Thursday, June 29, 2017

That Seattle minimum wage study has some curious results. [feedly]

That Seattle minimum wage study has some curious results.

http://jaredbernsteinblog.com/that-seattle-minimum-wage-study-has-some-curious-results/

[I'm outta town with very shaky internet access, but wanted to make a tiny bit of noise about this.]

I'm quoted in this story about a new paper on the Seattle minimum wage increase–it's in the process of phasing up to $15/hr–as follows:

"The literature shows that moderate minimum wage increases seem to consistently have their intended effects, [but] you have to admit that the increases that we're now contemplating go beyond moderate. That doesn't mean, however, that you know what the outcome is going to be. You have to test it, you have to scrutinize it, which is why Seattle is a great test case."

I still think that. But I also think something seems pretty "off" with the study, reviewed here by the WaPo.

–How could they get such job- and income-loss effects for low-wage workers in Seattle relative to their controls with such tiny wage effects? This is especially curious when considering the excellent point made by Schmitt and Zipperer, who critically review the Seattle study, that compared to Seattle's relatively high wage base, $13/hr isn't that far out of the usual range (be sure to read their critique).

–It seems extremely unlikely that increasing the min wg to $13 leads to job growth for those making >$19. I can't think of any labor market logic to that.

–The Seattle economy is doing really well, with solid job and wage growth amidst very low unemployment. I'd think that if the increase threw such a large wrench into the low-wage labor market as this study suggests, we'd see it in the broader economic statistics.

When you have an outlier study–their negative results are huge multiples of past research—with such unusual "internals," there may be something wrong. It could be the multi-establishment firms they left out, though if the increase is whacking smaller firms, that's a problem too.

So I suspect their control cohort—the other parts of the state that are serving as a control—is non-independent of the Seattle increase. This new study from Allegretto et al doesn't have the granular data available to the Seattle researchers but it uses what looks to me like a more credible control cohort and finds the Seattle increase to be having its intended effect.

Like I said, those of us who support out-of-sample min wg increases need to scrutinize the Seattle experience closely, and protect against confirmation bias. This time may actually be different. But you really don't want to make that claim based on one extreme outlier study with some eyebrow-raising quirks.


 -- via my feedly newsfeed

No comments: