Essays on Search Frictions in Financial Markets
- Author(s): Uslu, Semih
- Advisor(s): Weill, Pierre-Olivier
- et al.
This dissertation consists of three chapters about search frictions in financial markets.
Chapter 1: “Pricing and Liquidity in Decentralized Asset Markets”
I develop a search-and-bargaining model of liquidity provision in over-the-counter markets where investors differ in their search intensities. A distinguishing characteristic of my model is its tractability: it allows for heterogeneity, unrestricted asset positions, and fully decentralized trade. I find that investors with higher search intensities (i.e., fast investors) are less averse to holding inventories and more attracted to cash earnings, which makes the model corroborate a number of stylized facts that do not emerge from existing models: (i) fast investors provide intermediation by charging a speed premium, and (ii) fast investors hold larger and more volatile inventories. I also calibrate the model, demonstrate that it produces realistic quantitative outcomes, and use it to study the effect of trading frictions on the supply and price of liquidity. The results have policy implications concerning the Volcker rule.
Chapter 2: “Price Dispersion and Trading Activity during Turbulent Times”
I construct a dynamic model of crises in a decentralized asset market that operates via search and bargaining. The crisis is modeled as a one-time aggregate shock to uncertainty with a random recovery. The arrival of the crisis shock leads to an increase in both the volatility of asset payoff and the volatility of investors’ background risk. The equilibrium path for investors’ valuations, terms of trade, and the distribution of investors’ positions is characterized in closed form both during the crisis and during the recovery. Tractability of the model allows me to derive natural proxies for price dispersion and trading activity. I show that both volatility of asset payoff and volatility of background risk contribute to higher level of price dispersion during the crisis. Trading activity might be higher or lower depending on the increase in the volatility of background risk relative to the increase in the volatility of asset payoff, consistent with the “flight-to-quality” observations during extreme episodes. A flight to the asset market always starts with a “heating-up” in trading activity but a flight from the market might start with a dry-up or heating-up during the onset of the crisis. If the relative increase in the volatility of asset payoff is too high, a period of fire sales is triggered leading to a short heating-up before the complete dry-up of the trading activity. I calibrate the model according to the U.S. corporate bond market data and show that it captures the observations during the subprime crisis.
Chapter 3: “Endogenous Liquidity and Cross-section of Returns in Dynamic Bargaining Markets”
The empirical analysis of liquid/illiquid asset pairs reveals the existence of a return differential (liquidity premium) between those types of assets. The time variation in liquidity premia is delineated by the term "flight-to-liquidity," meaning that liquidity premia are higher during extreme market episodes. In this paper, I extend the search-and-bargaining model of Weill (2008) by allowing for risk aversion, to explain this observation. Risk-averse investors optimally allocate their limited budgets of search efforts to various assets. This extension allows me to examine the relationship between risk and liquidity of assets in the cross-section and over time. My model generates endogenous cross-sectional liquidity differentials corroborating much of the empirical evidence. Furthermore, I show that when asset payoffs are more volatile, trade surpluses are higher because idiosyncratic hedging quality differentials are wider. Higher trade surpluses lead to higher value of search, and in turn, higher opportunity cost of committing to a particular asset, especially to an illiquid one. Therefore, periods of high volatility are associated with a flight-to-liquidity.