Price Action Lab Blog

Premium Market Analysis


Models, S&P 500 Returns and Related Folklore

According to Investopedia, the Sports Illustrated Swimsuit Issue Indicator was first coined by the Bespoke Investment Group. This indicator suggested, when it first appeared, that when the cover model is an American the S&P 500 has a higher probability to generate a yearly return above its historical average return than when the cover model is a non-American   The recent article in Bespoke mentioned the fact that the cover model this year will be a Russian. I am sure no investor got scared and started selling based on this unique piece of information…

The Bespoke article calculated that when an American was on the cover of the issue, 82.4% of the time the S&P 500 return was positive and when a non-American was on the cover, 75% of the time the S&P 500 return was positive. I will add to those revealing statistics that the data also indicate that when an American was on the cover, the return was above mean return 61% of the time and when a non-American was on there, the return was below mean return 60% of the time. According to the limited sample that there is available of course.

Actually, the available sample is too small to be able to claim anything significant. There are just 33 data points available, with 17 Americans and 16 foreigners on the cover. Out of those 33 times, the return was above average 18 times due to 11 Americans and 7 non-Americans. The remaining 15 times that the return was below average, there were 6 Americans and 9 non -Americans on the cover. I am confident that when the sample increases, the observed bias between the American and non-American “influence” on returns will converge to 0. I am sure everyone understands that.

The point of this post is to investigate another aspect of this alleged correlation. In essence, this indicator implies that there is a correlation between the nationality of the model on the issue cover and the deviation of S&P 500 from average return. The yearly returns are shown below:

Source: Bespoke

I ran a Wald-Wolfowitz randomness test using the S&P 500 return numbers  shown above. The result was a p-value of 0.44834.  This value implies that very little or no evidence against randomness was found in the data. So we have this sequence of returns which we cannot find any evidence that it is not random and we try to correlate it with the nationality of the model on the cover of a yearly magazine issue. Well, I am sure most of us know that such correlations are bogus. It is just that articles like this give authors the opportunity to throw in a few pics of models and break the boredom inflicted upon us by a market that resists to take a break.

Subscribe via RSS or Email, or follow us on Twitter.


Technical and quantitative analysis of Dow-30 stocks and 30 popular ETFs is included in our Weekly Premium Report. Market signals for longer-term traders are offered by our premium Market Signals service.

Copyright Notice

Comments are Closed

Theme by Anders Norén