Summary - The Efficient Market Hypothesis and its Critics PDF

Preview

Summary

Article summary...

Description

The Efficient Market Hypothesis and its Critics

Introduction 









A generation ago the efficient market hypothesis (EMH) was widely accepted in financial economics o It was generally believed that securities markets were extremely efficient in reflecting information about individual stocks and about the market as a whole o The belief was the news spread quickly, and is incorporated into prices immediately o Thus, neither analysis of past price patterns, nor more fundamental analysis (e.g. on firm’s earnings etc.) would allow a greater return than a randomly selected portfolio of individual stocks This theory is associated with the random walk of stock prices o The logic of this idea is that if information is unimpeded and then tomorrows price change will reflect only tomorrows news o News is by definition unpredictable, therefore price changes must be unpredictable and random also By the start of the 21st century, the dominance of EMH had become far less universal o A new breed of economist emphasized psychological and behavioural elements of stock price fluctuations, and came to believe there was some predictability This paper examines the attacks on efficient market hypothesis, and the belief that stock prices are partially predictable o Will describe the major statistical findings and their behavioural underpinnings o Examine relationship between predictability and efficiency o Describe major arguments of those who believe markets can be irrational by examining the 1987 crash and the internet bubble o To conclude that stock markets are far more efficient and far less predictable than some academic papers would allow efficiency – efficient financial markets are those that do not allow investors to make above average returns without taking above average risks

a non-random walk down wall street 

a review of some of the patterns of possible predictability suggested by behaviour of past stock prices

short-term momentum, including underreaction to new information  original empirical work on stock market randomness supported the view that the stock market has no memory o the way a stock behaved in the past has no bearing on its future behaviour  Lo and MacKinlay, 1999 – there are too many successive moves in the same direction for stock market prices to be truly random, therefore there is some momentum in short run o Economists and psychologists in the field of behavioural finance agree that such short run momentum could be the result of feedback mechanisms o Individuals see a stock rising, and get drawn in, this creates more excitement and another round of price increase, thus the cycle starts again  Behaviouralists offer another explanation for this short run momentum, tendency for investors to underreact to new information o If the full impact of new information takes a while to be reflected in the stock price, this would provide explanation for the kind of serial correlation observed  Several factors should stop us from interpreting the empirical results above as an indication that markets are inefficient  While stock markets may not be a perfect mathematical random walk, one must be aware of the difference between statistical significance, and economic significance o Statistical anomalies giving rise to momentum are very small and would not give the opportunity for anyone paying transaction costs to profit from them  While behavioural hypotheses such as feedback mechanisms and underreaction to new information may seem plausible, the evidence of them in the stock market is often very thin  Key factor is whether any patterns of serial correlation are consistent over time o Momentum strategies appeared to produce positive returns during some periods of the late 1990s, but highly negative ones in the 2000s  Many predictable patterns seem to disappear one they appear in financial literature o Schwert, 2001 – two possible explanations  Researchers are always sifting through mountains of data; their normal tendency is to seek things that challenge perceived wisdom  Every now and again, a combination of a certain method and certain sample will produce something statistically significant  Alternatively, perhaps practitioners learn of the predictable pattern and exploit it to the point it is no longer profitable Long-run return reversals  In the short run, over days or weeks, the usual argument again EMT is that some positive serial correlation exists o However, many social studies have shown evidence of negative serial correlation (return reversals) over longer holding periods  Some studies have attributed this forecastability to the tendency of stock markets to ‘overreact’ o Debondt and Thaler, 1985 – investors are subject to waves of optimism and pessimism that cause prices to deviate systemically from their fundamental values and later to exhibit mean reversion o They suggest this is consistent with the findings of Kahnemann and Tversky that inventors are systemically overconfident in their ability to predict future fluctuations in stock prices or corporate earnings  These findings give some support to contrarian investment techniques





Such return reversals for the market as a whole may be consistent with efficient markets theory o Since stock returns must rise or fall to be competitive with bond returns, there is a tendency for when interest rates to go up, prices for bonds and stocks to fall, and vice versa o If interest rates revert to the mean over time, this will result in mean reversion that is quite consistent with EMT Moreover, it may not be possible to profit from the tendency for individual stocks to exhibit return reversals o Fluck, Malkiel, and Quandt, 1997 – found that stocks with very low returns over the last 3 years, had higher returns in the next period, and vice versa. Thus, they confirmed the statistical evidence of return reversals o However, they also found that returns for both groups in the next period were very similar, so could not confirm that a contrarian approach could yield higher than average returns

Seasonal and day-of-the-week patterns  A number of researchers have found that January can be a very strange month for stock markets o Returns from an equally weighted stock index have tended to be unusually high during the first two weeks of the year  There also appears to be a number of day-of-the-week effects  Also patterns around the end of the month and around holidays  However, the problem is that these patterns are not consistent from period to period

Predictable patterns based on valuation parameters  

It has been claimed that valuation ratios, such as price-earnings multiples, or the dividend yield of the stock market as a whole, have considerable predictable power This section will analyse that

Predicting future returns from initial dividend yields  Formal statistical tests have shown the ability for the initial dividend yield of the market index can account for as much as 40% of the variance of future returns for the stock market as a whole o In general investors have earned a higher rate of return from the stock market when they purchased a market basket of equities with an initial dividend yield that was relatively high, and low rates of return when stocks were purchased at low dividend yields o These findings are not necessarily inconsistent with EMT, dividend yields tend to be high when interest rates are high, low when interest rates are low  Thus, the ability of dividend yields to predict movements in the stock market may just be reflection of the market adjusting to general economic conditions  Furthermore, the predictive power of dividend yields has diminished over time o This may be partly due to the changing behaviour of American corporations with regard to dividends

Many would now rather issue a share repurchase programme than increase their dividends, thus making dividend yields less predictive Finally, it is worth noting that this does not work with individual stocks o One example of the is the ‘dogs of the dow’ strategy o This involves buying the ten stocks on the Dow Jones Industrial Average with the highest dividend yields  In some past periods this method outpaced the average rate of return  However, such funds generally underperformed the average during the 1995-1999 period 



Predicting market returns form initial price-earnings multiples  The same kind of predictability shown for dividend yields, has been shown for price-earnings multiples  Investors have tended to earn larger long horizon returns when purchasing the market basket of stocks with relatively low price-earnings multiples o Campbell and Shiller, 1998 – P/E ratios explain as much as 40% of the variance of future returns o Equity returns in the past have been predictable to a considerable extent  Recent experience of investors basing their decision of P/E ratios has not been good Other predictable time series patterns  Studies have found some amount of predictability for stocks based on various financial statistics o Fama and Stewart, 1977 – short term interest rates are related to future stock returns o Campbell, 1987 – term structure of interest rate spread o Etc.  Even if some predictability exists, it is far from clear if they can be used to generate profitable trading opportunities

Cross sectional predictable patterns based on firm characteristics and valuation parameters 

A large number of patterns that are claimed to be predictive are based on firm characteristics and different valuation parameters

The size effect  One of the strongest effects investigators have found over time is the tendency for small firm’s stocks to generate greater returns than large ones o Since 1926, small company stocks have averaged annual rate of return over 1% larger than those of large companies  Crucial issue is whether the high returns of smaller companies produce predictable patterns o From the mid-1980s through all of the 1990s there has been no added benefit from holding smaller stocks o In most markets larger capitalization stocks produced larger rates of return  It is also possible that some studies of smaller firms have been subject to survivorship bias o When a small firm collapses, data for it is no longer held, thus, only small firms that have stood the test of time, i.e. succeeded, have their data analysed

Value stocks  Several studies show that value stocks have higher rates of return than growth stocks o The most common method of identifying value stocks have been price-earnings ratios and price-book value ratios  Stocks with low price-earnings multiples tend to provide higher rates of return than those with high price-earnings ratios o Finding is consistent with Behaviouralists views that investors are overconfident of their abilities to project high earnings growth and thus overpay for growth stocks  The ratio of stock price-book value has also been found to be a useful predictor of future returns o Low price-book is another sign of value, and is also consistent with the behaviouralists view mentioned above  Fama and French, 1993 – concluded that size and price-book value together provide considerable explanatory power for future returns, and once they are accounted for, P/E adds little else o Such a finding raises questions about efficiency if one believes the capital asset pricing model o But these findings do not necessarily imply inefficiency, they may simply indicate failure of the capital asset pricing model to capture all the dimensions of risk  E.g. companies in some financial trouble are likely to sell at low prices relative to book value  We must also remember that the tests of published studies, even those done over decades, can still be time dependent, also return patterns of academic study may not be able to be replicated with real money o Over a period running back to the 1930s, it appears that investors would not have been able to make higher than average returns from purchasing value stock  However, it does appear that 1960-1990 may have been a unique period in which this was possible Equity risk premium puzzle  Another puzzle used to suggest that markets are less than fully rational is the existence of a very large historical equity risk premium that seems inconsistent with the risk of actual stocks o This finding may be the result of perceived equity risk being considerable higher at the beginning o the period, and on average equity returns being much higher than predicted by investors  It is easy to say 50-75 years later that stocks of the day were under-priced, but one must remember that an annual average of almost 6% in corporate earnings was not expected so soon after the great depression Summarising the ‘anomalies’ and predictable patterns  preceding sections have pointed out anomalies, and statically significant predictable patterns o however, these patterns and anomalies are not robust and dependable in different sample periods o furthermore, if these patterns do exist, they could self-destruct in the future, as many already have  suppose there is a January effect, what will investors do

they will buy in before the January effect, which will serve as an increase in demand, thus price will rise, and the January effect will eventually become the December effect the more potentially profitable a discoverable pattern is, the less likely it is to survive many patterns may simply be the result of data mining o given enough time and combinations of samples and methods anomalies will inevitably be found o furthermore, these types of findings are more likely to be published than those that maintain the status quo o

 

seemingly irrefutable cases of inefficiency 

critics of EMT state there are recent examples of stock market fluctuations that could not have possible been rational

market crash of 1987  behaviouralists would say that a drop in the value of the stock market by one third can only be explained by psychological considerations  it is of course, impossible to rule out entirely psychological causes  however more logical considerations can explain a sharp change in markets such as what occurred in October 1987  a number of factors could have changed investors views o yields on long term treasury bonds increased from 9% to almost 10.5% in the 2 months prior to October o congress threatened to impose a merger tax o secretary of the treasury threatened to encourage a further fall in the exchange value of the dollar  share prices can be highly sensitive as a result of rational responses to small changes in the interest rate and risk perceptions  this is not to say that psychological factors did not play a role, but it would be a mistake to dismiss the significant change in external environment internet bubble of the late 1990s  it could be said that the remarkable value assigned to internet and high tech related companies is inconsistent with rational valuation o even here though, no arbitrage opportunities were available to investors before the bubble popped  we know now that unsupportable and outlandish claims were being made about the internet o many of wall streets most respected independent security analysts were recommending internet stocks, stating them as being fairly valued  while there were no profitable and predictable arbitrage opportunities available during the bubble, an argument can be maintained that prices were incorrect for a time, and that too much capital may have flowed to internet and telecoms related companies o in this way the stock market may well have failed in its role as an efficient allocator of equity capital

other illustrations of irrational pricing  Palm Pilot/3com  Ticker symbol confusion

The performance of professional investors  

 



Surely if market prices were determined by irrational investors, it should be easy for professionals to spot and take advantage of errors in the market A remarkable body of evidence stands against this o Active mutual fund managers tend to underperform the market by about the rate of their additional expenses o Survivorship bias is present again  Poorly performing funds are merged into better performing ones, thus burying records of underperformers Throughout the last decade, three quarters of actively managed mutual funds have failed to beat the index Managed funds are regularly outperformed by broad index funds with equivalent risk o Funds that produce above average returns in one period are unlikely to do so in the next The record of professionals does not suggest that sufficient predictability exists in the stock market or that there are recognisable and exploitable irrationalities sufficient to produce excess returns

Conclusion   

As long as markets exist, collective judgements of investors will sometimes make mistakes o Some market participants are far less that rational The market cannot be perfectly inefficient of there would be no incentive for professionals to try and uncover the information before others do Whatever patterns or irrationalities do exist, they will not persist and will not be able to provide market participants with extraordinary returns...