Since September, we've seen markets in turmoil, with considerable emotion in play. Equity classes no longer have the correlations that were used to allocate assets efficiently. As a result, the bases for asset allocations may no longer be valid.

Nevertheless, the fundamentals remain. Equities have dropped in value, but eventually they're likely to revert to pre-Sept. 15 values. They appear undervalued now, making this an opportune time to rebalance.

How do you rebalance in these turbulent times? I've devised an approach, called opportunistic rebalancing, that squeezes more alpha out of rebalancing. This article first describes the process, then applies it to periods of volatility.

My four years of research on rebalancing led to a paper published in the Journal of Financial Planning (January 2008). I also had discussions with Chris Cordaro of RegentAtlantic Capital, Mark Balasa of Balasa Dinverno & Foltz, Tom Davison of Summit Financial Strategies, Tom Orecchio of Greenbaum and Orecchio, and Jerry Miccolis of Brinton Eaton Wealth Advisors.

The Basics

Let's start with some basic concepts. "You're in the Band" shows five asset classes with allocation targets and their rebalance bands. Any asset whose allocation has drifted outside the rebalance band is "out-of-band," which means the portfolio needs to be rebalanced.

In the chart, assets A and B are out-of-band. A is so far above its allocation that we anticipate a reversion to the mean—a drop in value. Rebalancing would have us sell A before it reverts. This is the sell-high concept. Similarly, B has dropped so low that we anticipate it will revert to the mean by rising in value. Rebalancing would have us buy B before the reversion—the buy-low concept.

Thus, the rebalance band we have defined is an implicit statement of the line above or below which we anticipate a reversion to the mean. Advisors commonly use a 20% relative band, so if an asset class allocation is 10%, the rebalance bands are 8% and 12%. Generally, if the asset class is above 12% we will sell high, if below 8% we will buy low. My research has shown that 20% relative bands are optimal.

You don't need to rebalance exactly to the target (the dotted line), though. The reasoning goes as follows: Suppose an asset, in this case A, is above the target by $1; there is clearly no need to rebalance (even if trading costs are free). Similarly, if C is off just a bit, we don't know if it is headed north or south. It's in what we call the noise band or the tolerance band. So there is a noise band near the target, inside which we cannot predict whether the asset will go up or down. When you rebalance, the change needs only to bring the asset class within this tolerance band.

To be practical, set rebalance bands to 20% of the portfolio target and set the tolerance band to 10% of the target. So if the target is 10% for a class, make the rebalance bands 8% to 12% and the tolerance bands 9% to 11%.

Thus, if A is over by $15,000 and B is under by $10,000, you can rebalance the portfolio by correcting A and B by $12,500 each, leaving A just above target and B just below target (but within the tolerance band). Then you would need to look at C, D and E. There are two advantages to not touching them: First, you would save on trading costs. Second, E appears to be on a momentum swing and is headed toward reversion land-but not yet. We don't want to run the risk of selling prematurely and losing the opportunity to reap a more significant benefit.

To summarize, the rebalance bands define a demarcation outside of which we expect reversion. The tolerance band is a no man's land, where asset classes could move up or down. The zone between the tolerance and rebalance bands is the momentum band; classes in this zone will likely head up if over, and down if under. If assets are in this momentum band, try to delay rebalancing to the target to preserve momentum and decrease trading costs. If only one asset class is out-of-band, you may have to correct with a complementary class that's still in the momentum zone.

What's the Frequency?

The next question is how often to rebalance. Most advisors do it quarterly, but the actual implementation may be random, owing to the exigencies of workflow. This randomness is unsatisfying.

"Look Frequently" shows the alpha offered by various rebalancing schedules. Looking at your portfolio every 10 business days is as often as you need. Ten days is frequent enough to catch most transient buy-low or sell-high opportunities without overtrading.