Reconsidering the efficient market hypothesis

The EMH plays a central role in investment theory and has done so since the seminal paper by Eugene Fama appeared in the Journal of Finance in May 1970. Fama did not claim to be inventing something new. He did what academics are supposed to do – consolidated and advanced prior knowledge.

Fama’s hypothesis appears in all leading textbooks on investment theory and corporate finance and is taught at all leading universities in courses dealing with business finance. Wits University is no exception.

More articles have been published on the EMH over the past 42 years than one cares to recall. So the question whether or not the EMH is flawed is of considerable interest to academics and practitioners alike.

If correctly understood, taught and applied, the EMH is not flawed. To appreciate this it is necessary to restate the theory and understand its context correctly. If this is done then most of the objections recede.

The EMH is a hypothesis, as the name implies, not a fact. More often than not the hypothesis is incorrectly stated and applied. The fundamental hypothesis is that, in an efficient market, prices reflect all relevant information needed to determine the price.

It is important to understand the reason why Fama suggested the EMH in the first place. If that is understood then it becomes clear why the EMH is unlikely to go away. It had been noted, over many years, that share prices seem to reflect information.

This correlation is self-evident and can be seen on any day of the week. Take for example recent events. China announced a downgrade in its forecast of GDP growth and immediately the price of equities in South Africa declined. It is thus clear that prices respond to information.

In this case the reason for the downward movement of SA equity prices can be understood. China is an importer of raw materials and South Africa a supplier. If the demand for raw materials abates, one can expect a decline in South Africa’s economy too.

This decline exhibits as an expectation of lower corporate earnings that immediately translates to lower share prices. If China infers it will import fewer raw materials it is not surprising to observe SA equity prices declining in value. New information changes market prices!

Take another example. America announced recently that its interest rates would remain very low for the foreseeable future. Immediately the rand strengthened. Why should this be so?

Investors with spare cash will look for higher returns than can be gained in the US – and since South Africa has higher interest rates – money flows out of the US into South Africa. The inflow causes the rand to strengthen.

There are times when it appears that new information does not change prices (or if it does, then not in the way which many expect). Occasionally a company announces very good profits and the price of the share remains unchanged or even declines. The problem in this case is that the market expects these profits and the information is already factored into the price.

Information influences market prices daily. It is therefore not surprising that someone argued that prices reflect information. The next logical step was to argue that in an imaginary market where all information is known, prices would reflect all known information! That imaginary market is called the efficient market.

Because the perfect market cannot exist different degrees of efficiency were proposed. Economists therefore reference weak form, semi-strong form and strong form efficient markets in their works.

Many studies have been conducted on stock markets around the world to determine how closely they approximate efficient markets. At Wits we have carried out numerous studies to determine the efficiency of the JSE, for example. Empirical evidence singles out the JSE as a semi-strong efficient market.

It is not possible to abandon the EMH since it is a statement of common sense of something that is observed all the time.

The EMH can be used to explain another phenomenon, the so-called random walk. In 1953 a famous statistician, Sir Maurice Kendall (1907-1983), decided to examine changes in weekly share prices.

What he and everyone expected to find is that a relationship existed between past and future changes. He expected to find a recognisable distribution. He found none. He expressed his findings (as Fama pointed out) in graphic terms.

"The series [of changes in the weekly prices] looks like a wandering one, almost as if once a week the Demon of Chance drew a random number from a symmetrical population of fixed dispersion and added it to the current price to determine the next week’s price”.

And thus another famous idea came into common usage: Share prices follow a random walk. This idea is widely discussed by Burton Malkiel, an economist from Princeton University, in his book Random Walk down Wall Street.

The random nature of changes in share prices can be illustrated by comparing the changes in prices today with the change from the previous day. Let’s compare, for example, the changes in an investment earning compound interest in a bank with the changes in share prices of, say, Absa Bank Limited.

This was done for each over a 30 day period and the results shown graphically below (on so called X-Y scatter diagrams).

The compound interest investment produces a neat "return” line. There is a 100 percent correlation between changes in the past and those in the future. In contrast the changes in the Absa share price do not show a neat relationship. This result surprised everyone in the 1950s. Why does such a vast difference exist?

The EMH explains this. Share prices on the stock market derive from available information – and new information arises all the time, randomly. The lack of correlation between past price changes and future changes can thus be explained. All one must do to test the EMH is to test if there is a correlation between past and future changes in the prices of shares.

If a strong correlation exists, as in the case of the compound investment, then clearly the market does not reflect any new information. The market is inefficient. If there is no correlation, then the prices are changing all the time as new information becomes available, and it can be said that the market is efficient.

One of the difficulties in interpreting the EMH is the failure to correctly understand it. Students often assume the information is about the firm itself. And so they assume the price of the share will change when information about the firm becomes known. This is not true.

The information is much broader as illustrated by our earlier examples. Either of these announcements – from China or the United States – would have influenced Absa’s share price despite having little to do with Absa.

A second error arises when students confuse value and price. Many students assume these concepts are the same, when they are clearly not. The EMH has to do with the movement of prices and not the value of companies.

It does not seem possible to abandon the EMH at this stage since it explains observable phenomena. It will have to be replaced with a theory which can explain these phenomena.