Carry is one of the three foundational systematic risk premia in futures and institutional portfolios, alongside trend following and mean reversion. This guide explains what carry is, how its signal is constructed across four major asset classes, the academic evidence establishing it as a persistent premium, the economic mechanisms that make it pay, its historical performance statistics, and — critically — the crash risk that defines its risk profile.
What Is Carry?
Carry has a precise, asset-class-neutral definition: it is the return an investor would earn if market prices remained completely unchanged. It is the yield generated by holding an asset, net of the cost to finance it. Another way to state this: carry is the expected return assuming no price change.
This definition applies universally. In currencies, carry is the interest rate differential between two nations. In commodity futures, it is the roll yield generated by the shape of the futures curve. In fixed income, it is the slope of the yield curve. In equities, it is the dividend or earnings yield. The signal looks different in each market, but the economic concept is identical.
A carry strategy ranks assets within a universe by their carry signal, goes long the highest-yield assets, and shorts the lowest-yield assets. It is a cross-sectional strategy — it cares about relative carry between assets, not whether any individual asset has positive carry in absolute terms.
Carry Across Four Asset Classes
Select an asset class below to see how carry is defined and measured in that market.
The Academic Evidence Base
Carry is not merely a practitioner heuristic. It has been rigorously documented across decades and multiple asset classes in peer-reviewed literature, with several foundational papers establishing both its existence and its economic rationale.
The definitive cross-asset paper. The authors proved that a unified carry signal predicts returns not only in currencies — where the premium had been documented for decades — but across global equities, bonds, and commodities as well. A diversified global carry portfolio combining all four asset classes generates significantly higher risk-adjusted returns than any single-asset carry strategy, because the macroeconomic drivers of yield differ across markets. This paper established carry as a universal phenomenon, not a currency-specific anomaly.
The commodity foundation. Demonstrated through an extensive dataset of commodity futures that historical returns were driven primarily by roll yield — the carry component — rather than by spot price appreciation. This finding explains why many long-only commodity indexes have dramatically underperformed simple roll-yield-capturing strategies: commodity exposure without carry optimization captures far less return than the underlying risk premia would suggest.
The foundational theory. Keynes argued a century ago that commodity producers — farmers, miners, oil companies — need to hedge future price risk and are willing to sell futures at a discount to expected spot prices, effectively paying a premium to financial speculators who absorb that risk. This premium appears as backwardation in the futures curve, and collecting it is the economic logic behind commodity carry. The theory predates all the modern academic literature and remains the most intuitive explanation for why carry pays in physical asset markets.
Individual asset class carry strategies have historically delivered Sharpe ratios in the range of 0.4 to 0.5 before costs — comparable to equity risk premia. A global diversified carry portfolio combining all four asset classes has been documented at approximately 0.8 to 1.0 Sharpe, reflecting the substantial diversification benefit. These are gross figures; transaction costs and management fees reduce net performance.
Why Does Carry Pay?
The persistence of carry premia across asset classes and time requires explanation. Researchers have proposed three non-exclusive mechanisms.
Risk Compensation
The mainstream economic interpretation: high-yield assets are riskier or less liquid, and investors require compensation for holding them. The yield is payment for accepting the possibility of large losses during bad economic states — when liquidity is scarce, credit is tight, and investors are forced to sell. Currency carry strategies, for example, tend to hold high-yielding emerging market currencies that are susceptible to sharp devaluations during global stress events. The premium is real, but so is the crash risk it compensates.
Hedging Pressure Theory (Keynes)
In physical commodity markets, producers have an asymmetric need to hedge. A wheat farmer with a crop in the ground has a natural short position — if prices fall before harvest, they suffer losses. They will pay to sell futures forward, accepting a price below the expected spot price in exchange for certainty. This selling pressure creates backwardation. Financial speculators who buy those underpriced futures collect the carry premium as compensation for absorbing the producer's risk. This is a structural, not behavioral, mechanism — it arises from the real-economy hedging needs of commodity producers.
Behavioral Explanations
A complementary family of explanations centers on investor psychology. Safe, low-yield assets — US Treasuries, Japanese Yen, Swiss Franc — attract systematic overbuying from investors seeking certainty or lottery-like protection against tail risk. High-yield assets — emerging market bonds, commodity futures in backwardation — are systematically undervalued because investors overestimate their risk or underweight their yield. The carry strategy exploits this persistent mispricing, collecting the excess yield until the behavioral bias corrects. This explanation predicts that the premium should be most reliable in markets where behavioral biases are strongest — which may explain why currency carry is particularly robust.
Implementation Mechanics
The carry signal must be operationalized differently for each asset class, but the portfolio construction logic is consistent: rank assets by carry, long the top tier, short the bottom tier, size by volatility, and rebalance periodically.
Signal Construction by Asset Class
Signal: Forward premium ≈ domestic interest rate minus foreign interest rate. Long high-yield currencies (e.g., MXN at 11%), short low-yield funding currencies (e.g., JPY near 0%). Rebalancing is typically monthly.
Signal: Roll yield = (Front contract price − Next contract price) / Front price × (365 / Days between contracts). A positive value means backwardation (positive carry); negative means contango (negative carry). Long backwardated markets, short contangoed markets. Robust models use the entire term structure slope rather than just the front two contracts.
Signal: Yield curve slope (e.g., 10-year minus 2-year). Borrow at short-term rates, hold long-duration bonds, collect the spread. Also includes roll-down: as a bond ages, its yield declines along a normally-upward-sloping curve, causing its price to rise. Suffers during aggressive rate-hiking cycles that flatten or invert the curve.
Signal: Dividend yield or earnings yield. Long high-dividend-yielding stocks or markets, short low-yield (growth-oriented) stocks. Highly correlated to the value factor and the underlying equity risk premium, making its diversification benefit within a broader portfolio more limited than the other three carry types.
Rebalancing
Carry signals evolve slowly — yield curves change gradually, interest rate differentials shift over months rather than days. Monthly rebalancing captures most of the signal without generating excessive transaction costs. Commodity carry signals may warrant weekly rebalancing given faster-moving term structure dynamics during supply shocks.
Historical Performance
The defining performance feature of carry is the diversification benefit from combining across asset classes. The macroeconomic conditions that drive high yield in currencies are different from those driving backwardation in commodities or slope in the yield curve. When one carry strategy suffers, others may be generating returns, smoothing the combined portfolio.
Koijen et al. (2018) documented that currency carry, commodity carry, fixed income carry, and equity carry have low to moderate cross-correlations — typically in the 0.1 to 0.3 range. This low correlation is the source of the diversification Sharpe improvement. When combined in equal risk-weighted proportions, the portfolio Sharpe substantially exceeds the average of the individual components.
The Carry Crash
The most important characteristic of carry — the one that defines its risk profile — is its negative skewness. The strategy generates steady, small positive returns the overwhelming majority of the time, then loses large amounts rapidly and suddenly when global liquidity dries up. Practitioners describe it as "picking up pennies in front of a steamroller."
Currency carry trades — predominantly long high-yield emerging market currencies and short the Japanese Yen and Swiss Franc — lost approximately 20 to 30 percent peak-to-trough within a matter of weeks during the acute phase of the 2008 crisis. The mechanism: as global volatility spiked, investors were forced to liquidate risk assets simultaneously. Crowded carry positions unwound in the same direction at the same time. The Yen surged as the primary funding currency was repatriated. Emerging market currencies collapsed. The carry trade did exactly what critics warned: it fell sharply at the worst possible moment.
The Swiss National Bank unexpectedly abandoned the EUR/CHF floor, allowing the Franc to appreciate by approximately 15–20% against the Euro within minutes. The Franc had been a widely-used funding currency for carry trades — traders borrowed cheaply in CHF to fund higher-yielding positions elsewhere. The sudden, massive Franc appreciation caused immediate, unhedgeable losses for any carry position with a CHF short leg. Several retail FX brokers became insolvent. This event illustrates how carry crashes can be instantaneous, with no opportunity to exit positions before the loss is realized.
The Mechanics of the Crash
Carry crashes share a consistent mechanism. During normal markets, carry is harvested steadily and positions accumulate. As more capital enters the trade attracted by stable returns, positions become crowded. When a volatility shock arrives — a financial crisis, a central bank surprise, a geopolitical event — investors simultaneously attempt to reduce risk by unwinding carry positions. The selling pressure reinforces itself: as prices move against carry positions, margin calls force further unwinding, which pushes prices further, which triggers more margin calls. The unwind is violent precisely because the crowding is a slow, invisible process and the exit is sudden.
The negative skewness of carry returns is empirically measurable. While equities exhibit negative skewness of approximately −0.5 to −0.7, currency carry skewness has been measured at approximately −1.0 to −1.5 — substantially more negatively skewed. This statistical fingerprint reflects the asymmetric return distribution: many small gains, occasional catastrophic losses.
Carry + Trend: Complementary Risk Profiles
Carry and trend following are the two most robust systematic risk premia documented in futures markets. Institutional systematic macro funds typically hold both strategies simultaneously — and the reason is their complementary risk structure.
Carry is fundamentally a short volatility strategy. It relies on market stability — it earns yield when prices stay in the same range, and it loses when large directional moves occur. Its return distribution is negatively skewed: steady gains, sudden crashes.
Trend following is fundamentally a long volatility strategy. It relies on sustained directional moves — it earns returns when markets trend strongly, and it loses when markets are flat and choppy. Its return distribution is positively skewed: steady small losses (the cost of being wrong about many small moves), punctuated by large gains during crisis periods.
The long-run correlation between carry and trend is low to slightly negative — typically in the −0.1 to +0.1 range under normal conditions. During market crises, this correlation becomes strongly negative. When carry crashes — as in 2008 — markets are experiencing the sustained directional moves (collapsing equities, surging bonds, surging safe-haven currencies) that are ideal for trend following. The two strategies are designed by their structure to perform well in opposite regimes. A portfolio combining them in equal risk-weighted proportions produces a substantially smoother equity curve than either strategy alone.
This is why the combination of carry and trend is sometimes described as the closest thing to a "free lunch" in systematic trading — not because it eliminates risk, but because it reduces the correlation of bad years between the two strategies significantly. The combined portfolio's worst years are milder than either strategy's worst years in isolation.
Key sources: Koijen, R., Moskowitz, T., Pedersen, L., Vrugt, E. (2018). Carry. JFE. · Gorton, G. & Rouwenhorst, K.G. (2006). Facts and Fantasies about Commodity Futures. FAJ. · Keynes, J.M. (1923). A Tract on Monetary Reform. · Lustig, H. & Verdelhan, A. (2007). The Cross Section of Foreign Currency Risk Premia. AER.