Efficient Market Hypothesis
The above writeups are interesting, but I'd like to have a more in depth look at this hypothesis - although many ask `does it work?' the answer is, as in so much of economics, `it depends.' Firstly, it's not right or wrong - the question we should be asking is `to what extent are markets efficient?' Let me start by saying a lot of what is written below is adapted from a discussion paper I wrote in 2003 for my school's Economic Society. This explains why parts of the discussion are slightly dry and verbose, although I have tried to alter the tone. I found the theory absolutely fascinating, and after I read a superb book by Paul Ormerod - `Butterfly Economics' - I began reading much deeper into the subject.
Firstly, what is the theory? Well taking a step back, what is price? Oscar Wilde once commented `A cynic is a man who knows the price of everything, and the value of nothing' although according to the efficient market hypothesis (I'll call it the EMH from now on) they are the same thing. That is, price = value. That means price is equal to the expected stream of future earnings of an asset - what in God's name does that mean? Well, tke a house price. The price would be equal to the rent of each of each following month. P = R1+R2+R3... where each variable is the rent for a month. Of course, this series would just go on forever, meaning a small addition has to be made. We need to remember that money in the future is worth less than money in the present. Why? Say I was promised £100 in 10 years. Forget inflation for the time being. One would need to save LESS than £100 now to get £100 in 10 years because of interest rates generally (almost always) being positive. I might only need £60. Hence, £100 in the future is worth less than £100 in the present. This means we need to offset each rent (each R) against interest rates. Obviously, each variable will get smaller and smaller (if you don't understand this ask me and I can explain) meaning the series is almost converging (I'm not a good mathematician so I may be wrong when I say that) - anyway, all this means that we can work out the price, P, and say that is the summation of all future earnings from the asset. Hence the price of a house is the expected earnings from the rent.
This doesn't only work for house prices. If we take a stock, it's price should be the summation of all future dividend payments, again, offset against interest rates to account for the fact that money in the future is worth less than money now (please note I'm not suggesting that dividends are the only source of `earnings' - many blue-chip companies have never offered dividends). Basically, as far as I can see, this price formation model holds true for all assets.
It is just like any valuation model - many will have heard of the Black-Scholes valuation model which could work out the price of options contracts. After all, investors need a way of working out the underlying value of an asset.
Now to make some definitions - let's call the expected stream of future earnings the `fundamentals' of the company. The price is, well, the price.
One last note on this topic - how do I reconcile this with the usual belief that price is where demand and supply interact? Well, for all the economists out there consider the axis on a demand/supply graph. Demand/Supply on the y-axis and Price on the x-axis. Hence y=f(x). This means Demand=f(Price) and Supply=f(Price). Both are functions of price, therefore the next question must be how is price determined for assets? See above!
What are the implications of all this stuff? Well, price should equal the fundamentals; therefore the price should only change when the fundamentals change. That may seem obvious, but it forms the basis of this discussion.
So What is it?
When Alfred Marshall, one of the 20th century's greatest economists, penned `Principles of Economics' in 1890 he was certainly not trying to suggest that `all things' were always equal. Rather, his approach of isolation and analysis of individual factors affecting, say, an asset price was all that was available to him mathematically. However by ignoring interaction between agents, he left his theories deeply flawed and a rebuttal of the Efficient Markets Hypothesis is, by logical extension, a denunciation of the principles of precise equlibria upon which neoclassical economics rests. If you have no idea what this paragraph means, don't worry.
The EMH was only put forward in 1965 by Eugene Fama:
An efficient market is defined as a market where there are large numbers of rational, profit maximizers actively competing, with each trying to predict future market values of individual securities, and where important current information is almost freely available to al participants ... at any point in time, actual prices of individual securities already reflect the effects of information based both on events that have already occurred and on events which, as of now, the market expects to take place in the future. In other words, in an efficient market at any point in timethe actual price of a security will be a good estimate of its intrinsic value.1
Although no more than a natural inference drawn from Marshall's work, the implications of this theory were hugely significant: that no security should be overvalued or undervalued at any point in time - if it is the former, then the perfectly rational agents in the system (conveniently equipped with all public information) would simply sell the security in question lowering its price to the equilibrium level. This level would be the true value of the company.
Take this example. Say the fundamentals told you that a house's true value is £100. Say the market price is £150. Would you buy the house? Well if you were rational you wouldn't since it's overvalued.
Many say that there is no scope for arbitrage in an efficient market, and this is correct - it is wrong to say profits can be made as a result of short-term swings. Why? Well say the fundamentals of a security were worth £100 and therefore the price was £100. If a new piece of news revealed the house was in fact worth £200, the new trading price would be £200. Sure, people holding the stock would have made a profit, but no-one could make a profit by buying or selling once the new piece of news was out, since the new price would be resolved instantly. Again, tell me if you don't understand any of this.
The second important point is that only randomly arriving, new information should affect the price. If the information were not new, it would have already been factored into the price and if it were not random, it could have been predicted and would have been worth someone's while to act upon. Thus if only random events can modify this strangely stable price, all future movements of price must be random and so technical analysis of these markets is useless, making a large number of CNN and Bloomberg analysts wholly redundant. Asset prices should follow a variant of `random walk theory'2 with a serial correlation of around zero.3 As Burton G. Malkiel points out
A blindfolded monkey throwing darts at a newspaper's financial pages could select a portfolio that would do just as well as one carefully selected by the expert. 4
Warren Buffet once commented, `I'd be a bum in the street with a tin cup if the markets were efficient.' Yet Mr. Buffet has made over $10 billion 5. The thought of Simian replacements probably doesn't worry him deeply. His success came about through an ability to identify undervalued assets - assets that according to the EMH should never have been undervalued for any significant period of time. Of course, a glance at the Financial Times will unequivocally show you supply and demand do not, in elegant neoclassical majesty, come together and settle at a perfectly stable level, only budging when random information suddenly arrives.
Let's look at an example now and see if the EMH is actually useful. Take a look at any good stock price site (try fool.com) and have a look at the graph of PeopleSoft shares.
Take a look at how they varied over May and June 2003. If the EMH is correct one would expect a relatively flat line with changes only when new information is revealed. In fact, if you have the graph up you'll notice the large jump occurring on June 6 - this can be attributed to the announcement that Oracle had made a £3.1 billion bid for the company. The aforementioned rational buyers calculated perhaps that the company would subsequently benefit from economies of scale and its dividend payments would increase. Having decided that the company's intrinsic worth had increased (the fundamentals we talked about), they realised it was undervalued and began to purchase the security until it reached a certain price level - the equilibrium. Thus it can be seen in reality that this new, random information had a large impact on the stock price and this is a pattern observed in most securities, with the price adjusting within a day of information being made available.6 Yet what can be said for the other fluctuations occurring? The wildly oscillating nature of the price contradicts what the EMH tells us, for no new information was made available on May 2 yet the price changed.
What's wrong with the theory?
At this point we can begin to pick holes in the theory, which one can see is empirically inaccurate. My objection is not with the general randomness of data, since even a more complex model would predict securities to follow a chaotic path (where trends may occur but their occurrence and duration itself is random - that is what chaotic behaviours is, right? Correct me if I'm wrong) with little scope for prediction. The true problem is that as pointed out, price (in markets that should be efficient) oscillates without a change in the fundamentals upon which the price theoretically rests. Kenneth Arrow, a Nobel laureate who contributed much to this field, called this an `empirical falsification' of the theory. That is, P != fundamentals in most of the data we look at.
In the rest of this discussion I'll refer constantly to `agents' - this refers to anyone buying/selling in any market, whether the market in question is for houses or securities. Firstly, the theory assumes that agents in the system act to maximise their utility and they do so by purchasing undervalued stocks and selling overvalued stocks. Even if we assume this agent has a magnificent ability to process huge amounts of public information (involving memorisation of every company's annual report, no doubt) and that this information is available to everyone (because, God forbid, insider information would never be available) who is to say the agent will act upon his valuation of the stock?
So how do markets work in real life?
An important concept that must be introduced here is the interdependence/interaction of agents - a phenomenon wholly ignored by the conventional models.7 If one is attempting to maximise gains by buying a potentially undervalued asset it is interesting that if other agents act irrationally it will not matter if a single agent starts to sell when he thinks the asset is overvalued - the other agents may continue to buy and as the price rises and rises (no matter what the "intrinsic value" of the asset) the lone agent will have lost out. In other words, the equilibrium price may not be the underlying value of the asset, but what agents think at what price other agents value the asset. Keynes illustrated this bizarre double guessing by using the famous analogy of a beauty contest.
Professional investment may be likened to those newspaper competitions in which the competitors have to pick out the six prettiest faces from a hundred photographs, the prize being awarded to the competitor whose choice most nearly corresponds to the average preferences of the competitors as a whole; ... It is not a case of choosing those which, to the best of one's judgment, are really the prettiest, nor even those which average opinion genuinely thinks the prettiest. We have reached the third degree where we devote our intelligences to anticipating what average opinion expects the average opinion to be. And there are some, I believe, who practise the fourth, fifth and higher degrees.
8
This means gains will not necessarily be maximised if you buy when the price is below the intrinsic worth of the company; what if others sell? Conventional theories suggest a system of negative feedback whereby if a price is too high, demand falls lowering the price until it is too low, when demand increases again. As with any biological negative feedback phenomenon (pH levels in the bloodstream, for example) the data in question might be expected to oscillate around a mean level. This is because agents in an economic system - like correcting systems in the human body - cannot gauge the precise quantity they need to buy or sell in order to restore the price to the equilibrium. Yet the oscillations observed in asset price data does not show this regular fluctuation around a mean point, and occasionally even fall unchecked into oblivion, or explode upwards to the delight of our schizophrenic agents. So we have two key problems: volatility unrelated to changes in fundamentals, and large rises or dips in price that should, according to conventional theory, be stopped.
An example of the EMH failing
In 1999 the value of securities on the NASDAQ began to rise rapidly without any hint of slowing down, and within months investors began their cry of `overvalued!' Yet these same investors continued to slam the `buy' button on their keyboards so hard that security prices soared upwards even further (until early 2000, at least). Alan Greenspan memorably called it `irrational exuberance.' Why this behaviour? As Paul Ormerod points out, `The emulation of the activities of others leads to trends being reinforced rather than reversed.'9 In other words, positive feedback reigns supreme at times and irrational behaviour abounds with individuals trying to guess what others are guessing; the very behaviour described by Keynes. Since rationality is the pillar upon which Marshall built his theories, the existence of irrational behaviour is an effective erosion of the bricks of conventional hypotheses. Yet this applies not only to asset markets but also to any other market presumed to be efficient.
The main shortcomings of the EMH are similar to those of the long-run competitive theories that focus exclusively on equilibrium outcomes while ignoring the entrepreneurial activity that generates those outcomes.10
Frank Shostack identified above the paradoxical nature of the EMH and therefore of any market which is presumed to reach an equilibrium state: if there is no long run incentive to buy and sell, why should investors do so? If they then withdraw with this reasoning, the markets lose all fluidity and the EMH fails immediately. Thus one can conclude that for the theory to hold true the agents in the system must not only be rational, but also refuse to believe in the EMH itself. Lest we forget, these are the very same agents who have access to every piece of financial information on the planet! Further weakening its usefulness, I would say that the EMH can only be considered at all when agents can only act on a local level, unable to witness the actions of others - this rather contrived assumption would mean the impact of positive feedback is reduced, and that agents may act hoping that their peers will buy and sell on non-price factors. However, since we are aware that agents do buy and sell, they clearly believe markets to be inefficient and since they can make profits over a period, they seem to be correct in presuming so.
The EMH is unable to explain crashes and large booms, as shown, and modern economic theories can easily comprehend this behaviour. Yet why the volatility? This can be attributed to the presence of two different types of agents in the system: fundamentalists and chartists - something neoclassical economics never even thinks about.11 The former will buy and sell on the basis of the price representing the underlying value of the asset. The latter will attempt to extrapolate information from past data believing that it affects future prices (putting aside whether or not this is an essentially self-fulfilling belief). As time goes on, different proportions of the market will believe in one approach or the other. As these proportions change randomly, we can see that the volatility of the asset price is only a reflection of the `underlying volatility of the proportion of different types of agent operating in the market.'12 That is, the manner in which the price is determined keeps changing therefore the price itself keeps changing. Since the former process is random, so is the latter - this reveals that although asset prices may indeed be close to impossible to predict accurately, it is for an altogether different reason.
Neoclassical economics craves to find the set of rules that agents follow yet it fails to address the fact that the precise rule being used at any given moment in time cannot be determined - is the majority of the market acting in the manner of a fundamentalist or a chartist? The answer explains the seemingly peculiar movement of asset prices, and the very presence of chartists undermines the principles of rational decision-making suggested by Marshall.
One objection to my analysis might be that I discount the effects of external factors such as interest rate: that is indeed true, but this is just irrelevant: say interest rates do rise. This would simply be factored into the fundamentals (because don't forget we're offsetting each fundamental against future interest rates) thus causing the price to change. That is, the consumers would still act rationally by buying less. Having said that, changing rates would cause changing fundamentals and therefore changing prices. These aren't the ideal conditions for investment meaning that if the EMH is shown to be a useful explanation of behaviour in a particular market it can also be an effective anti-interventionist argument. Having said that however, I have just gone to great lengths to persuade you that most markets are not efficient. Thus the interacting agents will still act irrationally even if not prompted to do so by a central bank. My point? Don't use the EMH as an anti-interventionist argument.
So does it work?
An economic model's primary purpose is to reflect the manner in which a real economy operates (I think this was a view first put forward by Milton Friedman) and it seems as though the (flawed) mathematical simplicity of neoclassical theories has blinded its extreme supporters. Although a contemporary model involving interaction between agents is a messy, unrefined business its empirical roots are far stronger. As Ludwig Boltzmann pointed out, `Elegance is for tailors,' - not for economists.
Other Notes
Something important: the degree of market efficiency is affected by the amount of information mobility - I'm not sure if that's a real phrase, but I refer to it as the amount of information available and how easily people can get to it as well as how many peopel have read it. Why? Well the more information there is, the more people know about the fundamentals and the quicker prices can adjust. Clearly, blue-chip stocks with higher market capitalisations will have a greater degree of market efficiency generally speakling - however these securities can also be the most affected in times of high positive feedback.
If anyone has anything to add, or if they diagree please msg me and I will post comments and my reponses over here. I'd love any feedback at all (no pun intended).
1Euegene F. Fama, "Random Walks in Stock Market Prices," Financial Analysts Journal, September/October 1965
2Maurice Kendall, "Analysis of Economic Time-Series,"
Journal of the Royal Statistical Society 1953
3You can consider this as the correlation of a variable with itself over successive time intervals - often called ACF or serial correlation. Ask me if you don;t understand this
4Burton G. Malkiel, "A Random Walk Down Wall Street," 1973
5Robert G. Hagstrom, "The Warren Buffet Way," October 1995
6http://www.westga.edu/~bquest/2002/market.htm, 7th paragraph
7Although Marshall recognised `tastes' as a determinant of demand, he failed to realise their significance and even more importantly he didn's allow tastes to affect each other - that is, he saw them as wholly independent
8John Maynard Keynes, "The General Theory of Employment, Interest and Money," 1963 (Page 156)
9Paul Ormerod, "Butterfly Economics," 1998 (Page 156)
10 Frank Shostack, "In Defense of Fundamental Analysis: A Critique of the Efficient Market Hypothesis" (Page 30)
11Paul Ormerod, "Butterfly Economics," 1998 (Page 19)
12 Ibid.
|
