In his online bio, Forbes contributing blogger Mike Patton claims to provide “simple, straight-forward and unbiased analysis.” Well, he got “simple” right, in the sense that his analysis is simplistic. As for “straight-forward” and “unbiased,” not so much.
Turning his analytical prowess toward the effects of Seattle’s $15 minimum wage, Patton looked at the city’s unemployment numbers and determined that “the early results are in:”
Almost six months have passed since the first wage hike (April 1, 2015). Although it’s early, thus far the data doesn’t bode well for supporters of this law. … The following graph contains Seattle’s unemployment rate from January 1, 2014 to July 31, 2015. I have marked two important dates and included the unemployment rate at those times. When the law was signed, May 1, 2014, Seattle’s unemployment rate was 3.70%. When the first wage hike occurred, April 1, 2015, unemployment was 3.0%. Since then, unemployment has risen steadily while the national average has trended lower.
“If I had to guess,” Patton haphazardly guesses, “I’d say the unemployment rate will likely trend higher for several years as businesses seek ways to mitigate the negative financial consequences of this law.”
Wow. I mean, just wow. We’ve seen a lot of shoddy analysis in the service of slandering Seattle’s $15 minimum wage ordinance, but Patton’s work is downright embarrassing.
Patton looks at Seattle’s low unemployment rate in April of both 2014 and 2015, and the subsequent rate jumps after the city first passes and then implements its minimum wage ordinance as “clear” evidence that Seattle’s “unemployment rate has been rising.” And apparently, he conducted this simple, straight-forward, and unbiased analysis by Googling “Seattle unemployment rate” and grabbing a chart from Ycharts.com (Ycharts is the second hit on Google’s results page after a post from even-stupider Forbes blogger Tim Worstall).
But any idiot clicking through that Google link to Ychart’s initial five year chart should instantly see an obvious pattern: Seattle’s unemployment always bottoms out in April, followed by an offsetting rise over the next two or three months:
What explains this extraordinarily consistent pattern? Ychart is reporting non-seasonally adjusted data (hence the parenthetical “NSA” in the header on Patton’s chart), and this pattern merely represents the natural seasonal rhythm of unemployment, not just in Seattle, but in Washington state and the nation as a whole.
That’s why when we usually talk about the unemployment rate (and just about every other economic statistic) we talk about seasonally adjusted numbers. And when we run the same data set through the US Census Bureau’s X-13ARIMA-SEATS seasonal adjustment software we find that Seattle’s unemployment rate has actually fallen since April:
Patton saw in the data what he expected to see—a substantial minimum wage disemployment effect—leaving him utterly blind to the obvious shortcomings of the non-seasonally adjusted data. His bio says that he is a financial planning advisor, but if this is what Patton means by “simple, straight-forward, and unbiased analysis,” I wouldn’t want him anywhere near my money.
The truth is, all this early analysis is a fool’s errand led by erroneous fools like Mark Perry, Tim Worstall, and Patton. Whatever the month-to-month ticks of the unemployment rate or the number of food service jobs, it will take years to determine the employment effect, if any, of Seattle’s $15 minimum wage. And even then there will be a healthy debate over how to properly analyze the data. In the meanwhile, anyone who asserts, based on a few months of non-seasonally adjusted data, that “it’s clear that the unemployment rate has been rising,” is either an idiot or a liar or both.