The Carry Trade: Benefits and Risks
I've gone from thinking that categorizing strategies was useful to thinking it is essential. Dividing trades into inefficiencies or risk premia helps when deciding how aggressive to be in sizing and also how suspicious to be about its decay in effectiveness. I've found the model based or situational dichotomy helpful with sizing decisions.
It is also useful to have a general framework of what works and what doesn't. This helps when thinking of places to look for new ideas. The best unifying concept I've found is carry. Carry is either the cost we incur or the income that we receive for holding a position, assuming no other price changes. In "Carry" Ralph Koijen, Tobias Moskowitz, Lasse Pedersen and Evert Vrugt looked at expected return in several asset classes as a function of carry and price change. Using data from 1972 through to 2012, they looked at equities (carry being dividends), bonds (coupons), currencies (literally the carry trade), commodities (storage and yields) and index options (the variance premium). They also divided carry into two parts. Passive carry refers to the income from holding a position and dynamic carry is how well carry predicts future price changes.
Their conclusions were:
Carry returns and total returns are positively related.
In equities, carry and value are positively related.
In commodities and bonds, carry and momentum are positively related.
Carry performs badly in recessions.
Most of the returns were due to dynamic carry.
Their carry based portfolio had a Sharpe ratio of 1.1.
This might not seem interesting, given that we already knew that value, momentum and the variance premium were good things but this is more general, and is applicable to situations where the Fama-French factors are not.