Volume 15, Number 2, 2017
Andreas D. Christopoulos
This paper presents a structural generalization for pricing commercial mortgage backed securities (CMBS) and their derivatives, CMBX. I compare results for the structural generalization with a reduced-form approach using identical data sets and analyses. My comparisons are made at both the loan and bond levels and cover the period November 2007 through June 2015 using $389 billion of loans serving as the underlying collateral for CMBX Series 1 through 8. The sole difference between the two modelling approaches is found in the set of conditions and methods for simulating the default event which together comprise the ‘default trigger’ that differ for each model. I statistically validate the default estimations and then construct an automated long/short trading strategy using the risk measure Theta to compare the impact of default estimates on investment and risk management decision making. The findings indicate the reduced form provides greater precision than the structural generalization in estimating default events and in assessing trading opportunities.