William J. Bernstein
Rebalancing Individual National Markets
In these pages Iíve extensively examined the returns and risks of rebalancing asset classes in a global portfolio. For those unfamiliar with the basics, the guiding principles are as follows:
- Rebalancing benefit is driven by low asset correlations and high asset volatility. The more volatile the asset and the lower the correlation with the rest of your portfolio, the better.
- The more similar your asset returns, the better. If the returns of one asset are regularly higher than another (as has happened with US and Japanese equity over the past 10 years) then it is quite possible to lose money rebalancing.
I thought that it might be instructive to compare rebalancing strategies for national markets in 3 different categories: Europe, the far east, and emerging markets.
Developed MarketsóNot Worth the Trouble
Looking at annual returns over the past 11 years, from 1989 to 1999, I examined portfolios consisting of the following 9 European markets:
Over the 1989-99 11-year period the annually-rebalanced equally-weighted portfolio of these 9 nations returned only 0.09% more than the unrebalanced portfolio (12.62% versus 12.53%). It was quite easy to find asset pairs with unfavorable rebalancing characteristics. For example, one might expect that the German/Austrian pair would be a poor one, because of the wide difference in returns and close correlation (0.67) between the two markets. Such turns out to be the caseórebalancing the two countries cost you 0.47% in annualized return. Only where the returns are reasonably close, such as the UK/Germany pair, is there a benefit, and even here it is a razor-thin 0.16%.
The same occurs in the Pacific Rim:
Here the equally-weighted rebalanced portfolio underperformed the unrebalanced portfolio by 0.29%, mainly because of the miserable showing of the Japanese market.
Finally, I looked at the returns for 11 emerging markets:
Here, over the 1989-99 11-year period annual rebalancing an equally-weighted portfolio earned 5.71% of excess return over the unrebalanced portfolio. And only 4 of the 55 possible national pairs (Korea-Malaysia, Korea-Mexico, India-Indonesia, and India-Korea) had negative rebalancing effects. Even the Indonesia-Turkey pair, with its almost 26% annualized return difference, benefited from rebalancing. In fact, it is just about impossible to put together a reasonable emerging markets portfolio which does not benefit significantly from rebalancing.
Why the difference in rebalancing effects between the emerging and developed markets? First, the volatility of the emerging markets is much higher than the developed markets, with SDs averaging about 50%. Since the rebalancing bonus is proportional to the variance of the asset, a doubling of SD results in a quadrupling of variance, and thus of rebalancing benefit. Second, correlations are much lower in the emerging markets arenaótypically about half of those in the developed world, providing yet another margin of benefit.
One area of concern with emerging markets rebalancing is trading costs. Although a significant problem, it is not a killer. The average annual turnover of the equally-weighted emerging markets portfolio was about 15%. Assuming that a round-trip costs about 6% in terms of spreads, commissions, and impact costs, the trading necessitated by annual rebalancing should cost less than 1%.
Practical Rebalancing Advice for the Small Investor
For the institutional investor, passively investing oneís emerging markets exposure in an equally-weighted (or otherwise fixed) portfolio of national markets is a no-brainer. But what is the small investor to do? There is only one equally-weighted indexed emerging markets product out there, and that is the DFA series of EM funds. And even it is not fully rebalanced, striving towards equal weighting only with inflows and outflows. Limiting "rebalancing" in this way is costs return, and the portfolios can get seriously out of wack. As of 9/30/99, for example, the DFA Emerging Markets I Fund had a 13.3% Korean contribution but only a 3.3% Indonesian one. This fund has outperformed the unrebalanced Vanguard Emerging Markets Index Fund by about 3% over from 6/1/94 to 12/31/99, or about half the theoretical 6% amount. It is interesting to speculate that this shortfall may be to the "inefficiency" of their rebalancing mechanism.
To invest in the DFA fund you will need to services of a financial advisor. What other options are available? Closed-end funds can be used, but most of these are actively managed, have high expenses, and in many case high turnovers, with their attendant trading costs. These beasts are also plagued with fluctuating premiums/discounts from NAV, and as Iíve previously noted this causes adverse portfolio behavior, since it increases correlation with the US market. A better option might be the ADRs of emerging markets nations, but managing a portfolio of a 20 to 50 of these would require considerable effort and attention, and would also probably run a fair amount of terminal wealth dispersion risk. (I.e., the risk that you may miss out on the best performers in each market, which is a real worry in a portfolio which only owns a few stocks from each country.)
My own opinion is that it is probably worth the 1% fee to own the DFA fund. (Disclosure is in order; Iím a registered investment advisor, and I employ their funds.) The ADRs are not a bad second choice, although going this route is quite a bit of effort.
Bottom line; for tax-sheltered investing with the developed nations itís best to go with Gus and his Vanguard Pacific and European portfolios. For the emerging markets rebalancing individual national allocations is quite worthwhile. Unfortunately there is at present no simple way for the small investor to do this; consider the above options.
And for taxable accounts Iíd forget about rebalancing altogether and keep it simple with Vanguardís International Tax-Managed Portfolio (VTMGX). Since VTMGX does not have any emerging markets exposure you may wish to add the Emerging Markets Fund, which is reasonably tax-efficient. The adverse tax consequences of portfolio rebalancing are on the order of a percent or two per year, and obliterate any rebalancing benefit.
Copyright © 2000, William J. Bernstein. All rights reserved.