Rodney Johnson | Wednesday, August 21, 2013 >>
Recently, The Wall Street Journal published a story about the Ghibli Curse. Ghibli is a production house that creates anime, essentially Japanese cartoons.
The basic premise of the curse is that when a Ghibli anime movie is aired on a Friday night, the currency-trading session immediately following the movie results in a rising yen and falling dollar. This, in turn, leads to a falling Nikkei, the Japanese stock market index.
This sequence of events is magnified when the Friday showing of the Ghibli movie coincides with the release of a U.S. jobs report.
There’s only one problem.
This entire concept is idiotic.
People who follow such nonsense are confusing correlation with causation. Here’s another great example I recently read in a blog…
Americans eat more corn dogs when it’s hot out. The correlation of higher temperature and a greater consumption of corn dogs has been strong for many years.
To cement the relationship, it’s also true that when it’s cold out, Americans eat fewer corn dogs. This relationship has also been in existence for years.
Armed with this information, a person can factually say the movement of temperature and the consumption of corn dogs is positively correlated.
However, no one in their right mind would argue that temperature “causes” a person to eat a corn dog. There is no cooling principle to corn dogs, where they provide relief from the heat.
Everyone knows that what “causes” people to eat corn dogs is their availability at fairs and festivals, which happen more frequently in the summer. If festival attendance suddenly dropped or if fairs quit being held, then corn dog consumption would also drop, no matter what the temperature.
Conversely, if we experienced unseasonably cool summers, we would not expect a big change in corn dog consumption because we’d still have fairs and festivals.
In the end, there is a correlation between corn dogs and the outside temperature (they have moved together in the past), but there is no causality (these facts alone do not give you any insight into the future).
So what about the correlation between Ghibli and the currency trading?
As history would have it, eight out of the last nine times a Ghibli show coincided with the release of non-farm payrolls, the data were weak. Seven of those times there was a sell-off in the U.S. dollar, a rise in the yen, and then a drop in the Nikkei.
When two things occur together seven out of nine times, it sure looks like a pattern. And therein lies the deception…
As human beings, we’re hardwired to recognize patterns because it helps us make sense of the world. I know that the correlation of these two data sets truly exists. I also know there is not one speck of causality in this relationship. What could it be?
Do statisticians at the Bureau of Labor Statistics (BLS) watch the schedule for Ghibli movies in Japan and “guesstimate” lower figures to coincide with certain broadcasts?
Do analysts who watch non-farm payrolls get really excited about upcoming Ghibli presentations and therefore make overly optimistic jobs forecasts, which lead to disappointment?
Of course not.
All this is stupid. Unfortunately, being stupid doesn’t preclude it from being seized upon by market participants.
This is where people have to pay attention. I don’t think many Americans are actually watching Ghibli presentations AND basing trades on them. But I do think people look for patterns where they don’t exist and clearly attribute causality when what they see is simply correlation.
Whenever you hear the notion that “each time X happens, then Y occurs,” ask yourself: “Is this just correlation, or is there truly a basis for causality?” If you can’t come up with a logical, well-reasoned explanation for causality, then don’t gamble your hard-earned money on the correlation continuing in the future.
That’s not to say the two things won’t continue to move together. It means that betting on it is nothing more than a roll of the dice. You just don’t know what will happen.
Ahead of the Curve with Adam O’Dell
I’ve written about the correlation (or lack thereof) between individual stocks many times.