This post is the second in an ongoing series focusing on incredibly idiotic stories in the newspaper.

Today’s victim is a story in today’s Wall Street Journal by Spencer Jakab titled Market’s Unlucky-Sevens Streak in Danger (please note that for copyright purposes, I can’t reproduce the story here).

The story states that, “Unless the rising stock market suddenly slams into reverse, a pattern for U.S. blue-chips that has held for 130 years is about to end.” and “For the past 13 times that a year has ended in seven, going back to 1887, the Dow Jones Industrial Average or its predecessor has suffered a sharp downturn between August and November.”

The story offers one explanation for a seasonal slump in the fall: “Seasonal patterns also were more powerful when an agrarian economy soaked up credit during the summer growing season, leaving money tight for the rest of the economy, including stock speculators. The period from November through April, by contrast, was a time of easy money.”

But the US hasn’t been an agrarian economy for a long time, and even if it still was, why would there be a causal relationship between a market decline and the year ending in “7”?

The punch line – there isn’t a relationship – it’s a spurious correlation.  Ronald Coase a British economist who won the Nobel Prize in Economics in 1991, is quoted as saying,  “If you torture the data long enough, it will always confess.” This is clearly a case of data torture.

A spurious correlation is a presumption that two variables (in this case the decline in the Dow Jones Industrial Average between August and November and the last digit of a year being seven) are causal when in reality they are not.

Tyler Vigen has written a book (which I have not read) appropriately titled Spurious Correlations where he states: “Provided enough data, it is possible to find things that correlate even when they shouldn’t.  The method is often called data dredging.”  On his website, he shows many correlations including:

  • Divorce rate in Maine and per capita consumption of margarine (99.3% correlation)
  • Total revenue generated by arcades and computer science doctorates awarded in the US (98.5% correlation)
  • US crude oil imports from Norway and drivers killed in collision with railway train (95.5%)

Clearly, there is no causal relationship for a decline in the Dow Jones Industrial Average and the year ending in a “7”. And therefore the article stating that the “Market’s Unlucky-Sevens Streak is in Danger” is absolutely preposterous.

The real lesson to take away from this story is be skeptical of what you read in the paper – question everything.

 

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *