January 19, 2018 at 11:40am to 1:10pm
Ives Hall, 115
B07 Tower Rd, Ithaca, NY 14853, USA
Giulia Brancaccio - Princeton University
Learning by Trading: The Case of the US Market for Municipal Bonds (joint w/Dan Li & Norman Schürhoff)
In markets where public information about fundamentals is limited, and trade takes place under conditions of asymmetric information, agents may rely on their trading activity to acquire information about the state of market fundamentals. Information acquisition, therefore, becomes an additional motive for trade. In this paper we use a combination of reduced-form techniques and structural analysis to characterize and measure experimentation motives for trade in the U.S. secondary market for municipal bonds. First, we provide reduced-form evidence that experimentation is a first-order motive for trade. To rationalize these facts, we design a dynamic model of trade in this market that allows for linkages between trading activity and information acquisition (i.e., experimentation). The model is estimated using detailed micro-data on trading activity on the secondary market for municipal bonds. We use the model to characterize the incentives to experiment. We find that dealers are willing to pay up to 15% of the intermediation spread to double the precision of their information about the state of fundamentals. Furthermore, we show that experimentation allows dealers to increase the precision of their estimate of the asset’s value by 25%, and we characterize the process of information diffusion across agents. Finally, we find that experimentation explains up to 10% of the volume of trade in the market.