The dissertation composes of two essays on credit risk in mortgage lending and securitization. The first essay studies the impact of credit rating agency interactions on credit rating quality. The second essay focuses on credit risk of high loan-to-value (LTV) mortgage.

The first essay is an empirical study of credit rating agency behaviors in asset-baked security market. Credit rating agencies serve a critical function in alleviating asymmetric information in financial markets. However, their independence and objectivity have long been a concern of investors and regulators. By analyzing the credit ratings of 17,889 subprime ABS bonds, I identify strategic interactions among credit rating agencies in rating assignments and rating changes. Regression analysis of the probability of rating changes shows that credit ratings are less stable when more rating agencies rate a bond. I distinguish two competing explanations for these higher probabilities, with loose original rating standards as the main cause. The number of original ratings positively correlates with credit support for all rating agencies during the period when subprime credit worsened, demonstrating that rating accuracy decreased when more rating firms rated a bond. Since rating stability can change, I test the impact of multiple ratings on credit rating quality to rule out this alternative explanation for a higher probability of a rating change. For Standard & Poor's, I show that a higher probability of a rating change did not result from a change in rating stability choice, since the probability of a subsequent rating change was not affected by the number of original ratings. For Fitch, the test does not eliminate the alternative explanation. Further analysis of the number of rating changes is consistent with results of the first test.

The second essay takes structured credit modeling approach to show theoretically how loan-to-value (LTV) ratio affects credit risk in mortgage and quantifies the credit risk of first lien mortgage and second lien mortgage. Default risk is derived implicitly. Optionality of defaultable debt results in an upward sloping credit supply curve in terms of a function of interest rate with respect to LTV. Current regulation in high LTV mortgage is shown to create a funding advantage in separating a high LTV mortgage into a lower funding cost first mortgage and a higher cost second mortgage.