27 Aug 2014

The EMH versus the CAPE

Efficient Markets Hypothesis, Scott Sumner 76 Comments

Robert Shiller’s CAPE (cyclically adjusted price/earnings) ratio is well above its historical average:

shiller-cape

Shiller himself has been pointing out that the only periods since 1881 where the CAPE has been this high were 1929, 1999, and 2007 (on the way up, we should say for precision).

Scott Sumner, proponent of the Efficient Markets Hypothesis (EMH), is unfazed. Here’s Sumner commenting on the CAPE chart:

Any fool can data mine and find spurious correlations. The real test is how they do out of sample. Robert Shiller became famous in 1996 with his “irrational exuberance” claim (or at least one degree of separation from famous, as it was when Greenspan repeated this claim that the public took notice.)

What do you see [in the chart of the CAPE]? I see a trend line that seems to have shifted upward from 16 in the mid-1990s, precisely when he made the irrational exuberance claim. The new average looks like about 25. And the current ratio is above 25. So the out of sample test was about as complete a failure as one can imagine. The model simply did not perform out of sample.

How in the world does Scott get that the CAPE “was about as complete a failure as one can imagine”? Now if the CAPE had leveled off since 1996, or if the stock market had risen at a steady growth rate with a bit of volatility, then sure, Shiller would have been totally wrong.

But that’s not what happened. A few years after his “irrational exuberance” call, the market crashed hard. Then when the CAPE zoomed back up [edited–RPM], the market again crashed hard, falling back to the historical average. What would the data have to look like, for Shiller to think he was on to something?

The irony is, no matter what happens, Sumner will still think he is right. If the stock market crashes 40% next week, Sumner will say, “That means nothing, EMH is still correct. If Shiller’s model actually had predictive power, people would’ve shorted the market this week.”

I’m not making a joke. That is literally what Sumner would say. And he wouldn’t even be crazy to do so; that’s the non-falsifiable world of the EMH. But what’s annoying is that Sumner seems to think he’s looking at the data with an open mind and then forming his conclusion, when in fact “believing is seeing” in finance as much as other fields.

UPDATE: I should clarify that having the market crash while the CAPE falls back down, is not itself a tautology. For example, if the government suddenly declared that all existing CEOs would be executed, that would probably make the S&P 500 fall substantially. But part of the mechanism would be that earnings themselves would fall sharply. The CAPE ratio is capturing the relationship of asset price to underlying earnings. (It goes back 10 years, adjusting for price inflation, rather than the more conventional current-year P/E ratio.)

76 Responses to “The EMH versus the CAPE”

  1. skylien says:

    The problem with looking at P/E’s is that it has two moving parts. You cannot conclude from a high P/E that the stock market has to fall. Earnings could rise as well, then a high P/E was just investors front running (predicting correctly) higher future earnings.

    Now I am not so sure anymore that the stock market now “must” be in a bubble as I was at least one year ago. What if earnings rise, maybe due to inflation?

    And the worst thing is that I actually think that we are currently having a genuine Schrödinger’s stock market, which is both at the same time, in a bubble and not in a bubble. The thing (poison as in Schrödingers example) that actually determines the condition of the stock market is the Fed. The Fed decides to prop it up or not!

    I think they will try to get a little correction, but then try to push it further higher, means we should see further tightening by the Fed and then easing again as soon as the stock market starts to correct. I think at some point they will lose control, but don’t ask me when.

    • skylien says:

      And you ask how in the world earnings should rise?

      1: Who can for sure say that those excess reserves never come through to the little guy on the street?

      2: Or maybe just like this:
      http://www.foreignaffairs.com/articles/141847/mark-blyth-and-eric-lonergan/print-less-but-transfer-more

    • Kevin Erdmann says:

      Why do you think loose monetary policy makes CAPE go up? It was below 10 in the late 1970’s when inflation was double digits, and it has been at very high levels during the 35 years since then, when inflation has been declining steadily. Why wouldn’t you conclude that tight money causes a high CAPE?

      • skylien says:

        I have not said that. What I said was that you cannot conclude from a high P/E that the stock market has to go down in order to have the P/E revert to a long term average. It might but it must not. Else that would be an easy bet to get rich if you only need to follow the P/E.

        Also as I have said the Fed can’t do what it wants villy nilly, it may lose control and therefore its ability to stimulate the stock market without creating inflation like in the 70ies.

        • skylien says:

          *it might, but it doesn’t have to* … those “must” mistakes..

          Do native English speaking people recognize the inconsistent use of the term “must” and “must not” that plaques people who learn English?

          😉

      • Major.Freedom says:

        Why do we have to consider consumer price inflation as the gauge of how tight or loose money is?

        It could happen that during one age, high *monetary* inflation affects mostly stock and bond prices, whereas in another age that same percentage of monetary inflation could affect mostly consumer good prices.

        • Bob Roddis says:

          Do you agree that those who are fixated on these CPI arguments are oblivious to the problems of economic calculation and Cantillon Effects caused by funny money injections?

          • Major.Freedom says:

            Yes, and regular old fashioned myopia. “If it foes not negatively affect my standard of living in the short run, then it is someone else’s problem.”

    • Scott H. says:

      This CAPE analysis is inadequate. The stock price is tied to the net present value of future cash flows. Those future flows are being discounted at some of the lowest anticipated interest rates in history. The result is a high value on the CAPE chart.

      • skylien says:

        It is not inadequate. Discounting and therefore interest rates do depend on expected (price) inflation. Therefore price inflation increases nominal earnings and lowers the PE (if the stock market stays flat or increases less than earnings), and it will pressure interest rates upwards. See 70ies.

        • Scott H. says:

          So P/E’s are dependent on future discount rate expectations. However, a graph that doesn’t contain any information about those expectations is still adequate? You seem to be contradicting yourself here.

          • skylien says:

            Have I said the graph is adequate? And adequate for what? What do you mean if you say this?

            I have said myself above that looking at high P/Es and concluding that the stock market therefore is poised to fall is dangerous, did I not? So it is certainly not adequate enough to conclude this.

            Also of course there is the possibility that the long term average of CAPE shifts generally since it generally depends on people’s subjective values… I am not denying that possibility. However I think it is not likely that it shifted from 15 to 25 now and therefore end of story.. Of course that is just my opinion as investor, at least I don’t see any reason for such a shift of that magnitude..

      • Yancey Ward says:

        True, but a lot of market optimists are combining this new valuation metric with old equity premium.

        Sure, it may make sense to have a higher PE today, but the lower returns are now baked into the price you pay.

  2. Luke says:

    “..if the government suddenly declared that all existing CEOs would be executed, that would probably make the S&P 500 fall substantially. But part of the mechanism would be that earnings themselves would fall sharply. ”

    Why would earnings fall?

    • Gamble says:

      Probably the opposite, right.

      Funny how bottom out anarchy is only good some of the time and top down hierarchy is hard to shake. I hate those darn pyramids.

      Walk like an Egyptian.

    • Major.Freedom says:

      “Why would earnings fall?”

      Because of the loss in value the CEOs can no longer provide.

      We’re not Marxists.

      • Gamble says:

        I don’t think you are a Marxist just because you question the value of todays CEO( some of them).

  3. Billy Pilgrim says:

    Major problem still remains (not merely computational), and in this Sumner appears to be correct – albeit indirectly. When you are making predictions out-of-sample, you should not include ‘predicted outcomes’ in the in-sample average, and you should also have a theory for what is to be considered a “historical average”.
    Relying on Shiller’s data (easily avaiable at his site): the historical average in the whole sample 1880-2014 is 16.5. If one wanted to ‘predict’ a bubble in the late Nineties, however, one would have referred to the historical average before 1999, which was around 15.5. And if you are predicting a bubble for the increase of the late-Eighties/early-Nineties, your “above historical average” is even lower, around 14.5 (I am using Shiller’s data).
    On the other hand, the “historical average” 1990-2013 is 25.3; the “historical average” 2000-2013 is *again* 25.3.
    Therefore latest ‘crashes’ never brought us back to where we started, and, strange as it may seem at first, in the last 25 years the “historical average” has remained considerably stable over time. Moreover, the noticeable increase in volatility of the last 25 years (std dev total sample: 6.6; 1880-1990: 4.6; 1990-2014: 6.9) could also be an indication either of a trending process or of a one-time permanent shock.

    All of this begs the questions: are we in bubble-time or is this just an above-the-trend movement? Which brings us to a second consideration: yes, EMH-totalitarians are stupid and we can all see that, but how much predictive power does CAPE have? Is Shiller right to infer anything at all from this graph?

    • Major.Freedom says:

      Bob isn’t saying CAPE is right, he is just saying that Sumner’s critique of it is wrong.

      I think Bob would say that both CAPE and EMH are not all they’re cracked up to be. Both are flawed.

  4. Maurizio says:

    Bob, I recently asked Sumner: suppose you come to believe that, in the absence of further QEs, stocks must _necessarily_ crash. In other words, that stocks cannot hold their current value, _unless_ there is a sequence of QEs, each larger than the previous one. At least in this case, would you then admit that stocks are in a bubble? And he said “no, it depends on why stocks crashed. Suppose they crashed because of falling NGDP expectations”. Isn’t this delightful?

  5. H. Kazemi says:

    There seems to be a misunderstanding as to what EMH means. It does not mean that the market can predict the future, that prices are accurate and that bubbles do not exist. Please go back to the source (Fama) and do not rely on what is being repeated in the media.

    Ignoring the data mining that is taking place with using CAPE, let’s devise a trading strategy using CAPE that would beat buy and hold on a risk adjusted basis. For example, if you go back and say that you will be long the market whenever CAPE is was below its mean known at that time and short the market whenever it was above it’s known mean, your performance would have been terrible. So CAPE does contain any useful information as far as this simple test is concerned. Once you come up with a credible evidence to show that markets are ignoring some useful information on a consistent basis, then you can question EMH.

  6. Daniel Kuehn says:

    The same is very much true of market monetarism. If it doesn’t work, it just means it was never tried in the first place. There is no world in which it could be tried and fail.

    You get frustrated with Keynesians for saying “it just needed to be bigger”, but at least our counterfactuals are generated from impact estimates in past data. Sure we’re copying and pasting those impact estimates to this crisis (with varying degrees of sophistication), but one thing we are definitely not doing is defining away the fact that we have to have a plausible counterfactual.

    • Daniel Kuehn says:

      I always have to add, of course, that you guys are equally frustrating for always saying “austerity just needed to be smaller”. At least we will whip out a multiplier estimate every once in a while when we say that 😉

      • Major.Freedom says:

        Cuz it’s better to BS than to admit ignorance.

    • Bob Roddis says:

      Your analogy is seriously flawed. Violent government intervention is based upon the unfounded (and totally bogus) assertion that the free market of voluntary peaceful exchange somehow fails and requires violent government intervention to fix the alleged failure. David Stockman has again shown that the Great Depression was nothing more than a long festering problem caused by violent government intervention in WWI. Indeed, your own paper is based upon a similar claim about the cause of the 1920 depression:

      2. The austerity depression of 1920–21

      During WorldWar I federal expenditures ballooned and although the new income tax was able to partially finance the war effort, most of the financing was done through federal borrowing and by the highly accommodating monetary policy of the Federal Reserve. The role of the Federal Reserve at this time was expressed unambiguously by the New York Federal Reserve Bank Governor Benjamin Strong, who told a Congressional committee in 1921 that ‘I feel that I, or the bank at least, was their [the Treasury’s] agent and servant in those matters’ and further added that the wartime inflation caused by the low interest rates maintained by the bank were ‘inevitable, unescapable, and necessary’ for prosecuting the war (Strong, 1930).

      http://factsandotherstubbornthings.blogspot.com/2013/01/poor-kid-was-just-couple-years-too.html?showComment=1359067576135#c3828640123045664561

      It would seem that you should be required to prove beyond any reasonable doubt that a) the market fails; and b) violent intervention will solve the alleged market failure.

      Since you cannot show that the market fails (and since non-violence is, in fact, the societal norm), you are left with making a bad joke about the alleged equivalence of your side calling for more economic violence by the government and our side demanding less or no violence at all. There is no such equivalence.

      • Philippe says:

        Could you please explain how spending money, or printing money, or reducing taxation is a violent action.

        • Philippe says:

          “Since you cannot show that the market fails”

          According to your definition, market ‘success’ is whatever occurs in a market. You define failure out of existence.

          “our side demanding less or no violence at all”

          That’s funny. I thought ancaps were not opposed to violence.

          • Richie says:

            “According to your definition, market ‘success’ is whatever occurs in a market. You define failure out of existence.”

            “Market failure” means that people like you are not pleased with the results of what economic agents agree to voluntarily, so you feel the need to “fix” the “problem” with threats of violence.

            “That’s funny. I thought ancaps were not opposed to violence.”

            Not all libertarians are an-caps, but I don’t expect you to know that.

            • Philippe says:

              so you agree with my description of your definition of market ‘success’: it is whatever happens in a market. A market cannot fail by definition, according to you. Right?

              “Not all libertarians are an-caps, but I don’t expect you to know that.”

              That is not relevant. Ancaps are not opposed to violence.

              • Bob Roddis says:

                The essence of both Marxist and Keynesian critiques of the market is that a free market leads to permanent levels of either poverty or unemployment. There is no theoretical or evidentiary basis for that assertion. Further, the success of the Keynesians has been based upon their use of social intimidation against their opponents and their complete refusal to engage the alternative theory of the Great Depression which is that it was an after-effect of the interventions of WWI (which cannot be blamed upon “the market”).

              • Philippe says:

                “The essence of both Marxist and Keynesian critiques of the market is that a free market leads to permanent levels of either poverty or unemployment.”

                Not so. First of all, a marxist would not refer to a capitalist market as being ‘free’.

                Keynesians of different kinds would tend to disagree among themselves on whether market economies have a “long-run” tendency towards full employment and general prosperity, whilst agreeing that market economies can fall seriously short in the “short run”. There is a related debate about the extent of ‘short-run’ effects on ‘long-run’ outcomes, with some arguing that the ‘long-run’ is just a series of short runs.

              • Philippe says:

                also, Keynesians don’t see the market economy and the public sector or state as totally separate things.. rather they are part of the same thing, which is society with its various institutions.

              • Major.Freedom says:

                Philippe:

                It doesn’t matter to Roddis’ point that Marxists would not use the term “free” to describe markets. You’re not engaging the substance of the argument, which is that Marxists claim permanent poverty for the working class in capitalism, which indeed does not have any empirical support. You’re just dodging this.

              • Philippe says:

                “Marxists claim permanent poverty for the working class in capitalism”

                But they don’t claim that such immiseration is the result of a “free” market economy, because they don’t agree that market economies are “free”. Instead they think that the immiseration of the poor is the result of coercive social and legal structures which benefit one class at the expense of another.

              • Philippe says:

                in other words, marxists blame the state, which they see as being a fundamental part of the capitalist society. They don’t consider ‘the market’ and ‘the state’ to be different things.

              • Richie says:

                ” A market cannot fail by definition, according to you. Right?”

                Depends on the definition of “fail.” If by “fail” you mean, I (Philippe) don’t like the outcome of millions of economic agents’ decisions voluntarily agreed to, so I need to use force to get the outcome I (Philippe) desire, then yes, the market “fails.”

              • Philippe says:

                to begin with, your whole ‘voluntary’ rhetoric is bogus.

              • Bob Murphy says:

                OK guys, I don’t want you having your standard argument about aggression on a post about the Efficient Markets Hypothesis. I’ll leave the previous stuff up, but I’m zapping further comments on this topic for this post. You can’t keep having virtually the same argument every post. It drives people away from the comments.

              • Joseph Fetz says:

                I must say that I actually agree with Murphy here. I stopped commenting here because of how the comment trend would always turn out. It’s one thing to have a principled position and then argue it on a post about EMH–even if it isn’t necessarily related, but is instead a response to another comment–it’s quite another when this happens on every post.

                As for EMH, all I have to say that it *is* a hypothesis, after all, not a theory in the true sense of the term. So it is essentially nothing more than a supposition that not only cannot be proved (due to its non-falsifiability), but it also doesn’t have the rigor of deductive reasoning to back it up.

                It’s sort of like saying, “this is what I believe, and you cannot tell me different”. In a way it reminds me of religion in that one cannot prove the negative, but at the same time the hypothesis put forth has no logical and/or empirical underpinnings to give it some meat.

                (Bob, I didn’t say that last part to offend you, I was just using an example. IOW, I think that EMH is more a matter of faith than science. It is one thing to have a hypothesis or faith about the creation of reality and life, because none of us could possibly prove such a thing empirically or logically; we instead have a belief or faith about such a thing. But in science we can prove things and form theories, so that is the point that I was trying to push. Oh and trust me, I’m not attempting to say that the scientific method is the best way to come to economic truths, but words do have meanings.)

          • Bob Roddis says:

            That’s funny. I thought ancaps were not opposed to violence.

            The entire statist defense to the NAP is based upon their alleged refusal to understand the clear difference between a) the initiation of force and violence and b) force and violence used to defend against the initiation of force.

            That’s it. That’s all they’ve got.

            • LK says:

              You can’t realistically expect to have any stable property rights or even law and order without a monopoly on law enforcement and justice.

              Rothbard’s private “justice” system is a joke:

              http://socialdemocracy21stcentury.blogspot.com/2013/06/rothbard-on-private-protection-agencies.html

              • Major.Freedom says:

                LK:

                “You can’t realistically expect to have any stable property rights or even law and order without a monopoly on law enforcement and justice.”

                You can’t realistically expect to have any stable property rights or even law and order without a market on law enforcement and justice.

                Wow, what profound insightful dialogue here.

              • Philippe says:

                ‘a market in justice’ does not make sense as a concept.

            • Philippe says:

              “The entire statist defense to the NAP is based upon etc”

              The ‘non-aggression principle’ referred to by ancaps is generally meaningless and irrelevant without a theory of legitimate property ownership and property rights, which is precisely what people disagree on.

              “NAP and property rights are closely linked, since what aggression is depends on what a person’s rights are. Aggression, for the purposes of NAP, is defined as the initiation or threatening of violence against a person or legitimately owned property of another.”

              http://en.wikipedia.org/wiki/Non-aggression_principle

              • Major.Freedom says:

                Philippe:

                “The ‘non-aggression principle’ referred to by ancaps is generally meaningless and irrelevant without a theory of legitimate property ownership and property rights, which is precisely what people disagree on.”

                Within “citizen ethics”, everyone agrees with almost all of what ancaps say about property rights. Don’t steal, don’t trespass, don’t initiate force against people’s persons, etc.

                The “disagreement” you are referring to is not what constitutes property rights, but whether or not those who are statesmen should abide by a different, indeed opposite set of property rights ethics.

                You say yes, they should.

                Ancaps say everyone should be equal under the law.

              • Philippe says:

                “everyone agrees with almost all of what ancaps say about property rights. Don’t steal, don’t trespass, don’t initiate force against people’s persons, etc.”

                No, because most people don’t think, for example, that taxation is stealing.

              • Major.Freedom says:

                That is not inconsistent with “almost all.”. If that is all you have, you’ve proved my point.

          • Major.Freedom says:

            Philippe:

            “According to your definition, market ‘success’ is whatever occurs in a market. You define failure out of existence.”

            According to your definition, market failure is whatever does not occur in your ideal world. You define market failure *into* existence.

            • Philippe says:

              economists have theories of what constitutes market failure and success.

              You simply state that market failure is impossible because whatever happens in a market is by definition ‘success’, and the best possible outcome.

              Your ideology is purely circular.

        • Major.Freedom says:

          Philippe:

          He is referring to the initiations of violence that enables these actions that seem peaceful when abstracted from the source.

          It would be like exasperatingly saying “Can someone please tell me how a thief merely spending money is a violent action? Seems to me it’s just an exchange. And can someone please tell me how the central bank merely hitting CTRL-P using the counterfeiting operation constitutes aggression?”

          You have to dig deeper than the top layer of atoms.

          • Philippe says:

            what you are saying is that you think taxation and other forms of existing law are illegitimate.

            So when Bob Roddis says that government spending money is “violence”, what he means is that he thinks the state, government and existing law are illegitimate and should not exist.

            As usual with you ancaps, all you do is use emotive and deceptive word games to express your deeply unpopular political opinions.

      • LK says:

        “Your analogy is seriously flawed. Violent government intervention is based upon the unfounded (and totally bogus) assertion that the free market of voluntary peaceful exchange somehow fails”

        You a “free market” based on unworkable Rothbardian anarchism and private justice? Where everyone or virtually everyone magically respects the law and property rights? lol..

        The empirical evidence of real world states where governments collapse shows us that Rothbard’s fantasy libertarian world doesn’t emerge. Chaos happens.

        This supports the inductive argument that it is unrealistic to expect stable property rights and law and order without a monopoly on enforcement of law and justice.

        Therefore the “free market” roddis is talking about cannot even be taken seriously because it is unlikely it would have any stable property rights in the first place.

        • LK says:

          Correction:

          You **mean** a “free market” based on unworkable Rothbardian anarchism and private justice?

          • Philippe says:

            there’s no such thing as ‘private justice’, though there are different beliefs about what justice is.

            • Philippe says:

              I think what you are referring to is the concept of ‘private law’.

            • Major.Freedom says:

              Philippe:

              there’s no such thing as ‘public’ justice’, though there are different beliefs about what justice is.

          • Bob Roddis says:

            Same old same old same old dodge of the issue of the Great Depression having been the aftermath of violent government intervention in WWI. Zzzzzzzzzz.

            You’ve lost. Give up.

            • Philippe says:

              economic expert bob roddis knows exactly what caused the great depression.

              • Major.Freedom says:

                Economic expert Philippe knows caused the Great Depression.

              • Philippe says:

                my nose caused the great depression?

        • Major.Freedom says:

          LK:

          “You [mean] a “free market” based on unworkable Rothbardian anarchism and private justice? Where everyone or virtually everyone magically respects the law and property rights? lol.”

          You [mean] a “social democracy” based on unworkable violent intervention and monopoly justice where the state is its own arbitrator? Where everyone or virtually everyone magically rejects property rights? lol.

          “The empirical evidence of real world states where governments collapse shows us that Rothbard’s fantasy libertarian world doesn’t emerge. Chaos happens.”

          The empirical evidence of real world states where governments exist shows us that your fantasy world doesn’t emerge. Oppression and impoverishment relative to the alternative happens.

          “This supports the inductive argument that it is unrealistic to expect stable property rights and law and order without a monopoly on enforcement of law and justice.”

          This supports the inductive argument that it is unrealistic to expect stable property rights and law and order without a market on enforcement of law and justice.

          “Therefore the “free market” roddis is talking about cannot even be taken seriously because it is unlikely it would have any stable property rights in the first place.”

          Therefore the “socialist market” you are talking about cannot even be taken seriously because it is unlikely it would have any stable property rights in the first place.

        • razer says:

          Please give me an example of market failure at the micro level? Which completely voluntary action between two participants do you think causes this failure? Please don’t hide behind abstract macro jargon. Show me how voluntary participation among acting agents is a failure. Maybe then I can accept the other premise you introduce — that your agents that can spot this failure have the insight to correct it. There, the burden of proof is on you. Show me this failure. Then we’ll get to the part about proving clairvoyance that your assumptions imply.

    • Major.Freedom says:

      “You get frustrated with Keynesians for saying “it just needed to be bigger”, but at least our counterfactuals are generated from impact estimates in past data.”

      But those past estimates of past data are themselves based on counter-factuals. And those counter-factuals are based on past estimates of past data that are also based on counter-factuals.

      Counter-factuals support the entire theory of Keynesianism. That counter-factual is of course philosophical.

  7. Bob Murphy says:

    I’m repeating it down here in case others missed it: The point of this post is the EMH versus CAPE. I’m deleting further comments about the nature of anarcho-capitalism etc. I don’t want to keep having the same arguments in every post.

    • Major.Freedom says:

      Oops, missed it.

  8. Benson says:

    Hussman Fund’s William Hester deals with the various critics of CAPE (spoiler no EMH). “Does the CAPE Still Work? ”
    http://www.hussmanfunds.com/rsi/cape.htm

  9. Jeff says:

    I suppose the first question is “what should CAPE be?” There’s no theory that states CAPE should be such-and-so value. The historical long-run average is one, but only one, basis for comparison. It’s not clear that it’s a good one, although it does fit within the 10-20 value range that many practitioners use as a rule of thumb of reasonable P/E values.

    There are reasons to believe CAPE higher than 16 is fine. For example, if earnings are less volatile than the historical average, a higher CAPE makes sense because the risk of holdings equities has gone down and thus prices are higher, even though the actual level of earnings may not have changed. Higher expected growth of earnings would have a similar effect.

    The problem, from an EMH perspective, is that there’s no model for what P/E should be. We have pricing models, and so forth, but it’s impossible to know which are correct or not. This is actually part of what’s known as the “Roll critique” in finance. Essentially, when we go to test market efficiency, we need a model for asset pricing. So every test of efficiency is a joint test: that whichever market you are testing is efficient, and that you have the correct model. So when the test fails, as they normally do, we don’t know why (inefficient market or wrong model?)

    In the current understanding of EMH, finance researchers have subdivided efficiency into three forms: (1) weak, (2) semi-strong, and (3) strong. The strength refers to what information is already impounded into the price of the asset. The way finance researchers understand efficiency is that if one can earn abnormal returns (returns in excess of the risk-adjusted expected return on the asset) using some specific information, then that information is not priced into the asset.

    Weak-form includes all historical price information. Thus if one can use historical price data to systematically earn abnormally high returns in a market, or on an individual stock, then that market would not be weak-form efficient.

    Semi-strong form includes all publicly available information. So if something is known by non-insiders, it is priced. This means certain anomalies, like the small-firm or value effect, should not exist. When we found those anomalies, we have a problem: is the market inefficient, or is our model deficient? Much of the asset pricing literature surrounds this issue. In general, finance researchers think it means we have a deficient model, but this is not universal.

    Strong form includes all information, public and private. Thus, if only the CEO knows s/he is retiring tomorrow, the stock price (under this form) would be priced. I can’t think of any finance researcher that thinks any market is strong form efficient, so this only exists as a theoretical category.

    EMH is not a tautology, and it’s not taught as such in finance. I can’t speak for macroeconomics. It’s a great way of organizing our thinking around markets, and gives a basis for research. It makes little sense to say “there’s no rhyme or reason to stock pricing, so let’s through away all our tools.”

    We also (finance researchers) don’t think efficiency is binary. It’s a spectrum: the more trading (liquidity) that exists in a market, the more efficient it’s going to be. But surprises happen, and one measure of efficiency is how quickly prices more in response to the surprise.

    Finally, finance researchers do not think the past five (or so) years are trustworthy when one thinks of asset pricing. The markets are so manipulated and there are so many new regulations, it’s not clear to what markets are responding. We actually have debates as to whether or not the data from these years should be included.

    • Major.Freedom says:

      The evidence is that the market is weak form efficient.

      This gels with the rationalist position that we humans ADD a component to all market activity and therefore even perfect knowledge of past events would not be able to provide sufficient information to be able to find a way to permanently beat the market.

      • Jeff says:

        Well I would think that without humans there would be no market activity… but I think your underlying point is correct: the act of trading is, itself, information.

        And the stock markets in the developed world tend to be semi-strong efficient too; a lot of anomalies, once discovered, disappear quickly because people arbitrage them away.

  10. Major.Freedom says:

    We should be careful not to put weight on the CAPE simply because it currently gels with the Austrian notion that we’re in a bubble.

    • Bob Murphy says:

      Right MF, my point isn’t, “Hey everybody, calibrate your portfolio based on CAPE.” My main point was that it seems to have plausibly served well in the sense of “warning” of market crashes, and yet Sumner is saying it has failed miserably.

      • Major.Freedom says:

        I feel kind of bad, because I didn’t mean to suggest that that is what your motivations were…

        It was more of a…well, I guess a way of communicating to everyone else that your argument is in fact what you just explained above…to say “We’re aware of potential bias, and so don’t mistake what Murphy said as just clutching at something that reinforces existing convictions.”

        I’m awkward sometimes….

        • Bob Murphy says:

          No problem MF I didn’t take offense. You didn’t say my post was rubbish!

Leave a Reply to Bob Roddis

Cancel Reply