Document category: Efficient Market Hypothesis (EMH)
[Return to category list]
[Return to category list]
"Random Walks in Stock Market Prices,"
Eugene F. Fama, Chicago School of Business Selected Paper Series.
Eugene F. Fama, Chicago School of Business Selected Paper Series.
This paper is based on on Fama's doctoral dissertation. It comprises 19 pages and is intended to be non-technical.
Everybody wants to be able to predict stock-market prices movements, and there are two main types of prediction theory.
This paper explains why, in an efficient market, such theories are without value.
An efficient market is one where large numbers of people seeking to maximise profit actively compete at trying to predict future share prices and all having access to the same information. In this situation, the price reflects all that is known and is expected about the share, and is a good proxy for intrinsic value.
In an uncertain world, there is a element of disagreement about intrinsic value, and the competition between buyers and sellers causes the price to wander randomly around the intrinsic value.
If there was a systematic reason for price movements, traders would be able to take advantage of the information and neutralise each other.
New information is reacted to instantaneously but there is a lag between the new information and establishment of a correct new intrinsic value. The size of temporary over- or undercorrection is random, as is the time to correct the difference.
Successive price changes will be independent - a random walk. A series of stock price changes has no memory. The future path of a share price is no more predicatable than a series of cumulated random numbers.
The paper concedes that the random walk hypothesis is unlikely to be an exact description of share price behaviour, but for practical purposes it is acceptable. A simple policy of buy and hold will be as good as any complicated method of timing trades.
The random walk theory has been tested empirically many times, both statistically and mechanically. The results are consistent and impressive. Fama states that he knows of no standard study evidencing important dependence in series of price changes.
However, an analyst will do better than buy-and-hold where he can identify non-negligible discrepancies between price and value, and predict the impact of events on value. The more analysts who are good at this, the narrower will be the discrepancy. Many analysts make the market efficient. The average analyst and the average investor will do no better than a randomly selected share of the same riskiness.
Conclusion
If, as the evidence suggests, the random walk theory is correct, chartist methods are akin to astrology and of no real value to investors. The challenge to the fundamental analyst is to demonstrate that his more complicated procedures are more profitable than a random selection.
Everybody wants to be able to predict stock-market prices movements, and there are two main types of prediction theory.
- Chartist methods are predicated on history, and stock-price patterns, repeating themselves, and being able to predict the future from past patterns.
- A fundamental analysis approach believes that an individual share has an intrinsic value, based on the earning potential for the share.
This paper explains why, in an efficient market, such theories are without value.
An efficient market is one where large numbers of people seeking to maximise profit actively compete at trying to predict future share prices and all having access to the same information. In this situation, the price reflects all that is known and is expected about the share, and is a good proxy for intrinsic value.
In an uncertain world, there is a element of disagreement about intrinsic value, and the competition between buyers and sellers causes the price to wander randomly around the intrinsic value.
If there was a systematic reason for price movements, traders would be able to take advantage of the information and neutralise each other.
New information is reacted to instantaneously but there is a lag between the new information and establishment of a correct new intrinsic value. The size of temporary over- or undercorrection is random, as is the time to correct the difference.
Successive price changes will be independent - a random walk. A series of stock price changes has no memory. The future path of a share price is no more predicatable than a series of cumulated random numbers.
The paper concedes that the random walk hypothesis is unlikely to be an exact description of share price behaviour, but for practical purposes it is acceptable. A simple policy of buy and hold will be as good as any complicated method of timing trades.
The random walk theory has been tested empirically many times, both statistically and mechanically. The results are consistent and impressive. Fama states that he knows of no standard study evidencing important dependence in series of price changes.
However, an analyst will do better than buy-and-hold where he can identify non-negligible discrepancies between price and value, and predict the impact of events on value. The more analysts who are good at this, the narrower will be the discrepancy. Many analysts make the market efficient. The average analyst and the average investor will do no better than a randomly selected share of the same riskiness.
Conclusion
If, as the evidence suggests, the random walk theory is correct, chartist methods are akin to astrology and of no real value to investors. The challenge to the fundamental analyst is to demonstrate that his more complicated procedures are more profitable than a random selection.