Carry is the return on an asset’s futures position when the asset’s price does not change over the holding period in question. It is applicable to any asset for which a synthetic futures position can be created. A security’s return consists of carry, expected price appreciation, and unexpected price appreciation.

The return and carry for a futures contract can be evaluated, and from current futures and spot prices, carry may be observed directly. Carry uniquely enables return predictability across any asset class, although the degree of predictability may vary.

Carry is a strong predictor of positive returns for each of the asset classes analyzed in cross section and time series. Carry strategies are defined as assuming a long position in one asset class and selling short another. Strategies that take a long position in assets with high carry and a short position in assets with low carry achieve strong risk-adjusted performance for each of the asset classes to which the carry trade applies. The Sharpe ratio averages 0.80. Other known return factors relate to carry returns but may fail to explain them. Serving as a centralized framework, carry appears uniquely able to capture return predictors from many different sources. The authors find that carry’s prediction strengths and ability to earn alpha hold up to numerous robustness tests.

Quantitative analysts would glean useful information from the authors’ findings. The ability to derive and apply carry to asset strategies could serve as a useful risk management and alpha generation tool.

Traditionally segmented by asset class, the literature on predictability of returns and the most common application of carry—namely, to the currency markets—informs the basis for the authors’ extension and application of the carry concept to other asset classes and subclasses. They construct, adapt, and apply the carry formula to global equities, commodities, finite maturity securities, global bonds, the slope of global yield curves, US credit market categorized by maturities and credit quality, US Treasury securities of varying maturities, and options. Depending on the availability of historical records, the statistical information set of carry and excess returns both within and across asset classes can be extensive, using data that may cover periods of more than 40 years for certain asset classes or sub-classes.

The analysis addresses how carry is connected to the various asset class returns, testing the applicability of uncovered interest parity and the expectations hypothesis against risk premiums. The authors define and construct a carry trade portfolio, applying different portfolio weighting approaches to which their test results are robust. Their inquiry considers carry returns both within and across asset classes and time. Through regression analysis, the inquiry considers the interrelationship between the carry return and the main return predictors for each asset class. Carry proves to be a strong predictor of returns, though for some assets more than others.

Correlation and factor exposure analysis reveals that carry strategies offer returns that exceed the asset classes’ local market returns without a great deal of exposure to individual asset classes themselves. Trading costs do not exert any meaningful influence on the authors’ findings.

Carry is a complex and not entirely explainable asset pricing phenomenon. Its anomalous nature seems unable to interpret downside, liquidity, and volatility risk entirely. Skilled investors could potentially exploit carry strategies to their advantage. Yet, some lack of data precision may be masking how carry is able to generate alpha. In any event, carry is a return component worthy of professional investors’ and academics’ further study in an increasingly fast-moving global investment climate where alpha is more and more elusive.