The myth of an efficient market, part 1/3 - Definition of efficiency

Deutsche Börse, Frankfurt, 20.6.2007

According to the efficient-market hypothesis the prices of all traded assets always "fully reflect" all information related to them. The most essential direct implication of this is that no one can continuously outperform the market. However, hardly any one suggests that the market would be entirely efficient, although on the other hand utter ineffectiveness would be an even more impossible idea. It is also clear that not all markets or submarkets function with the same degree of effectiveness.

Nevertheless, I'm surprised how strong the belief in market effectiveness appears to be in the economic field. For example according to the view of a Finnish doctor in Science and economics, the market would be 95 per cent efficient (Meklari 1 / 2010 in Finnish). No matter how you define this percentage and regardless of whether you look at the market through the obvious theoretical preconditions of perfect effectiveness or through practical experience in history, I find it very hard to believe in an efficiency of this magnitude.

This is the first part in a three-part series of blog entries. In this part I will have a look at the definition of effectiveness, whereas the latter parts will concentrate on the theoretical preconditions of perfect effectiveness and the level of effectiveness encountered in actual stock markets.

The definition of an efficient market

Eugene Fama is perhaps the most well-known spokesman for market efficiency. In 1970 he defined three forms of efficiency. The weak form only suggests that you cannot predict future prices for a traded asset based on past prices - and thus technical analysis is a futility. According to the semi-strong form all public knowledge is reflected in the price, so there's no point performing fundamental analysis. The strong form asserts that even insider information is fully reflected in the price.

In my opinion it's misleading to even define the weak form as efficiency. In practice, it only tells whether investors groundlessly base their choices on something that others have done earlier. The fact is that the price history of a stock has nothing to do with how sensible or valuable a specific company is right now. Ignoring the historical prices is certainly a part of the efficiency as a whole, but in itself it only measures how common irrational decision making actually is - in other words it's more like a measure of ineffectiveness rather than effectiveness.

The strong form on the other hand is rather absurd to begin with: how could insider information be included in the price if the holders of that information cannot even utilize the information without breaking the law? And if the insider information would be thoroughly reflected in the market, it wouldn't be insider information any more, now would it? Thus it could be stated that the semi-strong form is most open to debate on actual efficiency. However, rather than looking at three different forms of efficiency, I think it is more sensible to think of these three forms as different layers of the overall efficiency - layers which can be evaluated separately from each other. For instance, one practical implication of this is that in principle the efficiency on the "semi-strong / fundamental layer" could actually be higher than the efficiency on the "weak / technical layer". The fundamental and insider layers could also be considered as one single layer of informational efficiency, whereas the weak/technical layer could also be seen as a psychological and speculative layer.

Whether efficiency is viewed as a whole or through its layers, it seems to me that there is a common fallacy in its definition. It is true that as a consequence of perfect efficiency no one could ever outperform the market without good luck. However, since people are deficient, this reasoning doesn't work the other way around. In other words I find it rather peculiar, if someone tries to prove market efficiency by showing how no-one or merely few can actually beat the market. It's peculiar because already those few are a sign of inefficiency. But more than that, it's peculiar, because prices that fully reflect all information can only mean that the price of a traded asset is always truly "right", and that this price would include all possible information about the state of the company and its future prospects in a perfectly interpreted fashion. It's no wonder that not even the most talented investors can fully define this price in order to benefit from it - not even if they would have a limitless amount of time and resources at their disposal.

In the introduction an efficiency percentage was mentioned, so a definition for calculating one is in order. So let's say there is a "correct/true" price for each stock. Let's also think of a stock exchange and assume that we would have the information about the "correct/true price" of all shares. Since we're talking about the efficiency of the market as a whole rather than just the index, every single share needs to be taken into account separately. The preciseness of the price of a stock could be defined directly as a ratio between its "true value" and its market value. In order to get this value between 0 and 100%, this ratio should be calculated comparing the smaller value to the bigger one. The overall efficiency of the market could then be for example a weighted average (eg. based on company market values) of all stocks. As a formula this would look like as follows:

, where E is the Effectiveness as a whole, n is the number of stocks, w the entire market value of a company, and the letters v represent the "true" (t) and market-based (m) values of a single share.

The true value required here is in practice of course not explicitly definable. However, compared to the time when the market value was defined on the stock exchange, it still should be easier years afterwards. In this case there might of course be a tendency to hindsight bias, where this true value was based on something that could never have been predictable earlier. However, there are still many things that can be fairly judged in retrospect as something that could have been predicted, as long as those who were in the market would have just looked close enough and interpreted the information in the right manner.

In my opinion, when one considers a subject as theoretical as market efficiency, it cannot be justified by just assessing the performance of invidual people. Otherwise you could as easily state that no mammal can run 100 km/h, since no man is capable of that; or that a neighbouring galaxy does not exist, since man can't reach that. In other words, even though a person was not able to define the true price of a share let alone do that in adequate time to actually benefit from it, it does not mean that this true price didn't exist. After all, practical markets are only a manifestation of the inefficient decision-making of people who work under economical and even social restrictions, affected by irrational emotions, with limited information, limited time and limited comprehension in terms of actually interpreting the data at hand. Accordingly it is only natural that the stock prices are also defined at least somewhat inefficiently.

No comments:

Post a Comment