Refuted economic doctrines #1: The efficient markets hypothesis

I’m starting my long-promised series of posts on economic doctrines and policy proposals that have been refuted or rendered obsolete by the financial crisis. There will be a bit of repetition of material I’ve already posted and I’ll probably edit the posts in response to points raised in discussion.

Number One on the list is a topic I’ve covered plenty of times before (in fact, I was writing about it fifteen years ago

), the efficient (financial) markets hypothesis. It’s going first because it is really the central microeconomic issue in a wide range of policy debates that will (I hope) be covered later in this series. Broadly speaking, the efficient markets hypothesis says that the prices generated by financial markets represent the best possible estimate of the values of the underlying assets.

The hypothesis comes in three forms.

The weak version (which stands up well, though not perfectly, to empirical testing) says that it is impossible to predict future movements in asset prices on the basis of past movements, in the manner supposedly done by sharemarket chartists. While most of what is described by chartists as ‘technical analysis’ is mere mumbo-jumbo, there is some evidence of longer-term reversion to mean values that may violate the weak form of the EMH.

The strong version, which gained some credence during the financial bubble era says that asset prices represent the best possible estimate taking account of all information, both public and private. It was this claim that lay behind the proposal for ‘terrorism futures’ put forward, and quickly abandoned a couple of years ago. It seems unlikely that strong-form EMH is going to be taken seriously in the foreseeable future, given the magnitude of asset pricing failures revealed by the crisis.

For most policy issues, the important issue is the “semi-strong” version which says that asset prices are at least as good as any estimate that can be made on the basis of publicly available information. It follows, in the absence of distorting taxes or other market failures that the best way to allocate scarce capital and other resources is to seek to maximise the market value of the associated assets. Another way of presenting the semi-strong EMH is to say whether or not markets are perfectly efficient, they’re better than any other possible capital allocation method, or at least, better than any practically feasible alternative.

The hypothesis can be tested in various ways. First, it is possible to undertake econometric tests of its predictions. Most obviously, the weak form of the hypothesis precludes the existence of predictable patterns in asset prices (unless predictability is so low that transactions costs exceed the profits that could be gained by trading on them). This test is generally passed. On the other hand, a number of studies have suggested that the volatility of asset prices is greater than is predicted by semi-strong and strong forms of the hypothesis (note to readers – can anyone recommend a good literature survey on this point).

While econometric tests can be given a rigorous justification, they are rarely conclusive, since it is usually possible to get somewhat different results with a different specification or a different data set. Most people are more likely to form their views on the EMH on the basis of beliefs about the presence or absence of ‘bubbles’ in asset prices, that is, periods in which prices move steadily further and further away from underlying values. For those who still believed the EMH, the recent crisis should have shaken their faith greatly. But, although the consequences were less severe, the dotcom bubble of the late 1990s was, to my mind, are more clear-cut and convincing example of an asset price bubble. Anyone could see, and many said, that this was a bubble, but those, like George Soros, who tried to profit by shortselling lost their money when the bubble lasted longer than expected (perhaps long-dated put options would have provided a safer way to bet on an eventual bursting of the bubble, but Soros didn’t try this, and neither did I.)

More important than asset markets themselves is their role in the allocation of investment. As Keynes said in his General Theory of Employment Interest and Money, this job is unlikely to be well done when it is a by-product of the activities of a casino. So, if the superficial resemblance of asset markets to gigantic casinos reflects reality, we would expect to see distortions in patterns of savings and investment. The dotcom bubble provides a good example, with around a trillion dollars of investment capital being poured into speculative investments. Some of this was totally dissipated, while much of the remainder was used in a massive, and premature, expansion of the capacity of optical fibre networks (the fraudulent claims of Worldcom played a big role here). Eventually, most of this “dark fibre” bandwidth was taken up, but in investment allocation timing is just as important as project selection.

The dotcom bubble was just one component of a massive asset price bubble that began in the early 1990s and is only now coming to an end. Throughout this period, patterns of savings and investment made little sense. Household savings plunged to zero and below in a number of developed countries (including nearly all English-speaking countries) and the resulting current account deficits were met by borrowing from rapidly growing poor countries like China (standard economics would suggest that capital flows should go in the other direction). The massive growth of the financial sector itself, which accounted for nearly half of all corporate profits by the end of the bubble, diverted physical and particularly human capital from the production of goods and services.

Finally, it is useful to look at the actual operations of the financial sector. Even the strongest advocates of the EMH would not seek to apply it to, say, the Albanian financial sector in the 1990s, which was little more than a series of Ponzi schemes. They would however want to argue that the massively sophisticated global financial markets of today, with the multiple safeguards of domestic and international financial regulation, private sector ratings agencies and the teams of analysts employed by Wall Street investment banks is not susceptible to such systemic problems, and is capable of correcting them quickly as they arise, without any need for large-scale and intrusive government intervention. I’ll leave it to readers to make their own judgements (maybe with some links when I get around to it).

Once the EMH is abandoned, it seems likely that markets will do better than governments in planning investments in some cases (those where a good judgement of consumer demand is important, for example) and worse in others (those requiring long-term planning, for example). The logical implication is that a mixed economy will outperform both central planning and laissez faire, as was indeed the experience of the 20th century. More to follow!

118 thoughts on “Refuted economic doctrines #1: The efficient markets hypothesis

  1. Gene, you appear to be a candidate for a trade I proposed some time ago.

    I will agree that
    (a) the US is not a free-market economy, and its failures do not constitute evidence against the claim that a pure free-market economy is the best possible form of social organization
    (b) no other actually existing society is, or has ever been, a free-market economy, and no actual or conceivable events anywhere constitute evidence against the claim that a pure free-market economy is the best possible form of social organization
    (c) In discussion with parties to the agreement, I will not contest the claim that a pure free-market economy is the best possible form of social organization

    All I ask in return is that the counterparties to the deal agree not to advocate, oppose, criticise, or comment on any policy or political position that might actually be implemented, to ensure that the purity of the free-market ideal is not compromised by actual experience.

  2. Gerard and Steve,

    I agree with what you are saying.

    The EMH hypothesis is rejected just about everywhere we look. The focus of economists on price is also puzzling.

    There are so many more interesting questions to ask and to experiment with in markets than trying to justify an unjustifiable EMH.

    The scary part is that the Rudd fiscal stimulus announced yesterday appears to be based on the idea that governments can regulate unstable markets through monetary and fiscal intervention. Working on the causes of internal instability would seem to be a more effective cheaper approach.

  3. Gene,

    I think I am probably not on your team politically, but I get a kick out of being objective. It is the liberal in me, not your Aussie Liberal, but the real upside right liberal liberal.

    You are correct to point out that there was no pure laissez faire economy in the 20th century, against which to test the hypothesis asserted by purfessor. (You ought not end a sentence with a preposition.) The reason, I think, is that the laissez faire approach caused a global Depression pretty early in that century which prevented a full run of the experiment you would have conducted.

    So I have to agree with you that purfessor’s case against laissez faire is much more speculative than empirical. His reference to evidence is bogus.

    This reminds me of a news story in The Onion, America’s finest newspaper, which mentions a group of physicists who want to crash a planet into earth to test some of their hypotheses about particle physics. According to the article, only a collision of planet-size heavenly bodies would generate sufficient kinetic energy to test some of mechanics that interest them. In order to observe the effect up close, they propose earth be one of those bodies. They accuse people who disagree with them of being unempirical, a criticism I have some sympathy with.

  4. Steve van Emmerik,

    Actually I would suggest that bubbles are not truly in conflict with anything but the strong form of the EMH, and even there it depends on how you want to define your terms.

    I mean imagine tech stock prices start rising and everyone becomes convinced that not only will the continue to rise indefinatly but that others will continue to believe they will rise indefinately. Even though the prices on those stockes might sprial off towards infinity this doesn’t mean any investor has a strategy to do better than the average nor that any of them where behaving irrationally.

    In other words there may be a Nash equilibrium in bidding up the bubble if people have the right kind of utility functions. Sure, everyone knows it is going to crash but no one knows when and the asymptotically increasing valuations always make it worthwhile to keep playing the upside. If things really work out this way then the bubble can happen without anyone having any kind of arbitrage strategy even if they know it is going on.

    I don’t think the EMH is particularly useful in understanding this sort of more applied question anyway. The EMH is a wonderful theoretical device roughly analagous to the way a physicist might assume a person is a cylinder of water to do some calculation. It’s unfortunate that what is primarily an elegant way to think about certain ideal cases has become used so much in practical politics.

  5. I definitely think Geralds on to something there at 96. EMH reminds me on the best predictor of a shares price tomorrow is the shares price today… and that they follow a random walk (like a drunk in paddock!).
    What on earth good is that as a theory?. I thought it then and I still think the same now.

  6. One of the problems of economic theories taking over in trasury departments, is that you can almost guarantee the remedy will be taken too far (much further than the economist may have ever prescribed). McFarlane noted in his short book ‘in search of stability” that by the 1960s any slight deviation from full employment was given a good shot of Keynesian fiscal policy. Well, more intervention than Keynes, who unfortunately died in 1946, would ever have prescribed. Alas free markets, freedom to choose, and focus on monetary policy as a cure all appears to have gone much the same way. We need a “there is no economic silver bullet” theory.

  7. “Once the EMH is abandoned, it seems likely that markets will do better than governments in planning investments in some cases (those where a good judgement of consumer demand is important, for example) and worse in others (those requiring long-term planning, for example).”

    Wow, what an extremely reasonable and sane conclusion.

  8. I find this “refutation” completely unpersuasive. The existence of bubbles does not refute the semi-strong EMH, unless one can produce a prediction methodology that reliably foresees them based on publicly available information. There are several trite metaphors interspersed with the argument that reveal bad language (and bad thinking), e.g. the likening of the stock market to a casino.

    For a dissenting opinion check out Scott Sumner’s entry on his blog:

  9. I think you mischarachterize the semi-strong form (and even the strong form) of the efficent market hypothesis.

    These hypothesises are about the existance of strategies based on various sorts of information to outperform the market in general. However, what you seem to be looking at is whether things are predictable in the intuitive sense of the word, in particular the idea that one shouldn’t be able to predict a crash. However, one should certainly be able to have a model where everyone agrees that a certain stock will have value 0 at the end of the year but yet everyone’s expectation is that the stock price will increase so fast up till the point it crashes that the current price always reflects the expected future value of the stock.

    In other words everyone knows it’s a bubble and knows the bubble will certainly be over by the end of the year (a prediction) and even agree on the probability distribution for when the crash is but yet the bubble price is such a fast growing function that the gamble on getting out before the bust has sufficiently high expectation to justify getting out before the bust.

    Secondly I’m a bit troubled by the confusing language you use suggesting the EMH is about getting the ‘value’ of the underlying asset correct. Once again EMH is about the value only in the sense of expected future prices not about some underlying notion of what the real value of the thing is truly. Once again just because the prices spiral out of proportion to the intrinsic worth/usefullness of the asset doesn’t mean the expectations don’t work out the correct way.

    Finally, on a broader point I’m troubled by the sorts of empirical demands you are putting on a claim that isn’t truly a direct empirical statement. I mean obviously phrased the way you do all the versions of the EMH are false. Surely
    some sufficiently powerful alien computer or deity could take enough public information, the true laws of physics and enough computational power to derive better estimates of the stock market. However, it would be absurd to say this discredits the EMH.

    The EMH isn’t really an empirical claim the way quantum field theory is, but rather the claim that a certain kind of economic model is a good approximation to use when trying to infer things about real markets. However, I think on this notion you need to be a lot more careful about the way in which the approximation is supposed to be good.

  10. some technical notes:

    “the weak form of the hypothesis precludes the existence of predictable patterns in asset prices”

    Not strictly true. Firstly, it is asset _returns_ (price movements plus dividends/coupons) that the EMH applies to; thus you get around 100% predictable stock price movements when stocks go ex-dividend.

    Secondly, those returns are defined as excess returns over the risk free interest rate. Ignore that, and you get serial correlation in the returns due to serial correlation in the interest rates used to discount cashflows.

    And Thirdly, even those excess returns in theory and practice have a predictable component. This is because volality varies through time in a semi-predictable manner, and this variation induces corresponding changes in risk premia and hence asset returns, which are predictable using lagged volatility (though this point is little known outside the inner sanctums of finance theory).

    – Tim

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s