Forms of Market Efficiency

Forms of Market Efficiency

The capital market is considered to be efficient in three different forms: the weak form, semi-strong form and the strong form. Thus, the efficient market hypothesis has been subdivided into three forms, each dealing with a different type of information. The weak form deals with the information regarding the past sequence of security price movements, the semi-strong form deals with the publicly available information, while the strong form deals with all information, both public and private (or inside).

The different forms of efficient market hypothesis have been tested through several empirical studies. The tests of the weak form hypothesis are essentially tests of whether all information contained in historical prices of securities is fully reflected in current prices. Semi-strong form tests of the efficient market hypothesis are tests of whether publicly available information is fully reflected in current stock prices. Finally, strong form tests of the efficient market hypothesis are tests of whether all information, both public and private (or inside), is fully reflected in security prices and whether any type of investor is able to earn excess returns.

Empirical Tests of Weak Form Efficiency

The weak form of the efficient market hypothesis (EMH) says that the current prices of stocks already fully reflect all the information that is contained in the historical sequence of prices. The new price movements are completely random. They are produced by new pieces of information and are not related or dependent on past price movements. Therefore, there is no benefit in studying the historical sequence of prices to gain abnormal returns from trading in securities. This implies that technical analysis, which relies on charts of price movements in the past, is not a meaningful analysis for making abnormal trading profits.

The weak form of the efficient market hypothesis is thus a direct repudiation of technical analysis.

Two approaches have been used to test the weak form of the efficient market hypothesis. One approach looks for statistically significant patterns in security price changes. The alternative approach searches for profitable short-term trading rules.

Serial Correlation Test

Since the weak form EMH postulates independence between successive price changes, such independence or randomness in stock price movements can be tested by calculating the correlation between price changes in one period and changes for the same stock in another period. The correlation coefficient can take on a value ranging from —1 to 1; a positive number indicates a direct relation, a negative value implies an inverse relationship and a value close to zero implies no relationship. Thus, if correlation coefficient is close to zero, the price changes can be considered to be serially independent.

Run Test

The run test is another test used to test the randomness in stock price movements. In this test, the absolute values of price changes are ignored; oly the direction of change is considered. An increase in price is represented by + signs. The decrease is represented by – sign. When there is no change in prices, it is represented by ‘O’. A consecutive sequence of the same sign is considered as a run.

For example, the sequence + + + — — — has two runs. In other words, a change of sign indicates a new run. The sequence — — — + + 0 — — — + + + + has five runs; a run of three,

followed by a run of two + ‘s, another run of one 0, a fourth run of three — ‘s and a fifth run of four + ‘s. In a run test, the actual number of runs observed in a series of stock price movements is compared with the number of runs in a randomly generated number series. If no significant differences are found, then the security price changes are considered to be random in nature. 

Filter Tests

If stock price changes are random in nature, it would be extremely difficult to develop successful mechanical trading systems. Filter tests have been developed as direct tests of specific mechanical trading strategies to examine their validity and usefulness.

It is often believed that, as long as no new information enters the market, the price fluctuates randomly within two barriers—one lower, and the other higher—around the fair price.. When new information comes into the market, a new equilibrium price will be determined. If the news is favorable, then the price should move up to a new equilibrium above the old price. Investors will know that this is occurring when the price breaks through the old barrier. If investors purchase at this point, they will benefit from the price increase to the new equilibrium level.

Likewise, if the news received is unfavorable, the price of the stock will decline to a lower equilibrium level. If investors sell the stock as it breaks the lower barrier, they will avoid much of the decline. Technicians set up trading strategies based on such patterns to earn excess returns.

The strategy is called a filter rule. The filter rule is usually stated in the following way: Purchase the stock when it rises by x per cent from the previous low and sell it when it declines by x per cent from the subsequent high. The filters may range from 1 per cent to 50 per cent or more. The alternative to this active trading strategy is the passive buy and hold strategy.

The returns generated by trading according to the filter rule are compared with the returns earned by an investor following the buy and hold strategy. If trading with filters results in superior returns that would suggest the existence of patterns in price movements and negate the weak form EMH.

Distribution Pattern

It is a rule of statistics that the distribution of random occurrences will conform to a normal distribution. Then, if price changes are random, their distribution should also be approximately normal. Therefore, the distribution of price changes can be studied to test the randomness or otherwise of stock price movements.

In the 1960s the efficient market theory was known as the random walk theory. The empirical studies regarding share price movements were testing whether prices followed a random walk.

Two articles by Roberts and Osborne, both published in 1959, stimulated a great deal of discussion of the new theory then called random walk theory.

Roberts’ study compared the movements in the Dow Jones Industrial Average (an American stock market index) with the movement of a variable generated from a random walk process. He found that the random walk process produced patterns which were very similar to those of the Dow Jones index.

Osborne’s study found a close resemblance between share price changes and the random movement of small particles suspended in a solution, which is known in Physics as the Brownian motion. Both the studies suggested that share price changes are random in nature and that past prices had no predictive value.

During the 1960s there was an enormous growth in serial correlation testing. None of these found any substantial linear dependence in price changes. Studies by Moore, Fama and Hagerman and Richmond are some of the early studies in this area. Moore found an average serial correlation coefficient of — 0.06 for price changes measured over weekly intervals. Fama’s study tested the serial correlation for the thirty stocks comprising the Dow Jones industrial average for the five years prior to 1962. The average serial correlation coefficient was found to be 0.03. Both the coefficients were not statistically different from zero; thus both the studies supported the random walk theory.

Fama also used run tests to measure dependency. The results again supported the random walk theory. Many studies followed Moore’s and Fama’s work each of which used different databases. The results of these studies were much the same as those of Moore and Fama.

Hagerman and Richmond conducted similar studies on securities traded in the ‘over- the-counter’ market and found little serial correlation. Serial correlation tests of dependence have also been carried out in various other stock markets around the world. These have similarly revealed little or no serial correlation.

Much research has also been directed towards testing whether mechanical trading strategies are able to earn above average returns. Many studies have tested the filter rules for its ability to earn superior returns. Early American studies were those by Alexander, who originally advocated the filter strategy, and by Fama and Blume. There were similar studies in the United Kingdom by Dryden and in Australia by Praetz. All these studies have found that filter strategies did not achieve above average returns. Thus, the results of empirical studies have been virtually unanimous in finding little or no statistical dependence and price patterns and this has corroborated the weak form efficient market hypothesis

Empirical Tests of Semi-Strong Form Efficiency

The semi-strong form of the efficient market hypothesis says that current prices of stocks not only reflect all informational content of historical prices, but also reflect all publicly available information about the company being studied. Examples of publicly available information are—corporate annual reports, company announcements, press releases, announcements of forthcoming dividends, stock splits, etc. The semi-strong hypothesis maintains that as soon as the information becomes public the stock prices change and absorb the full information. In other words, stock prices instantaneously adjust to the information that is received.

The implication of semi-strong hypothesis is that fundamental analysts cannot make superior gains by undertaking fundamental analysis because stock prices adjust to new pieces of information as soon as they are received. There is no time gap in which a fundamental analyst can trade for superior gains. Thus, the semi-strong hypothesis repudiates fundamental analysis.

Semi-strong form tests deal with whether or not security prices fully reflect all publicly available information. These tests attempt to establish whether share prices react precisely and quickly to new items of information. If prices do not react quickly and adequately, then an opportunity exists for investors or analysts to earn excess returns by using this information. Therefore, these tests also attempt to find if analysts are able to earn superior returns by using publicly available information.

There is an enormous amount and variety of public information. Semi-strong form tests have been performed with respect to many different types of information. Much of the methodology used in semi-strong form tests has been introduced by Fama, Fisher, Jensen and Roll. Theirs was the first of the studies that were directly concerned with the testing of the semi-strong form of EMH. Subsequent to their study, a number of refinements have been developed in the test procedure.

The general methodology followed in these studies has been to take an economic event and measure its impact on the share price. The impact is measured by taking the difference between the actual return and expected return on a security. The expected return on a security is generally estimated by using the market model (or single index model) suggested by William Sharpe. The model used for estimating expected returns is the following:

This analysis is known as Residual analysis. The positive difference between the actual return and the expected return represents the excess return earned on a security. If the excess return is close to zero, it implies that the price reaction following the public announcement of information is immediate and the price adjusts to a new level almost immediately. Thus, the lack of excess returns would validate the semi-strong form EMH.

Major studies on the impact of capitalization issues such as stock splits and stock dividends have been conducted in the United States by Fama, Fisher, Jensen and Roll and Johnson, in Canada by Finn, and in the United Kingdom by Firth. All these studies found that the market adjusted share prices instantaneously and accurately for the new information. Both Pettit and Watts have investigated the market’s reaction to dividend announcements. They both found that all the price adjustment was over immediately after the announcement and thus, the market had acted quickly in evaluating the information.

Other items of information whose impact on share prices have been tested include announcements of purchase and sale of large blocks of shares of a company, takeovers, annual earnings of companies, quarterly earnings, accounting procedure changes, and earnings estimates made by company officials. All these studies which made use of the Residual analysis approach, showed the market to be relatively efficient.

Ball and Brown tested the stock market’s ability to absorb the informational content of reported annual earnings per share information. They found that companies with good earnings report experienced price increase in stock, while companies with bad earnings report experienced decline in stock prices. But surprisingly, about 85 per cent of the informational content of the earnings announcements was reflected in stock price movements prior to the release of the actual earnings figure. The market seems to adjust to new information rapidly with much of the impact taking place in anticipation of the announcement.

Joy, Litzenberger and McEnally tested the impact of quarterly earnings announcements on the stock price adjustment mechanism. Some of their results, however, contradicted the semi-strong form of the efficient market hypothesis. They found that the favorable information contained in published quarterly earnings reports was not always instantaneously adjusted in stock prices. This may suggest that the market does not adjust share prices equally well for all types of information.

By way of summary it may be stated that a great majority of the semi- strong efficiency tests provide strong empirical support for the hypothesis; however, there have been some contradictory results too. Most of the reported results show that stock prices do adjust rapidly to announcements of new information and that investors are typically unable to utilize this information to earn consistently above average returns.