Resumen de Modelat i predicció de la volatilitat de l'actiu i del patrimoni net de l'empresa
Volatility of firm’s assets is one of the fundamental variables in credit risk modeling and refers to a degree of fluctuation of firm's asset returns. While equity volatility and its time-series properties have been extensively studied in the finance literature, little is known about the time-series behavior of volatility of firm's assets. The main reason for such a gap in the literature lies in the unobservability of the underlying value of the firm’s assets. For that reason, many practical applications applied in credit risk modeling are based on the use of stock return volatility as a proxy variable of asset volatility. However, recent studies point out in the direction that equity volatility has different time-series properties when compared to asset volatility. Thus, the main goal of this research is to study the time-series properties of firm’s asset volatility, the eventual differences it presents with respect to equity volatility, and to provide an in-depth understanding of its most relevant features such as asymmetry and long-range persistence in the context of volatility modeling and forecasting.
This research is structured as follows: First, we examine the persistence properties (long memory) of firm’s asset volatility and its relationship with equity volatility. Second, we determine which model or models, considering the discrete-time family of GARCH (Generalized Auto-Regressive Conditional Heteroskedasticicty) models are the best to estimate and forecast conditional asset volatility. We analyze in detail the implications of asymmetry and long-range persistence on modeling and forecasting firm's asset volatility. Third, we provide a practical application by estimating a CDS implied firm’s asset volatility and by analyzing its time-series properties. Throughout this research we use as a baseline sample a sample of 52 nonfinancial iTraxx Europe companies during the 2004-2016 period, and estimate underlying firm’s asset values using different, commonly applied procedures.
