Who's Noisy Now?
In 1985 Fischer Black published a paper entitled “Noise”. In this he argued that much of the irrationality in financial markets could be explained by people trading without information: essentially using various spurious signals to decide when to buy or sell. In fact there’s an argument that without such people markets couldn’t work at all. The downside of this is that it’s difficult to separate the noise from the information.
Black predicted that there would be people who spend their lives trading on noise – so-called “noise traders”; people that Paul Krugman calls "idiots". Unfortunately these idiots generate so much noise that it’s very hard to determine the real information, which would suggest that if we're not careful, at root, we’re all idiots.
Noise, in information theory, is anything that makes the outcome of a communication process less predictable. If we think of a market as being a communication process then Fischer Black’s analogy means that anything preventing market prices from summarising all known information about a security is effectively noise. When the noise dominates the information signal then we can no longer discern the information being communicated.
Normally you’d expect the obvious information reflected in public sources such as annual reports and regulatory news filings to be reflected in stock prices. What Black’s model of noise trading suggests is that this isn’t always the case, because of the noise injected by information-less traders. Which would explain why fundamental analysis can reveal what technical analysis can’t. Of course, fundamental analysts can be idiots too, but they’re not always trading on noise.
Quite why noise arises in financial systems is an open question, because the classical economic view is that noise traders should be arbitraged away by the smart traders taking advantage of them. However, there’s reason to believe that this won’t always happen, because there are asymmetries in the way markets work. Andrew Lo suggests that irrational behavioral biases may be so strong as to overpower the available firepower that rational investors have available to them. Empirically he argues that this has to be true:
“One anecdotal type of evidence is the series of financial manias and panics that have characterized capital markets ever since the 17th century when tulip bulbs captured the imagination of the Dutch. From 1634 to 1636, “tulip mania" spread through Holland, causing the price of tulip bulbs to skyrocket to ridiculous levels, only to plummet precipitously afterwards, creating widespread panic and enormous financial dislocation in its wake.6 Other examples include: England's South Sea Bubble of 1720; the U.S. stock market crashes of October 1929 and October 1987; the Japanese real estate bubble of the 1990's; the U.S. technology bubble of 2000; the collapse of Long-Term Capital Management and other fixed-income relative-value hedge funds in 1998; and the current real-estate bubble in England. These examples suggest the forces of irrationality can dominate the forces of rationality, even over extended periods of time.”
This, of course, has a second-order impact. If rational investors know that there comes a point at which they simply can’t stand in the way of a tsunami of irrational money then they will be loathe to take on the risks associated with this. In affect the sheer unpredictability of noise traders introduces risks into the markets which discourage arbitrage (see Noise Trader Risk in Financial Markets). However, this leads to another question because, as this paper states:
“In the context of fluctuations in the aggregate market, we find the idea of privately informed investors somewhat implausible. While one can always think of a person’s opinion as private information, this seems like playing with words. Speaking of the private information of a market timer like Joe Granville—who himself insists that he has a “system” rather than an informational advantage—makes little sense to us.”
Noise Traders Unite
Indeed, it would seem likely that the majority of noise traders are more likely to be individuals – a finding in line with the results of this paper from Alexander Kurov. But, if the majority of noise traders are private individuals how do they move markets? It’s only when they form into a huge, correlated crowd that they can overrun the rational market investors. This was investigated in Systemic Noise by Brad Barber, Terrance Odean and Ning Zhu, who discovered that individual traders are surprisingly co-ordinated in their buying and selling of stocks:
“Auxiliary analyses establish five strong empirical regularities: (1) Individual investors buy stocks with strong past returns; (2) individual investors also sell stocks with strong past returns; this relation is greater than that for buys at short horizons (one to two quarters), but weaker at long horizons (up to 12 quarters); (3) individual investors’ buying is more concentrated in fewer stocks than selling; (4) individual investors are net buyers of stocks with unusually high trading volume; and (5) individual investors are net buyers of stocks with recent extreme negative and positive returns.”
Although it’s hard to say with certainty what’s going on the most likely co-ordination methods would seem be based on behavioral biases and the researchers finger the representative heuristic, attention problems and the disposition effect. So the representative heuristic may be generating similar beliefs based on extrapolation of past performance and attention-grabbing stocks may be front-of-mind of various reasons when it comes to buying. When selling the disposition effect, where we try to avoid losses and so preferentially sell our winners, may be overwhelming rational decision making.
Evidence for Fischer Black’s idea that noise traders may be dominating markets has been mounting for years. Fundamentally, however, this is based on an analogy with information theory, the study of how information is communicated. The basics of this were worked out in the 1940's by Claude Shannon and Warren Weaver, starting from Shannon’s classic paper, which defined communication as the various ways in which one mind could affect another. Unfortunately this is a rather simplistic way of completely modelling human interaction, as it assumes that the information is passively received by someone when, in fact, we tend to construct our own meanings based on a combination of our own experiences and the data we receive.
Nonetheless, if we wanted to push this analogy onward the classic way of improving communication efficiency in a noisy system is to add redundancy. Redundancy is not additional information, but it does work to reduce noise by allowing us to cross-check that the data received is correct. At the simplest level it's duplicating the information transmitted so that we can check for errors. If the signal to noise ratio of actual markets means that market prices don’t always provide useful information then we must look for apparently redundant sources of the same.
Patience and Deep Pockets
The research shows the truth of Keynes’ old maxim that markets can stay irrational longer than we can stay solvent. Rather than incorporating all available public information they may be signalling nothing but noise traders betting against themselves and (hopefully) slightly smarter investors betting on the behavior of the traders. This is not a game most of us we should want to play, because betting on other peoples' mental states is a risky business that requires constant attention.
Anyone not steeped in the moment by moment discovery of share prices needs to seek out real information rather than noise, to focus on value rather than price. However, to succeed you need to ensure you can outwait the irrationality, which may require patience and deep, deep pockets. Anything else is just behaving like an idiot.