Minimum wage and employment numbers are highly sensitive to the time period analyzed

See These Numbers Lend No Good Support to the Case for the Minimum Wage by Don Boudreaux of Cafe Hayek.

"GMU Econ PhD student Liya Palagashvili and I have studied some numbers that Barack Obama and plenty of pundits have seized upon as evidence that raising the minimum wage is good economic policy.  Here – in U.S. News & World Report – is our first attempt to expose the flimsiness of those numbers and the wrongheadedness of using them to argue in favor of raising the minimum wage.  (In this U.S. News essay we had only 600 words, but the pro-minimum-wage case that is built on these numbers is infected by several other errors that we had no space here to point out.)  A slice:

Second, these statistics are disturbingly sensitive to small changes in their starting and ending dates. It’s true that employment in the 13 states that raised their minimum wages in January was, on average, 0.85 percent higher in June 2014 than it was in December 2013. It’s also true that, over this same time span, employment in the 37 states that did not raise their minimum wages rose by only 0.61 percent. (These are the very figures that minimum wage proponents are now trumpeting.) But if we shorten this time span by just one month — looking now at January 2014 to June 2014 — we get a very different picture.

In June, the number of jobs in the 13 minimum-wage states was, on average, only 0.59 percent higher than it was in January, while, for the same time period, the number of jobs for the 37 states that did not raise their minimum wages was higher, on average, by 0.69 percent. Job growth since January (the month that these 13 states actually hiked their minimum wages) was slower in states that raised the minimum wage than in states that did not.

We don’t report these particular statistics as evidence that raising the minimum wage slows job growth. Our point instead is that finding simple trends, especially those that are highly sensitive to the time period analyzed, and then drawing policy conclusions is scientifically illegitimate. Using such a method makes it far too easy to cherry-pick from the data those numbers that support your preferred policy. So just as our comparison of January 2014 to June 2014 employment data is no evidence that raising the minimum wage reduces job growth, Obama’s comparison of December 2013 employment data to June 2014 data is no evidence that raising the minimum wage enhances job growth."