During financial crises, distressed companies often hold the troubled assets for a long time. For example, years after the crisis started since 2007, there is little sign that major banks have offloaded much of the alphabet soup of debt that caused repeated write-downs, including the mortgage-backed securities (MBS), collateralized debt obligations (CDO), etc. The inability or unwillingness to sell via the market channel raises the concern with market liquidity. Curiously, the banks’ burdens of troubled assets coexist with large amount of outside capital that can be a potential source of liquidity yet, for some reasons, has not stepped in. For example, as of the first quarter of 2008, there are $3.4 trillion, $5.3 trillion, and $7.8 trillion of money market funds, Treasuries (excluding Agency and Government Sponsored Enterprise (GSE)), and time/saving deposits respectively, according to the Federal Reserve statistical release (2008). As of the second half of 2007, the notional amount of interest rate/foreign exchange derivatives and equity derivatives totaled $382 trillion and $10 trillion respectively (International Swaps and Derivatives Association (2007)). Although many of the derivatives positions are for hedge purposes, the data suggest neither a lack of capital nor a lack of appetite for risk. What reasons then delay the capital from flowing to the trouble assets?
This paper uses a dynamic model of heterogeneous beliefs to study the provision of market liquidity. In this model, investors have different priors at the start of a downturn and update their beliefs based on Bayesian learning from common information. In another word, the investors agree to disagree (Aumann (1976)). Trade (liquidity provision) opportunity arrives when beliefs cross, i.e., when an initial pessimist becomes more optimistic than an initial optimist that holds the troubled asset.
The model points out two reasons that can delay the provision of market liquidity. First, for some investors, beliefs may not cross and may instead diverge. I.e., the pessimists become more pessimistic relative to the optimists. This is seemingly contradictory to the prediction from Bayesian learning that beliefs converge when more data are available. The reason is that the convergence occurs for the belief parameter, which is the recovery probability in the model. However, asset value relates to the expected length of the bear market, which is the inverse of the recovery probability. In a prolonged bear market, Bayesian investors update their recovery probability towards zero. This, under certain conditions, leads to divergence in perceived asset values. This finding is unique to prolonged bear markets or severe financial crises. Such divergence in valuations precludes some investors (pessimists) from providing liquidity. These pessimists buy few troubled assets and impose high financing cost to fund the optimists’ purchases. Belief dynamics constructed from analyst target prices for ten major banks indicate that such predicted belief divergence occurs for some investors during 2007-08.
Download
Option Value of Cash
