Date of Award
Doctor of Philosophy
This dissertation presents two papers that examine the efficacy of expected default frequency (EDF) when predicting bankruptcy events and whether short-selling utilization could be used as an indicator of default risk change. The first paper examines the efficacy of Merton's (1974) distance to default model as simplified by Bharath and Shumway (2008) to forecast a firm’s Expected Default Frequency (EDF) or probability of defaulting on debt obligations. Merton’s model, further developed by the KMV corporation, is based on the Black-Scholes asset pricing model. We apply the simplified Bharath and Shumway model by relating it with Cox’s proportional-hazards model (Cox, 1972) to forecast firm-specific expected default frequency (EDF) or probability of default. The accuracy of the Merton/Bharath-Shumway Model is further examined by including its input as one of the variables in a principal components-factor analysis to determine its ability to predict firm bankruptcy and its orthogonality in predicting firm default and potential bankruptcy. We also measure how much of the total variation is explained in a adjusted R squared decomposition analysis. The second paper investigate the relation between short interests of a firm’s common stock and commensurate changes in the firm's expected default probability/risk, using Compustat and NYSE data from 2007 to 2018. Default risk is measured by a simplified Merton distance-to-default (EDF) model, where we find that a firm’s short-interests predicts changes in default risk. Further, short-interest levels predict a firm’s likelihood of default as measured by it movement into the top default-risk decile. Thus, we recommend that investors, managers, and regulators employ short-interests as an early warning signal for a firm's potential default.
Dissertation or thesis originally submitted to ProQuest
Embargoed until 12/14/2023
Li, Huiyang, "EDF and short selling as indicator of default" (2023). Electronic Theses and Dissertations. 3045.
Available for download on Thursday, December 14, 2023