Scott Sumner  

Paul Krugman on inequality

Straw Men Rule... Paging Frederic Mishkin...

Paul Krugman reviews Robert Reich's new book in the New York Review of Books:

Something else began happening after 2000: labor in general began losing ground relative to capital. After decades of stability, the share of national income going to employee compensation began dropping fairly fast.
I'd call that a bit misleading. In fact, the share of national income going to labor has risen slightly in the past 50 years, from 68.0% in 1965:2 (the golden age of unions) to 68.1% in 2015:2. The only reason I say a bit misleading, rather than completely false, is that the share has indeed fallen since 2000.

In a comment section to this post, Matt Rognlie points to many flaws in the labor share data, which suggests the recent decline in labor's share is overstated. One issue is the problem of how to impute labor income to the self-employed entrepreneur. (He has several comments, read them all.) Both Rognlie and Kevin Erdmann have pointed to the fact that much of the recent strength in capital income is due to a rise in implicit rents in owner-occupied homes, not explicit "income" as most people think of the term.

All you have to do is look at the enormous decline in real interest rates on 10 year bonds, from nearly 7% in the early 1980s to less than 0.7% today, to realize that capital isn't doing all that well either.

Krugman continues:

Other evidence points indirectly to a strong role of market power. At this point, for example, there is an extensive empirical literature on the effects of changes in the minimum wage. Conventional supply-and-demand analysis says that raising the minimum wage should reduce employment, but as Reich notes, we now have a number of what amount to controlled experiments, in which employment in counties whose states have hiked the minimum wage can be compared with employment in neighboring counties across the state line. And there is no hint in the data of the supposed negative employment effect.

Why not? One leading hypothesis is that firms employing low-wage workers--such as fast-food chains--have significant monopsony power in the labor market; that is, they are the principal purchasers of low-wage labor in a particular job market. And a monopsonist facing a price floor doesn't necessarily buy less, just as a monopolist facing a price ceiling doesn't necessarily sell less and may sell more.

This is cherry picking the empirical results. First of all, studies of the minimum wage reach mixed results on the employment impact. But the more important problem is that the studies generally show higher minimum wages being passed on to consumers in the form of higher prices. (Again, I'm indebted to Matt Rognlie for pointing this out.) If firms really did have monopsony power, and as a result employment did not fall, then there should be no pass through of higher wages in the form of higher prices.

Krugman continues:

Once upon a time, around a third of workers in both the US and Canada were union members; today, US unionization is down to 11 percent, while it's still 27 percent north of the border. The difference was politics: US policy turned hostile toward unions in the 1980s, while Canadian policy didn't follow suit.
I'm tempted to ask what kind of story begins, "Once upon a time"? Union membership was nearly 35% of wage and salary workers in 1954. By the time Reagan took office it was down to 21%. Thus most of the decline in unionization occurred before Reagan took office. Indeed the share of union workers has been declining for 60 years. That's not to say Reagan's policies played no role, but they certainly were not the primary factor.

Despite these objections, I have some sympathy for Krugman's argument that rising inequality (especially at the top) is due to increases in market power. In my view that's partly due to the US economy shifting from a commodity and basic manufacturing economy, to a high tech economy where intellectual property rights form a barrier to competition. At the low end I think that immigration for Latin America has depressed the wage levels of less well educated American born workers, especially men.

My preference would be to address the inequality issue in four ways:

1. Have a lower proportion of low skilled immigrants and a higher percentage of high skilled immigrants---there are plenty in India and China who wish to come here, but also more than you might expect from Africa, the Middle East and Latin America.

2. Weaken intellectual property rights. I don't favor eliminating them, but I'd prefer to keep them only for entirely new inventions, not improvements of existing products. Copyrights need to be made much shorter.

3. Change zoning laws to encourage more building. This will be really hard to do; indeed I think things are likely to get worse, not better.

4. Replace income taxes with progressive consumption taxes and low wage subsidies. Eliminate cigarette taxes. Legalize drugs.

Thus I oppose the progressive agenda of high marginal tax rates on personal income, taxing corporate income, inheritance taxes, higher minimum wage rates, pro-union legislation and tighter regulation of business. Indeed I think less regulation of business, especially eliminating occupational licensing laws, would be beneficial. Our current regulatory regime is simply too complex for the less educated to deal with. Heck, it's too complex for me; I can barely do my taxes.

PS. I also have a new post on Krugman at MoneyIllusion.

Comments and Sharing

CATEGORIES: Income Distribution

COMMENTS (23 to date)
Simon Cranshaw writes:

Wouldn't preference 1 lower inequality within the US at the expense of increasing it globally? What would be the benefit of that?

Steve Fritzinger writes:

I agree with 2, 3, and 4. I don't think low skill immigration is that big a problem so we'll have to disagree on 1.

Regarding the evidence on minimum wage, I think Krugman's case is much weaker than typically stated.

Many MW studies Krugman cites are fatally flawed. From Card and Kruger on, they tend to start the study just before the MW increase takes place. Businesses, otoh, start adjusting as soon as the proposed increase starts gaining ground. The academics are 2 or 3 years late to the party.

roystgnr writes:

It's not just businesses that are allowed (forced) to adjust to proposed laws, either.

If some counties passed laws demanding the execution of people named Paul but nearby counties did not, you can see how the death rate of people named Paul (per-capita, not per-Paul) would in fact become much *lower* than in the counties where no such law was passed. That would still not be evidence that such laws were a good idea.

Edogg writes:

I think 1965 is a misleading point of comparison. The labor share increases to over 73 % by 1970 and looks stable for three decades. (I tried to use the same calculation you did in your linked moneyillusion post.) Krugman looks entirely correct in your first quote.

E. Harding writes:

"Eliminate cigarette taxes. Legalize drugs."

-How would that decrease inequality?

ThomasH writes:

I like your four and would add

5. Shift from wage taxes to a share of that progressive consumption tax (or at least a VAT) for financing Medicaid, Medicare and Social Security.

I'm uncertain about whether bequests should be taxed as ordinary consumption, at a preferential rate, or not at all. Your thoughts?

E. Harding writes:

There's one problem with a shift from wage taxes to progressive consumption taxes: wage taxes are very, very reliable at raising revenue. Will PCTs be so?

ChrisA writes:

EH - VAT is a good 1/3rd of total tax revenues in the UK, see attached informative chart.

So yes personal consumption taxes are quite a reliable way of collecting taxes.

Michael writes:
If firms really did have monopsony power, and as a result employment did not fall, then there should be no pass through of higher wages in the form of higher prices

That seems wrong. If I have a labor monopsony, but my market power is neutralized by a minimum wage, my marginal cost is the wage divided by marginal product of labor (ignoring other inputs, which don't seem relevant to the point). If many firms like mine, but with separate labor pools, compete in the output market, the output price equals marginal cost. Hence a rise in minimum wage affecting all of us and resulting in no employment change should yield a proportional rise in price.

Thomas B writes:

A switch to consumption taxes means a cut in purchasing power of already-saved money. People have saved that money from after-tax dollars (except for dollars in tax-deferred savings). Justice requires that they get a plus-up on savings from after-tax dollars, to cover the consumption taxes they will pay when they spend the money (or, alternatively, a rebate of their consumption taxes if they can show the money came from after-tax savings).

I never see any discussion of this in policy proposals advocating a switch to consumption taxes.

Yes, I realize it would be a complicated thing to do. But it would be the right thing to do. And not doing it would be very unfair.

Nathan W writes:

I can't fathom an argument in favour of eliminating tobacco taxes. Care to explain?

Garrett M writes:

E. Harding writes:

"Eliminate cigarette taxes. Legalize drugs."
-How would that decrease inequality?

I think the idea is that the demand for them is inelastic to price (steeply sloped demand curve), so all these laws do is force poor people to spend a higher percentage of their income on them.

Nathan W writes:

I can't fathom an argument in favour of eliminating tobacco taxes. Care to explain?

Really? You can't fathom? Did you really sit for a while and meditate on the issue before concluding that the view is incomprehensible?

Scott Sumner writes:

Simon and Steve, I don't see the level of low skilled immigration as a problem, I'd lower the proportion of low skilled immigration by raising the level of high skilled immigration.

I agree that the minimum wage probably costs more jobs than some recent studies have suggested.

Edogg, Krugman often says the post war period (1945-73) was the golden age for labor. It's also true that labor's share of national income during 1945-65 was about the same as today. I think he gives his readers the impression that it is not. I think that's a bit misleading, but I suppose some people may view it differently.

And again, it's not at all clear that the data are accurate, Rognlie showed the data may be substantially overstating the fall in labor's share.

E. Harding, These policies are quite regressive, as lower income people spend a much bigger share of their incomes on these products.

As far as wage vs. consumption taxes, wage taxes ARE a form of consumption tax. So it's not either or, they are the same. Rather both wage and consumption taxes are very different from income taxes.

Thomas, Since wage taxes are consumption taxes, they should be kept, but made progressive.

Taxes on bequests should be eliminated. Right now a billionaire who spends $500 million on a yacht is taxed far more lightly than a billionaire who spends $500 million on his children. That seems bizarre to me, but progressives like the system. They briefly put a tax on big yachts, but there were so many complaints that it was withdrawn.

Michael, If you are a monopsonist then a minimum wage lowers the marginal cost of labor, at least in the case where employment doesn't fall. (Someone correct me if I'm wrong, it's been 40 years since I studied micro.)

Nathan, Think about a struggling single mom who spends $1000 a year on cigarette taxes, that could go to food for her family.

Michael writes:
Michael, If you are a monopsonist then a minimum wage lowers the marginal cost of labor, at least in the case where employment doesn't fall. (Someone correct me if I'm wrong, it's been 40 years since I studied micro.)

Thanks for the reply, Scott.

I think you are correct if we are discussing a newly-imposed minimum wage which is just a bit above the previously prevailing monopsony wage. Or, equivalently, if we are considering an increase in the minimum when it was previously non-binding (i.e., below the monopsony wage). But not for an increase from one binding level to a higher one. In that case, as far as calculating marginal cost goes, all the complexity of monopsony disappears: the marginal cost of labor is the wage, both after the increase and before.

Michael writes:

Ah! I was forgetting the context! The increase in minimum has to move it from below the monopsony price to above, or else the impact on employment is the same as in competitive markets.

I hope this clarified for someone beyond me...

Floccina writes:

Labor force participation is already lower among those with less than a high school degree, how do advocates of higher minimum wages see this?

Swami writes:

The data on unemployment effects of higher minimums must consider the following:

1). The minimum wage does not directly affect cost of an employee. It increases the cost of an hour of an employee's time. These are not the same thing. The reasonable short term response to an employer (I should know I once was an employer) is not fewer employees, it is the same number of employees OR MORE each working fewer hours (which is now possible via market imbalance).

2). The second logical response is to upgrade the caliber of employees, from lower skilled to higher, thus benefitting the former at the expense of the latter. The impact on overall employment itself is going to be negligible (it is now a factor of reduced needs due to higher productivity), but not the impact on lower skilled employment levels.

3). The longer term responses include replacing lower skilled workers with automation. This occurs over years or even decades and thus is not measured in any short term study. Other longer term responses include lowering investment in the affected area, or raising prices.

The fact that studies show higher minimum wages is not substantially affecting unemployment is exactly what We should would short term. It is still counterproductive to the unskilled and to market efficiency overall.

Proper-thinking and honest economists need to re-establish control of this discussion. Those attempting to exploit the situation have defined the terms and are using it for political games.

Brett writes:

I think the biggest mistake made when assessing the minimum wage was in assuming the negative employment effects would be even distributed, when they're not - teen and minority employment is disproportionately affected, when there is a negative employment effect.

Although if minimum wages are leading to price increases instead of job losses . . . then you could say that it's working as a subsidy. Workers are getting higher wages, and the costs are being distributed across a broader population with more capacity to bear them. And that can be mitigated with productivity gains and technology over time.

Simon Cranshaw writes:

Thank you for the reply! That makes sense to me now.

G Wilson writes:

Well stated Scott.

Short-term, isolated experiments on the impact of an increase in minimum wage to low-skill labor are misleading as technology (robots, process automation, etc) have not had the time nor the market appeal to product developers to solicit a response. Expand the minimum wage increase to a national level then provide adequate time and we will see a major transformation on these services and the associated labor. Stop by your local grocery store self-checkout if you have doubts.

Intellectual property is a challenging issue - creating an environment where the creator has an opportunity to enjoy a "fair" return on investments in new developments. Our ability to rapidly and cost effectively innovate has evolved to the point where ROI is achieved in a much shorter period of time, so I agree that the length of time needs to be shortened. I would also propose that patents ONLY provide protection to entities actively delivering the product protected to the market, thereby reducing the marketability of the patent itself.

Lewis writes:

Actually Scott I think that things will get better on zoning laws. The reason is that their legitimacy is largely a judicial matter. Once an idea is generally understood by elites to be a good one---e.g., gay marriage---if the idea is even remotely related to a judicial issue then the courts will find a way to override legislatures.

In the 20th century, zoning was thought of as a good, sensible thing by the elite. For a while it was expressly meant to keep poor people and blacks out. Then when that wasn't fashionable it was there to solve environmental problems and promote "character," which is obviously vague enough to be anything. Consequently, the courts discovered that, all along, the "general welfare" clause had given local government a police power to protect a neighborhood's character, on-street parking or pretty much anything homeowners might want.

Today the tide is turning. The conventional wisdom about zoning among liberals "in the know" (e.g., the Economist, the NY Times, SF Chronicle, Mother Jones, etc.) is that a bunch of chauvinist NIMBY's are keeping their million-dollar homes expensive by keeping construction down. That is what we teach all the planning master's students here at my liberal school. The mayors of SF, LA and NYC are all into upzoning. In addition, there are now plenty of new legal options, such as covenants and homeowners associations, that elites can use to protect their property values in sophisticated courts and arbitration away from the public eye, not in sloppy council meetings.

Therefore, I expect judicial opinion will sour on zoning. The SCOTUS "disparate impact" ruling was a step in that direction. Liberal courts will discover that, all along, the freedom of movement and equal protection are incompatible with municipal-level zoning. Conservative courts will discover that it is tyrannical infringement on property rights that the Founders wouldn't have liked. In any case, it will be curtailed.

Ali Bertarian writes:

Scott Sumner wrote: "the share of national income going to labor has risen slightly in the past 50 years."

There is a flaw in the standard measurement of national income that exaggerates the quantity of national income that corporations take, and hence understates labor income. Corporate profits made overseas (now more than 30% of national income due to increased globalization) are counted as national income, but foreign labor costs are not counted as part of national income. See

GU writes:
1. Have a lower proportion of low skilled immigrants and a higher percentage of high skilled immigrants---there are plenty in India and China who wish to come here, but also more than you might expect from Africa, the Middle East and Latin America.

The problem with high-skilled immigrants is that they make life a lot harder for young high-skilled natives. To wit, they tend to (1) take up spots in colleges, graduate schools, and professional schools that might otherwise be filled by (reasonably analogous) natives; (2) they tend to move to the most expensive urban areas and create even more demand, further driving up the already insane price of living in such areas (the supply ain't going up any time soon, as alluded to); and (3) take jobs or drive down wages of high-skilled natives.

If your life is already "set" (as is the case for essentially all baby boomers), then sure, bring on the high-skilled immigrants, they'll spice things up and reduce the bargaining power of milennials in the labor market.

But if you're in your 20's or 30's and trying to establish something resembling an upper middle class life (a reasonable expectation for a high-skilled native), then high-skilled immigration is a giant slap in the face. More competition for higher education? Higher housing prices? Crappier job market? The milennials have nicer baubles and the Internet, but when it comes to the stuff of adulthood, it's never been harder to "make it." Increasing high-skilled immigration would only exacerbate that problem.

Comments for this entry have been closed
Return to top