Are currency traders the last ones to know? In the February 2014 draft of their paper entitled “Cross-Asset Return Predictability: Carry Trades, Stocks and Commodities”, Helen Lu and Ben Jacobsen investigate whether commodity and stock index returns predict currency carry trade performance. They consider equally weighted carry trade strategies that each month buy (sell) one-month forward contracts for the one, two or three currencies with the highest (lowest) beginning-of-month interest rates and hold to maturity. They account for bid-ask spreads and express profits in U.S. dollars. They evaluate the power of three commodity indexes (CRB Spot, CRB Raw Industrials Spot and CRB Metals Spot) and three total return equity indexes (MSCI All Country, MSCI World and S&P 500) to predict carry trade profitability. Using monthly levels of the commodity and stock indexes and monthly one-month forward rates and spot rates for the G-10 currencies during February 1988 through December 2011, they find that:
- Over the sample period, the New Zealand dollar, Australian dollar and British pound (the Japanese yen, U.S. dollar, and the Swiss franc) are most frequently the high-interest rate (low-interest rate) currencies. Conventional carry trades involving the one, two or three top and bottom currencies generate average monthly returns of 0.37%, 0.24% and 0.20%, respectively.
- There is a strong positive relationship between commodity index return three months ago and currency carry trade long-side profitability this month. For example, a one standard deviation movement in commodity indexes (about 2.8%) three months ago predicts an increase of 0.24 standard deviation (about 0.83%) in carry trade long-side profit this month.
- There is a positive relationship between stock index return this month and currency carry trade short-side profitability over the next two months. For example, a one standard deviation movement in equity indexes (about 4.5%) this month predicts an increase of 0.20 standard deviation (about 0.6%) carry trade short-side profit over the next two months.
- Currency carry trade timing strategies based on these predictive relationships generate significantly higher Sharpe ratios than comparable conventional carry trade strategies.
- Results are robust to controlling for a variety of other potential currency exchange rate predictors.
In summary, evidence indicates an exploitable flow of information from commodity and equity markets to currency exchange markets.
Cautions regarding findings include:
- While accounting for the bid-ask spread in currency carry trade portfolio reformation and timing, the study apparently does not account for broker fees or capital requirements.
- Consideration of multiple currency carry trade portfolios, multiple potential predictors and multiple lags between signals and results introduces data snooping bias, such that the best-performing combination likely overstates expected performance.
