(Listening notes: Veruca Salt- I watched Willy Wonka and the Chocolate Factory tonight. Just me, no kids, I don’t want to hear any snickering from you people;)
Economist: That looks like a $20 bill lying on the ground over there.
Other Economist: It couldn’t be. If it were, somebody would have picked it up already.
Or so one would think if the Efficient Market Hypothesis (EMH) is true. I’ll just provide a link to a couple of discussions on the hypothesis, because today, though I question it, I am really interested in the question of even if it is true, does it matter? Or at least in one particular case. First what is the EMH? I’ll send you to Jane Galt for a basic idea:
I think that what they’re saying is that the “strong” version of EMT — that market prices correctly incorporate all possible knowlege about the stock price — is incorrect. And in that I concur. But there are also “semi-strong” and “weak” versions of the theory.
The weak version says only that you can’t make excess returns–basically, profits above the economic costs of finding the investment opportunities–based on historical financial data. I think it’s safe to say that pretty much every economist thinks this is true, and that “chartists”, who attempt to predict price movements based on past trends, are doomed to fail in the long run.
The semi-strong version says that share prices adjust instantly to new public information, and you therefore can’t make money trading on that information. I think, again, that almost all economists would agree with this.
Now, I was at the University of Chicago Business School, intellectual home of efficient markets theory, during the stock market bubble. Even the strongest of market advocates among my professors didn’t try to argue that we weren’t in a bubble; they could read a P/E as well as anyone. About the only people trying to argue that prices were sustainable were my classmates who had taken out $100K of student loans so they could hold on to all of that valuable Webvan stock.
What my strong market efficiency professors argued was, not that prices reflected some metaphysically true value for the stock, but that the prices reflected all the information anyone in the market had, and that you therefore had no hope of being able to consistently generate excess returns.
So who cares? Well me, but it is a key in a discussion on fundamental indexes over at Greg Mankiw’s Blog:
According to Mr. Siegel, there is a “revolution†under way, a “new paradigm†in which the traditional indexes like the S.& P. 500 will make way for fundamental indexing, which constructs indexes based on measures like companies that pay dividends, rather than just a company’s size.
…….
Mr. Siegel says the central problem with traditional index funds, which are weighted by market capitalization, is that they overweight overvalued stocks and underweight undervalued stocks. Historically, value stocks outperform growth stocks, so an index should be constructed to invest in the cheaper value stocks rather than the expensive growth stocks.
……..
Mr. Bogle disagrees. “Beware when you hear about the new paradigm,†he said yesterday. “I think the claims they make are outrageous.” He refers to some of the data provided to back up the premises of WisdomTree as “data mining.â€
Greg after reading the piece says:
I am placing my bets with Bogle on this one. Bogle’s position has the virtue of an economic theory to back it up: the efficient markets hypothesis. The hypothesis is probably not exactly true, but it may be true enough to make it sensible for typical investors to follow its prescriptions.
I love Greg’s blog but I think he is dead wrong here. Two reasons. As explained here and here, Bogle’s argument in favor of indexes isn’t based on the EMH anyway. It is on transaction and tax costs. As an aside I do think it is interesting that by and large Vanguards own actively managed funds have outperformed the indexes, but obviously that alone doesn’t disprove Bogle’s cost or EMH rationale for prefering index funds.
Part of Greg’s mistake is that Siegel is slightly off base as to why fundamental indexes can work. The best work in the field has been done by Rob Arnott:
We can’t know what true fair value is. But we can know that every stock, every asset, every bond is going to be trading above or below what its ultimate true fair value is. Even the most ardent fans of the efficient markets hypothesis would say, “That’s reasonable. That’s reality.”
Now if every asset is trading above or below its true fair value, then any index that is capitalization-weighted, (price-weighted or valuation-weighted) is automatically going to have us overexposed to every single asset that’s trading above its true fair value and underexposed to every single asset that’s trading below its true fair value.
So this is the first time we’ve circled back to some concrete implications for the market. It means that capitalization-weighted indexes on which our entire industry relies, are fundamentally, structurally flawed and will inherently overweight every stock that’s above fair value and underweight every stock that’s below fair value.
This is the key. If markets are efficient it really only means that the errors are random and you cannot know which stocks are really trading above fair value or below. What we do know is that those stocks trading above fair value are going to be over represented, and those trading under fair value will be under represented. Therefore, indexes weighted by any factor that is not dependent on the value of the firm as represented by stock price will not be subject to that error. What that means is that the EMH supports fundamental indexing over market cap weighted indexing. Fundamental indexes have the theory behind them, and Greg is wrong here. See here and here for more on this including a good discussion and links to papers demonstrating this.
Siegel has more of an issue because some of his indexes are based on assumptions about low p/e stocks and other traditional value methods. I still believe they will work over time, but they do not all have the EMH behind them, but how valid the EMH is in the first place is another discussion. The great thing about Arnott’s versions of fundamental indexes is that they work whether the EMH is true or not. That leads me to point out the problem a naive faith in such theory might have as related by the (at least in this case) unfortunate Alex Tabarrok:
Indeed.
even if that’s the case(pricing errors), market participants would place a “premium” on low price to fundamental stocks (ie stocks likely to be “mispriced” on the downside). Another way of saying it is “What would happen if everyone owned Arnott’s index?” The market should discount the pricing errors, that’s what EMH says..
Obviously if everyone owns the index you have a problem. Of course that is a problem for any index, much more so with cap weighted indexes which exaggerate the effect in the growth of each individual security.
Your general point however doesn’t hold, or at least I don’t think it does, go ahead and let me know what you think. Since the fundamental indexes (I assume you clicked through to see what Arnott is talking about) take no position about what is or is not over or undervalued there is no simple way for the market to put a premium on the “low price to fundamentals” of the stocks or short the others.
For example, imagine a two company index. Let us assume the fundamental chosen is merely the number of employees. One has 2000 employees and the other 2500. In the index they will be weighted where the first company is 40% of the index and the second 60%. However, in the stock market the first is valued at exactly the same cap weight. That might be a true fair value, it might not. The index doesn’t give you any information.
All the fundamental index does is track the actual growth in the companies workforce. The increased return comes from capturing the growth in the value of the firms from that point. It assumes the errors are random. Its excess return over cap weighted indexes, if the EMH holds true, comes from removing the return drag from always overweighting overvalued companies and underweighting undervalued companies. Fundamental indexes take no position on which companies those are! So there is little to arbitrage.
To the extent that the EMH doesn’t hold true the excess return can be arbitraged away, but removal of the return drag still holds. The problem with single factor fundamamental indexes, such as in Siegel’s low p/e indexes or other style factor indexes, is that theoretically under EMH the effect you are describing will arbitrage away the excess return, so if you believe in the EMH one would worry about that, though of course if the EMH holds true then it should have already happened.
The same would also hold true for things such as small cap indexes. The EMH proponents have used small cap concentrations within a diversified portfolio to up return, they assure us the excess return has historically been due to higher risk levels. One might also hypothesize that it has historically been because they have been a lot cheaper than they should be, which also means that right now they may not be and future returns from this point will not exceed larger indexes. The mispricing has been arbitraged away.
Arnott is basing his index on several fundamental factors to capture the breadth of valuation indifferent factors that might signal a firms value to avoid any “value effect” which can be arbitraged away, or more importantly to this discussion may already have been.
I don’t know if I am explaining this well, so let me know.
I think in my brievity I lost some of my explanation.
people hold assets based on their perceived expected future return. The “arnott effect” if there is one should be incorporated into the expected return. if it isn’t please buy the fundamental index for me so my cap weighted index fund is moved back closer to the efficient frontier… ; )
my advice from someone who works on wall street selling these products, whatever you decide to do, keep your costs very very low.
Paul,
I think you, Bogle and Arnott and I are on the same page there, and if fundamental indexes cannot be had cheaply enough the removal of the drag from cap weighted indexes is certainly wasted. I am looking at it theoretically, not from a practical standpoint, though it seems to me the products based on it should be relatively inexpensive to construct.
By the way, have you read Arnott’s research or Mauldin’s synopsis of it? Unlike Siegel who can be accused of data mining, it looks to me as if Arnott has handled that issue rather well, so it seems to hold up both theoretically and empirically.
There’s a sleight-of-hand in the sales pitch for fundamentals-weighting that buyers should be aware of. The core of the argument for fundamentals-weighted investing revolves around the observation that “[in cap-weighted indexes] stocks trading above fair value are going to be over represented, and those trading under fair value will be under represented”. While that observation is true for cap-weighting, it’s also true for any portfolio…including fundamentals-weighted portfolios. By definition, any portfolio owns too much of stocks that are trading above fair value and too little of stocks that are trading below fair value. The only exception to that is a portfolio where we assume all stocks are trading at fair value (i.e., there are no errors). Since that observation applies to all portfolios (with the sole unrealistic exception noted) it is meaningless. You can make an argument that fundamentals-weighting or some other weighting scheme positions a portfolio to benefit from valuation errors, but if markets reflect that argument in prices then that weighting scheme becomes cap-weighting and the opportunity evaporates.
Eric,
Not true Eric, while each security will not represent the true value of each company, the error will not be exaggerated as it is in cap weighting. Each company will be ranked by its actual fundamental size in our economy. Thus while each company should be ranked differently, or at least weighted differently, the errors will be random.
Therefore, unlike with cap weighting, those that are overvalued carry no more weight merely for being overvalued. Those that are undervalued carry no less weight for being so. If the errors are random the return drag theoretically would be zero. If they are not random the the EMH says the non random errors, those recognizable by investors, will be arbitraged away.
Of course, over time the increasing and decreasing market value of the firms would drive weights above and below their fundamental weights, and over time they would begin to resemble cap weighting. However, that is why the index is rebalanced back to policy, in Arnott’s research, each year. Thus the cap weighting bias over time is corrected.
Pingback: dustbury.com
Lance,
Any portfolio owns too much of stocks trading above fair value and too little of stocks trading below fair value. The exaggerated valuation errors you associate with cap-weighting exist if your assumptions about valuation errors are met. You can make a case that those errors exist and persist, but you can also make the case that investors price stocks in a way that those errors provide no opportunity. My point is not that one cannot make a case for fundamental indexing, but rather that the statement that “cap-weighting over-represents overvalued stocks and under-represents undervalued stocks” does not make the case. The more important argument to make is why investors might fail to price stocks to reflect the opportunity that valuation errors present. That would explain how valuation errors can provide a on-going opportunity that investors can harvest with fundamentals-weighted or other valuation-tilted approaches. I think that case can be made, but it is a more subtle argument than a simple statement about how cap-weighting suffers from exaggerated valuation errors.
Eric,
How can that necessarily be true if the weight of the security is in no way related to the capitalization or price of the firm, at least as of the date of formation of the index? The index’s are valuation indifferent. You have said it has to be true, but you have not explained why. I admit I may be missing something obvious, so I am ready to be persuaded. Give me an example, or a link to one, because I am not getting it.
Lance,
Say you have a portfolio of 100 stocks that is equally weighted. If I suspect half of the stocks are trading above fair value then my 50% weight in those stocks is too high, and similarly my 50% weight in stocks that are trading below fair value is too low. I agree that if valuation errors are symmetric around some true fair value, then price (or capitalization or price/fundamental) will help investors identify valuation errors – i.e., stock with low price, low cap, or low price/fundamental are more likely to be undervaluation errors, and vice versa for stocks with high price, high cap, or high price/fundamental. However, the important question to answer before making that bet is whether or not you believe prices of stocks already reflect the notion that price (or cap or price/fundamental) is a predictor of valuation errors. I would concede that the empirical evidence suggests that markets do not fully reflect that tendency in prices, so there may well be value in trying to take advantage of those valuation errors. I guess my point is that the simple statement about the influence of valuation errors on cap-weighted portfolios makes it sound like it’s an arbitrage opportunity that’s true by definition, whereas I would suggest it’s really a much more uncertain bet that markets don’t – and won’t -reflect in prices the the influence of valuation errors.
I am not sure it is a bet at all. More importantly that isn’t what I took you to mean by it is definitional that an index will hold more of an overvalued stock than an undervalued stock. In the fundamental index case, which is based merely on size measures, the errors in an efficient market would be symetrical.
Maybe, but it seems to me that almost nobody holds that the valuations are perfect, what they hold is that we can’t know what they are and therefore cannot consistently profit from the errors. Fundamental indexes are not a way of profiting from them or any arbitrage, but of avoiding being penalized by them. The theory makes sense and the evidence backs it up. If you read Arnott’s paper they did a good job of avoiding the data mining charge.
Markowitz and Treynor are on board, so its flaws have been looked at by some pretty sharp folks, (Though not Sharpe himself, he just dogmatically says it can’t be true, just like Bogle.) I know, its an appeal to authority, but read the papers, Treynor’s is good.