Monday, March 23, 2015

The Benefit of Slowing Down the Rebalancing Process

One of the top Google “rebalance free lunch” results is an article quoting one of the brightest minds in finance, Clifford Asness, that I think most investors would agree with:

"Rebalancing is one of the few free lunches out there," said Clifford Asness, managing principal of New York hedge fund AQR Capital, in a recent interview with the Online Journal. "You're generally selling things that have gone up the most and buying things that have gone down the most."

While the article is from 2004 (and he may have changed his tune), I like that he felt rebalancing was a free lunch despite the strong research he has done over the years specific to momentum, something he acknowledges the rebalancing process eliminates in the same article:
 "Someone who doesn't rebalance is a tacit momentum investor." 

The key points in my mind as to why rebalancing may not be a free lunch:
  1. Momentum works: much research exists outlining that momentum can improve risk-adjusted performance (here is a great white paper by Cliff Asness himself)
  2. Relative outperformance doesn’t equal a relatively more expensive asset: quite simply, stocks SHOULD outperform over time as they can grow their earnings, while bonds pay a static coupon. So rebalancing doesn't always mean selling high and buying low.

It would be good enough if slowing down the rebalancing process did not materially hurt performance, as there are significant benefits associated with slowing down (lower transaction costs, tax benefits through delayed capital gains, and any behavioral benefit of looking at your account less often), but rebalancing less has actually improved average results specific to a 60% S&P 500 / 40% Barclays Aggregate Bond index since the latter’s 1976 inception with seemingly more potential outperformance than underperformance over any 36-month iteration.

Source: S&P, Barclays

1 comment:

  1. what would the un-rebalanced result be? what are the respective Sharpe's? the difference looks so small, though I imagine trading and tax costs weren't factored in, likely giving less frequent rebalancing a greater advantage