Date of Award

6-2019

Degree Name

Doctor of Philosophy

Department

Economics

First Advisor

Dr. C. James Hueng

Second Advisor

Dr. Eskander Alvi

Third Advisor

Dr. Peng (Roc) Huang

Abstract

The main objective of my three essays is to incorporate liquidity shocks and the linkages between the liquidity condition of financial markets into asset pricing and valuation models. The first essay focuses on the liquidity adjusted capital asset pricing model, while the second and the third essays examine the popular asset valuation model called the Fed model.

The first essay investigates the pricing of the commonality risk in the U.S. stock market by using a more comprehensive market illiquidity measure that can reflect the liquidity condition of different asset markets. This measure is given by the yield difference between commercial paper and treasury bill. In addition, consistent with the definition of commonality risk, I form portfolios based on the sensitivity of each stock’s illiquidity to the market-wide illiquidity. Using monthly data from January 1997 to December 2016 and the conditional version of the Liquidity-adjusted Capital Asset Pricing Model (LCAPM) estimated by the Dynamic Conditional Correlation approach, I find a significant commonality risk premium of 0.022% and 0.014% per year for 12-month and 24-month holding periods, respectively. This premium estimate is significantly higher than those found using the market illiquidity measure and estimation procedures from previous studies. These findings provide evidence that a security’s easiness in terms of tradability at times of liquidity dry up is extremely important. It is also higher than the excess return associated with other forms of liquidity risk. In addition, the paper finds a variation in the estimated commonality risk premium over time, with values being higher during periods of market turmoil. Moreover, estimating the LCAPM with the yield difference between commercial paper and treasury bill as a measure of market illiquidity performs better in predicting returns for the low commonality risk portfolios.

The second essay examines the inflation illusion hypothesis in explaining the high correlation between government bond yield and stock yield as implied by the Fed model. According to the inflation illusion hypothesis, there is mis-pricing in the stock market due to the failure of investors to adjust their cash flow expectation to inflation. This led to a co-movement in stock yield and government bond yield. I use the Gordon Growth model to determine the mis-pricing component in the stock market. In the next step, the correlation between bond yield and stock yield is estimated using the Asymmetric Generalized Dynamic Conditional Correlation (AG-DCC) model. Finally, I regress this correlation on mis-pricing and two other control variables, GDP and inflation. I use monthly data from January 1983 to December 2016. Consistent with the Fed model, the paper finds a significant positive correlation between the yield on government bonds and stock yield, with an average correlation of 0.942 - 0.997. However, in contrast to the inflation illusion hypothesis, mis-pricing in the stock market has an insignificant impact on this correlation.

The third essay provides liquidity shocks contagion between the stock market and the corporate bond market as the driving force behind the high correlation between the yield on stocks and the yield on government bonds as implied by the Fed model. The idea is that when liquidity drops in the stock market, firms' credit risk rises because the deterioration in the liquidity of equities traded in the stock market increases the firms’ default probability. Consequently, investors’ preferences shift away from corporate bonds to government bonds. Higher demand for government bonds keeps their yield low, leading to a co-movement of government bond yield and stock yield. In order to test this liquidity-based explanation, the paper first examines the interdependence between liquidity in the stock and corporate bond markets using the Markov switching model, and a time series non-parametric technique called the Convergent Cross Mapping (CCM). In order to see the response of government bond yield and stock yield to liquidity shocks in the stock market, the study implements an Auto Regressive Distributed Lag (ARDL) model. Using monthly data from January 1997 to December 2016, the paper presents strong evidence of liquidity shocks transmission form the stock market to the corporate bond market. Furthermore, liquidity shocks in the stock market are found to have a significant impact on the stock yield. These findings support the illiquidity contagion explanation provided in this paper.

Access Setting

Dissertation-Open Access

Included in

Economics Commons

Share

COinS