During the financial crisis of 2007 to 2009, there was a stunning difference in how assetmarkets were affected according to their infrastructure. Markets with centralized trading functioned rather well. To the contrary, in over-the-counter markets where trading takes place on a decentralized and ad-hoc basis and where assets are less standardized and arguably more opaque trading came to a halt. Most prominently, collateralized debt obligations, asset backed securities and commercial paper were traded only sporadically or not at all (see Gorton and Metrick, 2010). This prompted massive government intervention in the from of short-term liquidity provision and long-term asset purchases which led to a (partial) resurrection of trading in these markets over time.
Markets for structured debt financing experienced a sudden, unexpected deterioration of the average quality of assets and some observers have pointed to adverse selection as the source for these markets not functioning properly. However, the question arises whether the peculiar features of over-the-counter trading opaqueness of assets and search for a counterparty made those markets more prone to a market freeze.
If so, are these features important to understand the role and design of a government intervention in these markets? In this paper, we find that search frictions in asset markets compound adverse selection, making the market more fragile to unexpected shocks to the quality of assets. The trading frictions also determine the response of the market to a government intervention and thus influence the optimal design of a policy aimed at resurrecting trading in those markets.
To derive these insights, it is natural to combine two existing strands of literature on market mircostructure which have emphasized private information and trading frictions separately. Starting with Kyle (1985) and Glosten (1989), models of traders that are privately informed about the asset quality have been used widely to shed light on pricing and transaction costs in financial markets. More recently, a new approach to study price and trading dynamics has spawned from the work by Duffie, Garleanu and Pedersen (2005) that uses random search to model over-the-counter trading, but does not take into account that assets might be opaque.
