© 2017 Elsevier B.V. Carry returns have been widely observed in the FX market. This study exploits the common information embedded in several factors previously identified as relevant to carry trade returns. We find that the extracted common factor successfully models the time series and cross-sectional characteristics of carry returns. Empirical evidence is presented that the common factor produces smaller pricing errors than other well known factors, such as innovations of exchange rate volatility and the downside stock market excess return. Our results also suggest that stock market risk is somewhat segmented from FX market risk.