David R. Henderson  

Message to Michael: There's No Tooth Fairy

David Beckworth has a new podc... Inform Me...

In an article today, Shikha Dalmia goes after Democrats' attempts and proposals to raise the minimum wage by a lot. She quotes my thinking about monopsony. But she also made me aware of work by U.C. Berkeley economist Michael Reich that he and colleagues did for the Los Angeles City Council. So I looked at that work more carefully.

It contradicts basic economics.

One of Reich et al's arguments is that the added wages to low-wage workers will have a multiplier effect by increasing demand for goods and services. But how can this be? Those who keep their jobs will have more income. They'll also, presumably, have fewer benefits to the extent that employers can cut benefits. The reason: the mandated wage increase would not make them magically more productive and there was a reason their wages were what they were: they reflected low productivity.

But let's assume that benefits don't decline dollar for dollar with wage increases. So these lower-wage employees who keep their jobs will have more real income.

But who pays for that income? Employers. Employers will either take it out of hide, so that they will have less to spend, or will increase prices so that consumers have less to spend, or both. Reich et al discuss this but somehow conclude that the net effect is positive. They write:

To conclude, we find that the benefits of the proposed minimum wage law will largely outweigh the costs in Los Angeles City, and when the larger region is considered, the net impact of the law will be positive.

Their conclusion is driven by their model rather than by basic economics.

It's possible that Reich et al are assuming that some of cost to employers is borne by employers that have a span beyond Los Angeles or Los Angeles County, so that a firm operating in Los Angeles that is based in, say, San Jose, will pay some of these costs. So the reduced income the San Jose firm suffers does not result in reduced demand in L.A. But it will result in lower income and demand somewhere. If they are assuming that some of the cost is borne outside the Los Angeles County area, then their policy amounts to "beggar thy neighbor." It's not a policy for increasing overall real incomes.

How do we know that increasing the minimum wage will cause real incomes overall to fall? Because at a higher mandated wage rate, especially one raised a lot, the number of people employed will fall. With fewer people working, there is less output. With lower output, there is less real income.

If the effect that Reich et al discussed were dominant so that the gains in real income exceeded the losses in real income, there really would be a good reason, from a straight real income viewpoint (putting aside the morality of the issue), to raise the minimum wage to $20 an hour. But there's no tooth fairy.

By the way, for referral to evidence of some hanky-panky by Reich's co-author Ken Jacobs, see this.

Comments and Sharing

COMMENTS (25 to date)
Thomas Hutcheson writes:

Models with multipliers need unemployed resources. An while it still probably true that the economy is running a bit below potential (as seen from the employment to population 17-64 still below that of several years ago) it's a bit of a stretch to think that the extra wages of workers will be enough to bring these out. It would be better to raise the minimum wage conditional on the EITC not being high enough to achieve the same income. This would bring out the second best nature of the measure.

Anonymous writes:


What's your response to the research on happiness, which (as far as I'm aware) suggests that earnings have very little impact on happiness, but that there is a stark difference in happiness between the employed and the unemployed?

Anonymous writes:

@David Henderson

I've had a question brewing for a while, making explicit the argument that ThomasH and others have hinted at many times. Economically ignorant pundits often treat supply or demand curves as being totally inelastic: that you can have the government enforce a price, and the result will be the same quantity is supplied as before, just at the new higher or lower government mandated price. That's obviously silly, but - at least, the argument goes - don't economists often make the opposite mistake, by assuming that all curves are perfectly elastic?

To use a topic that there seems to be more economic consensus on than minimum wages, I'll take rent control. What will be the effect of rent control? Maybe it will be a noticeable drop in the supply of housing, but how do we know? Does the argument that obviously rent control is a bad idea depend on the assumption that supply of housing is very elastic? If so, is that a reasonable assumption, and why? Is it because all curves become more and more elastic the further into the future you measure them - and in which case, how far in the future should we look? Or is it something else? Or is this argument based entirely on empirical evidence, gathered on the real effects of real rent control - and if so, how do we know how it generalizes? Maybe governments should lower rents just a little, and this will have no measurable effect on the supply of housing? Why not?

I'm speaking here as an observer of economics who has never studied it seriously. What's the counter-argument you and other free marketeers would make against the claim, "Absent empirical evidence we have no way of knowing whether a supply or demand curve is likely to be elastic or inelastic, hence no way of knowing whether a price floor or ceiling will have a positive or negative effect"?

Patrick R. Sullivan writes:

[Comment removed. Please consult our comment policies and check your email for explanation.--Econlib Ed.]

James writes:

Amazing. Same fallacies over and over again. One would think that to be consistent with all the hand-wringing about special interest groups, the left would at least nod to the fact that this guy has organized with unions, who have then asked that they themselves can get jobs below minimum "to compete". The fallacy of labor v capital is a mirage that ideologues can hide behind while they do real damage to the most vulnerable of our society--some principles they have.

Tim Niblett writes:

I'm a computer scientist, not an economist, but your 'basic economics' seems wrong to me.

I can see that more money going to a low paid worker means less to the customer, employer (or supplier). But surely what is done with the extra money is important?

Let's assume that you can spend it (with some time lag), save it or invest it. I would have thought the best for the economy, at the moment when demand is low, would be for the low paid worker to have it and spend it immediately.

I can see an argument that overall employment might be lowered, but my understanding of the imposition of a minimum wage in the UK is that that didn't happen. Again, just from basic logic, there is likely to be a big difference between a universal requirement to pay a minimum wage and individual employers raising wages -- the equilibria could be completely different.

If 'basic economics' says that this is nonsense then, purely based on my (no doubt incomplete) logic, I'd say that basic economics is a pretty shallow and unrealistic model for the world as it is. Is it feasible for you to test your assertions against data?


Noah writes:

Naturally, GDP would grow to infinity if everyone were paid $0 an hour.

Or maybe increasing the minimum wage to its 1965 level is wrong, because clearly a minimum wage of $100 an hour would not work. Hence, return thinking that GDP would be infinite if wages were $0.

Jon Murphy writes:

Here's the thing: even with the rosy estimates, they find " a cumulative net reduction in GDP of $135 million by 2017 and $315 million
by 2019, or 0.1 percent compared to a scenario with no city minimum wage increase.
These effects on the level of economic activity correspond to a cumulative net reduction in
employment in Los Angeles City of 1,552 jobs by 2017 and 3,472 jobs by 2019, or 0.1 and
0.2 percent of all employment, respectively." (page 2).

This suggests to me that, should we use basic economics, the losses will be even worse than he is currently predicting.

Dustin writes:

"Their conclusion is driven by their model rather than by basic economics."

What is "basic economics" based on? Even if basic economics is more generally accepted, this argument amounts to 'my model is better than their model'. I was hoping that your cutting rebuttal would include a reference to some peer reviewed literature that unequivocally demonstrates just how rigorously tested and validated your model basic economics is.

From the DOL

Myth: Increasing the minimum wage will cause people to lose their jobs.

Not true: In a letter to President Obama and congressional leaders urging a minimum wage increase, more than 600 economists, ..." (to long to copy the complete response)

I frankly have no idea what the truth of the matter is, but I am entirely comfortable with uncomfortable questions of an orthodoxy that has failed substantiate its own claims.

J Mann writes:

1. If I had to guess, I'd imagine they have a story involving velocity - that because minimum wage workers are likely to spend a greater part of the transferred income as consumption rather than investment, monetary velocity increases -> underpants gnomes -> profit! (I admit that I'm vague on the last part.)

Alternately, maybe it's David's hypothesis -- that the authors imagine that low wage workers spend a greater portion of their income locally.

2. Growing up, I was always told that Henry Ford started the middle class revolution by increasing his workers' wages. Is that story still generally accepted, controversial, or generally rejected?

Jon Murphy writes:


What is "basic economics" based on? Even if basic economics is more generally accepted, this argument amounts to 'my model is better than their model'.

Well, no. It would be incorrect to describe "basic economics" as a model. You're confusing cause and effect. The model arises out of basic economics, not the other way around. Basic economics is based upon tried-and-true observations, theories, and reasoning. Models are created to allow us to visualize and make predictions based upon basic economic theory.

LK Beland writes:

The paper seems pretty straightforward. An increase in the income of low-wage workers will slightly increase prices, decrease profits and decrease land rents.

The author basically claim that profits and land rents flow, to some extent, out of the LA area, while low-wage workers mostly live and spend in the LA area. This seems to make sense. Of course, the beggar-thy-neighbor critique may certainly be valid.

Now, as to whether reducing land rents and profits and increasing low-wage workers income throughout the US would be stimulative, I would guess that it largely depends upon your preferred macro model specifications.

(Also, from the paper, it seems that this minimum wage raise might actually help low-income adults, to the opposite of claims that it will help teenagers from middle-class families.)

invisible finger writes:

[Comment removed pending confirmation of email address. Email the webmaster@econlib.org to request restoring this comment. A valid email address is required to post comments on EconLog and EconTalk.--Econlib Ed.]

Benjamin Cole writes:

I think California needs free markets in labor and in property development.

I look forward to a series of posts on the evils of property zoning and the need to abolish single-family detached housing districts.

Imagine the boom in California if 50-story condo towers could be erected in high-income neighborhoods--- that is where developers want to build.


Jon Murphy writes:


I can see that more money going to a low paid worker means less to the customer, employer (or supplier). But surely what is done with the extra money is important?

Excellent question. What's going on here is what we economists call the Broken Window Fallacy.

Essentially, it is easy to see that the increase in the income of some workers is fueling economic activity. They are not taking that money and burying it in a jar; they are spending or saving some of it. That generates economic activity; this is where the multiplier effect essentially comes from (it's more complicated than that but I don't want to bore you with a very long lecture).

However, what remains unseen is how that income would have been spent/saved by it's original owners: the customers, business owners, etc. As you rightfully note, they are now out some income they would have otherwise enjoyed. Because of this, there is the lost multiplier from their lost economic activity. Transfer payments like this (that is, actions that cause just transfers of wealth rather than the actual creation of wealth) don't have a multiplier effect (unless one assumes those transfers, in their original owners' hands, would be buried in a jar).

The Original CC writes:

Correct me if I'm totally missing something, but wouldn't some economists argue that the decrease in employee turnover is a real economic gain? In other words, if 10% of the minimum wage employees are "reshuffled" every month, that could create a deadweight loss. If you force every minimum wage employer to pay more, that loss should be diminished (since turnover would be lower).

I understand that turnover isn't always unproductive (lots of people leave one job for a job that's a better fit), but surely within the fast-food industry, turnover could be largely wasteful. At least I thought that's what some people were arguing.

Puzzled writes:

If you want to decrease turnover at your business, you pay more. Pushing up wages at all companies will not decrease turnover, though. Why should it?

We could increase the minimum wage and pass a law saying you can't lay people off so it doesn't increase unemployment. Then we'll get increased demand and increased prices, making the whole thing rather pointless, but it still would not have an impact on people leaving their jobs.

Dustin writes:

Jon Murphy,
I immediately thought of a basic supply and demand model of labor and assumed this is what David had in mind. He does make a prediction, after all, which is what models are for:

"How do we know that increasing the minimum wage will cause real incomes overall to fall? Because at a higher mandated wage rate, especially one raised a lot, the number of people employed will fall. "

Says who? Maybe David is not referring to any functional model and is only considering the underlying reasoning, in which case the argument becomes 'how I and some other people envision the economy working is better than your empirical model.'

Stephen writes:


'What is "basic economics" based on? Even if basic economics is more generally accepted, this argument amounts to 'my model is better than their model'. I was hoping that your cutting rebuttal would include a reference to some peer reviewed literature that unequivocally demonstrates just how rigorously tested and validated your model basic economics is.'

Do you honestly have doubts that people react to prices? Do you honestly need to read this in a peer-reviewed journal to understand that it is true?

This isn't rhetorical. I'm really asking you these questions.

Dustin writes:

I am inclined to think that an increase in the minimum wage will lead to transient job losses. But I'm wrong about lots of things, and I don't read academic blogs to confirm my priors. Until there is unequivocal basis to support the reasoning that minimum wages lead to job losses, it is just a theory that may well be incorrect. And in the meantime, I've no qualms with academics exploring alternate theories. The minimum wage movements across the US should provide fertile data for on the impact of minimum wage increases.

But so what of my view? An appeal to "basic economics" is a common sense fallacy. Asking for some real science should not be controversial.

Stephen writes:


Price theory is not a common-sense fallacy, it's widely accepted theory with a tremendous amount of empirical and even laboratory evidence. If you want evidence particularly related to price floors on labor, I recommend reading the Neumark and Wascher meta-study on minimum wage, which concludes that most studies do show employment disincentive effects.

If you have a generally accepted theory with a lot of support, as we do with price theory, and you want to disprove that theory, then you need very good evidence. Extra-ordinary claims require extra-ordinary evidence. A somewhat-murky empirical record does not nearly rise to that level.

On the other side of the coin, you don't need to post a complete bibliography every time you reference such a theory. That doesn't make it fake science, though.

David R. Henderson writes:

Price theory is not a common-sense fallacy, it's widely accepted theory with a tremendous amount of empirical and even laboratory evidence.
Well said.

Dustin writes:

Correct, "basic economics" is not a common sense fallacy. However, appealing to it as a rebuttal is the very definition - let's call it the 'common economic sense fallacy'.

Anyway, my general point is being lost. Trained economist are familiar with basic economic theory. There are probably better ways of rebutting their empirical work than referring to unsubstantiated theory that the economist is already aware of. For example, critique the substance of the work itself.

Separately, here is Neumark in 2015:

Recent research shows conflicting evidence on both sides of the issue. In general, the evidence suggests that it is appropriate to weigh the cost of potential job losses from a higher minimum wage against the benefits of wage increases for other workers.


or John Schmitt in 2013:

The weight of that evidence points to little or no employment response to modest increases in the minimum wage.


Suffice it to say, convincing empirical support for the effect of a minimum wage is lacking and more empirical work is needed. We should also be dispassionate about results that conflict with our priors.

Stephen writes:


What do you believe a common sense fallacy is? Your understanding of it seems to be different than mine. Simply citing "basic economics" isn't a common sense fallacy, because it's a shorthand way for referencing the reason and evidence that underlie it. Furthermore that's not even what happened in this blog post, because David Henderson even went on to elaborate on the ways that the results do not comport with traditional economic models. A common sense fallacy would be saying "the minimum wage must cause job losses, because that just makes sense!" It might be true, but phrased in that way is not logical.

Moreover, price theory is by no means unsubstantiated. If it were, I'm going to go out on a limb and guess that every economist wouldn't be as familiar with it as you claim they are. It's actually extremely well supported.

Finally, I agree that we should dispassionately consider evidence that conflicts with our priors. However, that includes accurately assessing the prior probability and weighting the evidence appropriately. Like I said before, extraordinary claims require extraordinary evidence. The existence of dis-employment effects associated with the minimum wage is not some 50/50 coin-flip, because the claim that they do not exist contradicts the well-established and reasonable theory. Disproving them requires very clear and convincing evidence, which is a standard we are not close to approaching. From the same Neumark letter you cited:

"Recent research using a wider variety of methods to address the problem of comparison states tends to confirm earlier findings of job loss. Coupled with critiques of the methods that generate little evidence of job loss, the overall body of recent evidence suggests that the most credible conclusion is a higher minimum wage results in some job loss for the least-skilled workers—with possibly larger adverse effects than earlier research suggested."

Dustin writes:

My point is that we should not dismiss empirical research simply because it does not comport with commonly held economic views.

For the matter at hand, further work is needed as the real-world impact of a minimum wage hike extremely unclear. Sure, price theory says that minimum wages will cause job losses. However, the DOL (supported by ~600 economists) says this an outright myth. Some studies (such as the one I linked to earlier as well as the study that is subject of David's post) suggest there is little or no measurable effect. Narayana Kocherlakota just had a post in Bloomberg claiming that under the right conditions a boost to the minimum wage may actually increase AD and stimulate a recovery. Clearly there is no consensus.

Again, let's evaluate the research on its own merits and not how well it aligns to Intro to Econ. If the findings support price theory, great. If not, great.

Comments for this entry have been closed
Return to top