Super Sunday is here. In the interests of public service, we here at the IJSF blog want to warn you about the dangers of the Super Bowl Stock Market “Predictor.” George Kester, a finance professor at Washington and Lee University has recently updated the results of this “predictor” for the first time since the Journal of Finance inexplicably decided to squander 6 valuable pages of journal space on this drivel in 1990. In his defense, Professor Kester points out the spurious nature of this statistical artifact.
Here’s how it works: the conference affiliation of the Super Bowl winner is alleged to predict stock market performance for the upcoming year. If the winning team was affiliated with the original National Football League, the market will go up over the next year; if the winning team was affiliated with the old American Football Conference, the market will go down. Here’s the test and results:
Kester constructed a back-test with a beginning portfolio of $1,000 that he invested in S&P 500 stocks or Treasury bills, depending upon which team won the Super Bowl.
“Interestingly, over the entire history of the Super Bowl, my wealth would have been more than twice as great had I used this strategy rather than a passive buy-and-hold strategy with the S&P 500,” he said. “I took brokerage costs into account whenever I sold T-bills and invested in stocks and vice versa, and I also included dividends on the S&P stocks. The dollar values of the portfolios at the end of 2008 would have been $43,000 for a buy-and-hold strategy and $105,000 for the Super Bowl market-timing strategy.”
Of course the original predictor can’t be applied to today’s game because the Saints are not an original AFL franchise, but why let facts get in the way of a good story? I posit that one could select any event with a binary outcome observed annually and claim this “predicted” market outcomes in some way (my choice would be Punxsutawney Phil). An ex post analysis like this would have a good chance of finding some positive correlation.