Essays on commercial mortgage-backed security

Structured finance products including Commercial Mortgage-Backed Security (CMBS) suffered tremendous losses during the 2008 financial crisis. My dissertation consists of three chapters that contribute to our understanding of the causes of the crisis. My first chapter is an empirical study on potential misrepresentation of CMBS. Although CMBS suffered large scale losses during the past financial crisis, currently, this segment of the structured finance market has almost recovered to its pre-crisis level. While evidence was found regarding the systematic misrepresentation of loan quality information for residential mortgages, there was no evidence of large scale misreporting for CMBS. This paper examines important financial variables reported in financial documentation of commercial mortgages such as Underwritten Net Operating Income (UW NOI). I find that, prior to the financial crisis, UW NOI was consistently over-estimated by an average of 7.8%. This overstatement lead to Loan-to-Value ratio and Debt-Service Coverage Ratio being misreported as 67.1% from 84.2% and DSCR as 1.72 from 1.59. The levels of aggregate over-estimation substantially differed among originators and the variations explained the performance differences between originators. Each 1% increase in over-estimation resulted in a 20% higher likelihood in delinquency. The ratings issued by rating agencies failed to capture the adverse impact from over-estimation on CMBS performance. The second chapter of my dissertation studies the CMBS credit rating market using a strategic interaction model. The 2008 financial crisis that arose in the mortgage market has brought renewed attention to the failure of the credit rating mechanism. Using Bloomberg data, I conduct a structural analysis of strategic credit rating behaviors in the Commercial Mortgage-Backed Security (CMBS) market. This chapter models the CMBS credit ratings as strategic behaviors that reflect the peer effects from other rating agencies. Peer effects are incorporated through the estimation of market “beliefs” about the ratings. We establish semiparametric identification of the model by exploiting an exogenous equilibrium shift due to the financial crisis. Moreover, the model is estimated using a two-step estimation procedure. The empirical results strongly support the presence of positive peer effects. By including peer effects, the fitness of our model has been significantly improved. The third chapter examines the entrant-related consequences in the CMBS credit rating market after the financial crisis. I find that the entrant has given more lenient ratings than the incumbents. Among securities that obtained ratings from both entrant and incumbent rating agencies, 13.8% are granted a higher rating from the entrant than the incumbents from 2011-2014. In addition, deal level and loan level analyses further provide evidence that the entrant granted CMBS with 2.25% higher AAA-rated portion while the underlying loans in these CMBS are 10% more likely to become delinquent than other rating agencies. The lenient ratings from the entrant coincide with the sharp increase in the entrant's market share.