When OTC markets become illiquid and dealers fail to buffer flows, institutional investors effectively face each other directly in the market. They can observe each other’s actions and position changes. For example, if large investors make offers to sell under illiquidity, the market expects to become “over-positioned” and will avoid bids at a fair price or even put in offers. In equilibrium investors transact at prices below true value and exacerbate initial negative shocks.
This post ties in with the subject of “price distortions and market liquidity”, a synopsis of which can be found in our summary of price distortions.
The below are excerpts from this rather technical paper. Headings, links and cursive text have been added.
The gist
“This paper studies the interaction between financial fragility and over-the-counter (OTC) markets… In developed financial systems, investors participate in asset markets as part of financial institutions which trade assets (often over the counter) on their investors’ behalf and provide liquidity (withdrawal options) to investors.”
“When the search friction is severe…[the implied lack of liquidity causes] run equilibria—where investors announce low valuation of assets because they believe everyone else in their financial institution is doing the same…Run equilibria are more likely in OTC markets with severe search frictions.”
The model
“I build a model embedding the main ideas of financial fragility discussed in the…literature, into a dynamic model of over-the-counter (OTC) markets. The model allows me to study the connection between this market structure and the existence of runs [self-reinforcing price dynamics] in the financial sector.”
“Investors are divided in groups and their preference shocks are privately observed. There are two types of agents: investors and dealers. There is an asset in fixed supply. Investors derive utility from holding this asset and dealers do not…Financial institutions and dealers trade assets in an over-the-counter market. The financial institution problem is to maximize the ex-ante utility of its own investors, taking as given what other agents in the economy are doing. Investors face two frictions. They may not meet a dealer to trade in the inter-dealer market when they need, and their investment decisions are distorted because dealers extract part of the gains from trade in the bargaining process.”
“In the periods where the financial institution does not meet a dealer, it efficiently reallocates assets among the financial institution investors. And in periods where the financial institution does meet a dealer, it minimizes the amount of assets traded with him, which reduces the rent he extracts.”
N.B.: If an institutional investor does not find a dealer to buffer its flow (by warehousing the risk on balance sheet), it can only effectively transact with other institutional investors, even of the flow is still channeled through a broker-dealer.
The findings
“I show that a truth-telling equilibrium exists if the bargaining power of dealers is sufficiently low…the truth-telling equilibrium is the unique equilibrium if the financial institutions’ probability of meeting a dealer is high enough. The literature on over-the-counter markets associates this probability with the degree of search friction in the economy.”
N.B.: A truth telling equilibrium is a term from game theory describing a situation where agents fare best by being truthful regardless of what others do. In the context of this model they then have an incentive to put in bids or offers in accordance with their true valuation of an asset and their true preferences.
“When the probability of meeting a dealer is not high, however, non-truth-telling equilibria also exist…In the “run” state, investors in a subset of the financial institutions announce low valuation for holding the asset independently of their true preference shock.”
“What generates incentives for a run in equilibrium? The answer lies in the liquidity insurance provided by the balanced team mechanism. This mechanism can be interpreted in the following way. Investors agree to buy or sell assets from other financial institution investors in case they have the desire to sell or buy assets and it is costly (or unfeasible) to trade it with a dealer…[This is] a form of liquidity insurance that is also embedded in the contract offered by several real-life financial institutions. The price an investor pays or receives for these assets is based on the expected welfare impact of his own announcement [by visibly making a bid or offer in the market and hence revealing asset value changes from his perspective].”
“When investors truthfully announce their preference shocks [i.e. transact in accordance to the true changes of an asset’s value from their perspective], this liquidity insurance improves the allocation of assets in the financial institution by equalizing the marginal utility of consumption of asset dividends. But this insurance also has another effect. The amount of assets purchased by a particular investor is increasing the other investors’ desire to sell the asset. Hence, when an investor believes for sure other investors will announce they want to sell their assets, he expects to have a lot of assets and, since there is decreasing marginal utility from holding assets, he expects a low marginal utility from holding more of those assets…Under this belief, he may prefer to untruthfully announce [his own] desire to sell his own assets as a way to avoid buying assets at a price that is higher than the marginal benefit he gets from buying those assets. And, of course, if all investors have this same belief, they all prefer to announce the desire to sell their assets—confirming their initial beliefs in equilibrium and generating a self-fulfilling run on the financial institution.”
“The runs have several implications for market outcomes. The price of the asset drops in the run states. Trade volume initially spikes, following a ‘fire sale’ by the financial institutions under a run, then collapses. The average bid-ask spread can go up or down with runs, depending on the fraction of financial institutions under a run, where the bid-ask spread in the model is the difference between the price in a financial institution-dealer meeting and the price in the inter-dealer market. When about half of the financial institutions are under a run, the average bid-ask spread increases in the run state. Otherwise, it decreases.”