Preference-Based Strategic Currency Hedging

This article considers the value to a given investor of volatility reduction through currency hedging relative to taking an unhedged position. Included is a decision framework to help international investors analyze how the potential risk reduction benefits of hedging interact with the investor's beliefs about the currency risk premium, risk aversion, and investment constraints to determine an optimal currency-hedging strategy. In our framework, investors choose from nine hedging strategies, each of which may be optimal under a specific set of assumptions. The article concludes that a unitary hedge strategy is best for investors who do not forecast currency returns. For those who make strategic currency-return forecasts, our aggregated results suggest an optimized hedge strategy using expected return inputs from an asset-pricing model. Today, the majority of large institutional portfolios are invested internationally. With investment in global markets comes the challenge of establishing a policy for managing currency risk—which poses several difficult problems. Currency risk usually increases the volatility of portfolio returns, and the manager may address this issue by reducing or eliminating currency risk through hedging. If bearing currency risk brings positive average returns, however, hedging may reduce the portfolio's expected return. The article addresses several questions:Should currency risk be hedged?If so, how much of the risk should be hedged?How does the optimal choice relate to the portfolio manager's risk preferences and beliefs about a currency premium?If the portfolio manager uses expected currency returns to choose the hedge ratio, what estimation method works best?The article organizes a preference-based decision framework to assist international investors in choosing a currency-hedging strategy. Empirical evidence on the performance of the recommended alternative hedging strategies lends broad support to the decision framework.The article begins with descriptions of different approaches to hedging. The simplest type of hedge involves a currency position equal in magnitude and opposite in sign to the equity position—that is, a “unitary” or “100 percent” hedge. A less-restrictive strategy uses least-squares regression to obtain minimum-variance hedge ratios. These hedge ratios take into account the correlation between the underlying equity and the currency. In the multivariate formulation of minimum-variance hedge ratios, currency cross-hedging is allowed. A more general form of currency hedging involves solving for the optimal currency weights in light of fixed equity positions. Known as “partial optimization,” this approach results in optimal currency weights that include a hedging component and a speculative currency component. Finally, the most general form of hedging is joint optimization of currencies and equities.In practice, most international asset managers or their overlay submanagers use more restrictive strategies than joint optimization. Even partial optimization is often not implemented in its most general form: Managers often place constraints on the magnitude of currency positions. Academic research offers managers some help, but it frequently focuses on unconstrained problems for a general investor. Of more interest to practitioners is the analysis of constrained hedging for specific investor types.We used 17 years of data to examine the out-of-sample performance of nine hedging strategies. For investors who do not forecast currency returns, we tested and compared the following hedging rules: no hedge, a unitary hedge, a minimum-variance hedge ratio, and a constrained minimum-variance hedge ratio. Although theory suggests using a minimum-variance hedge, we found that this strategy is empirically indistinguishable from using a simple unitary hedge.For investors who are willing to incorporate estimates of expected currency returns in their hedging decisions, theory suggests using a hedging strategy involving partial optimization based on equilibrium returns. We tested and compared a set of hedging rules for these investors—no hedge, a unitary hedge, a minimum-variance hedge ratio, and preference-specific optimized hedge ratios. As inputs into the optimization problems, we tried three approaches to generate the expected currency returns—historically based expected returns, equilibrium-based expected returns using the international asset-pricing model, and zero expected returns. The results supported the superior performance of the theoretically motivated strategy of using optimal hedge ratios based on equilibrium expected returns.