 |
Perhaps at the very heart of the schism that divides passive index investors is the slice and dice vs. total stock market debate. Namely, does it make more sense (cents?) to buy the entire market or to divide it into four sectors and buy equal amounts of each. This article is a chronicle of an IndexFunds.com discussion board debate that covers all the issues. If you have difficulty understanding the thread of the discussion read a primer that was also posted to the discussion board for newcomers to the debate Is Slice and Dice a Sector Bet? If that single line isn't enough to open up old Discussion Board wounds, I don't know what is. An index fund reader, Mark Murway, brought today's J. Clements article in the Wall Street Journal to my attention. Clements essentially says, in a very Boglesque manner, that anything that doesn't mirror the market is a sector bet. He does this while recognizing the flip side of the debate and remarking on Fama/French (FF) and quoting William Bernstein as counterpoint. It all brings us back around to the eternal question: is 4X25 Slice and Dice a "sector bet" or "diversification?" Leaving myself open to a ferocious rebuttal from the Dimensional Fund Advisors (DFA) advisors that frequent the boards here, I tend to support Clements and believe the former. For those of you new to this debate, please see Jeff Troutner's Slice and Dice and Total Stock Market (TSM) articles, for the Slice and Dice angle, or any Bogle book for Total Market. Anything else does seem to fly in the face of Efficient Market Theory. This DOES NOT preclude you from avoiding stocks you believe have obscene evaluations... I fully understand the flip side...that the Total Stock Market's large company quarter of the market is DFAs largest decile (or less). Slice and Dice puts more money on certain parts of the market than the market itself does. Is Slice and Dice a Sector Bet? Maybe I'll take on the Index Funds TSM opus myself at some point. I will not be posting my home address at the bottom of that article... Re: Is Slice and Dice a Sector Bet? First, an admission: Jonathan and John Bogle are historically correct. It is nearly impossible to come up with a more mean-variance efficient portfolio than the market, at least by size deciles, over the past 75 years. Another admission: for most of that period, they've been correct. However, this time I think it's different. Just because you believe
in the EMH doesn't relieve you of the responsibility of estimating
future returns. If anybody in the audience believes that we're
gonna see 8% real stock returns over the next few decades, then
I'm looking out the window at a really neat bridge I can let you
have cheap. To paraphrase Warren Buffett, if returns came from
history books then the wealthiest folks would be librarians. When
I look at large growth stocks I can't see much more than 2%-4%
real long term returns going forward. (And I'm in passable company--Fama
and French's latest working paper concludes that the Gordon Discounted
Dividend Model is a more accurate predictor of stock returns than
historical returns. And right now it says 3% real. FWIW, Buffett
concurs.) While I don't avoid large growth, I simply cannot stomach
their current cap weighting. Re: Is Slice and Dice a Sector
Bet? Could I be wrong? You bet. As Jonathan implies, this kind of judgement call involves industrial grade hubris. But cap weighting is an active choice. And right now it does not make a lot of sense, to me at least. BTW, one small quibble with Jonathan. He's arbitrary about US-Foreign
allocation, but is quite strict with his US cap weighting. And
yet there's a lot more daylight between 20% and 40% foreign than
there is between "the market" and slice-and-dice. Re: Bernie B. Post But the preference for the market is that in theory, and over long enough periods, in practice, it turns out to be the most mean-variance efficient portfolio. Re: Is Slice and Dice a Sector Bet? 1) detractors say TSM is a large growth index. but what's growth? what's large? adopt a different definition, and you could decide TSM is really a small, value index. 2) fama-french are arguing that small and value outperform because they are more risky. (of course, this is the tautology of efficient markets theory in action: if it's associated with return, it's got to be a risk factor.) nonetheless, if advisers are going to draw intellectual strength from the fama-french research, they should be intellectually consistent i.e. they should state that overweighting small and value is riskier. but i don't see many folks doing this. instead, they're saying TSM is riskier. 3) a TSM fund will be far more tax-efficient than slicing and dicing the market with sector index funds, because you won't pay taxes as stocks graduate from one index to another. 4) you can carp all day about the market being overvalued or dominated by growth or too heavily tilted toward tech. but whatever it is, it is the sum total of investors' best judgments based on all currently available information. it takes a lot of hubris to decide that everybody else is wrong and that you are right. i'm not dumb enough to consider myself that smart. Re: Clements Post 1) I think you would have hard time finding defintion of small
value that would meet your criteria. The generally accepted definitions
and quite logically so are the CRSP definitions, they are not
arbitrary and used by most academics. Now if you don't like the
top 3 deciles as value and bottom 3 as growth -split them in half
and you get very similar results.splitting large and small into
halfs is also logical.While you may debate the Ff risk factors,
it is very clear that their definitions of risk due great job
of predicting returns-therefore, it is logical to use those defintions
in deciding how to allocate assets. 2)IMO you are mixing two issues,most of us, including me state
value is risk factor and is compensation for risk-and if you tilt
to value then you accept that risk (basically credit and liquidity).
Now -must also keep in mind Markowitz work showing that you can
add risky asset classes with low correlation and reduce portfolio
risk, then you also have the RTM and rebalancing benefit. So bottom
line is some of the risk of value as stand alone asset class is
offset by its low correlation to the growth component. You also
avoid the all eggs in one basket issue and the long periods of
underperformance that can occur liek 66-88 when lg underperformed
one month cds. Everyone loves lg ex post t-wonder what they were
saying in early to mid 80's? 3) This issue is correct on the surface-however -most investors
can split their assets putting least tax efficient in tax deferred
account and most tax efficient in taxable.Then you also have addition
of TM funds (and in value classes to by DFA).this offsets much
of the issue-and you can rebalance with new cash too. 4)Prices of all assets may be correct-and I fully agree that market is very humbling mechanism.But that does not matter here IMO-the issue is controlling risk and too much in one big LG basket unless compensated for by higher expected returns (which no one argues for) makes no sense to me at least. Foolish slavery to gods of market. In other words it is simply issue of diversification making sense. Re: Is Slice and Dice a Sector Bet? At today's valuations, I must admit that overweighting SV is a very tempting bet to take. But then again, it sure seemed like a pretty tempting bet a few years ago too, and anybody who took it then isn't smiling today for it. Re: Lommel post Actually, what you are doing is betting that the asset class that dominates the market at the time of purchase will provide better returns than the other available classes over the longterm. Of course, a bet is just a bet. You don't know in which order the asset classes will finish. "Any other weighting is a BET that you can do better than the average dollar in the US market." No, it is an honest and sincere attempt to keep the regret factor from driving you crazy. You may not have overweighted the asset class that eventually "won," but at least you didn't overweight the worst of the four. "I think this is what people mean when they say that slice and dice is a sector bet. Beating the average dollar is a zero sum game, so for every guy who actually does beat the average dollar, there will be a guy who doesn't." There is probably some truth to this, but actually you are winning with 25% of your portfolio, losing with 25%, and possibly matching the broad market with the remaining 50%. You may or may not end up with a result that matches TSM. I would think that, even if you come in with less, the margin would not be large. In any event, it wasn't winning or losing that motivated you, but knowing that the extremes of overperformance and underperformance were blunted. Re: Lommel post Re: Swedroe post Re: Adams post Re: Adams post Re: Is Slice and Dice a Sector Bet? Also TSM is the BET not 4 x 25 which is by DEFINTION style NEUTRAL making no bets- just looks different than the market -so you have tracking error-which should IMO be irrelevant- though it is a big psychologicval risk for most investors. I simply cannot understand how anyone can say that a style neutral portfolio is making a bet -but a portfolio that is only 6% small and 8% value is NOT . IMO simply illogical- MR. SPOCK I think would agree. Re: Swedroe post Re: Kandell Post Re: Kandell Post Re: Swedroe post Larry, notice we don't really disagree. My argument in previous post was that one can't just blurt "you're not neutral" as is done here in the tsm versus s-a-d debate. One has to instead argue *for* the "vantage" (for want of better word) that lies behind one's standard. In other words, one needs to argue why *asset class* should be one's neutral point, as opposed to *the market*. Notice, you've done just that: you've argued that asset class is a better measure of neutrality than the market because "exposure to factors of size and value...determines virtually all returns." The implication is that since we're interested in returns, and since size and value, for the most part, determine that, then those are what bias should be measured against. I'm not as confident that size and value are the primary constituents of returns in a fundamental way--but I accept your argument. At least its assumptions are all out on the table. We have a focus of argument, if you will. Those arguing for the market as one's standard I think have to show that it has some heavy reason behind it and/or that asset classes don't determine returns. One way to go with that--with which I'm sympathetic--is to argue for behavioural theories of price. I know there's lot of academic evidence for asset classes, but they have trouble explaining the last decade Re: Kandell post |
||||||