Michael Edesess has published an article today on rebalancing. It goes through lots and lots of iterations of rebalancing, including many different asset classes, time periods, rebalancing intervals, etc. Then it compares the returns of all these different possibilities.
If you are a big believer in mean reversion, then I guess you might expect to see a positive “rebalancing bonus.” On the other hand, momentum traders would probably think that rebalancing could only damage returns.
It seems to me that the article is really missing the point of rebalancing. One of the first imperatives of a financial advisor is to keep the client invested in a way that meets the suitability standard. That is, you make a determination with the client to hold a specific portfolio, comprised of a variety of asset classes at specified percentages. By definition, the asset classes will have differing expected returns and risks, and their returns over time will vary compared to each other. If the agreed upon portfolio is not maintained by rebalancing, then you are no longer meeting that suitability requirement as time passes.
In addition to the problem of drift from model portfolio, there is also the issue of risk. As your portfolio changes to some previously undefined and unanticipated portfolio, the risks change, and almost certainly to the increasing side – that is, your highest returning assets, which are often also the highest risk assets, will come to dominate the portfolio. The possibility of other methods of risk management is mentioned in the article, but not tested.
So whether or not there is a “rebalancing bonus,” advisors whose clients hold more than one asset class will need to make sure that the portfolio is maintained as agreed upon by the client. We do this by rebalancing.