No one can reliably beat the market: that follows from the efficient market hypothesis, which is closely associated with the US economist and Nobel laureate Eugene Fama. Prof. Dr. Stefan May, head of investment strategy and product development at quirion, explains what lies behind the theory and what it means for how you invest your money.
When you follow the ups and downs of prices, along with the frequently shifting interpretations of trends, you don't always associate the behaviour of the stock market with “efficiency”. What does the efficient market hypothesis actually mean?
The word “efficiency” quickly tempts you to think of predictability. So let's start with what the efficiency hypothesis does not mean: “efficiency” specifically does not mean that experts with special expertise can calculate future price movements in advance, or otherwise foresee them. Instead, the assumption is that all available, price-relevant information is already reflected in current prices. Exactly what will happen in the future remains unknown. That is one of the core messages of the work of the US economist Eugene Fama from the 1960s and 1970s. Even though much of it is often misunderstood, his work still has a major influence on capital market research today.
How did Fama arrive at his efficiency hypothesis?
He was neither the first nor the only one to describe markets as efficient. As early as the mid-1940s, the Austrian economist Friedrich August von Hayek described the price structure of a market economy as a kind of information system. Prices signal shortages or surpluses. Market participants observe prices, form expectations and act accordingly – a particularly efficient mechanism for reconciling the various interests at play. Fama's efficient market hypothesis deepened this idea and applied it specifically to the capital markets.
According to Fama, there are three different versions of the efficiency hypothesis. With the thesis of “weak efficiency”, Fama very early on took a stand against so-called chart analysis, that is, forecasting price movements on the basis of particular price patterns or analyses. Because such patterns are visible to everyone and can be exploited, they contain – according to the weak variant of the efficiency hypothesis – no further usable information, usable in the sense of a successful forecast.
What about analysing fundamental data such as a company's earnings position?
That brings us to the “semi-strong” version. It holds that not only chart information, but all publicly available information is already priced in – insofar as it is price-relevant at all. If you accept this, then analyses of the economy and of companies likewise offer no additional knowledge about future price movements.
Finally, in the strongest variant of the hypothesis, it is assumed that, beyond chart-based and publicly available information, prices also reflect insider knowledge. But that can of course only hold true if insiders, despite the ban, actually trade on the basis of information available only to them.
To what extent do the different variants of the efficiency hypothesis reflect the current state of the science?
The semi-strong version is largely uncontested in academia, with minor qualifications. Fama himself pointed out that something like “momentum” exists. This means that stocks with strong price performance in the recent past tend to show greater positive momentum. That actually contradicts the efficiency hypothesis, according to which price movements themselves contain no usable information. However, empirical studies suggest that so-called momentum strategies, which aim to exploit the effect directly, are hardly worthwhile. The effect is too weak and the cost of the strategy is too high. In that sense, the efficiency hypothesis holds true after all. It merely denies the existence of information in price movements that can be usefully exploited for forecasting.
Unfortunately, the correctness of the efficiency hypothesis cannot be tested directly and empirically in any of its variants. This means you cannot really verify it, but “only” falsify it – that is, potentially disprove it by testing its implications.
And what would disprove the hypotheses?
If one of two questions could be answered in the affirmative. The first: is there a strategy that systematically and reliably beats the average market performance? Decades of empirical research into active portfolio management consistently show that this ultimately does not work. There have, time and again, been strategies that managed to do so – even over longer periods. But in all these cases it was shown that the successes were, with very high probability, down to chance. An analogy with games of chance like the lottery is entirely apt here: just because there are always people who draw the right six numbers, no one concludes that there is a usable winning strategy behind it. This conviction persists only in the securities business. Because there have been, for example, individual active funds with successful strategies for a while, people readily conclude that active management must be successful in itself. That such successes are inevitable purely because of the sheer number of active funds is overlooked in the process.
The second question would be: are there statistically significant connections between the price changes of yesterday, today and tomorrow? That, too, is not the case, at least not over time horizons that are relevant for investors. Academics speak of a “random walk property” of price movements. Past changes in prices mean nothing for future movements. Prices have, so to speak, no memory. That is why no one – and I emphasise no one – can give you a reliable “point forecast” for any market. If I were to name the DAX index level at year-end correctly today, it would only be a guess. It does not become knowledge, because it comes true by chance.
What does all of this mean for investment strategy?
That you should orient yourself towards the long-term market return. Attempts to beat it regularly go wrong. The focus on indices and the popularity of ETFs are closely tied to the efficiency hypothesis. The ETF market, however, has become very diverse. Some players want to capitalise on the popularity of ETFs and, in doing so, pursue active strategies after all. When products are marketed with a buzzword like “research enhanced”, for instance, that is just another attempt to lure investors with supposedly special knowledge. Such strategies have no place in our portfolios. We want to be very close to the market, to mirror the “global equity market” as precisely as possible. We don't rely on guesses about which stocks will be tomorrow's winners, and we trust the overall development of the market.
Can you actually rely on how the market develops?
The market economy is based on the principle that expected risk and expected return belong together: the higher the risk, the higher the reward for it, that is, the expected return. Ultimately, you can rely on this relationship. But unfortunately there are “good” and “bad” risks. So-called unsystematic risks, such as unfavourable developments in individual companies, sectors or regions, are not “rewarded” by the capital market. The reason: they could be eliminated through the broadest possible international diversification. That is why taking on such risks is not “compensated”. They are unnecessary and therefore “bad” risks.
So-called systematic risks, by contrast, are appropriately rewarded in the form of a higher expected return, which is why they are also the “good” risks. The reason: even the best possible diversification is unable to eliminate these systematic risks. Compensation for taking them on is therefore ultimately unavoidable. A global portfolio spread as broadly as possible across international markets contains only systematic risks. That is why it is superior to other portfolios in terms of the relationship between return and risk. This is another key finding of theoretical and empirical financial market research.
