William J. Bernstein
Over the past nine years, Iíve observed dozens of asset classes, simulated hundreds of portfolios, and probed thousands of funds. And as I look back, Iím happy that my conclusions have worn pretty well. Most of them, of course, were slam-dunks, like an emphasis on diversification, small and value exposure (implementing these two factors across multiple markets), rebalancing, and a passive approach. I lay no claim to originalityómost of the work was not my own, merely a reformulation and application of the work of others.
Even then, I didnít bat a thousand. As I look back at the Efficient Frontier oeuvre, two lulus stand out. In 1999 I criticized a piece by Mark Hulbert on the "January Effect," which I then thought had gone out with Bretton Woods and the Bee Gees. And in 2000 I knocked Barclays foreign ETFs (WEBS, as they were known at the time, now "country" iShares) because of their tracking errors, which had all been negative relative to the appropriate MSCI indexes. On both counts, I was proven wrong. So to the estimable Mr. Hulbert and to Barclays, I offer my apologies, as well as the following stochastic Hail Mary and financial Our Father.
First, the January Effect (JE): the theory that most of the small-stock premium seems to occur in January (never mind the reason). Six years ago, the JE did appear to have tapped out. When I brought this to Mr. Hulbertís attention, his assistant ventured that maybe its large magnitude before 1980 was due to survivorship bias, or perhaps that some of it was due to the value effect; actually, itís been long recognized that the value premium is also large in January. In the past several years, however, January has been a dandy month for small stocks, so I decided to revisit the issue. Below is a summary of the JE for both small and value (this time using the Fama/French small and value factors) both here and abroad over sequential ten-year periods:
7/26-6/35 (9 Yr)
For each period, the first numerical column represents the average small premium for the ten Januarys, and the second column the average premium in the remaining 110 months; this repeats in the third and fourth columns for the value premium. The last four columns repeat the entire sequence abroad, for which we only have 30 years data.
What to make of all this? In the first place, whereas in the U.S small has attenuated somewhat, it hasnít gone away, and abroad, the very limited data suggest that it remains robust. Whatever these data show, it is clear that the obituary for the small stock JE has yet to be written. As to the value JE, the jury is still out. The last decade has been a real clunker in this regard both here and abroad; more importantly, value seems robustly positive (remember to multiply the fifth and eight columns by eleven to get its real annual magnitude) the rest of the year. Is the JE for either small or value of any practical use? I doubt it. In the case of small stocks, a JE of at best 2%-3% is easily wiped out by the transactional expenses of buying in December and selling in February, and in the case of value stocks, their positive return the rest of the year militates selling value stocks at all.
Next, the country funds. Here, a more clear-cut conclusion can be drawn. When I surveyed the Barclays offerings five years ago, their performance was indeed awful: all 17 of the funds underperformed their respective MSCI indexes, three of them by more than 3% per year. I concluded that Barclays couldnít transact their way out of a paper bag. But on revisiting their performance after the September 1999 cutoff, the picture has turned around 180 degrees:
Eleven of the sixteen tracking errors (TE) are now positive, and only one of the negative values is larger than the fundís expense ratioóvery impressive indeed.
What went right at Barclays? There are two possible, and not mutually exclusive, explanations. First, the company may simply have cleaned up its transactional act. The second reason is more complex and has to do with the nature of some foreign indexes, those that contain single stocks constituting more than the five-percent limit of the fund (mandated by the 1940 Investment Company Act). During the mega-cap mania of the late 1990s, this would have caused many of the funds to underweight the best performers in the index; after 2000, the opposite would have occurred. As of press time, Barclays has yet to respond to my requests for clarification on these issues.
Do I believe that single-country ETFs have a place in most portfolios? For a variety of reasons, such as their negligible value and small loading, expenses in the 80 to 95 basis points range, and overall complexity of portfolio execution, I do not. (If youíre going to take that much trouble with your portfolio, youíre certainly going to be breaking down the foreign equity portion of your portfolio into multiple style boxes, of which the twenty or so iShares country funds would constitute just a few percent, namely the "large market" or "large growth" component.) But if, for whatever reason, entertainment most easily coming to mind, you want to own them, I now think they are well executed enough to use.
A little humble pie is good for the diet. The best I can hope for is that my future consumption of it will come as cheaply as the above entrees.
Copyright © 2005, William J. Bernstein. All rights reserved.
The right to download, store and/or output any material on this Web site is granted for viewing use only. Material may not be reproduced in any form without the express written permission of William J. Bernstein. Reproduction or editing by any means, mechanical or electronic, in whole or in part, without the express written permission of William J. Bernstein is strictly prohibited. Please read the disclaimer.