Asset Value Correlation Bounds for Firms with Foreign Exchange Exposure

Abstract

This article deals with asset correlation estimation among firms with foreign exchange exposure. In most credit risk models the exchange rate risk is ignored, i.e. the borrowing firms are supposed to have assets denominated in the same currency as their debt. In reality, this is often not the case and if the asset portfolios of two borrowers are exposed to foreign exchange risk the correlation between these asset portfolios is biased upwards. The size of the asset value correlation bias depends on the time-series behavior of the two borrowers’ asset values and an empirical assessment of the size of this bias is therefore non-trivial since the asset values of the borrowers are non-observable. In this article, using a new way of estimating asset values, we attempt to estimate this bias for a set of major S&P500 firms in various industries. Our empirical findings support the theory and we find typical asset correlation estimates ignoring the exchange rate risk to be significantly biased. We therefore suggest that risk managers compute upper and lower bounds to the asset correlation estimates as a matter of routine and that if the exact asset correlation is required, it is estimated through a careful assessment of the foreign exchange exposure of the borrowing firms.