February 8, 2012

Residual Momentum

"Residual Momentum" by David Blitz, Joop Huij and Martin Martens is available on SSRN. It was published last June in the Journal of Empirical Finance.

It's an interesting paper in that they determine momentum by the relative strength of the residuals after regressing stocks on market risk factors. They throw out beta and alpha and look only at the residuals, which represent each stock's idiosyncratic (non-market related) risk. In comparison to total return momentum, residual momentum earns the same return with about half the risk. I suspect  results would also have been good had the authors only removed market return as a risk factor, rather than also excluding size and value factors.

According to the authors, residual momentum is more consistent across different economic environments and has much lower drawdowns than total return momentum. The authors show that from 2000 through 2009, total return momentum had an average loss of 8.5% per year, while residual momentum gained an average of 4.5% per year.

I should point out that total return stock momentum has been stronger since then. Its average annual loss on the long side since 2000 is now around 0.5%. There are also better ways to use momentum than with individual stocks, as my momentum papers demonstrate.