| Popis: |
My dissertation aims to understand the impact of financial market frictions on stocks' expected returns. It contains three chapters.Chapter 1 documents that firms borrowing from the public bond market on average have 5.7% per annum higher subsequent stock returns than firms borrowing from banks. The return spread is larger, about 9% per annum, among less profitable and small firms. Neither the q-factor model nor the Fama-French 5-factor model can explain this return spread. Bonds are difficult to renegotiate in bad states due to the large number of creditors, whereas bank loans are easy to renegotiate. Borrowing from the bond market makes a firm riskier than borrowing from banks all else equal. On aggregate, the percentage of the corporate bond in the total corporate debt has superior in-sample and out-of-sample performance in predicting the equity premium, which is consistent with the interpretation of the risk associated with sources of debt.Chapter 2 examines asset returns in a production economy where firms face two types of aggregate uncertainty, a productivity shock and a financial shock. Borrowing constraints reduce firms' choice set when facing productivity shocks. Exogenous shocks to the financial market distort firms' optimal production plan due to the constraint on firms' working capital. The amount of systematic risk rises, comparing to the standard business cycle model. I develop a quantitative dynamic stochastic general equilibrium model to evaluate the impact of financial uncertainty on equity risk premium. Calibrated to the US data, the model generates sizable equity premium and stable risk-free rate while matching moments of aggregate economic quantities.Chapter 3 investigates the predictability of factor premium. While the excess return on the aggregate stock market is difficult to forecast, some macroeconomic variables can reliably predict the premium on factors that are used to explain the cross-section of stock returns. In particular, both long-term and short-term government bond yields have significant predicting power with respect to the premium on profitability factors and momentum. A variable that represents the volatility of the stock market can also predict momentum. Using the method of bootstrapping adjusted r-squared proposed in Welch and Goyal (2008), we show that the predictability is significant both in-sample and out-of-sample during the period of January 1968 to December 2016. Our findings suggest that the linkage between stock returns and macroeconomic variables is more robust at a disaggregate level |
