Comparisons and extensions of structural and reduced form approaches to the pricing of commercial real estate securities and loans in the financial crisis (2007-2010) and the recovery (2013-2014)

To date, the ~$1Trillion CMBS sector in the US does not actively utilize widely accepted advanced derivatives valuation methods. In the absence of risk neutral values for CMBS it is proposed here that risks of default were neither correctly anticipated nor priced in the Crisis (11/2007-12/2010) nor in the Recovery (1/2013-3/2014), thus far. If schisms between market and model prices enable one to secure excess returns then one may reasonably question the weak form efficiency of the CMBS sector. To investigate, I apply four model approaches (structural form, reduced form, generalization of calibrated simulation, and a special case of the generalization) in both the Crisis and the Recovery using two representative loan and bond samples on a daily basis. The key findings are: First, statistical analysis demonstrates the need for the generalized approach. The special case is misspecified and inadequate to the task of modeling CMBS default risk. Second , although the structural form yields results in keeping with the generalization, it too is insensitive to risks associated with loan characteristics, borrower behavior, and bond pricing. Third , the reduced form represents a comprehensive and better approach than all others. Building off details that characterize the generalized approach, the Cox Process of the reduced form has embedded within its design the capability to accurately evaluate complex economic relationships that govern the timing and amount of loan defaults. As the reduced form economy is robust, accurate pricing at the bond level is an immediate consequence, given accurate implementation. Finally , evidence indicates a sizable disconnect between fair value and market pricing with differentiation amongst the models. Trading tests and statistical analyses suggest an inefficient CMBS market evidenced by the earning of excess returns in backtesting. This dissertation provides valuable insights pertaining to CMBS risk estimation, the pricing of those risks, and CMBS market efficiency.Ph.D.Includes bibliographical referencesIncludes vitaby Andreas D. Christopoulo