The first paper ("An Analysis of REIT Credit Default Swap Pricing") first devise a closed-form solution of a non-arbitrage pricing model with observable factors in the default hazard rate to value Credit Default Swap (CDS). Then, I conduct panel regression to examine the explanatory power of REIT-specific and macroeconomic covariates in explaining the cross-sectional variation of CDS spreads. The high level of R-squared from the regression highlights the role of my list of fundamentals in determining credit risk. The second paper ("The Flow of Credit Risk Information among REIT Securities") discovers the credit risk information flow among REIT stocks, bonds, and CDS markets. In general, information flows from stocks to CDS, and then to bonds. However, there is a reversal of information flow around credit rating downgrades, where CDS leads stocks. Furthermore, I find evidence that large banks active in the CDS market can exploit private information obtained through their direct lending relationships. I conclude that the CDS market appears to be the primary market for trading on REIT credit risk information. The third paper ("The Predictability of REIT Index Returns") applies various machine learning and deep learning models to predict the out-of-sample REIT Index returns. Compared with traditional OLS method, machine learning algorithms significantly improve the predictability of REIT Index returns. To exploit the economic significances of machine learning models, I create a practical investment strategy which produces a substantial profit.