The wit and wisdom of Eugene Fama

…in hindsight, every price is wrong.

That’s Eugene Fama, from Fama vs. Thaler.  I enjoyed this part too:

Twenty years ago my criticism of behavioral finance was that it is really just a branch of efficient markets, because all they do is complain about the efficient-markets model. I’m probably the most important behavioral-finance person, because without me and the efficient-markets model, there is no behavioral finance. I still think there is no full-blown testable behavioral asset-pricing model.

I have more sympathies for behavioral finance than that, still the dialogue is worth reading in its entirety.  Judged as a debate, Thaler loses.

Hat tip goes to Allison Schrager.


Whether or not it's true, the efficient market hypothesis is one of the best aids to clarity of thought. It should be studied by philosophers.

It should be studied by theologians.

Are you a financial advisor or wealth manager?

Efficient markets may in reality be as rare as perfect competition, but it is the baseline assumption for every active fund manager. You can't profit off of exceptions to a rule if there isn't a rule. Too many inefficiencies are not exploitable because of transaction costs, noise, and speed of adjustment. I've made my living off of recommending passive investment. I have no desire to be and don't envy the profits of hedge fund managers.

EMH is useful.

The problem is a political one when charlatans claim the market is actually fully efficient, to justify how things are now; failing to ask: How can we make them more like the efficient ideal?

To understand the limits of the EMH one has to understand the original framing of the hypothesis. It was observed that prices followed a random seeming walk (fat tails and Mandelbrot came later). It was observed that investing was difficult. Pricing was opaque. Similar companies traded at different P/Es, etc.

So enter the EMH. Whatever the prices were, they must be a best guess. Beyond that they must represent a rational synthesis of news and conditions.

Then it was asked, what is the proof of this hypothesis? The original observations were repeated.

If you think a hypotheses can be proved by previous observations you might need some refresher in basic logic. We could have started instead with an IMH. We could have said that prices reflect a best, but probably bad, guess. We could have said that prices were always some opaque combination of rational analysts and emotional response.

We didn't though. It took decades to come to this point, where Tyler says Fama wins the debate, by making the EMH into an IMH.

Clever dude, but it rankles that it buries decades of error.


Great philosophy of science but weak knowledge of finance: EMH has numerous consequences in asset pricing & corporate finance beyond 1960s formulation (Shiller's initial work debunking the EMH in the 1980s relies on these consequences). Also, as you may be suggesting...EMH not wedded to any particular physical distribution of asset prices.

If all you're saying is that Fama's original 1965, and 1969 formulations of the hypothesis were either tautological or incorrect, then I agree. Roy (1973), and Lucas (1978) demonstrate this conclusively and Fama's "corrected" (1976) version of the hypothesis is ad hoc and yields no empirical test afaik.

rando, your second paragraph is my complaint, yes.

There are many on this page who do not get the tautology even now, nor how the EMH survived by becoming something else.

In this new Fama discussion the EMH becomes Behavioral, something it never was before, and Fama wins by jumpung out there and incorporating all of Shiller, leaving Shiller with nothing to add.


Entirely fair but this is a weakened version of the efficient markets hypothesis. Stronger forms can (mis)lead one into backing predictions about the future out of asset prices and that is where the real problem is. If all Eugene Fama ever said about finance is that you should invest in index funds, there wouldn't be much of a debate.

It's very useful, so long as you are on the lookout for the exceptions that will bite you in the butt. Of which there are many.

I don't think philosophers would find it very interesting. They would say "hey, but there are thousands upon thousands of easily demonstrated reasons to understand that it is incorrect", and then quibble about these numerous faults endlessly. I don't imagine you've actually accomplish the philosophical meditation on the human experience of efficient markets, or any such thing, in the sense of them becoming additionally sympathetic to the arguments.

In short, the effect would probably be opposite to desired.

Possible exception to the rule (if you're correct): philosophy of science professors. Many love complexity/chaos theory, and can happily talk all day about the degree of (in)efficiency of any complex adaptive system.

It's not a yes/no question. I nominate EMH is the easiest thing to understand that is yet widely misnderstood.

Do you understand what it means? Or do you also make it mean something new, so it can be true even when it is false?

The mid-1960s was a turning point in research on the random character of stock prices. In 1964, Cootner published his collection of papers on that topic, while Fama’s (1965) doctoral dissertation was reproduced, in its entirety, in the Journal of Business. Fama reviews the existing literature on stock price behaviour, examines the distribution and serial dependence of stock market returns, and concludes that “it seems safe to say that this paper has presented strong and voluminous evidence in favour of the random walk hypothesis.”

Do you actually believe that prices are random walks, with large and small variations triggered by large and small news events?

We are going to win so big so big we're going to win!

Philosophers already study it under its other name: the just world fallacy.

Thaler won the "value stocks" round.

Seems to be a lack of clarity on what it means to say a market is "rational." If it doesn't mean the market gets it right then how can rationality even be tested?

A market full of perfectly informed and perfectly "rational" (NPV maximising or utility flows maximising) people will not make a "rational" market. Essentially, it denies many market imperfections (there are many) and assumes that people know everything and can always evaluate exactly what they want among a set of options

I don't think that is what it is about. The price is what all the participants conclude based on the information present. If information is hidden either by design or by circumstance, the price will reflect the information available. As information becomes available, the price will adjust.

If you start with this assumption, then you can find ways to arbitrage information flows in time. Or you can look at for example the complex structured finance products that can be charitably described as opaque and recognize them for what they are; a means to arbitrage lack of information. Eventually everyone figured out what those mortgage backed securities were worth as information became apparent in time.

Also consider that a working market isn't dependent on the participants all being rational or having perfect information. In fact it works in spite of those things. Corrections aren't that the price was wrong; it means new information became available. An efficient market is quick to adjust prices to new information.

Behavioral economics has something to add to this in the sense that it describes the rather perverse way that we as humans assimilate new information. It isn't a straight line. Information becoming available may really mean that the information finally cannot be ignored. I would call this perceived value.

"Corrections aren’t that the price was wrong; it means new information became available."

This is actually a vestige of the original EMH, and one that has been disproved. Or "new information" is changed to mean sometimes bad information, sometimes emotional swings, which really destroys the concept from within.

I think there was a case recently where company X had sold its stake in company Y. Suddenly there was good news for Y, it's stock spiked, but so too did X's. It took a couple days for the market to realize that X no longer held Y. WTF does the EMH mean in that context? That the market "efficiently" made a stupid error?

So you are prescient? You know the future?

Why don't you assume, for the sake of argument, that at every moment in time over those transactions the participants acted rationally based on the information that they possessed. It could have been bad information. It could have been information that contradicted every assumption the participants had and was rejected as noise. It could very well have been someone arbitraging information that they had but no one else had.

So what do you call it? An act of God? Fate?

An efficient market punishes those who don't respond to information. It could be that the information was available, but because of a series of decisions it was ignored. Oops, someone lost money, and suddenly that type of information is now important and not ignored.

I read somewhere that MBA students challenge themselves by making decisions where they have 30-40% of the information required to decide. The article then said that in the real world people are forced to make decisions based on less than 10% of the required information.

In retrospect, all prices are wrong.

To lead with "So you are prescient? You know the future?" is also a EMHer subterfuge.

Did I ever say I did? No, what I actually said was that markets were *more* unpredictable than the original, 1965, EMH implies.

Two more books which address your points. I've read them, they helped:

Against the Gods: The Remarkable Story of Risk: Peter L. Bernstein

The Misbehavior of Markets: A Fractal View of Financial Turbulence: Benoit Mandelbrot

Perhaps this is more on the mark :-)

Fama is one twisty character. Whatever our understanding of markets become, Efficient Market Hypothesis means that.

Hilarious. Zero chance of simple lack of understanding on part of anon.

See above.

Suspicion... CONFIRMED!

Brian, I suspect you are a good old time "there are no bubbles" EMH adherent. No one else on this page is that extreme, which is why you make obscure digs and don't lay out what you know in your heart to be an old, dead, position.

Nope. I just know a lost cause when I see one. Watching you thrash around insde your own head day after day here has become surprisingly amusing.

I love it. I toss in another quarter, you do another little dance.

You use Marginal Revolution as a therapist's couch. You come in, lay on the couch, and start, often making obscure reference to one of hundreds of comments from prior threads that basically summed up the issue du jour, as if the whole MR world is hanging on your every comment.

Good for you. I also comment on the Internet primarily to sort my own thoughts rather than as any quixotic effort to change minds.

Seriously Tyler, if Fama wins it is by yielding everything that was taught as Fama 20 years ago:

The point is not that markets are efficient. They’re not. It’s just a model. The question is, “How inefficient are they?” I tend to give more weight to systematic things like failure to adjust completely to earnings announcements, or momentum, than to anecdotes, which are curiosity items rather than evidence.

It's just a model, so it's right again, even if it's wrong ...

"There’s a difference between anecdotes and evidence, right? I don’t deny that there exist anecdotes where there are problems. For bubbles, I want a systematic way of identifying them. "

I prefer higher grade evidence than anecdote, but that does not mean they have no epistemic value.

Its a model that is useful when we ask: WHY IS IT WRONG?

In my ideal world there is peace and love, goodwill to all and EMH is true. To attain our Marginal Revolution we must learn why we do not have these things.

A good book length treatment on why it is wrong is:

I really liked Thaler's suggestion about sunk costs.

How efficient is the Tokyo housing market? Last week Tabarrok linked to a study which found that, despite a rising population, Tokyo housing prices have been flat, unlike the housing markets in NYC and San Francisco where prices has been escalating. The study suggested that housing prices have been flat in Tokyo because national housing policies in Japan encourage construction of new housing units, unlike in NYC and San Francisco where local zoning and other laws limit construction of new housing. So the housing market in Tokyo is efficient, right? At its peak in the late 1980s, real estate prices in Tokyo were 350 times more expensive than comparable land in Manhattan. 350 times! By 2004, residential real estate in Tokyo was only worth of 10% of its late 1980s peak, while the most expensive land in Tokyo’s Ginza business district had fallen back to just 1% of its 1989 level in the same year. Today, is the housing market in Tokyo efficient or was the bursting of the real estate bubble in Japan so traumatic that the Japanese haven’t recovered from the experience?

Ok, start from the assumption that the markets were efficient, and that the buyers and sellers were doing the rational thing at the time.

What does that tell us? We know that real estate markets are finance dependent; people buy what they can borrow. If I remember correctly, this was at a time when Japanese manufacturing was at it's peak, there was enormous amounts of money coming into the country as a result of the vigorous and successful export market. Where did all that excess cash go? It is obvious in retrospect.

Why did the prices correct? Likely there was a change in the finance flows.

It is very tempting to think that there was a common madness that blinded otherwise smart people to reality. What efficient market modelling tells us is that there are reasons, rational reasons at the time for the actions that in retrospect seem insane.

Up till a month ago there was an extraordinary price increase and volumes in the real estate market in Vancouver. At any level, a small old house on a narrow lot isn't worth $4 million, except that someone was willing to pay that much. Rational? It doesn't seem rational unless you are a Chinese businessman with connections who figures that it is a good time to cash out, and where do you put your cash?

In other words, if Tokyo prices were an instance of irrationality brought about by pet rocks or some SMOD playing with the minds of people, then why did the same thing happen in Vancouver earlier this year? Maybe these are in fact rational decisions by market participants. That is a far more interesting consideration. What mixture of world market conditions, local market conditions, financing practices, etc. created this seemingly irrational event?

Again, the EMH becomes quite threadbare when stretched to cover mis-pricing cycles.

Are you aware of the studies on ponzi schemes? I think there are a few cases where in less regulated countries real ponzis were attempted. Researchers went in, to discover why people invested. To their surprise, it wasn't that people were deceived. They knew these schemes could not hold, they just planned on getting out in time.

Just to be clear, since these were ponzi schemes there was no underlying value investment, it was purely speculation of price *without* value.

So what does an EMH believer do, say that the ponzi investment was rationally priced all the way up the curve, and at the crash too, because at that point the market discovered new information that it was deflating?

" it doesn't seem rational to me and I have a preference for government intervention in markets so my opinion is what it is not because I think about the evidence but because I have an underlying bias"

"It is very tempting to think that there was a common madness that blinded otherwise smart people to reality."

"Maybe these are in fact rational decisions by market participants. That is a far more interesting consideration. What mixture of world market conditions, local market conditions, financing practices, etc. created this seemingly irrational event?"

That is just The Bigger Fool theory. Indeed it is when the madness becomes common that bubbles form. Rinse and repeat.

Isn't part of this because the population is not growing?

Another way to express the efficient market is to say that it is working as designed. The collection of regulation, incentives, tax structures, participant assumptions, etc. are creating an outcome.

The demand for financial/investment assets is upward sloping. Demand rises with price. That's testable. Soros got rich off the idea. You can see it in action in everything from stock prices (down to sectors and individual names) to Chinese houses.

The market is efficient in the sense that it is always a weighted reflection of the current thinking/information in the market. Much of the thinking is not rational.

Any money manager who has to deal with the public ends up doing what the public wants because when you lose money following the herd you can blame the herd, but when you lose money going your own way, you are to blame. So managers tend to be risk averse in a business sense, but end up taking on a lot of risk because they are afraid to move apart from the herd.

Maybe it is smart to follow the herd, because a herd of animals running together is efficient at fending off predator attacks. But sometimes you have the situation which the Native Americans exploited, which is the herd will run off a cliff and kill everyone in it.

The question is not whether markets are efficient or inefficient. The question is how do we make markets more efficient and thus more mutually beneficial for everyone? Fama's original paper ( outlines three models of efficiency- weak form (historical price information only), semi-strong form (historical price information with publicly available fundamentals such as earnings), and strong form (whether firms have access to monopolistic access to information). In the paper he explicitly states that the EMH does not seem to apply to the strong form model. In other words, insiders and specialists may have monopolistic access to information, which would make markets less efficient. This is why EMH says only that prices fully reflect *publicly* available information. EMH does not include private information. In my humble opinion, markets can only be efficient EMH includes private information; i.e. markets can only be efficient if natural language communication between individuals is also efficient.

I think the efficient market hypothesis also includes limits on information, regulatory behavior, etc.

The CUBA fund could probably be explained by both. Us humans take a little time to process information, but the first thing that happens when the president makes an announcement like that on Cuba - people look for ways to invest in CUBA. That being the most efficient and obvious ticker symbol gets bid up. Eventually people look up some data on it and realize the price is overvalued and it falls again. So part of it is incomplete information. No one can immediately arbitrage the spread or shot the fund because of mark-to-market risks. So the first guys who think about shorting it have to wait for momentum to change on it because they really don't know how far that short-time period of punters buying cuba will last. So there are limits to immediate arbitrage.

Also the CUBA fund may have been better set-up to get into cuba fast. If they had a regulatory edge built in and could be the first fund to buy CUBA that could be well worth paying a premium for it. Since it would take a bit of time for western companies to buy up the cheapest assets and get get CUBA valued inline with the amount of demand for investments.

It takes markets time to get the right price. Its not like every investor is constantly logged into their investment account. For most its not the efficient response as their day job pays better than trading.

Does E Harding have a brother? More than 25% of the comments here belong to one guy.

The EMH gets under my skin, sorry. The group mind failed when it accepted the EMH, and it fails still when it does not recognize how post EMH our world really is.

The guy who said that paying $100 to buy four $20 bills is arbitrage won the debate? Really?

If you need to break a hundred and fill up your tank, it could save you $50.

Late the conversation here, but Robert Shiller's Nobel Prize lecture, in which he directly confronts the EMH is excellent:

The interview above of very interesting, but would be far more interesting if every sentence was annotated with an estimated time scale for the example or idea in that sentence being referenced. Otherwise, this conversation can devolve into people talking past each other.
Fama: 'EMH makes sense in this case' [t= 1 sec]
Thaler: 'EMH doesn't make sense in this case' [t = 10 years]
Fama: 'EMH has gotten better in this case' [t1 = 30 days, t2 = 2 minutes]
Thaler: 'EMH was clearly wrong in this case' [t1 = 20 years, t2 = 5 years]

Comments for this post are closed