The Zurich Axioms By Max Gunther
Previous Chapter | Next Chapter
Beware the Correlation and Causality Delusions
There is an old story about a fellow who stands on a street corner every day waving his arms and uttering strange cries. A cop goes up to him one day and asks what it’s all about. “I’m keeping giraffes away”, the fellow explains. “But we’ve never had ay giraffes around here”, says the cop. “Doing a good job ain’t I?” says the fellow.
It is characteristic of even the most rational minds to perceive links of cause and effect where none exist. When we have to, we invent them.
The human mind is an order-seeking organ. It is uncomfortable with chaos and will retreat from reality into fantasy if that is the only way it can sort things out to its satisfaction. Thus, when two or more events occur in close proximity, we insist on constructing elaborate causal links between them because that makes us comfortable.
It can also make us vulnerable, but we don’t usually think of that until it is too late.
I’ll give you a personal example. Many years ago, before Frank Henry and I had talked much about the Zurich Axioms. I made a little money jumping back and forth between IBM and Honeywell stock. Honeywell in those days was heavily committed to building big, general-purpose computers and was a direct competitor of IBM to a much greater extent than is true today. Over a period of eighteen months or so, I noticed that the two stocks’ prices often moved in opposite directions. When Honeywell climbed for a few weeks, IBM would be dropping, and vice versa. I put a little money into what I thought was a very smart parlay: ride Honeywell up a way, jump off, buy IBM at a low point, ride that up ….. and so on.
It worked tolerably well a couple of times. I should have realized it was working only because I was lucky, but I wasn’t that smart in those das. I thought it was working because ……. because …… well, I constructed a causal relationship to fit the phenomenon I had been witnessing.
I theorized that there were a lot of big investors – mutual funds, insurance companies, and wealthy private plungers – who periodically shifted enormous mountains of cash from IBM to Honeywell and back. When Honeywell announced an attractive new product or made some other good move, all those hypothetical fat cats would sell off IBM stock in order to load themselves up with Honeywell – and vice versa. This rigged-up hypothesis, if true, would explain the opposite motions of the two stocks’ prices.
Was it true? Almost certainly not. The truth undoubtedly was that the seemingly orderly price movements had been caused by events that coincide by pure chance. These events were random and unpredictable. The fact that those opposite price moves had occurred a few times in the past was no indication, and should never have been taken as an indication, that they would recur in the future. But my rigged-up causal relationship made the whole minuet seem more orderly than it was, and confidently bet too much money on it.
I bought a bunch of Honeywell at what I thought was a low point. Whereupon Honeywell and IBM both plunged together like a pair of ducks with their tails shot off. Before I understood what was happening and abandoned my illusion of order, I had lost about 25 percent of my investment.
Unless you can actually see a cause operating, really see it, regard all causal hypothesis with the greatest skepticism. When you observe events happening together or in tandem, assume that the proximity results from chance factors unless you have hard evidence to the contrary. Always remember that you are dealing with chaos and conduct your affairs accordingly. As the Axiom says, chaos is not dangerous until it begins to look orderly.
Because so many people in the money world are so desperate seeking orderly patterns, places like Wall Street generate steady streams of ideas about possible causal links between this and that. Some of these postulated links seem plausible to many, others only to a few. But all of them have some kind of allure for that order-loving organ the human mind, and every one of them probably has meant trouble for somebody.
For example, one set of perceived links – laughed at by some, taken seriously by others – has to do with a phenomenon known as the Republican First-Year Jinx. Since the early decades of this century, the stock market has consistently slumped in the first year of every term served by a Republican President – first terms and second terms alike. It happened to Herbert Hoover once, Dwight Eisenhower twice, Richard Nixon twice, and Ronald Regan (as of this writing) once. It even happened in the first twelve months of Gerald Ford’s irregular three-year term.
The first question is : Why? And the second question is: What should an investor do about it, if anything?
The most likely answer to the first question is that the phenomenon has been caused by random events having nothing to do with the political party of the newly inaugurated President. Chance correlations with market movements are dime a dozen, and this is one of them. It is like the Super Bowl Omen – the peculiar fact, often noted around Wall Street, that the market invariably rises in any year when January’s Super Bowl game is won by a team tracing its origins back to the old National Football League. The Super Bowl Omen is great fun to talk about, but nobody seriously thinks a causal relationship exists between the football game and the stock market. The correlation just happens, that’s all. And so it is with the Republican First-Year Jinx.
But there are investors who insist on making something orderly out of it. Their theory is that the President’s Republicanness causes the market to dip in his first year in office causes it how? Well, you can take your pick of hypothesis. One notion is that the Republican Party, billing itself as the party of business prosperity, raises people’s financial expectations to an unrealistically high level. When they don’t get rich instantly on Inauguration Day, they get disgruntled, and the backwash of disappointment swamps the stock market.
That is one theory. There are others. There is no need to waste our time on them, for none should be taken seriously. All are examples of the way in which people rip up phantom causal links to explain observed phenomena. And all are examples of the way in which a causal link, once invented and accepted can make a phenomenon look more orderly than it probably is.
Which can be dangerous, as we’ve seen. If you believe the President’s GOP-ness causes the stock market to slump then you perceive an orderly series of events and may feel pushed to take action on them. You become like Professor Fisher, seeing patterns that aren’t really there.
Maybe the Republican Jinx will operate true to from in the future, and maybe it won’t. It began by chance and one day will end by chance. No predictions can be made about it one way or the other. It is, in fact, only another part of the chaos.
Guard against imagining causes when you can’t actually observe them at work, and you will save yourself a lot of grief. Have fun at the Super Bowl game – but if the wrong team wins, see your bartender, not your broker.