A STABLE COINTEGRATED VAR MODEL FOR CREDIT RETURNS WITH TIME-VARYING VOLATILITY

摘要

In order to forecast the daily returns for corporate bonds with a given credit quality (credit rating) and a given maturity, we set up a framework based on cointegrated vector-autoregression. We assume the variables of the model to follow a stable law as they exhibit peakedness and heavy-tailedness. For the residuals, we observe time-varying volatilities (volatility clustering). When dealing with Value-at-Risk (VaR) applications, the forecast of conditional volatility and covariance is crucial. Therefore, aside from unconditional stable modeling of the dependent residuals, we examine two different volatility models for these, we compare the predictive accuracy of the multivariate stable GARCH(1, 1) with the constant correlation matrix and the stable Exponentially Weighted Moving Average (EWMA) model.