A Smart Way To Look At A 'Dumb' Investing Strategy

by: James Picerno

Expecting rebalancing to juice portfolio returns is a "dumb idea," according to Forbes. "Rebalancing is fine if all you are trying to do is sleep better at night. But the idea that it increases your expected return is balderdash." This can be reasonable advice… up to a point. But all-or-nothing blanket statements don't usually fare well when it comes to evaluating investing strategies for the real world. That's surely the case when reducing the pros and cons of rebalancing, and all the accompanying nuances, to a one liner.

Yes, it's a mistake to assume that rebalancing will always and forever boost performance. But that's not the same thing as saying that it never elevates returns or that it's foolish to consider the possibilities. Indeed, no sooner does Forbes tell us that thinking of rebalancing as a performance enhancer is "dumb," the article challenges its own advice by recognizing the opposite:

So far in the 21st century, stocks have bounced around while going nowhere and bonds have done pretty well. Someone who rebalanced over the past 12 years would indeed have earned more than someone who allocated his money only at the beginning and then stood pat.

But we're told that these positive stretches for rebalancing are misleading because different time periods produce different results. Fair enough. But the same can be said for every investment strategy. Nothing wins all the time. Whether it's a value-based stock-picking or index-based momentum, perfection eludes everyone in money management. Pointing this out is at once accurate and worthless. In order to properly assess a given strategy, and decide what's useful, what's not, and how to tell the difference, a deeper review is required.

Nonetheless, Forbes claims to have found the smoking gun on this topic: A 2010 Vanguard study sliced and diced several rebalancing strategies for the last 80-plus years for a simple US stock/bond mix. "Just as there is no universally optimal asset allocation, there is no universally optimal rebalancing strategy," Vanguard concludes. That's roughly what numerous rebalancing studies from other sources show, as I discuss in some detail in Chapter 6 of my book Dynamic Asset Allocation: Modern Portfolio Theory Updated for the Smart Investor. There's near universal agreement that rebalancing is productive if not essential, but debate rages over the details and what to expect.

The failure of rebalancing to deliver a constant and unending premium may inspire some to see the strategy as offering only risk management advantages. Should we agree and leave it there? You can, and do quite well in your investing travels as a result. But there's more to say. Much more, in fact. I'll spare you the gory details, but a few observations are in order, if only to remind ourselves that the last word on this topic shouldn't be framed by provocative headlines.

First, the Vanguard study. Useful as it is, studying what is effectively a two asset class strategy - US stocks and bonds - across eight decades may not be as definitive as it appears. In the 21st century, investors are no longer limited to a binary choice in asset allocation. The major asset classes that are available in ETFs, by my accounting, stack up to more than a dozen basic choices these days. There are many more, of course, if we subdivide the standard choices further. As a result, the return/correlation matrix is quite a bit broader than the Vanguard study allows. In turn, a wider opportunity set may offer more encouraging results for rebalancing than history suggests. Keep in mind too that the past 80 years may not be a guide for the next 80 - or even the next decade.

What is clear is that restricting your rebalancing plans to, say, two assets unnecessarily limits the potential for enhancing return (and managing risk, for that matter). It's long been clear that if we restrict asset allocation to two assets, and one asset delivers considerably higher performance over a long period vs. the other, rebalancing's advantages will be squeezed if not eliminated entirely. Perold and Sharpe formally outlined the details some 25 years ago and so this is hardly news at this late date. The practical result is that if you have a high degree of confidence that asset X will deliver superior results vs. Y, you can dispense with asset allocation and rebalancing. The problem, of course, is that the future's uncertain.

The issue, then, is how to optimize asset allocation and rebalancing. Again, there are no quick fixes, but there's a useful framework to consider. For example, William Bernstein demonstrated rather persuasively that, as he put it, "the intrinsic rebalancing potential of any asset pair is the difference between its mean variance and covariance." He also recognized that there is no secret formula to earn a rebalancing bonus at all times. But neither is it beyond the pale to say assume that modestly higher returns are possible, even probable, under the right circumstances - circumstances that are more likely to be "right" when favoring broad diversification across relatively low-correlated assets.

A new study revisits the issue and investigates if portfolio rebalancing is a stable source of alpha. Once again, the answer is "no." There are still no free lunches, but that's not to say that all's lost. As the authors explain:

In this paper we proved it is possible to generate alpha through portfolio rebalancing. We reviewed the various contributions on this subject and we validated the formulae developed by academics and practitioners that aimed to quantify this alpha (rebalancing bonus). Portfolio rebalancing does not always generate a positive alpha, in contrast to what the aforementioned formulae show. It is essential, in fact, for the securities to have specific characteristics. Our findings confirm that the presence of high volatility and low correlation helps in generating a positive rebalancing bonus but they act mainly as amplifiers. The essential condition is the presence of relative mean-reversion. Identifying securities that share these three requisites: relative mean-reversion, high volatility and low correlation remains a challenge.

Depending on how you define "alpha" may influence your thinking on rebalancing as well. The concept of rebalancing for purposes of harvesting tax losses, for instance, lays the groundwork for expecting what some analysts refer to as after-tax alpha.

There's also the issue of relative performance to consider. As one example, it's well established that equal-weighting has an encouraging history of raising realized performance vs. market-value weighted strategies, all else equal. What's true for strategies within a given asset class tends to apply to asset allocation too.

When you study why equal weighting tends to do better, you'll find that the trail leads back to rebalancing. In order to maintain an equal weight, it's necessary to regularly rebalance the portfolio. Left untended, every portfolio eventually reverts back to market weights.

In fact, if you look at any strategy that appears to deliver superior results, the fundamental source of the enhancement is usually - you guessed it - rebalancing. The triggers for why and how you rebalance vary, of course, and the strategy is marketed under numerous labels. But when you pull back the curtain, you'll probably find some form of rebalancing for strategies that are something other than passive.

Can you count on equal weighting to prevail? No, of course not. Nothing rises to that, the standard in the money game. But neither should we dismiss the potential for equal weighting (and its main benefactor, i.e., rebalancing) to raise returns. If you're betting the farm on no less over a short period, you may be setting yourself up for trouble.

A better assumption is to see rebalancing generally as a risk-management tool, first and foremost, with the potential to elevate returns. How much return enhancement should we expect? The answer varies widely, depending on the strategy and the targeted assets. For a broadly diversified portfolio across the major asset classes, 50 to 100 basis points over market-value-weighted portfolio has been typical over the last decade or so. It could all evaporate tomorrow, of course. But that's true of alpha generally. Assuming otherwise truly is stupid. Rebalancing, however, is still smart, for several reasons.