PsyFi Search

Tuesday 9 February 2016

Pssst ... Wanna Invest in a Conspiracy?

The Great Con

There is a great, massive, hideous conspiracy at the heart of modern financial system, designed to steal the ordinary person's money.  Like all of the best conspiracies, it's hidden in plain sight.

And the conspiracy is this: there is no system, no one knows what's going on and no one knows what's going to happen next. Fortunately we can now prove this, although perhaps not quite in the way intended.

Order from Chaos

Some of us are particularly prone to believing in conspiracies. Back in 1966 Richard Hofstader noted that it's the people who feel most powerless who are most likely to ascribe to conspiracy theories as a way of understanding complicated events. The idea that the world is a disorganised place in which distressing things can happen more or less by accident is something that many people find hard to accept.

Conspiracy theorists are not - at least in general - pathologically ill: too many people believe in conspiracies for this to be some kind of mental illness.  However, there is evidence that an inclination towards such theories makes people more likely to be radicalized and less likely to good citizens, so it's become an area of interest for governments.

Psychologists theorize that conspiracies are a way for the disenfranchised to make sense of society and that they're linked to the relative powerlessness of the communities and individuals involved. Humans have an in-built tendency to seek to exert control over their environment and, if placed in a situation where they can't, will often invent illusory narratives to explain what's going on.

Fishing for Control

The anthropologist Bronislaw Malinowski described these self-same behaviors over half a century ago among fishermen from Papua New Guinea. Those who fished in safe, shallow waters and who generally much more in control of events weren't much prone to superstition. But those who fished in deep, dangerous waters had lots of ritualized practices, none of which made any difference to whether or not they were in control, but which presumably served to help them feel like they were.

In The Influence of Control on Belief in Conspiracy Theories, Jan-Willem Van Prooijen and Michele Acker showed that if people felt they were in control then they were less likely to believe in conspiracies. However, they also showed that the belief in a conspiracy behind a specific threat to society - in this case the Y2K bug - was predictive of a wide range of beliefs in other conspiracies, including the annoying habit of the US Air Force to repeatedly cover up evidence of UFO incursions:
"Our reanalysis of the Y2K-bug data indicates that a genuine control threat predicts people’s susceptibility to unrelated conspiracy theories, such as belief in a UFO cover-up, and the belief in a Kennedy conspiracy theory. Indeed, these relations hold up even after controlling for trust in the government, and the specific belief in a Y2K conspiracy theory. These findings underscore the external validity of the control–conspiracy belief relation and suggest that societal threats to control are associated with a generally suspicious, conspiratorial mindset."

This research backs up longer term programs run by Jennifer Whitson and Adam Galinsky, who have repeatedly shown that if they can induce feelings of helplessness in research subjects those individuals then tended to see false patterns in random data, a problem familiar to anyone who has spent any time among private investors.  They also demonstrated a preferences for superstitious behaviors, linking good but random outcomes with equally random prior events - the "lucky sock" syndrome.

In a market related experiment, Lacking Control Increases Illusory Pattern Recognition, Galinsky and Whitson were able to show that in a stable market people were able to accurately estimate the number of good or bad comments about a company that had infrequent updates. However, in a volatile market their judgement was badly impaired, and the net result was for significantly less investment in the company. In essence, the market conditions framed investors' expectations by causing them to link unfavorable news to an under-reported company.

More generally, as we saw in Deep Time and the Fallacy of Frequency the same researchers have shown that illusory pattern recognition in times of great uncertainty may be an adaptive benefit. As Galinsky and Whitson state:
"Increased pattern perception has a motivational basis by measuring the need for structure directly and showing that the causal link between lack of control and illusory pattern perception is reduced by affirming the self."

So at least it causes people to try and engage with their environment rather than simply curling up in a ball and hoping the nasty market volatility goes away.

Quantifying Conspiracies

Anyway, now someone has figured out a way of determining whether or not any given conspiracy is likely to be real or not. In On the Viability of Conspiratorial Beliefs David Grimes has come up with an equation to measure this which assumes the probability of any conspiracy collapsing can be modelled on a Poisson distribution. What the discussion argues is that the probability of any conspiracy being revealed increases as time goes on and with the number of people in on the know.

So Grimes estimates that over 400,000 people would have to stay schtum to keep the "fact" that the Apollo 11 mission was a hoax quiet, and a similar number to protect the climate change "fraud". Meanwhile over 700,000 people are implicated in keeping the cure for cancer under wraps. None of these are very likely anyway, but the research uses three actual exposed conspiracies to figure out the real value for the probability of a conspiracy failing, including the PRISM spying scandal exposed by Edward Snowden.

Conspiratorial Analogues

Now I don't have any substantive issue with this piece of research - I think it's amusing and it's a valuable way of debunking the conspiracy theorists. It might even be reasonably accurate, although frankly three data points isn't anywhere near conclusive, but it doesn't really matter if it isn't because precision isn't the point. However, what it is, in a simple form, is an analogue of how mathematical models are generated in finance.

You start with a guessed distribution (usually looking suspiciously Gaussian) and a whole lot of data points. You then muck about with the model parameters until you get an outcome that looks like it fits most of the data and you then use the model to predict future events. And this works quite happily until something new happens and the whole thing collapses. This usually happens shortly after everyone has become convinced that the model really describes the real world.  But it doesn't.

Dumb Models

In financial markets the real conspiracy is that everyone goes around pretending that they have some idea about what's going to happen or even why some things have happened. In general they don't, but to admit it would be to firstly reveal the great conspiracy and secondly expose everyone to the frightening possibility that no one's in control: which would create even more conspiracies as people desperately tried to exert control over the uncontrollable.

The truth is that no one's in charge, and no one has a damn clue how markets or stocks are going to move. All explanations are post-hoc rationalizations based on intuition and dumb modelling. The only solution is to ignore it all, treat it as noise and focus on the few things you can control. Buy quality as cheaply as you can, diversify heavily and be patient. Everything else is a giant con.

Honest. Trust me.


No comments:

Post a Comment