Tuesday, June 27, 2017

The Seattle Study: Increasing the Minimum Wage as a Way to Boost High Income Jobs

As labor market mavens all know by now, the University of Washington team chosen by the city of Seattle to evaluate its minimum wage law has issued a new report.  This one is particularly juicy since it covers the increase from $11 to $13 an hour, which moved Seattle into new territory, beyond what has been studied elsewhere.  The report makes much of its use of Washington State data which include not only numbers but also hours worked, allowing (in this respect) a more precise analysis of the effect of changes in the statutory minimum on employment.

The headline result is that the elasticity of hours worked to changes wages actually paid is in the vicinity of -300%.  The key paragraph is this:
Our preferred estimates suggest that the Seattle Minimum Wage Ordinance caused hours worked by low-skilled workers (i.e., those earning under $19 per hour) to fall by 9.4% during the three quarters when the minimum wage was $13 per hour, resulting in a loss of 3.5 million hours worked per calendar quarter. Alternative estimates show the number of low-wage jobs declined by 6.8%, which represents a loss of more than 5,000 jobs. These estimates are robust to cutoffs other than $19.  A 3.1% increase in wages in jobs that paid less than $19 coupled with a 9.4% loss in hours yields a labor demand elasticity of roughly -3.0, and this large elasticity estimate is robust to other cutoffs.
This has got the labor econ blogosphere quite excited: finally, after years of published studies that largely downplayed the labor demand disincentive effects of minimum wage laws, a report has been issued that finds immense negative effects—vastly larger in fact than those that have appeared in the past.

The backdrop to this, of course, is the economic performance of the city of Seattle itself, which has been about as strong as any city in the country.  During the period of the latest minimum wage increase Seattle has experienced essentially full employment, as reflected in an unemployment rate of about 3%.  Thus, any negative impact in one part of the city’s economy had to have been offset by positive impacts elsewhere.

And this in fact is also a finding of the Seattle minimum wage study, although its authors don’t mention it.  They restrict their sample of affected workers to those in the low wage labor market, and they employ a range of cutoffs to see how truncating the sample in different ways affect their results, but their empirical methods intrinsically apply to workers at all wage levels.  This is because their strategy for identifying employment impacts is to use various control groups, actual or synthetic, and compare employment between Seattle and these controls before and after the change in the minimum wage.  The employment impact is whatever comes out of this comparison.

Well, guess what?  The treatment-versus-control methods that generate employment losses in the lower-wage segment of the labor market are nearly exactly offset by employment gains in the higher wage segment.  This stands to reason, because Seattle as a whole is doing great: it hasn’t suffered overall from the rise in minimum wages, so dips in some parts of its economy imply bumps in others.

The calculations that make this explicit are performed by Ben Zipperer and John Schmitt of the Economic Policy Institute.  They do this for the overall city economy and also for the restaurant sector.  With its narrower focus, this second approach is especially informative.  The key paragraph here is:
The spurious results are clear in the case of the restaurant industry, as we illustrate in Figure B, where the authors’ own methodology and estimates imply that the Seattle minimum wage increase caused an incredible 20.1 percent growth in restaurant jobs paying above $19.00 per hour. While this number is not directly reported in their paper, it can be precisely inferred from their other results. To make this inference, we first note that when Jardim et al. focus on the restaurant industry, they estimate that the minimum wage increase to $13.00 caused restaurant jobs paying less than $19.00 hour to fall by an average of 10.7 percent (see their Table 9, averaging the estimates provided for the employment fall in 2016). At the same time, they find that the minimum wage caused essentially zero change in the number of all restaurant jobs, regardless of their wage rate. Because jobs under $19.00 comprised 65.4 percent of the restaurant industry prior to the first minimum wage increase (see Jardim et al.’s Table 3 for the 2014Q2–2015Q1 period), and because these jobs shrank by 10.7 percent while overall employment held steady, it follows that Jardim et al.’s estimates imply that the Seattle minimum wage increased the number of restaurant jobs paying over $19.00 per hour by about 20.1 percent.
Remember that these employment estimates are based on comparisons of Seattle to its constructed controls; measured differences are assumed to be due solely to the minimum wage hike that took place in Seattle but not in the comparitors.

One interpretation is that the Seattle study’s methodology didn’t sufficiently control for factors that have caused upward movements in wages (moving workers out of lower and into higher wage categories) in Seattle compared to other communities.  That’s what the EPI folks think.  I prefer to take the results at face value: by increasing the minimum wage we can, by some currently unknown process, cause a big upward shift in wages, not just around the minimum, but all the way up to the stratospheric reaches of the labor market.  That negative elasticity for the lower-paid is fully offset by a positive elasticity for the middle and upper class.



rosserjb@jmu.edu said...

Interesting, Peter, and certainly offsets much of the commentary going on all over the labor econoblogosphere indeed. The restaurant finding is consistent with earlier studies going back to the original Card and Krueger study that started all the whooping that maybe increasing the min wage does not necessarily lead to job losses, with that being a study of fast food restaurant industry specifically.

Sandwichman said...

I don't suppose it ever occurred to the University of Washington team that the pay range from $13 to $19 is only 75% of the range from $11 to $19? Have they not ever heard of a wage ladder? Increasing the minimum wage might result in some compression of the increments in a wage ladder but to ASSUME 100% compression of all the rungs on the ladder would be either disingenuous or brain dead.

Sandwichman said...

A few points, digging deeper into the report.

First, the reference to the "$13 minimum wage" is misleading because the study looked at establishments in which a good proportion of the low wage employees were earning under $13 an hour because they were with small employers or because their employer provided benefits. Around 5% of employees in the establishments analyzed were UNDER $13 an hour in the third quarter of 2016 and fully 22% of the low wage workers were under $13 an hour.

Second, the employment fluctuations indicate a clear seasonal trend with the summer quarters being the busy season. Folks, this is Seattle. It rains. But 2015 was an unusually sunny year. Officially Seattle's hottest on record!!! Their baseline quarter is 2014.2.

The 9.4% loss of hours is based on their comparison with a synthetic control group. Whether this control is superior to a simple year-over-year comparison is debatable. This is NOT a randomized control trial and there is simply no way of knowing what the unknown variables are. By comparison, year over year comparisons for Seattle yields an average loss of hours of only about 4.2% between 2015 and 2016 even with 2015 being the hottest summer on record. Regardless, taking as given their 9.4% hours loss and the 3.1% wage increase they estimate (again apparently by comparison with the synthetic control group) the lucky synthetics get the great privilege, compared to their Seattle compatriots, of working an additional 12 hours a month at the fabulous "overtime" rate of sixty-eight percent of their standard wage. The study portrays this as a great hardship for the Seattlites.

"Importantly, the lost income associated with the hours reductions exceeds the gain associated with the net wage increase of 3.1%. Using data in Table 3, we compute that the average low-wage employee was paid $1,897 per month. The reduction in hours would cost the average employee $179 per month [for 12.3 extra hours], while the wage increase would recoup only $54 of this loss, leaving a net loss of $125 per month (6.6%), which is sizable for a low-wage worker."

$125 divided by 12.3 = $10.22 an hour or 68% of the $15.02 average wage of the under $19 earners.

One thing I couldn't get clear in a back and forth email exchange with one of the researchers is whether they considered different low wage caps for the before and after treatments. The reasoning is this: if the average wage goes up 3.1% in the under $19 an hour group, what happens to the hours of those making $19 in the fourth quarter of 2015? They evaporate, along with the hours of everyone making over $18.45 in the fourth quarter of 2015. That's about 7% of hours disappearing through pure bracket creep, assuming a more or less even distribution of hours of work at each wage increment between $11 and $19. Seven percent is a mighty big number when your talking about a 9.4% decline in hours.

My informant kept repeating that they tried "different wage cutoffs" but never explicitly confirmed that they tried wage thresholds that differed between the before and after cohorts, which was my question.

AXEC / E.K-H said...

The minimum wage debate: a showpiece of economists’s hereditary idiocy
Comment on Peter Dorman on ‘The Seattle Study: Increasing the Minimum Wage as a Way to Boost High Income Jobs’

“In order to tell the politicians and practitioners something about causes and best means, the economist needs the true theory or else he has not much more to offer than educated common sense or his personal opinion.” (Stigum)

Fact is that economists do NOT have the true theory. Fact is, in methodological terms, that economics is axiomatically false. The lethal blunder comes under the label of microfoundations or as Krugman put it: “most of what I and many others do is sorta-kinda neoclassical because it takes the maximization-and-equilibrium world as a starting point.”

The methodological blunder of the minimum wage debate consists in partial analysis and microfoundations. This type of analysis NEVER leads to results that can be generalized, but always to results that change from place to place and from time to time. So, this type of analysis (i) runs directly into the Fallacy of Composition, and (ii), remains forever inconclusive.

Interim result: The traditional microfoundations approach is as false as one can get and has to be fully replaced by the macrofoundations approach.

The correct macro employment equation#1 is reproduced on Wikimedia.#2 From this objective-structural-systemic relationship follows inter alia:
(i) An increase of the expenditure ratio rhoE leads to higher employment L (the Greek letter rho stands for ratio).
(ii) Increasing investment expenditures I exert a positive influence on employment.
(iii) An increase of the factor cost ratio rhoF=W/PR leads to higher employment.

The complete structural-systemic employment equation is a bit longer and contains in addition the public sector and the foreign trade sector.

Item (i) and (ii) cover the familiar arguments about how effective demand affects employment. Item (iii) embodies the macroeconomic price mechanism. It works such that overall employment L INCREASES if the average wage rate W INCREASES relative to average price P and productivity R and vice versa.

From this follows:
(1) The average wage rate has to be prevented from falling because this leads to rising unemployment and deflation. One possibility is to fix a minimum wage rate which increases over time. Note that this is a SYSTEMIC necessity and has NOTHING to do with social policy.
(2) The minimum wage rate has to be implemented nationwide. To implement it locally or for certain branches is absolutely counterproductive.
(3) The implementation has to be done intelligently. It is, for example, stupid to kill the marginal firms with the introduction of a nationwide minim wage.
(4) Given their track record of idiocy, economists have to be kept out of further discussion and implementation.

Minimum wage policy has to be carried out under the macroeconomic condition w greater than p+r+pr, that is roughly speaking, employment increases if the increase of the average wage rate w is greater than the increase of average price p and productivity r.

To make local and partial minimum wage increases will in eternity lead to inconclusive results and keep a bunch of incompetent scientists busy with senseless debate, inconclusive empirical studies, and brain-dead blog posts.

Egmont Kakarot-Handtke

#1 For details see ‘Keynes’ Employment Function and the Gratuitous Phillips Curve Disaster’
#2 Wikimedia https://commons.wikimedia.org/wiki/File:AXEC36.png

bbk said...

Even if the very lowest wage workers were "hurt" by the Seattle minimum wage, why does that mean the policy is a failure or inherently "bad"?

We make policy choices which directly harm the lowest skill, lowest paid workers all the time with the idea being the changes will eventually make the low skilled workers better off even if that happens indirectly through transfers or via the actions of those directly helped by the policy change.

Why couldn't that happen here too, assuming the study's results are accurate? If those workers making higher wages pay more in sales tax because they consume more from their higher pay then the city/state's larger tax revenues could be transferred in part to those who were hurt by the increase in the minimum wage by paying for skills training or just a direct monetary payment. Everyone's a winner!

rosserjb@jmu.edu said...

Oh, Egmont, how hilarious. At least you gave us an actual policy bottom line from your half-baked theory based on a wacko view of protit: if wage increases are sufficiently high, then employment will rise, although you are totally against any local min wages differing from national ones, which, needless to say is being widely ignored by lots of politicians without much input from all the idiotic economists. Do you thus oppose the Seattle increases in minimum wage?

Of course, in your usual unscientific manner, you have completely missed the boat on what is going on here. There is some theoretical debate, with your position resembling that of some Keynesian economists, while orthodox mainstream neoclassicals hold to the old standard that raising the min wage will necessarily reduce employment of low wage workers.

What has happened, with this latest round, has been that this has turned into an empirical issue, a matter of scientific testing of hypotheses, something you seem to know nothing about. There had long been the mainstream textbook consensus that higher min wages hurt employment, and then along came Card and Krueger with their empirical study questioning this, at least in the case of fast food restaurants in New Jersey and Pennsylvania some years ago. The empirical debate has been hot and bothered ever since, with more data, different economstric techniques (something that is either beneath you or beyond your ken), with this latest study from Seattle, the latest entry. Howver, you have nothing of any use to say about this study, just an attempt at a general answer without a shred of any empirical, that is, scientific, backing. We are not surprised.

AXEC / E.K-H said...

Barkley Rosser

Note that the EconoSpeak admin=your sidekick Sandwichman has deleted since Jun 24 the following posts:

The minimum wage debate: a showpiece of economists’s hereditary idiocy

Note on Marshall’s Magic Wand

Economics and the Fallacy of Insufficient Abstraction

The role of labor and business in a well-organized society

In these posts and the references you find the detailed refutation of your unqualified blather.*

Egmont Kakarot-Handtke

* See also cross-references Employment
and cross-references Incompetence

rosserjb@jmu.edu said...

I read your stuff before it was posted, and I have bothered to read a bunch of your other stuff as well. It is unbelievably vacyous and pompouse, all of it. But,for the record, I do not approve of deleting your stuff, no matter how worthless it is.

BTW, you love to call people "idiotic" and "unscientific." Let me simply note on the first point, while they may be wrong, the most serious thinkers in the schools of economic thought that you dismiss because they do not share your obsession with retained esarnings, these people are not or were not "idiotic." People like Adam Smith, KarL Marx, Leon Walras, John Maynard Keynes, Friedrich Hayek, Paul Samuelson, Joan Robinson, and Kenneth Arrow were all brilliant people. So calling them "idiotic" simply makes you look beneath idiotic, aside from just plain rude.

AXEC / E.K-H said...

Barkley Rosser

I agree with you about the brilliance of Joan Robinson which is encapsulated in her assessment of economics: “Scrap the lot and start again.”

The others, who are brilliant in your eyes, will not even make in a footnote of the history of science. With regard to Adam Smith I concur with Schumpeter: “… he had no such ambitions; in fact he disliked whatever went beyond plain common sense. He never moved above the heads of even the dullest readers. He led them on gently, encouraging them by trivialities and homely observations, making them feel comfortable all along.” If this is you definition of brilliance you are probably one of the dullest readers.

For an assessment of the rest of your list see:

Marx, the moron

Walras is long gone

How Keynes got macro wrong and Allais got it right

Hayek and other informationally retarded proto-economists

The father of modern economics and his imbecile kids

How Arrow pushed economics over the cliff

Economics is a failed science and those you call brilliant messed it up.

Egmont Kakarot-Handtke

Jerry Brown said...

"The minimum wage debate: a showpiece of economists’s hereditary idiocy"

Egmont do you ascribe your idiocy to inheritable exogenous factors or to endogenous variables?

AXEC / E.K-H said...


“Few people, and least of all we economists ourselves, are prone to offer us congratulations on our intellectual achievements. Moreover our performance is, and always was, not only modest but also disorganized. Methods of fact-finding and analysis that are and were considered substandard or wrong on principle by some of us do prevail and have prevailed widely with others.” (Schumpeter)

Egmont Kakarot-Handtke

rosserjb@jmu.edu said...

What is really needed for these min wage studies, and I have so far seen nobody suggesting it or offering any data comparable to it, would be panel studies that followed individual workers so that we could determine whether those low wage workers who lost jobs due to the min wage hike simply became unemployed or if they got some of those new high wage jobs, of which many more were created than the number of low wage jobs that were lost. Most of the commentary by those making a big whoop of this study are implicitly assuming that the number of those previously low wage workers who got any of the new more highly paid jobs is zero, so that they simply took the full hit of that lower total income going to the low wage jobs, with that also assuming that they became unemployed, although, of course, it is possible some of them also went across city lines to get jobs in neighboring municipalities, another possible outcome.