The credit risk of firms in the DJIA is often underestimated because of biased asset correlation assessments that arise when corporations’ assets are denominated in currencies different from their liabilities. The bias, which fluctuates broadly over time, even briefly assuming negative values, is affected by exchange rate volatility and fluctuations in the correlation of firms’ asset values with exchange rates.
Financial institutions spend considerable time and money assessing not only the credit risk of individual borrowers but also the default dependency between borrowers. A contributing factor to the riskiness in both of these relationships is the value and correlation of the borrowing firms’ assets, which, in turn, can be affected by exchange rate movements. The risk arises when there is a currency mismatch between a firm’s assets and liabilities, and the greater the mismatch, the greater the credit risk. The author’s research provides insight into the increased credit risk by quantifying the effect that exchange rate movements have on the asset correlation bias. His key finding is that this bias varies considerably over time and that, contrary to previous studies, the bias is sometimes negative for short periods.
How Is This Research Useful to Practitioners?
The author builds on previous work that revealed the relationship between currency risk and credit risk by testing the robustness of this relationship over time, focusing on the influence of the exchange rate movements of various currencies on the 30 firms in the DJIA. He finds the average asset correlation bias to be significant during 2000–2013, but it can take on a wide range of values and even be negative for brief periods.
Risk managers and company treasurers should take note that although both the volatility of the exchange rate and the correlations of firms’ asset values with the exchange rate affect this bias, it is more sensitive to exchange rate volatility. Specifically, the author finds that, all else being equal, for every 1% change in exchange rate volatility, the asset correlation bias changes ±1.51%; for every 1% change in the asset value–exchange rate correlation, the asset correlation bias changes ±0.23%. But in terms of economic impact, the greater amplitude of asset correlations make these correlations more prominent.
Equally important is that when financial markets are stressed (such as in 2008), the flight to quality and the safe-haven mentality can unintentionally exacerbate the underestimation of credit risk by undervaluing firms’ asset correlations. The bias is greater for certain currencies, but the effect is prevalent in all non-US-dollar currencies to some degree.
How Did the Author Conduct This Research?
The sample includes the 30 firms in the DJIA from 3 January 2000 to 29 January 2013. Currencies selected for the currency exposures are the euro, the British pound, the Japanese yen, the Chinese renminbi, the Argentine peso, and the US dollar. Daily stock price and exchange rate data are obtained from the Datastream website, and firms’ debt levels are obtained from Aswath Damodaran’s website.
The author’s methodology for computing firms’ asset correlation bias differs from previous studies in several respects, notably in the method he uses to estimate the market value of firms’ assets and in the leverage adjustment (downward) he applies to financial firms. When estimating firms’ asset values, he considers several methods but ultimately settles on the popular Merton framework. In that framework, asset values are derived from firms’ balance sheet data and stock prices by using option-based valuation methodology. The rationale for adjusting financial firms’ leverage is to ensure comparability with nonfinancial firms. The equation for estimating the asset correlation bias for firms with a currency mismatch (i.e., between a firm’s assets and liabilities) is a linear function of the bias for firms without a currency mismatch. The author applies this equation first to theoretically predict and second to empirically validate the asset correlation bias in firms. The results are mutually consistent.
I think the equations that estimate the sensitivity of the asset correlation bias to exchange rate volatility and the asset value–exchange rate correlation will be of interest to industry practitioners. Given the size and creditworthiness of the firms in the DJIA, I suspect financial institutions would view any credit exposures to these firms different from credit exposures to smaller, less ubiquitous firms. I would be interested to know whether financial institutions overlay any hedging strategies (e.g., futures or forwards) on loan portfolios that contain companies with material exchange rate mismatches or simply rely on the diversification benefits of the portfolio approach as a hedge.