Valuation-Informed Indexing #27: Not One Study Supports the Claim That the Market Is Efficient

Updated on

by Rob Bennett

Not one study supports the claim that the market is efficient.

Not one.

Millions of people believe that there are studies supporting this claim. I don’t dispute that for a second. But people believing that there are studies and there actually being studies are two very different things. It’s important that those trying to understand how stock investing works keep the distinction in mind.

To understand how it is that so many smart and good people have come to believe that there are studies showing that the market is efficient when in fact there are none, you need to appreciate the historical moment in which the efficient market hypothesis was put forward. We have not been trying to learn about stock investing through academic studies for centuries now. This practice is something new. My understanding is that it did not become common practice to study stock investing scientifically until the 1960s.

So we’ve been at this for about fifty years. That’s hardly enough time in which to arrive at conclusive findings on any aspect of the subject being studied. Fifty years is barely enough time to develop tentative theories regarding how stock investing works. We’re still in the early stages of our long journey of discovery. We are still building the foundations of the building.

Unfortunately, we don’t enjoy the luxury of being able to hold off on making investing decisions for another 100 or 200 years, at which time humankind’s understanding of the subject matter will be much deeper and more nuanced. We all need to make investing decisions today. So we all feel a temptation to jump to conclusions about investment research, to put new findings to practical use years or even decades before the research has been subject to the scrutiny and testing and challenge needed for it to earn the designation “science.” This is what happened with the research that many to this day believe presents findings showing that the stock market is efficient.

In the 1960s, when University of Chicago Professor Eugene Fama was doing research that he believes shows that the market is efficient, it was not common for a distinction to be made between short-term timing (changing one’s stock allocation with the expectation of seeing a benefit from having done so within a year or two) and long-term timing (changing one’s stock allocation in response to big valuation shifts with the understanding that it make be 10 years before the change pays off). Benjamin Graham was aware of the importance of the distinction in the 1930s. But I think it would be fair to say that Graham was many decades ahead of the conventional investing wisdom in most of his thinking on the subject. To just about everyone else, the word “timing” and the phrase “short-term timing” were synonymous.

That’s not been true for 30 years now. In 1981, Yale Economics Professor Robert Shiller showed that, contrary to the claims of those who believe in the Efficient Market Hypothesis, long-term timing always works (or at least it always has for as far back as we have records of stock performance). This finding stands the conventional wisdom on its head. If long-term timing works, everything we learned about how stock investing works during the Buy-and-Hold Era (in which it was claimed that, since timing doesn’t work, there is no need for investors to lower their stock allocations when prices reach insanely dangerous levels) is wrong. A distinction that few in the field were even aware existed at the time that research said to support the Efficient Market Hypothesis was done has become of major importance in the new environment created by publication of Shiller’s amazing findings.

For the 30 years since Shiller’s research was published, we have seen a strange reluctance by most in the field to explore its implications. Why? A big part of the explanation is that many believe in the Efficient Market Hypothesis. That’s science, isn’t it? There are studies. The studies are backed by data, the hard stuff. No?


There is no data showing that the market is efficient. There is data showing that short-term timing doesn’t work. The idea that the market is efficient does not necessarily follow from any of the studies. The idea that the market is efficient is an attempted explanation of the data, not itself a finding. The two are not the same thing. A research finding can be said to be the product of science. An attempted explanation of a research finding can never be anything more than subjective opinion until it is tested in follow-up research.

Short-term timing doesn’t work. That’s a research finding. The idea that the market is efficient is (or at least once was) a plausible explanation of this finding. An efficient market is one in which the price is always right or at least as close to right as it is possible to get (an “efficient” market is one in which all factors known to have a bearing on price have been taken into consideration by the investors setting the price). If the market price is always right, it is not possible for any investor to gain an edge by analysis of the factors affecting price. Any insight he might develop was taken into consideration by the market before he developed it. It is only by gaining an edge that investors can hope to engage in effective market timing. The idea that the market is efficient does indeed explain why short-term timing doesn’t work.

But it is not the only possible explanation.

An alternative explanation is that the market is totally inefficient, that prices are driven almost exclusively by investor emotion and thus are almost entirely irrational and random and essentially meaningless. For timing to work, effective prediction of prices must be possible. The actions of a crazy man are highly unpredictable. If reason plays little role in setting stock prices in the short term, short-term timing would not work.

When a scientist is faced with two alternative explanations for a phenomenon, he tests them to see which one the data supports. The way to test whether the explanation for why short-term timing doesn’t work is that the market is highly efficient or highly inefficient is to determine whether overvaluation and undervaluation are meaningful concepts. In an efficient market, there cannot be overvaluation or undervaluation. An efficient market is by definition a properly priced market.

This is just the test that Shiller did in 1981 (and which has been repeated many times in the three decades since). He found that both overvaluation and undervaluation effectively predict long-term returns. Both overvaluation and undervaluation are meaningful concepts. The reason why short-term timing doesn’t work is that the market is highly inefficient and almost entirely controlled by emotion in the short term. There is no data-based reason to believe that the market is efficient.

There is no non-data-based reason to believe that the market is efficient either. The price paid for an asset affects the value proposition obtained by it in purchases of every other asset that exists on Planet Earth. The only reason why anyone ever believed that it worked differently with stocks is that Fama’s attempted explanation of the data (not the data itself!) argued that the market is efficient. Now that the attempted explanation has failed its test, we are all free to return to having confidence in our commonsense belief that the price we pay for stocks affects the long-term return we obtain from them.

There is today no reason to believe in the Efficient Market Hypothesis.

Rob Bennett developed a unique portfolio allocation calculator. His bio is here.

Leave a Comment