Asset allocation decisions often require taking a view on whether one month's price action will reverse or continue in the subsequent month. The recent evidence is not convincing in either direction.
Sixty asset classes, one clean result: most month-to-month return relationships are noise, and the handful that aren't come from liquidity mechanics, not investor psychology. Commodities and short Treasuries persist because of curve structure, not conviction. REITs mean-revert for the same reason bond-like equities do. The data has an alibi. Trend followers still need one for their edge :)
This was pretty interesting, but leaves me with questions about how trend following can be successful. Yet the SocGen CTA index is positive and the DBMF ETF which includes equities as one of the trends they replicate has been very successful, including this year. Perhaps they use a more complex algorithm which does allow you to predict prices from trends?
I don't know the CTA track record off the top of my head. But I would guess they use much more sophisticated algorithms, like trend over multiple horizons, not just month-over-month, different portfolio construction techniques, and also they probably use forecasting variables which I did not consider.
Sixty asset classes, one clean result: most month-to-month return relationships are noise, and the handful that aren't come from liquidity mechanics, not investor psychology. Commodities and short Treasuries persist because of curve structure, not conviction. REITs mean-revert for the same reason bond-like equities do. The data has an alibi. Trend followers still need one for their edge :)
This was pretty interesting, but leaves me with questions about how trend following can be successful. Yet the SocGen CTA index is positive and the DBMF ETF which includes equities as one of the trends they replicate has been very successful, including this year. Perhaps they use a more complex algorithm which does allow you to predict prices from trends?
I don't know the CTA track record off the top of my head. But I would guess they use much more sophisticated algorithms, like trend over multiple horizons, not just month-over-month, different portfolio construction techniques, and also they probably use forecasting variables which I did not consider.