Currency risk management for international investors : a thesis presented in fulfilment of the requirements for the degree of Doctor of Philosophy in Finance at Massey University, Palmerston North, New Zealand

Abstract

When investors hold assets priced in foreign currencies, they are exposed to foreign exchange (FX) risk. This thesis comprises three independent studies, each exploring a different aspect of currency risk management. For investors in the US market, while FX volatility increases the risk of unhedged international investment, its covariance with the portfolio reduces investment risk for most short-term holdings, providing a natural hedge. For Australian, Canadian, UK, and South Korean investors, the net risk contribution from FX risk reduces short-term investment risk without currency hedging, providing a full natural hedge. In many cases, the additional risk reduction from adopting the optimal hedge is not statistically significant relative to the zero hedge when investment-implied foreign currency exposure provides protection, except for long-term investments. The existence of a natural hedge removes practical barriers associated with currency risk for international investors by mitigating the need for costly hedging instruments in the short to medium term. In addition, our findings show that CIP violations do not affect the results of forward contracts and money market hedges; investors can choose whichever suits them. Adopting the dynamic conditional correlation generalised autoregressive conditional heteroskedasticity (DCC–GARCH) framework improves overall currency risk management performance compared to simple hedging and static optimal hedging strategies. Notably, when the currency return involves the British pound (GBP), the return series consistently requires a GARCH model with an asymmetric term. To accurately estimate conditional variances, investors should select a univariate model aligned with each asset's risk profile. Across the seven univariate models considered, four were selected as optimal for different return series. This finding underscores the pitfall of relying on any single model. Instead, investors should identify candidate models based on data characteristics, then select the best-performing model for each series. The DCC model adequately describes the dynamic correlations among assets for the asset, currency, and sample period chosen in this study, without the need for the ADCC model. The performance of the dynamic model shows that it minimises investment risk, requires less currency exposure, and improves hedged returns compared with the static model, especially for Canadian investors. Currency risk management is also effective for home-biased investors whose domestic assets comprise a large part of their portfolios. For investors in Canada, the European Union (EU), the United States (US), the United Kingdom (UK), Australia and South Korea, optimal currency exposures are similar regardless of whether a home-biased or diversified portfolio is adopted. In contrast, optimal currency exposures shift noticeably between the two portfolio comparisons for investors in Japan, Brazil, Indonesia, and South Africa. Currency hedging is effective at a similar level for Canadian, EU, UK, US, and Australian investors in terms of risk reduction. This finding carries a significant implication: for Canadian, EU, UK, and US investors, the choice between a home-biased and a diversified portfolio becomes irrelevant when currency risk is actively managed. In addition, we investigate the role of gold and bitcoin as alternative assets in improving the optimal hedge and protecting against the uncertainty of traditional safe havens. Adding gold futures contracts to the currency hedging strategy results in a positive optimal demand for gold among diversified investors, but it does little to reduce US dollar (USD) exposure or overall risk. However, it does improve hedged returns across selected countries. Conversely, the optimal demand for bitcoin remains small for currency risk management.