"Default Prediction Around the World: The Effect of Constraints on Pessimistic Trading"
Assistant Professor Mark Maffett, University of Chicago USA
6th Sep 2013 11:30 am - 12:30 pm Room 214/215 Economics and Business Building (H69)
Research examining cross-country differences in the ability of market participants to accurately assess a firm's likelihood of default using publicly available sources of information is virtually non-existent. This paper examines one potential source of such variation, constraints on pessimistic trading (i.e., trades made in anticipation of future price declines). On average, predictive accuracy is significantly greater in countries where pessimistic trading is less constrained. This relation is further identified using time-series variation in restrictions on short selling and the introduction of put option trading. Consistent with trading constraints limiting the extent to which prices reflect publicly available default risk information, the direct incorporation of accounting information in the default prediction model leads to a larger improvement in accuracy where pessimistic trading is limited. Finally, although fewer constraints consistently lead to more accurate identification of actual defaults, during periods of heightened macroeconomic uncertainty, default prediction models in countries with fewer¿¿ pessimistic trading constraints inaccurately classify a greater proportion of non-default observations.