Pages

2011-07-13

The myth of an efficient market, part 3/3: Actual markets

In the previous part I described the preconditions for perfect market efficiency. In principle even with all its shortcomings, a market could get close to perfect efficiency even if none of the preconditions are really met. But has it been like this in the real-life markets? How have the markets acted in every day situations or under special circumstances? How would this manifest itself as a percentage? And will the market of the future be different from the market of today?


Examples from the last decade

Let's have a look at the recent history. Even though it's an extreme example, it's hard not to mention the dot-com bubble. Normally when a company has a 10% possibility of being worth 100 billion in the long run, and 90% possibility of failing miserably and being worth nothing, its value should be about 10 billion. However in the dot-com bubble the value in a case like this was often closer to that best-case scenario. Thus, during the dot-com bubble many companies were overvalued with a percentage of thousands or even tens of thousands. This could be seen not only in the valuations of single stocks but also very strongly on the index level. Less technology-driven indices had much less of a bubble, but for example NASDAQ was overvalued by hundreds of percent.

Then there's the current crisis that started from the U.S. housing market. Before the crisis stock prices rose to ungroundedly high levels, from which they dove to levels so low that they were even less founded on reality. One of the reasons for the gravity of the dip must have partially been in the efficiency factor number 4: external economical factors. Due to these special circumstances also many small investors saw an excellent opportunity buy shares at very low prices.

Moreover, the way this crisis has affected the European economy with all its troubled national economies starting from Greece, is also largely caused by market inefficiency. Had the debt market functioned efficiently, Greece would have had to pay higher interest for its loans many years earlier. But this never happened until it really had to happen. Greece was deceitful when joining the EMU and as a euro country it got cheaper loans than it should have received based on its economy. Since it took much longer to get to the point of rising interest rates, the economical fall is now of course a lot worse.



Everyday situations

In addition to the aforementioned episodic examples there are many more customary situations, where inefficiency is evident. The human restrictions related to time become obvious, when companies announce their quarterly reports. It's not uncommon that the first reaction will be disappointment eg. from too low quarterly earnings, due to which the stock price will plummet, but after the report is analyzed properly, the more essential factors start looking good, and the price will rise. Or the whole thing the other way around. Of course this is extremely short-term deviation, but what makes it interesting is that a quarterly report can set the price course for a significantly longer period of time. The price may slowly change in one direction, even if no new concrete information is brought to attention, and before the next quarterly report is announced, a plummet of 5% may have changed to a rise of 30%. Of course one could argue that the knowledge of absence of news is also knowledge. This is true, but it rarely justifies any significant price changes. All this can be seen as shortcomings in the efficiency preconditions 1-3 and 5.

In addition to these there are other generally recognized phenomena such as the May phenomenon - which can be summarized by the sentence "sell in may and go away", which somewhat applied to this year as well. Stocks with cheap key figures (which are commonly called value stocks) have also usually given better results than "growth stocks". If the markets truly were efficient, this couldn't be true.

Stock prices also more or less constantly vary without any clear reason. Of course given the extremely complex world and web of knowledge around us, this could be justified by new small pieces of knowledge. In practice the situation of the company or the market still shouldn't change significantly by minute.


Keskisuomalaisen kurssikäyrä
A special case are stocks that are particularly involatile and may have very large spreads between their sell and buy assignments. Here's a graph of a small cap OMX-H stock that suddenly rocketed 20% on 24.11.2010 from 17 euro to over 21 without any apparent reason. From the perspective of a regular private investor the change was also significant with a worth of over 30 000 euro. After that the spread was also that huge. Next day the price almost plummeted back to its earlier level, which means that the seller did a good deal there. Of course this is just a small stock in a somewhat peripheric market, but still shows an example of how unefficient the market can be.



The percentual efficiency of the market

So what do we get, when we look at all the practical data we have and try to put it into the the formula given in the first part? It's fair to assume that under full efficiency a stock market index should have a quite steady and balanced progress - not a straight line but not that far from it either. Still, also in an efficient market the stocks of individual companies would of course vary more than the entire index. Thinking like this, the efficiency of many stock market indices would have gone down to something like 30-40% during the dot-com bubble. By using some kind of more merciful weighting methods the efficiency could have been a bit more, around 50%, but that's still very inefficient.

The more recent bubble and recession don't seem much more efficient either. Even though the market started to be more aware of all the problems and risks related to the subprime loans, the market still continued its bullish trend for a while. In the end, most major indices, like DAX, NASDAQ and Dow Jones had gone down over 50% percent from their peaks reaching a bottom from which they have afterwards recovered close to 100% in two years reaching almost the level before the crisis! Of course unlike with the dotcom-bubble there are now more real issues instead of just extreme speculation about huge future profits. Still, even as the western economies are still facing severe problems, the course of events is extremely hard to consider as being even close to efficient. From overvaluation the situation progressed to a clear undervaluation, and once again, depending on the weighting method and other factors the efficiency levels might have been just a bit over 50 percent in many markets.

When thinking about efficiency on a more general level, we should of course consider a longer period of time. Between overvaluation and undervaluation there needs to be a moment when the stock market index momentarily "fully reflects all information". When the index is close to this level, the misvaluations of individual stocks take a bigger role in the overall efficiency. The stocks that are most actively bought and sold are usually of course relatively more efficiently priced as well, and when defining overall effectiveness these could actually have a bigger weight in it. Still, even if this would make the stock selection efficiency (in other words the relative price of different stocks, when comparing them with each other) have an efficiency of let's say 90 percent, the overreactions in the turns of economic trends seem so big that this would still make the overall efficiency much lower to about 70 percent.

When you consider Fama's efficiency forms more like layers as I suggested in the definition part, there is the possibility that efficiency on the information layer would in fact be stronger than efficiency on the psychologic-speculative layer. If the market pricing is divided into macro and micro level efficiencies (in other words the efficiency to detect the business cycle correctly and the efficiency to detect the price of a single stock correctly respectively), the psychologic-speculative layer can be considered as taking a bigger role as a cause for inefficiency on the macro level than it does on the micro level. Still, in practice the layers cannot be fully separated: even if the role of psychology and speculation grows, information efficiency always has a significant part in all of it. Nevertheless, if security pricing inefficiency on the macro level is enough to lower the efficiency to a level of 70 percent, one cannot neglect the non-informational causes for inefficiency.



True efficiency or just human efficiency?

It's probably clear by now that I don't find the market particularly efficient. Still, the level of efficiency is largely a matter of how you define "full reflection of all information". For a full reflection just having a large group of people isn't enough: all information will never be fully reflected in all prices. Still, the information could in principle be reflected to the extent that no individual could ever outperform the market. In practice, as long as the investment decisions are done by humans instead of machines, not even this can be possible. The reason for this is that then there would be no motivation to explore investment possibilities, and that in itself would already lower the overall efficiency from this level that would be humanly efficient.

I suppose when people talk about market efficiency, they are in fact talking about this human efficiency, even though in my opinion "the full reflection of all information" would require a lot more. In principle we can distinguish three different levels of efficiency. From the weakest to the strongest they are as follows:
The actual current market efficiency
< Theoretical full human efficiency
< Theoretical real full efficiency
Even the full human efficiency would already require a huge and extremely efficient organization without any personal goals to evaluate the prices of even the smallest of stocks. In addition the efficiency prerequisites 4 and 5 should always apply, since an individual investor can block these factors. If human efficiency would be for example 75% and the actual market efficiency 65% of the theoretical real full efficiency, the actual market efficiency compared to the human efficiency would be 87%. At good times this efficiency could reach even the 95% that was mentioned in the introduction.




Are the markets getting more efficient?

The market is always wrong
In practice the market is always wrong - it's only a matter of how wrong it is and how easy it is to pinpoint this error. There may be even big efficiency differences between different indices or markets and in somewhat peripheric indices like the Helsinki OMX-H the efficiency is probably lower than in bigger indices. As time passes efficiency might also grow. For example as the awareness of the January and May phenomena has grown, their effect may have become lower - or on the other hand sometimes these can be seen as self-fulfilling prophecies which would actually result in the opposite. What's more important, however, is that internet with all its information, interaction and the ease of trade should create a certain kind of intelligence of the masses. On the other hand as the amount of players on the playground grows, so might the amount of fools and lemming behaviour.

The dynamics of efficiency will probably still experience many changes. Information efficiency will still probably grow at least in terms of information available. However, as the world economy has become more and more connected, there is always a need for even more information and a possibility for more unexpected emotional bursts. Whereas movie critics for example evaluate a static fully finished product, the different parties in the stock market need to evaluate a constantly living, massive and hugely dynamic web of securities. And what's more, these stock market "critics" - investors, analysts and media - all live in interaction with the market, which in turn may cause self-fulfilling speculation and chain reactions.

All "critic parties" influence the market movement not only in the price formation of securities but also in the real economy. Thus, even absurd and originally inefficient psychologic-speculative reactions may through their own influence result in those reactions being partially sound and even expressing efficiency in a way. When you end up in a market that's been hit by a overreaction, even the tiniest straw can then break the back of that psychological camel, and once again the the entire market will turn from bull to bear or vice versa, and all this affects the real economy as well.

In a way, in addition to business cycles, one could define efficiency cycles. When efficiency gets too low, the motivational precondition for efficiency will get higher, and efficiency will grow. And when efficiency gets too high, the motivation gets lost and inefficiency takes over. By evaluating how efficient the market at a given time is, an investor might get clues on how worthwhile it is to investigate various investment possibilities. When efficiency gets low, it's time to strike.

No comments:

Post a Comment