Another nail in the coffin of the anti-EMH model(s)

Every few months I foolishly post a defense of the EMH, knowing that it will enrage my readers and cause me to spend endless time responding to critics.  My basic argument is that the question is not whether the EMH is “true” (which to a Rortian like me is a rather ambiguous concept) but rather whether the EMH is useful.  I think it is, and I think the anti-EMH is not useful.  The EMH underlies much of my research, and much of the analysis of this blog.  If I had started blogging in September 2008 my reputation would be higher than it is, as I’d be given credit for pointing to the Fed’s mistakes in real time.  All because the markets told me so.  (Of course if I’d started blogging in 2005, I’d have looked like a fool in 2008.)

The anti-EMH doesn’t seem useful.  Even if there are bubbles, low paid federal regulators won’t spot them.  We can’t base policy on the assumption that regulators will prevent sophisticated investment banks from making the wrong investments.  And indeed almost all levels of regulation were cheerleaders for the housing boom–hence the housing bubble is a massive failure for the anti-EMH model.  The anti-EMH model is not useful to regulators.  (Of course regulation can be used to offset moral hazard, but that has nothing to do with the EMH.)

And it’s also not useful to investors.  In the past I’ve argued that mutual fund excess returns are not serially correlated.  Just because someone does well one year, doesn’t mean he’s more likely to do well the next year.  It’s just luck.

My commenters insist that this is the wrong test.  They say the managers of mutual funds aren’t very smart–that the Smart Money manages hedge funds.  OK, so how has this ultra-smart money done in the past few years?  From The Economist:

Since a low point for global stockmarkets in March 2009, returns have been rather modest for hedge-fund investors and bondholders, while equity investors have done much better. According to the Hedge Fund Research index, a closely watched performance measure, total returns to investable hedge funds were just 19.6% in the 26 months to May 2nd. In the same period holders of rich-world government bonds saw total returns of 21.2%. Investors in stockmarkets in the developed world fared well in comparison. They enjoyed returns of more than 114% in the period covered. Since the beginning of 2011 equities have risen by 10%, while hedge-fund investors have seen total returns of only 1.4%.

Wow!  They must have been following the advice of Nouriel Roubini.  I’m sure glad I don’t invest with hedge funds, my 401k has almost doubled since early 2009.

PS.  Shiller should have been out in the streets screaming BUY in March 2009.  Instead he was kind of wishy-washy.



46 Responses to “Another nail in the coffin of the anti-EMH model(s)”

  1. Gravatar of spencer spencer
    14. July 2011 at 08:45

    the problem I have with hedge funds is that it is hard to believe that we recently had a massive shift in the number of very small managers that were able to successfully exploit another independent large jump in market inefficiencies.

    I guess we don’t really need the increase in inefficiencies.

    I suspect that if you risk adjust hedge fund returns you would find an even worse record.

    Maybe what we are doing is just paying some managers much more to take bigger risk.

  2. Gravatar of spencer spencer
    14. July 2011 at 08:47

    Sorry, that should be smart managers, not small managers.

  3. Gravatar of ssumner ssumner
    14. July 2011 at 08:50

    Spencer, Perhaps, the bottom line is that no one can beat the market consistently, even Warren Buffet lost a ton of money in 2008.

  4. Gravatar of Hyena Hyena
    14. July 2011 at 08:57

    Why then do hedge funds do so well in gathering investors and their heads so well remunerated? I’ve never really understood this; if a simply index that could be compiled by a seventh grader after a little instruction outperforms them, how does this business survive?

  5. Gravatar of Charlie Charlie
    14. July 2011 at 09:01

    For anyone curious, the index is here ( Over the longest period listed (60 months) the index is at 4.8% and the S&P 6.4%.

    Hedge funds tended to lose less in 2008 than the S&P but lost ground after.

    Anybody have an update on the Warren Buffet vs. hedge fund of funds long bet?

  6. Gravatar of K K
    14. July 2011 at 09:27

    OK, but why are you benchmarking against the market?  Have you checked hedge fund beta over the last couple of years?  If they have been zero beta then in an efficient market they have zero expected return (with the risk free rate at zero). If hedge funds are zero beta and have some positive alpha (no matter how small) then an optimized portfolio should hold some hedge fund exposure. At least run a correlation with the market (stocks, bonds, whatever) to back-of-the-envelope the beta.  Then see if you are able to say anything statistically significant (positive or negative).

    Also, anybody who is paying hedge fund style fees to invest in positive beta hedge funds is a fool since beta is free.  So if you do find positive beta in your index then what you have is a crappy-hedge-fund index.

  7. Gravatar of John hall John hall
    14. July 2011 at 09:28

    Part of the problem is that most people don’t really understand what EMH is about. The most common straw man I see is that EMH says prices are always at their fundamental values. Technically, EMH is more saying that information (assuming away liquidity/transaction costs) is quickly absorbed into stock prices. I think of it like EMH says that stock prices reflect participants expectations, but it doesn’t really say that these expectations ex post will be correct. For instance, the market priced in continued gains for tech stocks during the bubble. While these were the best guess of those at the time (weighted by how much money they had), ex post they were wrong. I don’t see that as inconsistent with EMH. Of course, that also makes it a very difficult concept to argue against.

    I suppose my biggest problem with EMH is that whenever someone can come up with a way to earn some money over time in a way that is not correlated with the market, a finance professor calls it a risk factor and say all you’re really doing is taking some other risk. This is all over the double-sort literature. For instance, high B/M stocks outperform low B/M stocks, but if you have a portfolio overweighting high B/M stocks then all you’ve done is taken some additional value factor risk. Call it alternative beta, but if you’re exposed to factor risks that are not correlated with someone’s portfolio, then I still think it’s adding value.

  8. Gravatar of foosion foosion
    14. July 2011 at 09:32

    The problem with the EMH is what exactly it means. In the beginning it seemed to mean that prices are right. It quickly became that it’s very hard to beat the market. That should have been obvious (at least, everywhere outside Lake Wobegon).

    Do any managers have skill? There just isn’t enough data to answer that question with a reasonable degree of statistical significance.

    What experiment could we run that would disprove EMH?

  9. Gravatar of Steve Steve
    14. July 2011 at 09:49


    The correct tests are:
    Do small pools of capital outperform larger pools of capital?
    Do funds with higher manager ownership outperform funds with lower manager ownership?

    In regard to your previous post, I agree that EMH is the only useful view for a MACROeconomist. EMH needs to be the null hypothesis especially for policymakers. And this is consistent with my above tests of EMH. A policymaker is the ultimate example of someone managing a big pool of capital (the entire economy) with no ownership stake in the outcome.

    However a micro guy who sees a twenty dollar bill on the ground should pick it up before someone else does. There are lots of little opportunities and infrequently bigger ones. That’s why you get hate mail whenever you talk about the EMH.

    The value of EMH is very context dependent. And I don’t buy the semantic defenses of the EMH either. Suppose I make one million a year trading stocks, but I am exceptionally talented and I work very long hours. Is that million really my “salary” and the markets are in fact still efficient? That might be a useful way for a macro guy to think about it, but that’s not the conventional meaning of EMH.

  10. Gravatar of Gregor Bush Gregor Bush
    14. July 2011 at 10:36


    But what was the beta and/or volatility of the HFRX index? I would imagine that the HRFX index vastly outperformed the S&P 500 from early 2008 to early 2009. It almost has to be the case because there always many funds that are effectively short the market (this is not the case with mutual funds).

    Also, many hedge funds are long-short strategies and have low beta (think long-short currency portfolios). Looking at their performance in the aggregate isn’t particularly meaningful because you know that it will be roughly zero.
    A hedge fund that generated positive returns while having a negative beta which would have been quite useful to have in your portfolio. It would have been nice, for example, to have been long CHF against the EUR, in addition to being long the MSCI DM index. The information ratio is the key measure of performance for hedge funds, not returns.

  11. Gravatar of Benjamin Cole Benjamin Cole
    14. July 2011 at 10:37

    Another superb Sumner post. The anti-EMH’ers may have a few points, but really, in the long run, EMH rules. The fact that EMH rules means that the Fed can stimulate w/o causing bubbles. If the Fed runs too hot, then all markets would inflate about the same. If a housing goes up more, it is for reasons of the market, not the Fed.

    As Sumner points out, you had Wall Street sharpies buying and selling huge tranches of residential MBS and CMBS deep into the danger zone. Okay, so the Fed is supposed to know that is a bubble? When Wall Street guys in the market do not know it?

    I fear the price of suppressing every “bubble” or hint of inflation is we do a japan.

    Listen I would much, much rather live with some “bubbles” and some 3 percent inflation, than go into the freezer for 20 years, ala Japan.

    Japan is the worst option available–and the one we are taking.

  12. Gravatar of caveat bettor caveat bettor
    14. July 2011 at 11:22

    Scott, what is your Sharpe ratio since 2008 (or even further back)? (If you prefer Information ratio or Modigliani-Modigliani, that would be fine, too).

    Good hedge funds hedge, which means they should occasionally underperform if they are mitigating some of the tails.

    I am a little sensitive to this, because, yes, I am up single digits 2011 YTD. But in my trading strategy, I am averaging 25% since 2007, and was up 19% in 2008 (of which i am rather proud).

    Without more information, I’d probably prefer sticking to my knitting, outperforming the stock market for a fraction of the market’s volatility over a 3-to-30 year timeframe.

  13. Gravatar of Charlie Charlie
    14. July 2011 at 12:09

    For commenters that want to talk betas (or sharpe ratios), it’s actually not enough. The EMH would just say that we should then look at the next moment, kurtosis. In fact, several studies have looked at hedge fund kurtosis (tail risk) and found it much higher than equities.

    It’s fully consistent with EMH for one to give up either mean return or take on tail risk to lower betas.

  14. Gravatar of John hall John hall
    14. July 2011 at 13:27

    Regarding Charlie’s point, Rockafellar and Uryasev have developed a version of CAPM that incorporates tail risk using CVaR deviation (basically you subtract out the mean and calculate the expected loss in the tail) instead of standard deviation.

    One thing they find is that implied risk aversion obtained from option markets data is not constant over time. In some periods, the market can be characterized as caring about only the bottom half of the distribution, but during crisis periods they put much more weight on the tail. Hence, a dynamic strategy that adjusts portfolio allocations to become more risk averse (in the tail risk sense) when the market is risk averse can result in better portfolio performance than a constant risk aversion strategy.

    Generally speaking, there is significant evidence that paying attention to regimes can improve portfolio performance. For instance, Ang and Bakaert look at Markov Regime-switching and global asset allocation and find positive results (I have done the same thing and found similar evidence). Since you are taking advantage of time-varying means and variances, you really aren’t taking on excess kurtosis risk. In fact, you are more often than not limiting your kurtosis risk by being more conservative in the very times when tails are widest (a high-volatility regime).

  15. Gravatar of K K
    14. July 2011 at 13:33

    Charlie: “The EMH would just say that we should then look at the next moment, kurtosis.”

    Yup.  And if that doesn’t work out, hey, there’s always higher moments. The ultimate unfalsifiable theory.

    The EMH: markets are just magically efficient, beyond the skills of even the smartest managers.  Any deviation from that state, no matter how small, are immediately discovered and arbitraged away by agents of such superhuman intelligence, that the cost to them of discovering said inefficiencies are zero.  Therefore, they provide this service for free.

  16. Gravatar of malavel malavel
    14. July 2011 at 14:49

    I buy the EMH arguments intellectually. But it’s really hard to buy them emotionally. I sold all my stocks right before the 2001 and 2008 crashes. And I like feeling smart by believing I can outsmart the market.

  17. Gravatar of Tom Powers Tom Powers
    14. July 2011 at 14:51

    Who really knows whether the hedge fund performance is good? It’s pretty unclear what risks they are exposed to, and to determine whether someone is beating the market you have to compare to a passive portfolio holding all the same types of risk.

    Some commenters pointed out that many funds are not exposed to market (beta) risk, so they don’t need to make as high returns to be beating the market.

  18. Gravatar of Charlie Charlie
    14. July 2011 at 14:53


    If you don’t care about the third moment, why do you care about the second?

    Also, no superhumans are required. The returns to correcting market inefficiencies should be equal to the returns for a similarly skilled person doing something else. It’s actually a bit of a puzzle* that in so many studies the average manager doesn’t earn their fee. The average manager may have no value added at all.

    *albeit a puzzle with plenty of answers

  19. Gravatar of Scott Sumner Scott Sumner
    14. July 2011 at 15:23

    Hyena, People are hardwired to believe the successful are smart, that it’s not just luck.

    K, OK, technically that’s right. But I was reacting to critics of the EMH, who use studies like Shiller’s. These guys claim the stock market is not a random walk, and we can know when to invest. Invest when the P/E is low, and stay out of the market when the P/E is high. Well it doesn’t seem to have helped either the hedge funds or Shiller, who were both out of the market during a huge run-up. That tells me that it’s not easy to tell when stocks are overvalued.

    John, Good points, and keep in mind that “fundamental value” is a very vague term. Does it mean perceived fundamental value? If so, perceived by whom? Does it mean what people think at the time, or what we know after the fact?

    foosion, There is only one rational way to test the EMH, and it’s already been done. Test for policy or investment implications. Try to see whether others have found flaws in the EMH. If they have, then excess returns should be serially correlated. And they aren’t, or are to such a small amount that it has no investment implications. This suggests the EMH is true, or false but close enough to be a useful approximation of reality. I’m fine with either one.

    Steve, You said;

    “The value of EMH is very context dependent.”

    I’ve never denied that people find occasional $20 bills on the ground. I’ve said the anti-EMH models have no policy implications for the government, and no investment implications for ordinary people. Buy index funds. But people don’t want to hear that. Academics think government bureaucrats know better than Goldman Sachs whether mortgage backed securities will fail. That’s what much of the liberal blogosphere keeps saying. The say the housing bubble was obvious, the regulators should have popped it. And Wall Street keeps telling ordinary investors that they can invest their money better than index funds.

    Gregor, See Charlie above–he says the hedge funds have underperformed over 60 months.

    Benjamin–Yes, monetary policy does not create bubbles. The 1929 “bubble” occurred in the only deflationary business expansion of the 20th century. And even that was blamed on “easy money”!!

    Caveat, I’m embarrassed to admit that I’ve outperformed the S&P by quite a bit–which totally undercuts my EMH advocacy. 🙂

    Also see my answer to “k” above–that was my real point, I should have made it clearer.

    Charlie, Interesting that you and the other commenters all know much more about the EMH than I do. These are just “fun” posts. My serious argument is:

    1. Anti-EMH has no policy implications.
    2. The publication of data-mined “anomaly” studies in top journals is a disgrace to the profession.

    I really don’t know anything about alphas and betas, beyond that fact that you are supposed to earn higher expected returns when your asset adds to risk (relative to the overall market) and lower returns when it reduces risk. Is that right? I’m not defending that part of the EMH, which I see as a separate issue.

    Of course the EMH is literally false, AS IS EVERY SINGLE OTHER THEORY IN THE SOCIAL SCIENCES. But because brilliant minds like my commenters keep looking for $20 bills, it is close enough to reality to be really useful for policymakers, academics and ordinary investors.

  20. Gravatar of Scott Sumner Scott Sumner
    14. July 2011 at 15:25

    malavel, That’s exactly how I feel. Well not exactly, because I haven’t been as smart as you, but I have outperformed.

    Tom, Yes, but see my first response to K.

  21. Gravatar of Matt Waters Matt Waters
    14. July 2011 at 15:41

    Without reading through the comments, I will say I made my dissertation in the next post before seeing this post. My argument probably does go to a straw man, like you say.

    As a tool, the EMH might be alright. It’s very expedient for research to look at the market price rather than digging through footnotes about some adverse writedown in Latin American markets. And even the most anti-EMH person would say the stock markets do absorb all information…eventually.

    The timeline is important though. To say new information was absorbed quickly in 1999 equities is silly. It was absorbed, but I wouldn’t call it quickly.

    The “truth” on the EMH, whatever that is, is probably somewhere between Greenspan circa 2006 and DeLong. In any case, the implicit idea that the EMH holds and therefore there are never any fundamental mispricing can be quite dangerous. Banking regulators aren’t perfect, but if the government has to backstop their operations, I would rather we try to have a lot of built in cushion in case they make mistakes in prices. When leverage can go to 33:1, the banks’ pricing of assets basically has to be perfect or the bank quickly becomes insolvent.

  22. Gravatar of MarkS MarkS
    14. July 2011 at 15:42

    That’s rich. You claim to have outperformed the S&P 500 yet you’re saying that no one can beat the market.

    Last time I stopped in here you were lecturing people on moentary policy just minutes after admitting that you are “not an expert” in banking. Ha.

    Pull that chart out another year. You’ll notice that hedge funds only fell about 20% in 2008 while the S&P fell 50%. There’s your outperformance right there. The whole point of a hedge fund is to reduce risk. The returns since 2005 vs the S&P are far better than your cherry picking.

    How do you take yourself seriously?

  23. Gravatar of Mike Sandifer Mike Sandifer
    14. July 2011 at 15:44


    Does the idea that so many try to beat the market, despite a lack of evidence as you see it they can undermine rational expectations theory at all?

  24. Gravatar of Mike Sandifer Mike Sandifer
    14. July 2011 at 15:48

    I’m guessing it doesn’t, since the collective market decisions that matter.

  25. Gravatar of Jon Jon
    14. July 2011 at 16:15


    The limitation of the graph is that it does not make an attempt to risk-normalize. This relates directly to the suspicious choice of starting year. Its not simply a question of whether hedge-funds should have piled into equities. That’s a hindsight question. Would that have been a better approach a priori? That’s the kicker.

    The very claim implicit in the name ‘hedge-fund’ is that risk-adjust return is better not simply the return.

  26. Gravatar of bill bill
    14. July 2011 at 16:31

    Another interesting Shiller note. I once read that he started giving a speech in 1995 that evolved into his book, Irrational Exuberance, by 1999 or 2000. Like the old saying, he wasn’t wrong, just early. But really lucky that the speeches were not nearly as widely heard as the book. In fact, I’m fairly sure that anyone who got out of stocks in 1995 did worse than people that stayed in stocks all the way through.

  27. Gravatar of Floccina Floccina
    14. July 2011 at 16:50

    And indeed almost all levels of regulation were cheerleaders for the housing boom-hence the housing bubble is a massive failure for the anti-EMH model.

    I would like to that shouted from he house tops.

    All this shows that we need not a more regulated but a more robust monetary system. The current system has bad feedback problems. The failure of banks is contractionary which leads to more bank failures, IMO that sort of system cannot work very well.

  28. Gravatar of John hall John hall
    14. July 2011 at 17:01

    Part of the problem with alpha/beta is that most people aren’t specific enough or don’t really know what they’re talking about. The way it has been used in this discussion is largely as an ex-post performance measurement. It has been used correctly in these discussions (regress returns on risk factors, significant intercept indicates a fund has outperformed after accounting for risk).

    However, CAPM is really a theory about ex ante expected returns. The theory is that stocks with higher betas are riskier and have higher ex ante expected returns (there is evidence that ex post stocks with high betas have periods that underperform, though in periods like the late 90s this wasn’t true). The evidence for traditional 1-factor CAPM is weak and there’s a lot of literature coming up with a better factor model (most notably Fama and French in 1992/3).

    That being said, no actual theory is required to get a beta.
    Meucci explains it well:

  29. Gravatar of Charlie Charlie
    14. July 2011 at 17:13

    “1. Anti-EMH has no policy implications.”

    I agree. Just the fact that EMH passes so many tests makes it a useful first-order approximation.

    “2. The publication of data-mined “anomaly” studies in top journals is a disgrace to the profession.”

    I would agree, but I’m entering a Ph.D. program in Finance. How else am I supposed to get through my dissertation? 🙂

  30. Gravatar of Lorenzo from Oz Lorenzo from Oz
    14. July 2011 at 18:07

    Scott: my recent comments make more sense if you realise I am reading your posts from latest to earlier. I seem to be furiously agreeing with you. Oh well, at least they are my honest immediate reaction 🙂

  31. Gravatar of Rien Huizer Rien Huizer
    14. July 2011 at 19:17


    We have monetary policy that relies on certain properties of the financial system. We have a financial system that consists of largely regulated firms, but often not regulated from a “prudential” point of view.

    The financial system used to raise capital for business but now the situation is that financial system assets and liabilities are largely (although often indirectly, via asset managers) in the hands of consumers. We are going to have macroprudential regulation covering primarily the banking system and important differences in consumer protection. This on top of a system that is already one of the most regulated in the world (and hence very beneficial to professionals).

    I suspect there are areas of tension between what would constitute an effective macroprudential system and efficient macroeconomic policy (to mean at least predominantly using the monetary component). An effective US macroprudential approach would go a long way in protecting the global financial system against systemic risk, would be protecting the taxpayer against opportunistic behaviour of licensed firms and would reduce the externalities of asset disposal (“fire sales”). All of that by forcing regulated firms to manage their capital on an anticyclical basis (and making it very difficult for investors and managers to understand financial firm value given the business cycle bias in the structure of returns that would apply largely across the board). The discussions preceding Basle II touched on procyclicality (compared to the previous combination of capital adequacy and accounting rules) but it was put in the too hard basket.

    Consumer protection (if effective) would have had important effects during the past decade because it would have (again if effective) kept construction activity at a lower level and related activities also, and probably caused less of a wealth effect. Hence maybe another anticyclical device with maybe lots of unintended consequences. And another element of friction in financial markets.

    Question: to what extent would (effective) macroprudential regulation (and an anticyclical consumer protection bias) change the operating environment of monetary policy. Also wrt efficient markets: would a large section of the stock market (financials) become more suscep[tible to analysis (by modelling bank shares as a business cycle derivative)?

    Of course I have serious doubts that macroprudential regulation can be made to work, especially in a country with such a large shadow banking system, but if one would assume an effectice solution, what would be the consequences fro macroeconomic management?

  32. Gravatar of caveat bettor caveat bettor
    14. July 2011 at 19:47


    I’m embarrassed by quite a bit, too. For instance, I believe that volatility is not risk. But standard risk measures are all about volatility, not so much about uncertainty (which is what I am more interested in hedging, quite frankly). I am embarrassed to have asked you to give me a measure that I have already discounted.

    All the greats have had their hats handed to them at some point: Buffett, Soros, Robertson, etc. But they are still the best players. Wouldn’t EMH purport that there are no good or bad players, they all break even, like a 70’s Kenny Rogers Billboard-topping song? I’m embarrassed to have to call Buffett and Soros great; I think Malone is even better, but then Lord Acton was right: most great men are not good men.

    I still think that it matters that Bill Gross consistently makes money, and Meriwether, Taleb, Schiff, Madoff et al do not. EMH doesn’t think that matters, right?

  33. Gravatar of Andy Harless Andy Harless
    14. July 2011 at 20:14

    The phrase “anti-EMH” has rather a straw-mannish sound to it (though maybe it’s a straw man that people actually believe in). We don’t talk about anti-Darwinian selection or anti-general relativity. Usually if you want to argue against a theory, you either have to argue that it’s completely useless or to provide an alternative that you say is better. I don’t think anyone can seriously argue that EMH is completely useless, but there are alternative theories (noise traders, rational bubbles, various principal-agent problems, etc.). Any reasonable argument has to be about alternative theories and their specific implications. Just to say something like, “Regulators should act to deflate bubbles before they get out of hand,” is silly unless you can say something about how regulators might be able to detect bubbles. But some people, I think, do have more sophisticated arguments.

    The publication of data-mined “anomaly” studies in top journals is a disgrace to the profession.

    I don’t agree with this. Taking the EMH as a working hypothesis is one thing, but academics also have a role in making the market more efficient. I believe (though it would be hard to prove) that some of the anomalies are genuine inefficiencies, the discovery of which has the effect of correcting them. The classic example is the January effect. The January effect is a very robust finding for the period before it became well known. After it became well known, it ceased to exist. You can argue that it never really existed and it was just noise in the data, but I don’t find that explanation plausible. I would say that people who published papers on the January effect actually performed a valuable service (especially for people who need to sell stock in December).

  34. Gravatar of K K
    14. July 2011 at 21:05

    Charlie: “The returns to correcting market inefficiencies should be equal to the returns for a similarly skilled person doing something else.”

    Should, yes. But in reality there are large barriers to entry. You need significant amounts of capital to do some forms of arbitrage efficiently. So returns are higher than in other endeavors (eg string theory). Doesnt resemble the EMH very much to me.

    Scott: it seems we are in near total agreement then. But what you are describing is a very weak version of the EMH that’s pretty hard to disagree with (though some people apparently do, as you point out in subsequent posts).


  35. Gravatar of Nick Rowe Nick Rowe
    14. July 2011 at 21:05

    Following on Andy’s second point: the anti-EMH hypothesis is useful because people who believe anti-EMH help make EMH true(er).

  36. Gravatar of K K
    14. July 2011 at 21:13

    “Efficient” just means “beats me”.

  37. Gravatar of Greg Ransom Greg Ransom
    15. July 2011 at 00:38

    I started blogging in 2002 and was talking about the distortion in the time structure of production (i.e. in housing production) before 2 years were out. Between that time and 2008 I had posted countless posts on the “housing bubble”, on Fannie Mae, on Bernanke’s bogus early 2000’s “deflation scare, etc., on the Greenspan put, on Greenspan’s rejection of the idea of a natural rate of interest or the idea of bubbles, etc. In 2005 i was specifically pointing to an artificial boom that was headed for an inevitable bust, linking to Calculated Risk, the Mises Econ Blog, the Housing Bubble blog, and many other “early” callers of the artificial boom/inevitable bust.

    Scott wrote,

    “Of course if I’d started blogging in 2005, I’d have looked like a fool in 2008.)”

  38. Gravatar of Greg Ransom Greg Ransom
    15. July 2011 at 00:42

    There is no single “Efficient Market Theory” there are a cluster of different and related empirical results, idealized “models”, and math construct stipulations.

    It really helps no one to pretend otherwise, and to conflate and confuse and jumble together apples and oranges into a fruit pie mess.

  39. Gravatar of Greg Ransom Greg Ransom
    15. July 2011 at 00:46

    Anyone know a good history of ideas on the development of the manyndifferent EMH and RE strands?

    I’ve read a hodge podge of various pieces of all this, but have read nothing close to comprehensive or definitive.

  40. Gravatar of TheNumeraire TheNumeraire
    15. July 2011 at 00:57

    Beating the S&P 500 has not been as difficult as it seems for most of the past dozen years. If you are a small fund or individual investor you can more easily accumulate and are more likely to have a heavier weighting of small and midcap stocks, which have greatly outperformed the cap-weighted S&P 500 almost year-after-year.

    One of my favorite market commentators, Eddy Elfenbein, wrote a brief on this last week;

    The Wilshire 4500 Rally

    From the article;
    “From October 8, 1998 to yesterday, the total return of the Wilshire 5000 is 99.78%, but the total return of the Wilshire 4500 is 218.29%.”

    “To give you another example, we’re coming up on the 10th anniversary of 9/11. Measuring from September 10th, 2001″” so we’re including a bear market that lasted for another 18 months””to yesterday, the total return of the Wilshire 4500 is 140.85%. Annualized, that comes out to about 9.4%. That fact probably would have surprised a lot of folks in the aftermath of 9/11.”

    Elfenbein is right; those returns would surprise an awful lot of people, especially those in the popular financial media, who have constantly been repeating the “stocks have gone nowhere for the past decade” meme.

    Even within the S&P 500, the smaller companies have been outperforming the biggest companies. The S&P 500 equal-weight index (which was first compiled in 2003) has been outperforming the cap-weighted S&P 500 considerably.

    I’ve personally been aware of the trend since about late 2005. I did a little digging after discovering my own portfolio was doing astonishingly well and many of my friends were boasting about their alleged stock picking prowess. Turns out we were all heavily or exclusively invested in non-S&P 100 stocks.

    Furthermore, to tie this market factoid into the endless political debates, I feel this small-company outperformance of the past 10 years contradicts the deeply held belief of the political left that wealth has been flowing exclusively to a small group of superwealthy, politically connected capitalists. It appears to be just the opposite.

  41. Gravatar of Scott Sumner Scott Sumner
    15. July 2011 at 07:03

    Matt, You said;

    “Banking regulators aren’t perfect, but if the government has to backstop their operations, I would rather we try to have a lot of built in cushion in case they make mistakes in prices. When leverage can go to 33:1, the banks’ pricing of assets basically has to be perfect or the bank quickly becomes insolvent.”

    Many people falsely believe that the EMH implies banks should not be regulated. Not true. It suggests some types of regulation are appropriate, and some aren’t. It does say regulators shouldn’t try to predict bubbles, but it also says they should force banks to behave more conservatively (because of moral hazard created by government insurance.)

    MarkS, Charlie says the S&P did better.

    And I attribute my success to luck.

    Mike, No, because if no one tried to beat the market, the market wouldn’t be efficient. So some people have to try to beat the market, to make it efficient. The rest of us should buy index funds.

    Jon, The point of this is to show that hedge funds don’t know when the market will go up. That’s the only point. I’m not claiming they should have piled into stocks. I’m saying if the claims of the anti-EMH people like Shiller are correct, they should have piled into stocks.

    Bill, He was wrong, early is wrong. You may have heard that story over here, I repeat it often.

    Thanks Floccina.

    John, Thanks for the link.

    Charlie, Make your dissertation a critique of data mining. Argue that finance professors should not look for anomalies, but rather should look for evidence that market participants have found anomalies.

    Lorenzo–don’t worry, my posting will slow down soon.

    Rien, There are two issues here. If monetary policy is NGDP futures targeting, then you want to remove all regulation of banking. If it isn’t, then financial crises can throw monetary policy off course, a la 2008. In that case you do the sorts of regulations you suggest, things like 20% minimum down payments. That would have greatly reduced the housing bubble/crash.

    more to come . . .

  42. Gravatar of Scott Sumner Scott Sumner
    15. July 2011 at 07:07

    cavaet, No, as this post shows, the EMH predicts that out of every million investors, one will have one-in-a-million streaks of luck.

    For some reason I can’t attach the link, but you can Google: “Being there themoneyillusion.”

  43. Gravatar of Scott Sumner Scott Sumner
    15. July 2011 at 07:08

    Andy, You said;

    “The classic example is the January effect. The January effect is a very robust finding for the period before it became well known. After it became well known, it ceased to exist. You can argue that it never really existed and it was just noise in the data, but I don’t find that explanation plausible.”

    I couldn’t disagree more strongly. Anti-EMH types like Shiller insist that their models work even after being discovered. I disagree. In contrast, the pro-EMH side says it’s all data mining and the anomalies won’t hold up. I don’t see where you have any basis for believing the January effect was real. None at all. Even if no such anomalies really existed, data mining studies would turn up a million of them, for every 20 million patterns examined. So the evidence from the anomaly studies is of zero scientific value. And with modern computing it’s not hard to quickly evaluate lots of patterns.

    The better test is whether there is evidence that market participants have discovered anomalies, and kept them secret. And that’s where the EMH really shines. For instance, mutual funds that do well one year tend to mean revert the next. If they had really found an anomaly, excess returns should be serially correlated.

    Thanks K.

    Nick, I sort of agree, but I think those people are the market participants (like Buffett), not academics. I think all this stuff about rainy day effects is just superstition. (Yes, the AER did a paper about how stocks did worse on rainy days, significant at 10%!!!)

    Greg, And did you correctly predict in 2005 that the Aussie housing bubble would not burst? And all the European housing bubbles that did not burst?

    TheNumeraire, I agree that after the fact it is easy to spot which market segments will, I mean did, outperform the S&P 500.

  44. Gravatar of Greg Ransom Greg Ransom
    15. July 2011 at 22:47

    Scott, I wasn’t following any of that — not on my radar, not part of my reading.

    “Greg, And did you correctly predict in 2005 that the Aussie housing bubble would not burst? And all the European housing bubbles that did not burst?”

  45. Gravatar of Mike C Mike C
    16. July 2011 at 10:56


    I know of proponents of an anti-EMH theory that is better than Shiller’s, although based on it. It is espoused by the writers at Zeal LLC who did make a buy call on 6 Mar 2009, They agree that expectations drive makets, but show how predictably incorrect expectations result from greed and fear. In addition to the longer secular cycles of P/E valuation that could last up to 34 years, they look at cyclical cycles that are predictable based on an index’s value relative to its 200 day moving average and the use of the VXO as a fear gauge. They also take advantage of the inverse correlation of commodity performance, particularly precious metal miners, to long-term stock P/E valuation. Following them since 2004 really peaked my interest in economics and has converted me to their position on the EMH (they have achieved 51% annual realized gains on stock trades since 2001). That interest in economics led me to your blog, where I have read every entry and many of the comments. It is excellent on macro policies, and I hope this information will influence your position on EMH.

  46. Gravatar of Scott Sumner Scott Sumner
    17. July 2011 at 16:43

    Mike, Sounds good, let me know when to sell.

Leave a Reply