We’ve been discussing the “Nonfarm Payrolls % change from Recession End” for quite some time now. Each time we revisit this chart, I am struck by how starkly different this recession/recovery cycle is from the preceding business cycle(s).
A reader sends in this Ned Davis chart (below), which looks at the issue from a somewhat different angle. NDR observes that the first interest rate hike does not occur on average until employment ticks up by 4% from the cycle low (The cycle low in August 2003 was payroll employment of 129.789 million).
Davis further states:
“If employment needs to rise 4.0% before the Fed takes the punch bowl away from the bull party, this would imply a rise of 5,191,560 new jobs. At 112,000 per month (last month’s employment rise – the best since December 2000) we would need to wait 46 months before the Fed hikes rates.”
Additonally, rate hikes aren’t nearly as fatal to the markets as many assume them to be: Davis notes that “stock prices continue to rise for 1, 3, 6, 9, and 12 months later, on average.”
Today’s New York Times has an OpEd titled “The Medals Don’t Matter.” It’s by Jake Tapper, who is a well regarded ABC News Correspondent (formerly of Salon). The article reaches the conclusion that voters do not care about the military service of their Presidential candidates.
To reach this feat of logical deduction, Jake focused primarily on the 1992, 1996 and 2000 Presidential elections (and the 2000 GOP primary), and the Military Service of each candidate.
There are many, many analytical errors in his approach, sample size being the most obvious. But let’s focus instead on a very common logic error which seems to catch most people unaware:
Controlling for a single variable instead of many when analyzing complex systems.
I would be oversimplifying the situation were I to call this error, well, a mere oversimplification. But that’s what lay at the heart of this fallacy: Taking an extremely complex and dynamic issue — who won the Presidency and why — and then boiling it down to a single, and in this small sample, mostly minor issue. The author might as well have based it upon how many letters were in the men’s first and last names.
Presidential victories are the result of a far more nuanced and multi-faceted set of factors. This issue deserves to be examined in far greater depth . . .