The Seattle Minimum Wage Study

The Seattle Minimum Wage Study, a study supported and funded in part by the Seattle city government, is out with a new NBER paper evaluating Seattle’s minimum wage increase to $13 an hour and it finds significant dis-employment effects that on net reduce the incomes of minimum wage workers. I farm this one out to Jonathan Meer on FB.

This is the official study that was commissioned several years ago by the city of Seattle to study the impacts of raising the minimum wage, in a move that I applauded at the time as an honest and transparent attempt towards self-examination of a bold policy. It is the first study of a very high city-level minimum wage, with administrative data that has much more detail than is usually available. The first wave (examining the increase to $11/hr) last year was a mixed bag, with fairly imprecise estimates.

These findings, examining another year of data and including the increase to $13/hr, are unequivocal: the policy is an unmitigated disaster. The main findings:

– The numbers of hours worked by low-wage workers fell by *3.5 million hours per quarter*. This was reflected both in thousands of job losses and reductions in hours worked by those who retained their jobs.

– The losses were so dramatic that this increase “reduced income paid to low-wage employees of single-location Seattle businesses by roughly $120 million on an annual basis.” On average, low-wage workers *lost* $125 per month. The minimum wage has always been a lousy income transfer program, but at this level you’d come out ahead just setting a hundred million dollars a year on fire. And that’s before we get into who kept vs lost their jobs.

– Estimates of the response of labor demand are substantially higher than much of the previous research, which may have been expected given how much higher (and how localized) this minimum wage is relative to previously-studied ones.

– The impacts took some time to be reflected in the level of employment, as predicted by Meer and West (2016).

– The authors are able to replicate the results of other papers that find no impact on the restaurant industry with their own data by imposing the same limitations that other researchers have faced. This shows that those papers’ findings were likely driven by their data limitations. This is an important thing to remember as you see knee-jerk responses coming from the usual corners.

– You may also hear that the construction of the comparison group was flawed somehow, and that’s driving the results. I believe that the research team did as good of a job as possible, trying several approaches and presenting all of their findings extensively. There is no cherry-picking here. But more importantly, without getting too deep into the econometric weeds, my sense is that, given the evolution of the Seattle economy over the past two years, these results – if anything – *understate* the extent of the job losses.

This paper not only makes numerous valuable contributions to the economics literature, but should give serious pause to minimum wage advocates. Of course, that’s not what’s happening, to the extent that the mayor of Seattle commissioned *another* study, by an advocacy group at Berkeley whose previous work on the minimum wage is so consistently one-sided that you can set your watch by it, that unsurprisingly finds no effect. They deliberately timed its release for several days before this paper came out, and I find that whole affair abhorrent. Seattle politicians are so unwilling to accept reality that they’ll undermine their own researchers and waste taxpayer dollars on what is barely a cut above propaganda.

I don’t envy the backlash this team is going to face for daring to present results that will be seen as heresy. I know that so many people just desperately want to believe that the minimum wage is a free lunch. It’s not. These job losses will only get worse as the minimum wage climbs higher, and this team is working on linking to demographic data to examine who the losers from this policy are. I fully expect that these losses are borne most heavily by low-income and minority households.


I've always said if you want a real free lunch improvement in the lives of the working men and women, you should advocate a minimum wage of $100 per hour.

We probably live in a nonlinear world, and effects will be different at each minimum in each local economy.

But certainly this should give pause to anyone who thinks you can just pick a higher number and get a broad benefit.

"If we make being poor illegal, people will stop being poor."

We don't do that overtly in America, but that isn't to say it is stamped out. I think Alex did a piece here on the nickle and dime penalties put on the poor, and how they add up. There was also an article recently about a southern state that arrested people for small crimes, and then sent them to work below minimum wage, which included cleaning crews for the state capital. Or the governor's mansion?

It's the like government said "what are people's complaints about private prisons, and how can we use that as an instruction manual?"

I was hoping for a twitter link. What's up with the mediamatters talking points?!?! Give us random tweets!

No it wouldn't. Poverty is relative.

If we gave everyone $1,000,000, within a very short time some people would have 100 Million, some would have 50 others 20 and some would have nothing. It is how life is.

If you said it was illegal for someone to be poor then you are implying that the state would have to give them cash, to stop them being poor. Very quickly people would set about qualifying for the cash.

It won't work

They kind of actually did this in Russia after the 1991 revolution under Yeltsin. They distributed "vouchers" - essentially shares of stock in the state owned industries to the public in order to privatize them. Almost immediately, con artists developed various ponzi schemes and scams, accumulated people's shares, and within a few years, ownership of most industries ended up in the hands of the "oligarchs", many of whom were former criminals from the black market.

At least that's how I remember it, I'm sure the details aren't exact and there was more going on than just that.

Something much like this happened in the 1790s when Alexander Hamilton declared that Continental debts would be paid off in full ... AFTER enough of his friends had gathered up enough of those debts at pennies on the dollar. It then and there established the Full Faith and Credit of the US Government, a development celebrated in history classes everywhere ever since, even as said Government is now borrowing itself into a pit from which it will NEVER be able to climb out.

Ignorance is strength, quoth Orwell.

But certainly this should give pause to anyone who thinks you can just pick a higher number and get a broad benefit.

Lots of things should give people who believe that pause, but that's not one of them.

"I've got a headache."

"Take two aspirin. That will help."

"Well, if two will help, I guess I'll just swallow the bottle. That'll fix things."

Uhhh moronic conservative economic analysts can't analyze

There is such a thing as a Free Lunch!

It actually expands employment in a virtuous circle! Win win win win win.

Unless you believe in free lunches, supply = wages = demand = m6 = 335xi = f-22 = 45. QED.

Is paying consumers higher wages so they can pay more for lunch a free lunch? Or is cutting wages in order to have lower income workers buy more lunches at the same price a free lunch?

Is gdp bought by labor costs the free lunch, or is gdp bought by labor costs that are a smaller fraction of gdp price the free lunch.

Ie, is gdp paid for with labor costs equal to 80% of gdp, or 50% of gdp, not a free lunch?

Ie, If I offer to pay you $1 too fix lunch that I will sell to you for $2 so you get a free lunch, will you take the job?

A mother makes her child a pbj: economic activity? What if the child makes it himself?

EPI's first point isn't so good as it completely misses the argument in the U of Wash paper. The more important claim in the U of W paper (rather than EPI focusing on whether the statutory increase is "large") is actually the measured increase in wages is smaller than the statutory increase. Using the statutory increase biases your elasticity down, which is the problem with previous research according to the U of Wash study.

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. ---EPI

Well, as I always say, why is the topic always problems of the minimum wage and never the problems of propertY zoning?

"Uhhh moronic conservative economic analysts can’t analyze"

Just the idea that there are such things as "conservative economic analysts" should give us pause! Or "liberal economic analysts".

I like how they put the gist of the article in link text itself, so that nobody need even be bothered to click on it to find out what it says.

If wages were raised as much as wages were increased from 1933 to 1969 from 1981 until today, ie, 36 years of FDR paying too much to labor vs 37 years of Reaganomics, then

Production Workers Hourly Compensation
(nominal dollars)
1933 $0.44
1970 $3.72

1981 $10.00

The equivalent hourly compensation would be $85 per hour.

The minimum wage in 1969 was $1.30, cut in 1966 from the FDR signed law's minimum of $1.60, or 1.30/3.72, so the minimum wage would have been $3.49 in 1981in "theory", and the actual minimum in 1981 was $3.35.

So, if Reaganomics results in higher growth than FDR/Keynesian economics, the current production wages should be $85 with a minimum wage of $30 no more onerous than the minimum in the era of FDR economics.

While the minimum of $30 today is not $100, it's based on a production wage of $85 which is close to your $100 an hour.

Only if you argue that consumer spending is inversely related to wage income can you argue that growing wages slower results in higher economic growth. If that is your theory, should all wages be driven to zero including your own wages, in order for you to be better off consuming much more and thus driving higher gdp growth?

Biggest test of success might be to ask minimum wage workers (employed or not) whether they like it (can always wait for better paying job and hang on to it). What you really want is a true market based labor price which can be set only by fairly balanced bargaining between ownership and labor unions -- coming to California (and others) soon:
* * * * *
In BALLOT INITIATIVE states, it typically takes only 5% of the number signatures of registered voters of how many voted in the last governor's election to put your initiative on the ballot. (OR, CA, MO, MI, OH, OK, CO, NE ND, SD, MT)

Check the numbers of who should line around the block to sign an initiative making union busting a felony:
-- nationally, bottom 45% income share has dropped from 15% to a penurious 10% over two generations (as per capita income doubled).

Does that mean bottom 45% are ahead absolute terms: 66% of twice as much? Not across the board: incomes are sorted on a slope. That leaves 15% behind in absolute terms: why we have a $7.25/hr fed min wage -- down from $11/hr (adjusted) in 1968.
* * * * * *
A little labor price perspective:
(2013 dollars)
yr..per capita...real...nominal...dbl-index...%-of


Apparently my charts did not transfer correctly. :-(

"Check the numbers of who should line around the block to sign an initiative making union busting a felony:"

Does this mean the supporters of this measure want to make the Constitutional right of Freedom of association a crime?

And Seattle's economy was red hot during the period in question.

Looks like the demand curve slopes downward, who knew?

Kshama Sawant., Econ Ph.D. NCSU seems to have missed that lecture. Among others.

"The 'high road' Seattle labor market and the effects of the minimum wage increase - Data limitations and methodological problems bias new analysis of Seattle’s minimum wage increase" - at

"I want those people to get paid more."


"And I want people I hate to pay for it."

Well, now we have a problem.

We can make people at the low-end get more money, but we can't make them into middle- or high-income earners.

"Increase EITC. Pay for their health care via government. Reduce employment taxes." Those are readily understood. "Make a third-party value their labor more." That one is out of our power.

Are you sure the Seattle conception was not "we'll all pay for it?"

Surely most voters are not themselves minimum wage workers, thus they are more payer than payee.

They just didn't expect their own marginal decisions (fewer coffees each month?) to work out this way.

In one mental model, the employee is worth a huge amount of money, but the rapacious employer uses power imbalances to squeeze all that out of the employee and pay them little.

In another mental model, different employees have different values, and employers make a conscious decision to use $11 labor over $15 labor.

If you are committed to the first model, there is no way a minimum wage backfires. You are just stopping the employer from exploiting the employee.

But if that model is wrong, and the second model is right, then when you force employers to choose the $15 labor, well, the employer can probably adapt. They get a slightly better quality of employee. It wasn't the ideal decision for them, but they'll spend less time dealing with problem employees.

And that employee whose labor is now illegal to offer? He is now unemployed. Maybe he can retrain and become a better employee using his time off. If you spend your time around high-achievers, you will be likely to believe that given a break the employee will pour energy into training and becoming more valuable. If he's already in his late 20s, though, he's done all the education he wants. School is torture for him. Either his labor is not likely to significantly increase in value, or any increase in value is going to come from on-the-job training, gaining the workplace experience necessary to make him more valuable to his employer.

Rather than peering into everyone's lives to determine who really needs this help, just reduce employment taxes and create a wage subsidy. The people are still being supported by the government, but they are providing most of their own support, and are engaged in their community, and are kept out of trouble, and display a behavior for their children that work is normal.

I'm just saying that these increases are usually accompanied by a claim that "your coffee will go up by X cents," or some such. People who voted thought they wouild buy that coffee and sustain those jobs.

If fewer hours are worked, and employers are not leaving money on the table, it is because those increased prices did actually reduce sales.

Was there any loss of sales documented?

The median reader of thinkprogress buys 0 Big Macs a year. Telling them those prices are going to go up and they'll be perfectly happy. The poor shouldn't be eating at McDonald's anyway!

It's just another example in the long line of making poor people illegal. Make their food illegal, make their housing illegal, make their income illegal. At the end, you will have no poor people (because they all moved away) and thinkprogress can report on the massive success.

If I were discussing minimum lot sizes as a way to keep out the poor out of my school district, this would be obvious and offensive to them. But if I say your labor has to be worth $15/hour to work in my city, somehow I'm a hero. Even though it has the safe effect: keep out the riff raff.

Now you are confusing me. I thought I had the market analysis in place.

If sales are not reduced, why a reduction in hours? Are you now assuming incompetent managers and previous featherbedding?

It could be less sales. Or it could also be a forced productivity increase. There are many possibilities here. They could have replaced two $11 workers with one $15 worker and $8 machine and made up the difference someplace else: maybe they did increase prices and the customer base absorbed it just fine, but that still resulted in disemployment to the $11 workers.

From a macro POV, that's not that bad, since productivity numbers will go up. But if it was sold as a way to help the low-value worker, this looks like a botch.

With wage subsidy, while I'm not sure how much of the benefit will go to the low-value employee, I at least know the change will be positive and not negative. I'm sure some surplus generated by a wage subsidy would go to employers, but 1. the point is to help the low-value worker, so the fact that someone else benefits doesn't bother me too much, and 2. the economy is competitive, so other employers can show up and bid away that advantage. (I'd like hiring to be as easy as possible. Let me go to my state's DOL website and say "I'm hiring Joe Blow for $8/hour for 8 hours today.")

This hints at the problems of a future world with even better automation. The better the automation the worse the minimum wage effect. Of course an EITC faces an uphill battle for the same reason.

An EITC is less expensive than having full wards of the state, at least.

The Seattle conception is to drive the least capable out of the city. And that is something they are willing to pay for.

I'd be interested in a comparison of that study and the recent Berkeley study ( Why such different results?

We seem to have another case of dueling economists...who should a layman believe?

If Professor Tabarrok is going to "farm out" his commentary to Jonathan Meer and Meer says, "...*another* study, by an advocacy group at Berkeley whose previous work on the minimum wage is so consistently one-sided that you can set your watch by it, that unsurprisingly finds no effect.", shouldn't one of them explain the differences between the two reports?

Why is the Berkeley report "one-sided" and the NBER report not?

Note that the Seattle City Council signed onto the NBER's general methodology when the law was passed.

Exactly. The NBER's was real science. Pre-existing hypothesis preregistered, methods and all. The Berkeley study is just the usual post hoc HARKing.

Yep. It should be noted that one could just as easily say Meer's work is similarly "consistent" and thus influenced which study he regards as by "an advocacy group" (a great boogeyman phrase if I've ever seen one) and which is by legitimate researchers.

No, actually one could not just as easily say that

The article that Alex quoted states that the Berkley group is a minimum wage advocacy group. You can't simply write it off simply because of who wrote it - but it does warrant being skeptical of their methodology.

Yes and no. It is only a "minimum wage advocacy group" in that its research has generally supported a minimum wage. So it's kind of a case of "OUR conclusions are clean, and THEIR conclusions are ill-gotten, because they differ from ours."

Not to sanction the Berkeley analysis per se (haven't read either yet), but let's recognize mud-slinging for what it is.

Meer suggests the Berkley study was commissioned explicitly as a corrective antidote to the SMWS, and it's hard to disagree. The Berkley study focuses on the food-service industry only, for which the SMWS also found no effect, and then scales its findings to apply to all sectors. It seems purposefully designed to find and exaggerate neutral or positive effects only. And IRLE doesn't exactly have a reputation for dispassion on this subject. Which isn't to say their research should be dismissed out-of-hand but it does warrant heightened skepticism, I think. As a layman I find the SMWS more credible.

Not having looked at either study or the credibility of either group, they had different data available. The Berkley group followed a methods that's got a longer trackrecord but only looks at a subset of low-wage jobs. The method lauded in the Facebook post Alex quotes looks at a larger subset of low-wage jobs. That can be better or worse depending on the characteristics of the extra data. Typically more data is better, but if the additional data has unexpected characteristics, it can lead to weird conclusions.

On one hand, the study discussed in the Facebook post produces a result that runs against much of the published literature. On the other hand, theory (and common sense) tell us that if you keep raising the minimum wage, you'll eventually reach a point where it really starts hurting employment at the floor.

Prior to reading anything about this, my prior would've been pretty diffuse on where the point that the minimum wage started biting was for large cities. Somewhere between like $9 and $20. Now I put a bit more weight on the $11-$13 range. There will be more studies, both using data from Seattle and from around the country where other cities, counties and states have upped their minimum wage. I expect to see a mixed bag throughout where it helps low-wage workers a lot in some places and in others you see stuff as bad as the headline for this paper.

I find the notion that other groups shouldn't study this data from different angles to be absurd. There are so many ways to slice this stuff that the only way you can show that your result is robust is if you get five or ten groups approaching the data from different angles and finding largely the same thing. You don't find truth by doing the one, perfect analysis. Truth is found by finding the trend that emerges from multiple good approaches.

If truth is a consensus, then there is no truth (a conclusion I support, by the way)

Ummmm...only if "there's no truth" is a consensus conclusion?

"I don’t envy the backlash this team is going to face for daring to present results that will be seen as heresy."

This is the reality of 2017, the triumph of wishful thinking.

What to make of these data? I'm tentatively working on a theory, it sounds crazy, but imagine the concept of a "price floor", i.e. the government says X cannot be bought or sold for less than $Y. What might that do? All very mysterious.

As an academic economist, my experience has been that publishing a paper that upends conventional wisdom is generally much better rewarded than one that confirms it, regardless of the political implications. This career incentive easily outweighs the "heresy" concern.

This paper is in fact a clear example of this. Its surprising result (in that is contradicts much of the past literature) is exactly why is getting it so much play right now.

If it found no effect, we probably wouldn't have heard about it. Case in point: did anyone here hear about last week's Berkeley paper on the subject until today?

How does this not just confirm conventional economics? Who didn't know about supply and demand curves. There have been papers out there that go against this, but most people know they were advocacy papers.

"(in that is contradicts much of the past literature) is"

That's not true. The past literature all points to rises in minimum wage above the local effective minimum wage are associated with upwards pressure on low wage unemployment. The effect is most pronounced with teenage workers and minorities.

alex (if I may) - if you're going to blog about this - as you should - you should also blog about kansas experiment. otherwise you look like an ideologue/partisan hack. just saying/fyi. if you have legit explanation for why kansas was not a disaster, great, i'd love to lean/be enlightened. i don't recall seeing any reference on MR yet. i appreciate how open-minded MR is in general..

This is some Grade A hackery right here.

Always fun to listen to party-bots when their price-control schemes fail. Again.

A quick google finds an article saying that the Kansas City minimum wage won't hit $13, where Seattle seems to be having trouble, until 2023. So it's rather too early to be blogging about research into the consequences of the Kansas City minimum wage hike.

How much of a role did workers deciding to work less hours factor into all this? In general, it doesn't seem like any of these economist pundits want to place any value at all on free time, hence the constant horror at "less hours worked". There's probably a much better way to judge the costs and benefits of 9% less hours worked vs. 3% more hourly pay, than labeling it as an "unmitigated disaster".

In order for this to be the case, you'd need to assume that minimum-wage workers are people who have more than enough money(so much that they're willingly reducing their income), and significant control over their own hours. In other words, they're the exact opposite of the people a minimum wage is intended to protect - high-income people with good working conditions do not need special government programs to protect them.

I don't mind people working fewer hours and getting more leisure - it's a decision I may well make myself in some contexts. But I want them to have the choice, and this sure doesn't look like them having a choice.

Yes. It is wonderful! They have all that extra time to enjoy eating cake!

Low wage people don't choose to work less. They typically don't have much choice of jobs, and are juggling multiple part time jobs.

That's an interesting point. Juggling multiple part time jobs sounds unpleasant. If I were doing that and my wages increased I might quit one job to get the time back, even if the wage increase didn't totally balance out the lost hours. Especially if that job caused child care or transportation costs that made the effective wage even lower.

The problem is that the workers didn't get the memo from Robert Reich. They were supposed to spend their increased earnings at their employer's so that the increase would pay for itself.

The supply / demand curves still work. What a surprise.

oh a study about minimum wage in Marginal rev... i wonder which way the study will find

"The numbers of hours worked by low-wage workers fell by *3.5 million hours per quarter*. This was reflected both in thousands of job losses and reductions in hours worked by those who retained their jobs."

This begs the question, what were those workers doing for those 3.5 million hours and who/what is doing them now? Are we looking at a forced productivity increase? Perhaps ZPM employee downsizing? Perhaps businesses that were close to the edge of profitability had to shutdown? There are a lot more options, but it is an important question before we declare "the policy is an unmitigated disaster" although I agree that it is a disaster for a lot of people.

It sounds like what you're saying is, we can have open borders and low wages, or lower immigration and higher wages. I wonder which one people will choose?

That's a false choice and has nothing to do with the study quoted.

It's the subtext of: "I fully expect that these losses are borne most heavily by low-income and minority households."

When discussing wages keep fully in mind that Tabarock looks favorable on India as a good place to be. It IS relevant. He wants to turn the USA into India.

If you believe that the demand for unskilled labor is very elastic then a small change in minimum wage will cause a large change in employment but a large exogenous change in employment (immigrants) will only slightly depress wages. If you think it's inelastic then the reverse.

In a sense this result shows that we're in the possible universe where high wages (though a high minimum wage) are impossible but open borders are possible. But in another sense it shows that opening borders wouldn't tend to lower wages from where we are now.

This paradox is resolved, iirc from labor Econ (the most idiotic and mathematically challenged area of Econ) due to labor market segmentation and the labor supply elasticity of native unskilled labor being large.

Demand elasticity of labor for unskilled native labor is large (and non linear). Make a minimum wage of 15 an hour and there will be disempmoyment in the official figures. However, many employers will switch to illegal labor (note to racists, this is not necessarily immigrants) and pay cash under the table at the market rate (actually lower than the market rate). Employers face a two tiered labor supply system, as the price floor ( or reservation wage, e.g. Fruit picking) of native wages rises, the relative cost of hiring illegals/natives under the table goes down.

As welfare increases: Medicaid, section 8, disability, etc the official reservation wage goes up for native unskilled labor. Many more work under the table.

Here's a question -- to what extent would we expect high local minimum wages to affect gentrification? Aren't mid-range and low-end businesses (that both hire and cater to lower wage workers) much more likely to be affected than those on the high-end? If so, over time, wouldn't we expect poor citizens to pick up and leave as the businesses they patronize start to disappear? One of the complaints I've heard consistently is that poor people are driven out, in part, as rising rents cause neighborhood places to be replaced by the kinds of upscale shops and restaurants their new wealthier neighbors prefer. Shouldn't minimum wage increases and rent increases have similar effects?

Probably less so than rent. The employees of those high-end stores can commute.

The expectation is that there will be 100% gentrification of the poor, resulting in social mores esthetic tastes and political views in line with Manhattan's Upper Westside.

Of course that's not the declared intent or expectation of politicians and voters supporting high minimum wages in Seattle. But if gentrification is the result and Seattle becomes ever more like San Francisco, will supporters regret their choices? Or will they be happy with a gentrified city full of high-end service businesses that can afford the higher minimum wages with most poor people and their jobs and the déclassé businesses that cater to them all pushed off stage to the periphery? If you wanted to keep poor people out of your city, wouldn't a very high minimum wage be an effective strategy (especially because it can be sold as a means of helping the poor, and the poor themselves will probably even buy the explanation and never make the connection)?

Seattle is the dark heart of white supremacy in America.

Speaking of gentrification.
The other day I saw a sign reading "Gentrification is colonialism" - on the window of a trendy bookstore/coffee shop - in an Italian neighborhood in America.

I'll just leave you to ponder the levels of irony inherent in that.

The white man's burden.

If a minimum wage makes sense, wouldn't a maximum one as well? Robinson Cano of the Seattle Mariners is making over $45,000 an hour to wander around between second and first base while occasionally popping out to centerfield. He could probably get by just fine on a more egalitarian $10K per hour.

Why stop at "wages"? Lots of Seattle-ites have substantial unearned income, too. And even greater true/economic income.

That said, there are serious proposals to limit the range of wages an org can offer, typically pitched as a CEO pay limit.

That said, there are serious proposals to limit the range of wages an org can offer, typically pitched as a CEO pay limit.

Every previous attempt to limit CEO pay has backfired completely. But I'm sure this one will work.

"That said, there are serious proposals to limit the range of wages an org can offer"

"Serious". LOL

Sure, but why should he? Does it really make the world a better place if some billionaire baseball team owner makes the $35,000 per hour instead? Because it's not like ticket prices will go down.

One thing that confuses me is the extremely low (3.2%) unemployment in Seattle over this time period coupled with the supposition of massive job losses. Are we to suppose that the unemployment would be even lower than this had the minimum wage law not gone into effect? "Full employment" is conventionally thought of as somewhere between 3 and 5 percent, IIRC. So we're saying Seattle is at full employment but also lost thousands of low-paying jobs?

The story seems more complex to me than the Facebook post Alex is quoting would indicate.

Check out table 3. Looking at Single-Site Establishments, the number of jobs and hours worked under $19 an hour did fall, while overall jobs and hours worked rose. It's likely that increased employment in the group averaging over $40 an hours numerically overwhlemed job losses in the $15 an hour group, though these are almost certainly completely different groups of people with rather different demographics.

Is it possible that pushing up the minimum wage also resulted in some people at $17 or $18 per hour being pushed over $19 per hour and thereby falling into a different category - people such as team leads and more experienced workers who employers felt should get at least the same raise as the minimum wage workers they work with?

City level unemployment rates are kind of bullshit for the purpose you're talking about--especially in expensive cities. If five poor people lose their jobs and are forced to move out of town, and nothing else happens your city level unemployment is unchanged, and your average earnings per capital just went up! It's awesome right?

I'm certainly no labor economist, but in Seattle, year-over-year unemployment is down, total employment (# of workers, not the rate) is up are up, and low-wage work is, per this study, down. I don't see how changes to the minimum wage can fully explain that particular cocktail. Once you start picking off bits and attributing it to minimum wage while ignoring the rest, you're in very murky water.

Could this just be a case of other forces in the local economy (namely, high rents making low-wage employment unaffordable in the city, even at the higher minimum wage combined with a red-hot job market for those making higher wages) entirely swamp whatever impact the minimum wage law had? Ironically, this would kinda square with the Berkeley study's results.

"I’m certainly no labor economist, but in Seattle, year-over-year unemployment is down, total employment (# of workers, not the rate) is up are up, and low-wage work is, per this study, down. I don’t see how changes to the minimum wage can fully explain that particular cocktail."

The Seattle economy is booming for Tech workers. However, the minimum wage was raised enough to overcome the general booming economy and result in fewer low wage workers. That's a pretty simple explanation that matches the data.

Can you imagine the effect of this rise in minimum wage as it goes to $15 per hour? And if the economy in the region experiences a recession?

When you have to pick out two concurrent trends and divide the outcomes up by which trend effects which outcome, the overall explanation suffers. Surely there's cross-contamination. The point is that the story is not as clean as is being portrayed.

An interesting counterfactual would be what if the region was experiencing a recession now instead of hypothetically after future wage hikes. Well, certainly this paper (and OP) would attribute the recession to the hike in minimum wage! But since there's a booming economy, they ignore inconvenient measures of city-wide economic performance (however meritorious they are) and look at only those that agree with their priors.

Is this paper...and/or Meer's polemics...likely to change any policymaker's mind about minimum wage legislation?

"Finally, the mechanisms activated by a local minimum wage ordinance might differ from those
associated with a state or federal increase; it is reasonable to expect that policies implemented at
a broader geographic scale offer fewer opportunities to reallocate employment in response."

Probably not.

"These findings, examining another year of data and including the increase to $13/hr, are unequivocal: the policy is an unmitigated disaster."

This finding seems completely premature. I think we should hold out till the minimum wage hits $15/hr in 2021 before we can determine the boundary conditions for "unmitigated disaster".

While the paper has some limitations, it does hit the literature pretty hard:

"Our analysis reveals a major limitation of conventional elasticity computation methods,
however. When comparing percent changes in employment to percent changes in wage,
conventional methods assume that the impact of a minimum wage policy on wages is equal to
the statutory increase in the minimum. This is often a necessity, as analysis is performed using
datasets that do not permit the estimation of policy impacts on wages themselves. We show that
the impact of Seattle’s minimum wage increase on wage levels is much smaller than the statutory
increase, reflecting the fact that most affected low-wage workers were already earning more than
the statutory minimum at baseline. Our estimates imply, then, that conventionally calculated
elasticities are substantially underestimated. Our preferred estimates suggest that the rise from
$9.47 to $11 produced disemployment effects that approximately offset wage effects, with
elasticity estimates around -1. The subsequent increase to as much as $13 yielded more
substantial disemployment effects, with net elasticity estimates closer to -3.


Most prior studies compute employment elasticities by dividing regression-estimated
percentage changes in employment by the percentage change in the statutory minimum wage.
Applied in this case, this method would use a denominator of 16.2% (i.e., ($11-$9.47)/$9.47) for
the first phase-in period, and 37.3% ($13-$9.47)/$9.47) for the second. The conventional
method clearly overstates the actual impact on wages given that many affected workers’ wages
are above the old minimum but below the new. This method is also unsuitable for evaluating the
impacts on workers who began over the new minimum wage but are nonetheless affected by
cascading wage increases (defined as the range of either $11 or $13 to $19 per hour). In column
2 of Table 8, we use the conventional approach for computing employment elasticities and find
estimates in the range of -0.08 to -0.28 (averaging -0.20). This range is high but not outside of
the envelope of estimates found in prior literature (see Appendix Table 1).44
Thus, computing
the elasticity based on the Ordinance’s impact on actual average wages suggests that the
conventional method yields substantial underestimates."

For all those curious about the "dueling reports," the Washington Post has a pretty good breakdown of both sides and the controversy. The headline is click-baity, but I found the contents of the article quite useful for thinking through the issues raised.

There is no controversy.

The study omitted multi-location businesses? That would mean chains. How the HECK do you omit chains from a minimum wage (!) study?

"The Washington researchers said they couldn’t track multisite employers because those businesses don’t identify the specific locations where their employees report to work. However, in separate surveys, where the researchers talked to both types of business, the ones with multiple locations were actually more likely than single-site employers to say they had cut jobs and hours when the wage went up. That suggests that there may have been an even greater job loss for low-paid workers in those bigger businesses. ."

Its a data limitation. From the paper:

"The data identify business entities as UI account holders. Firms with multiple locations
have the option of establishing a separate account for each location, or a common account.
Geographic identification in the data is at the account level. As such, we can uniquely identify
business location only for single-site firms and those multi-site firms opting for separate
accounts by location.13
We therefore exclude multi-site single-account businesses from the
analysis, referring henceforth to the remaining firms as “single-site” businesses. As shown in
Table 2, in Washington State as a whole, single-site businesses comprise 89% of firms and
employ 62% of the entire workforce (which includes 2.7 million employees in an average
Multi-location firms may respond differently to local minimum wage laws. On the one
hand, firms with establishments inside and outside of the affected jurisdiction could more easily
absorb the added labor costs from their affected locations, and thus would have less incentive to
respond by changing their labor demand. On the other hand, such firms would have an easier
time relocating work to their existing sites outside of the affected jurisdiction, and thus might
reduce labor demand more than single-location businesses. Survey evidence collected in Seattle
at the time of the first minimum wage increase, and again one year later, increase suggests that multi-location firms were in fact more likely to plan and implement staff reductions.14 Our
employment results may therefore be biased towards zero."

Another article on the study:

Last line of abstract:

"We estimate an effect of zero when analyzing employment in the restaurant industry at all wage levels, comparable to many prior studies."

So, isn't the paper consistent with higher minimum wage leading to a shift from low-wage to non-low-wage jobs?

"So, isn’t the paper consistent with higher minimum wage leading to a shift from low-wage to non-low-wage jobs?"

Absolutely. But with a loss of jobs near the minimum wage level. IE employers higher fewer higher wage employees with higher productivity.

"IE employers higher fewer higher wage employees with higher productivity."

My point is, the "fewer" is not implied by the paper. Effect on total employment is 0, not negative.

True, I was thinking about the overall effect on employment, whereas you were referring specifically to the restaurant industry.

Kinda of sorta. They do suggest that now firms are hiring workers with experience rather than workers who need training from scratch. However, I think more importantly, they also argue that using an industry like restaurants as your low wage population will attenuate your results. In contrast they can identify who is actually low wage:

"Using restaurant or retail employees or teenagers as proxies for the entire low-wage labor
market might lead to biased minimum wage effects. Intuitively, a sample mixing jobs directly
affected by the minimum wage with others for which the price floor is irrelevant would generally
skew estimated impacts towards zero. Isolating one industry, such as the fast food industry, may
lead to downwardly biased wage and employment effects due to heterogeneity in wages in the
industry (i.e., some workers whose wages are above the minimum wage will be misclassified as
belonging to the “treatment” group). The estimates capture the minimum wage’s net effects on
all restaurant employees, not the effects on low-wage employees, which would likely be
stronger. Similarly, using teenagers may lead to artificially large employment estimates as this
group omits other low-wage workers, particularly those that have a stronger attachment to the
labor force and are full-time full-year workers, for whom the wage-elasticity of demand may be smaller. On the other hand, since some teens earn wages well above the minimum, including
them in the sample would lead to artificially low estimates of the impacts for that demographic


"Column 1 of Table 9 repeats the main results findings from Column 1 of Table 6, and is
included as a point of reference. Moving from Column 1 to Column 5 of Table 9, we make one
change at a time to evaluate the sensitivity of our results to various modeling choices. In
Column 2, we use the same specification as in Column 1, but restrict the analysis to hours in
low-wage jobs in Food Services and Drinking Places (NAICS industry 722). The results are
comparable to those shown in Column 1 for all industries – if anything, the results show larger
decreases in hours, particularly when the minimum wage was raised to $11, suggesting roughly a
7% decline, although two of these three estimates are insignificant. Moving from Column 2 to 3,
we switch the focus to headcount employment, the outcome used in most prior literature. Again,
these results are quite comparable, suggesting that nearly all of the reduction in hours worked by
low-wage workers in Food and Drinking Places is coming from a reduction in jobs rather than a
reduction in hours worked by those who have such jobs.
In Column 4, we shift from examining low-wage jobs to all jobs in the restaurant
industry. Here we see a dramatic change: the effect on all jobs is insignificant in all quarters and
averages precisely 0.0% in the last three quarters when the wage increased to $13 per hour.
Thus, by using the imprecise proxy of all jobs in a stereotypically low-wage industry, prior
literature may have substantially underestimated the impact of minimum wage increases on the
target population. Finally, column 5 returns to evaluating effects on total hours, but now for all jobs in NAICS 722. While the estimates continue to be insignificant, they are now more
negative, averaging -3.3% in the last three quarters. This result is consistent with Neumark and
Wascher’s (2000) critique of Card and Krueger (1994). "

I should have noted above that this is essentially the difference with the Berkeley study or appears to be so.

Apologies, I left out the key point in my comment. Meant to say:

"Isn’t the paper consistent with higher minimum wage leading to a shift from low-wage to non-low-wage jobs, and no net employment effects?"

1. Individuals' total income may have dropped, but those individuals' also worked fewer hours. From their perspective it could still be a net win, depending on how many people prefer $125 less per month in exchange for N fewer hours worked. Do we know if they're taking those lost hours and using them to work 2nd jobs outside the city of Seattle? If so, then income may actually have increased. (But only because those jobs existed outside the city).

2. If jobs were lost, and the people who previously occupied those jobs can now collect unemployment, it could be a net win for the city in the sense that a bunch of new money is flowing in from the state coffers. Ultimately bad for everyone, but possibly a rational move for the City of Seattle.


Let's remember working is normally a cost, not a benefit, especially for jobs in that income range.

Knee-jerk you say? Hm...

The trouble with all minimum wage research is trying to figure out whether anyone doing it is reasonably neutral.

Could this simply be a mechanical result? Even if there were no disemployment effect at all, a minimum wage increase would nevertheless decrease the number of "low wage" workers by definition through wage increases (and hence decrease the number of hours worked by them). In other words, this result is also consistent with zero employment impact, or even the small positive one found in the Card/Kreuger paper.

Before someone tries to attack me as being a "hack", I would note that I am opposed to minimum wage increases.

According to Table 2 of their paper, they drop large chain businesses (think McDonalds and Applebys) from their analysis, which account for ~40% of employment in the industry. These excluded firms are on average 5 times larger than the ones included in their analysis.

I think it is quite likely these firms moved in and swooped up market share as small restaurants cut hours (classic regulatory capture). If true, the analysis is incomplete, as it only includes the (negative) impacts on small, shrinking firms and ignores the (positive) impacts from the big companies.

I would like to see their analysis if they didn't drop big firms. I bet it would attenuate their result substantially, if not eliminate it.

“The Washington researchers said they couldn’t track multisite employers because those businesses don’t identify the specific locations where their employees report to work. However, in separate surveys, where the researchers talked to both types of business, the ones with multiple locations were actually more likely than single-site employers to say they had cut jobs and hours when the wage went up. That suggests that there may have been an even greater job loss for low-paid workers in those bigger businesses. .”

If the numbers to back that claim up aren't included in the study, is that quote anything more than an anecdote?

"The University of Washington study excluded workers at companies with multiple locations—meaning McDonald’s, Starbucks, and the other big and small chains that account for about 40 percent of the overall workforce and a huge number of minimum-wage jobs—narrowing the scope of the results, Zipperer and Schmitt noted. The study also seemed to imply that the minimum-wage hike caused a boom in high-wage employment, a seemingly impossible feat. (It seems unlikely that a business would have reacted to a pay hike for a minimum-wage worker by paying many of them $19 an hour, after all.) It in addition had no way to tell if Seattle’s employers were switching to contractors, as opposed to employees, to avoid some provisions of the minimum-wage law; if that had happened, those workers would have dropped out of the data set."

I would also point out that Seattle has seen some of the largest rent increases in the country over the last few years.

We know that increased demand drives up prices for real estate. How much of the rent hikes in the bottom quartile of the real estate market can be attributed to increased nominal purchasing power created by minimum wage hikes?

That's funny, because the last time I checked Seattle's unemployment rate had dropped to only 3.4%, down from 4.8% the last year. Not much of a job disaster

Brilliant, assuming minimum wage workers live in the city and don't commute from lower price areas.

Guliani-esque in function if not form. If we raise the minimum wage to 20 an hour, the icky poor will leave.

I can't wait for the Voxsplainer about the rising school segregation that tautologically follows.

The Seattle-Tacoma-Bellvue MSA encompasses several thousand square miles surrounding the actual city of Seattle.

I run a store in Reno, NV. It can't decide whether to make or lose money from month to month, but due to long term hope and a love of the industry we choose to keep running it. Over the years, I've dropped about 40 grand into it and between myself and my partner we have worked there without pay for about 16,000 hours (I am semi-retired from other profitable businesses). At this point our time investment is down to about 100 hours per month, but we are paying multiple part timers to be there, and we are only paying minimum wage. It's all the store can afford. There are some efficiencies we could add that would allow us to increase pay to $13/hr, but by doing so we would have to end certain services that, while loved, are not making money yet.

Thus, I'm one of those business owners who can claim that a higher minimum wage would potentially end my business; however, I still believe people should have more opportunities. Right now the only people who can afford to work part time at my store are people who are using it to "earn a bit of extra cash." Of the companies I own, my favorite is the one which wouldn't survive without free labor or cash infusions. There is more to life than profit. My profitable businesses are far less emotionally satisfying than the un-profitable one. This is why I am a HUGE believer in high taxes beyond $60k (with major writeoffs) and a Universal Basic Income. Every day I see people slaving at jobs they hate. How many of them would contribute to society doing what they love if they weren't afraid for their survival?

Higher minimum wages are the less evil choice, but I think all of society is blessed when increased social programs require people EARN less to have a good life.

"Every day I see people slaving at jobs they hate. How many of them would contribute to society doing what they love if they weren’t afraid for their survival? Higher minimum wages are the less evil choice,..."

"There are some efficiencies we could add that would allow us to increase pay to $13/hr, but by doing so we would have to end certain services that, while loved, are not making money yet."

You seem to be saying that you are Evil.

I live in downtown Seattle and am self employed. I don't put stock in any of the studies so far because the economic boom here is so large and widespread (throughout the entire metro area) that I doubt most economic theories have much if any predictive power at the margins. Sure, housing is expensive - it's a boom town and it seems that a lot of people want to move here. Basic supply and demand. But the employment situation is complicated by several factors that are quite different in WA than in other states - no state income tax, but Employment Security and L&I taxes that tend to favor companies with small numbers of highly compensated employees. Seattle is also phasing out the "square footage" tax that they had for several years. So until they make comparisons to another boomtown with similar demographics and similar tax environments these studies are just extrapolation on extrapolation.

I will say that the idea that there were millions of hours lost and hundreds of millions in losses doesn't pass the smell test. Businesses close and businesses open all the time, but most business owners do not appear to be suffering right now. It's quite the opposite.

The losses are for the low-wage workers, not the businesses

The data is obviously inconclusive. Seattle, San Fran and LA should raise the minimum wage to $20 an hour. Those cities are so expensive it's only fair. Then we will get some real data. /Sarc off. If people really believe minimum wages help the low income people just keep turning that dial up.

Mississippi has the mandated Federal minimum wage of $7.25. Those 3 cities should easily be able to afford 2 to 3 times that rate. Why wait until 2021 or later to raise the rate to $15 per hour, why not raise it to $15 January 1st of 2018?

I have an actual question:
I just don't understand any world in which reducing people's hours save more money than firing one of them. I would totally understand if the result was 1 in 10 people lose their jobs, but I don't understand what manager in the world would instead decide to keep the overhead of 10 employees and reduce their work time by an hour. It doesn't make any sense at all.

Is there any business in the world where it is cheaper to have 10 employees working 9 hours versus 9 employees working 10 hours?!?! Even if you just factor in the hassle of scheduling one more employee.

This simple fact makes me really question the entire study.

Just a theory.

You have a restaurant. Your busiest hours are from 6-10. You structured your shifts from 4-12 knowing that for the first and last 2 hours of the shift, your employees wouldn't be working that hard, but it would be worth having them there at the lower wage.

Now your labor costs have jumped 30%. So you start some of your staff at 5pm and send them home at 11pm.

You don't fire some, because you still need the head count to cover the peak time rush. However you still cut your labor bill.

This was very much how things worked when I worked as a pizza delivery man for Dominos.

They started letting people go when business slowed down, but there always was some discretion by the manger about when to start cutting. Sometimes they would start to early and you would get slammed and other times too late and you would just be standing around.

There was always some optionality in keeping folks around a little longer. At a higher wage you have an incentive to cut earlier.

Agency bias? Nobody liked firing people.

"Is there any business in the world where it is cheaper to have 10 employees working 9 hours versus 9 employees working 10 hours?!?!"

Sure, if there are hour thresholds that kick in and require some added benefits, such as employer provided health care insurance. Also, if the government kicks in money per employee, a high head count may generate higher tax deductions.

Joshua Grass has stumbled on a surefire way for a major baseball team to save on payroll.

Instead of fielding 9 players for a 3 hour game, just have 6 players play 4.5 hours.

I believe the minimum wage is bit too low due to various reasons. I do Forex trading only and with broker like OctaFX, I find it easier with their ECN account which can be operated with as low as 20 dollars and comes with no issues over slippage, re quote or any such issues, so it really keeps me relaxed and comfortable with doing thing and also boost me with doing things and allows me to perform really nicely in all situations.

I'm not sure that's the best synopsis of the study. That writer is obviously very emotional about the issue.

For example, it shows statistically insignificant changes in labor demand until 2016 Q1. Consistent with the idea that the raise to $11/hr did not have a strong negative effect. So it might be simply that the point of unbalancing of forces is a bit lower than some people thought. I'm also not too concerned about the controls as some people, as a major goal of a minimum wage is to prevent booms from distributing wealth unequally. At the very least, the minimum wage system didn't do much to stop that natural concentration of the rewards to booms, so seems like a less effective policy to satisfy those goals.

The only way out is if workers in low-wage industries didn't just drop hours, but rather picked up more hours in other industries that better capture the boom time gains. If I were a minimum-wage-true-believer I'd start collecting data from households not industries. I'd build an app and have people put in their total pay each month and the source of that pay, or maybe set up check-cashing places and offer them a small decrease in the check cashing fee to use an image of the check in research.

actually do both. You can use the check-cashing to calibrate the self-report, but the app would be a lot easier to get large data volumes, and from non-check sources.

"For example, it shows statistically insignificant changes in labor demand until 2016 Q1. Consistent with the idea that the raise to $11/hr did not have a strong negative effect."

The rise to $11 was in 2015 Q2. There were immediate effects and then statistically large effects by 2016 Q2. So, I while it did not have a strong negative effect, it clearly had an effect.

That being said, the rise to $13 had a more pronounced effect per "The Seattle Minimum Wage Study" and only took effect for workers at large companies without medical benefits starting 6 months ago.

I would expect the $14 per hour raise in Jan 2018 and the $15 per hour raise in Jan 2019 to have an even greater effect. By 2021, it will be in effect for all workers in the city.

This will be a boon for commercial automation as the tech companies race to deploy ordering kiosks and low end robotics.

$13/hr took effect Jan 1st 2016. There were two-quarters of statistically insignificant effect inbetween the raise to $11/hr, then Q1 2016 is the first statistically significant effect. There certainly appears to be some non-linearity in these effects.

"$13/hr took effect Jan 1st 2016."

That's for Large Employers only. The Small Employers (which this study was examining) went to $12/hr in 01/2016 and then $13/hr didn't take effect till 01/2017.

Correction: The study specifically says $13/hr in 01/2016. So, I assume that the study was using Large Employers with single site locations. So, the $13 for 01/2016 is correct.

Looks to me like there could be some serious problems with the credibility of The study:

I went over this study with my Microeconomics class this morning. It fits extremely well with the elasticity and total revenue models that are taught right after supply and demand.

The lower end of the labor demand curve is inelastic while the upper end is elastic. From the elasticity measurements on page 31 of the study, increasing wages from $9.47 to $11 showed something between unit elastic to slightly elastic. Hence little to no change in total income. Moving from $11 to $13 moved further up the curve and had an elasticity of about -3. Hence total income fell.

All of that is perfectly consistent with basic economic models.

I would suspect that other locations and other time periods would have both different demand curves and economies operating at different places on the curve. Hence different elasticities and results.

It would be interesting to know if an economy was left alone where it would end up on the labor demand curve. Would it consistently end up at a unit elastic point or somewhere above or below? Perhaps there is no equilibrium.

Comments for this post are closed