Wow, talk about a straw man! This is an absurd statement. Those other options can be valid, and occur, AND employers can reduce staff. It can be (and probably is) a combination of all these things.

My argument applies. The claim, supposedly supported by this study, is that increases in minimum wage do not affect unemployment rates. Considering a $10,000/hour minimum wage is clearly a reductio ad absurdum argument meant to test this assertion. I am not saying that anyone is considering a minimum wage that high, merely that in these corner cases, albeit absurd, unemployment would certainly be affected. So, the assertion is false: unemployment can be driven by minimum wage. At what level this affect becomes statistically measurable is a different argument entirely, as is whether we "should" do it.

What I was getting at about there being other factors is that unemployment is a volatile measurement that is affected by many factors. It's quite possible that there *is* an affect, but smaller than the measurement certainty of the overall "unemployment" number. For example, if we imagine something like 2-4% of workers are actually employed at minimum wage, and then we say that raising it has a 1% increase of unemployment for those workers, then this is only .02-.04% of the employment total. It's quite possible that measurement methods won't show this affect and conclude there is "no correlation", when in fact there is one. Another example is how "unemployment" is defined, which is really mostly defined today as people collecting unemployment benefits. If a worker isn't on unemployment, they aren't "unemployed", which is why some people like to look at U6 instead of this number. Consider you generally need to work a certain length of time and at a certain number of hours before you become eligible for unemployment. Imagine a hypothetical situation where the minimum wage rises and employers fire all the recently-hired minimum wage workers, most of whom don't yet qualify for unemployment and thus are not considered. Finally, there are macro trends that affect the (un)employment rate much more than minimum wage. Yes, I understand that linear regression techniques are meant to eliminate those from analysis, but for small affects on small populations, this can be quite difficult to do.