[andrewla]

I can't seem to find the underlying paper, but statistics can be deceiving -- if the pool of "low-wage" workers was reduced because some of them, due to the boost provided by the minimum wage increase, became "medium-wage" workers, then this statement might not have a negative meaning.

It does not appear that the survey involved has the degree of resolution to make the statement that "the individuals comprising the pool of low-wage workers earned $125/month less on average after the change", then it is problematic, but it is far from clear that this is what the study says.

The GP's quote seems to capture the only hard data given -- that employers employed as many people at higher wages as they unemployed at the lower wage, meaning a net increase in wages paid by employers given that unemployment did not change.

[kevinnk]

http://www.nber.org/papers/w23532?utm_campaign=ntw&utm_mediu...

Total payroll fell, so your explanation is not consistent with the data.

[andrewla]

Thanks -- that was enough information to lead me to [1], which is one of the author's PDF of the paper to avoid the paywall.

Here's the initial justification for the $19 bucket:

> While the preponderance of evidence suggests that a low-wage threshold slightly above the statutory minimum poses little risk of miscoding jobs as lost when they have really been promoted to higher wage levels, in our preferred specifications we report findings based on a relatively conservative $19 threshold. The $19 threshold is roughly twice the initial value of the minimum wage, a level beyond which cascading effects are less likely to occur.

And here's the payroll falling statement (from the abstract):

> Using a variety of methods to analyze employment in all sectors paying below a specified real hourly rate, we conclude that the second wage increase to $13 reduced hours worked in low-wage jobs by around 9 percent, while hourly wages in such jobs increased by around 3 percent. Consequently, total payroll fell for such jobs, implying that the minimum wage ordinance lowered low-wage employees’ earnings by an average of $125 per month in 2016.

Table 3 contains more of the breakdown, but does not refute my point -- the assumption about the threshold and mobility around that is not obvious, and the "reduction in payroll" only makes sense if we restrict the payroll changes to just $19 and hour or lower, the exact thing that I am pointing out as potentially a bad assumption. Under the alternative model, payroll reduced for that group of jobs (less that $19/hour) because that payroll was still spent (potentially on the same people) but at a higher wage class.

[1] https://evans.uw.edu/sites/default/files/NBER%20Working%20Pa...

[joveian]

Thanks for the non-paywall link! It looks like a useful part of a broader analysis (and maybe more states should require the collection of hours worked data).

If it isn't the same people being employed at higher wages, it is still interesting that a sufficiently large pool of more skilled workers exists to take those jobs.

I wish they did the same all wage levels analysis for all workers that they did for the restaurant industry. Without that analysis it isn't clear if the net zero overall effect applies broadly. If it does, that would seem to indicate that wages in general are being held artificially low due to the availability of low wage labor.

Also, since it is a significant increase in a relatively short time, I wonder if the results of the change have stabilized yet. Hopefully they will repeat the analysis in the future.

It is interesting that one of the tests they thought might show a difference ended up failing their falsification test. I wonder what is going on there.

Since housing rental prices have been going up rapidly in Seattle, I wonder if their result could be explained by a sufficient number of low wage workers fleeing the Seattle area and moving to other parts of Washington (or a slower influx of new low wage workers than previously). Maybe something else to look into for someone who knows where to find that kind of data (if it is gathered).