This post will probably annoy almost everyone, as I will argue that both the left and the right misunderstand the issue of wages and productivity. Here’s Alphaville at the FT:
It’s well established that since the 1980s, as workers become more productive, wage growth hasn’t kept pace. The marginal productivity theory of wages suggests that it would. As an observation, this is uncontroversial. The Commerce Department’s Bureau of Labor Statistics produced a completely readable 14-pager about it last year. The current CEA is definitely thinking about wage growth; it reported in September that wages are growing, if you measure them correctly.
The second sentence in misleading, as the increased productivity he refers to is a higher average product, not a higher marginal product (MP). That doesn’t mean the theory that workers are paid their MPs is exactly correct, I suspect it’s only an approximation of reality. But you’d need other data to disprove this theory.
It’s also worth noting that the increased share of national income going to capital since the 1980s is due to the increased return on residential housing, a capital good. The people who own their own homes earn a return on that investment, in the form of housing services, that is higher than in the 1980s—even as a share of GDP. Matthew Rognlie provides some evidence:
It you are worried about the “typical worker”, you should focus on the fact that average wages have risen faster than median wages, due to greater inequality. Thus the pay at the very top has increased much faster than at the median wage level. That’s the actual concern.
On the other hand, I believe that some on the right are too willing to wave away concerns about wages by suggesting that the wages = MP of labor theory somehow proves that the existing wage distribution is fair. It may or may not be fair, but I don’t believe that the wages equal MP of labor theory has any bearing on this question. For instance, merely by changing the laws on intellectual property rights you will impact the distribution of wage income. But the appropriate IP laws are a complex issue, about which even free market economists do not agree. There is no obvious straightforward way to think about the marginal product of labor; it depends on many institutional factors.
Some people probably visualize something like the following thought experiment. In 1820, a hardy pioneer and his family move out to a remote valley in Montana, where they build a cabin, plant crops and hunt for animals. There are no other families nearby. Doesn’t that family earn their marginal product? Yes, but that’s not why our intuition tells us that their compensation seems fair. Rather the perceived fairness comes from the fact that they also earn their total product. They earn 100% of their valley’s GDP (minus depreciation of the cabin.)
Back in the real world, workers do not earn 100% of GDP minus depreciation. Furthermore, the marginal product of a worker who cooperates with many other workers and many other machines doesn’t necessarily match our intuition as to what the term “productivity” means. There’s nothing in marginal product theory to prevent a scenario where one man owns all the capital, and earns 99% of national income, and the other 1% is divided between 150 million workers on the basis of the MP of each worker. That’s obviously not likely to occur, but it’s not ruled out by the theory. And that’s very different from the intuition behind the Montana pioneer.
I believe that utilitarianism provides a better way to think about wages and productivity, at least for economists. A sound public policy will involve a set of labor market regulations, and income redistribution, than maximizes the well being of society as a whole. Concepts like “fairness” and marginal productivity of labor are not particularly helpful.
The counterargument is that this approach is too “dry” for the average person, who has very strong intuitions about fairness. Even worse, the average person has cognitive illusions about economics, which lead them to grossly underestimate the power of incentives. So if we take a dry utilitarian approach, the public will be led to support minimum wage laws, rent controls, soak the rich taxes, tariffs on imports, and all sorts of other destructive public policies, because it looks like these boost (domestic) aggregate utility. To fight this bias, we need to create a public mythology of the hardy pioneer earning his marginal product, and then extend that myth to the point where almost all workers are earning their “just deserts”. This mythology would also give owners of housing units a “right” to charge whatever rent to people who voluntarily choose to use their property. Consumers have a “right” to buy goods from anywhere in the world. Immigrants have a right to live where they wish.
Thus the deontological approach to economic rights and fairness can be defended as a sort of correction to the cognitive biases in the average person’s economic thinking, which leads even right-wingers to be inappropriately fond of government intervention. Even many of American conservatives favor socialist programs such as Medicare and trade barriers, and oppose immigration.
But I’m an economist, not a politician, so I’ll keep fighting battles using the utilitarian ethical framework. I’ll trust my readers at Econlog to not suffer from these common cognitive biases. In my view, utilitarianism (properly understood) leads to a world that is 95% libertarian, and 5% policies such as carbon taxes and low wage subsidies.
READER COMMENTS
David S
Nov 2 2018 at 2:59pm
Good post as usual Scott. Shouldn’t we be using compensation instead of wages though?
Scott Sumner
Nov 3 2018 at 11:05am
David, Sorry, I forget to answer your question. Yes, total compensation is best.
Ahmed Fares
Nov 2 2018 at 4:32pm
“It you are worried about the “typical worker”, you should focus on the fact that average wages have risen faster than median wages, due to greater inequality.”
There’s a nice graph at the bottom of the linked page below which shows that dispersion between average and median wages.
https://www.ssa.gov/oact/cola/central.html
I trot it out whenever someone thinks that the productivity gains are going to capital, when in fact they’re going to other workers, which the median does not pick up on, but the average does.
Also this from Tim Worstall:
“To provide a crude example, let’s say we’ve two groups of workers, computer programmers and burger flippers. Back in 1968 they both earned the same amount (look, this is an example, I am allowed to construct unrealistic examples to illustrate a point) and today they earn markedly different amounts. We can also note that the technology of burger flipping hasn’t changed over this time but the technology of computer programming has, markedly. The flip side of that statement is that the productivity of burger flippers is the same as it was and that of computer programmers is higher than it was.
Now, if we want wages to track productivity then we would want programmers’ wages to rise as against flippers’. Because as a society we want incentives for people to move into those higher productivity jobs: getting people into more productive lines of work is the very definition of us all becoming richer. So, programmers’ wages rise along with their productivity, flippers’ don’t.
And average productivity has obviously risen as well, because it’s the average of the productivity levels of our two groups. But there’s absolutely no reason at all in there why burger flippers’ wages should rise because programmers become more productive. So, that outrage over the minimum wage not tracking average productivity is misplaced. We should only be worrying about average wages tracking average productivity.” —Tim Worstall
Robert EV
Nov 5 2018 at 11:50am
Two things:
The network effect: Programmers save time when buying a meal from the burger flippers (vs making their own), so part of this savings is attributable to the burger flipper.
Even in burger flipping there are productivity increases (due to technology), but it seems to me much of this goes to keeping retail prices inexpensive instead of to the laborers or capital owners. There’s a corresponding productivity increase for computer programmers thanks to OSes, utilities, libraries, and computer language development that the individual programmer did nothing to make, except the computer programmers seem to be taking some of those productivity gains as wages. One could argue that this is rent seeking by the computer programmers.
The wage system does treat burger flippers and computer programmers fundamentally differently with respect to technological productivity increases.
Robert EV
Nov 5 2018 at 11:52am
I had my “two things” in a numbered list, but the numbers disappeared on posting.
test 1
test 2
after test
Ahmed Fares
Nov 6 2018 at 1:59pm
Robert,
I agree that the wages of unskilled workers would rise but the amount would depend on the ratio of skilled to unskilled workers. More skilled workers pulls up the wages of the unskilled, but if the unskilled are large in number, than that would hold their wages down, given that they’re competing for those unskilled jobs.
Basically supply and demand.
Scott Sumner
Nov 2 2018 at 5:00pm
Thanks David.
Ahmed, I don’t have any problem with people being “outraged” over the small pay gains for low wage workers, just that their outrage should not be linked to marginal productivity theory, partly for the reasons explained in the Tim Worstall quote you provide.
Thaomas
Nov 2 2018 at 6:44pm
I don’t know of any growth models that imply that the technical change parameter that they rely on to produce PC growth necessarily has a value that would make the MP of any or all categories of labor grow at the same rate as total output. Sure it was nice when it sort of happened and would be nice if it happened again, but there is no “mystery” about it not happening. The dog did not bark because there was no dog. 🙂
If we want to transfer income to low paid workers (I do) we have a well tested instrument, the EITC.
Mark Z
Nov 2 2018 at 8:47pm
I don’t think you’re presenting the best argument for the position with which you’re disagreeing. Appealing to marginal productivity (at least to me) has nothing to do with intuition or fairness, and everything to do with competitiveness. If markets are competitive, then workers’ compensation should equal marginal productivity, and if they don’t, it suggests markets aren’t competitive, and that potentially state intervention is needed to correct this. This is why supposed wage stagnation is so commonly discussed simultaneously with monopolies.
If one accepts that more competitive markets are generally more efficient, i.e., they maximize the welfare of society, then whether workers are earning the marginal product of their labor is an indicator of the how optimal wages are.
“A sound public policy will involve a set of labor market regulations, and income redistribution, than maximizes the well being of society as a whole. Concepts like “fairness” and marginal productivity of labor are not particularly helpful.”
It seems you’re just exchanging ‘fairness’ for some hypothetical objective definition of utility, which our statesmen are supposed to be able to measure and optimize from their commanding heights, which I’d say is, in practice, no more grounded or helpful than any deontological conception of fairness. Economists should, imo, eschew such concepts and stick to letting subjective preferences determine (or rather, define) people’s well-being, rather than imagining there is some objective ‘social’ utility function they can optimize irrespective of the revealed preferences of society’s many individual participants.
“There’s nothing in marginal product theory to prevent a scenario where one man owns all the capital, and earns 99% of national income, and the other 1% is divided between 150 million workers on the basis of the MP of each worker. ”
Perhaps not in marginal product theory, but if such an arrangement is suboptimal, wouldn’t Coase theorem (theory?) tend to preclude such an outcome?
Scott Sumner
Nov 2 2018 at 8:55pm
Mark, I don’t follow your comment. I certainly understand that W = MP assumes competitive markets, but that’s not what I’m discussing here. My point is that we don’t know if workers are being paid their MPs, contrary to the claims in the article. If they are not, then there might be some role for public policy to correct that situation, as you suggest. But that’s all very hypothetical.
My point about “fairness” is that MP doesn’t have the implications for fairness that many people assume it has. I understand that policies of redistribution involve lots of difficult choices, involving hard to measure concepts like utility. In principle, however, that’s how we should be thinking about the problem.
Mark Z
Nov 2 2018 at 9:55pm
I agree that MP has no implications for fairness. If you’re saying that MP can’t really be measured, then perhaps we can simply assess whether the market for labor is competitive, and reason that, if it is, workers are probably on average being compensated equal to their marginal product, assuming competitiveness is easier to measure than MP. But if we can’t measure MP, and we also can’t measure any variable that would predict whether workers are being paid the equivalent of their MP or not, then I’m not sure there’s anything we can do other than throw up our hands in futility anyway.
Or are you rather saying that you can only measure MP given some prior set of (perhaps arbitrarily determined) parameters (your example being IP laws)?
Scott Sumner
Nov 3 2018 at 11:06am
Mark, AFAIK, MP is very hard to measure. This is basically the same problem we face with antitrust law, except employers with market power are monopsonists rather than monopolists. Most firms have at least a bit of monopoly power in product markets, but only a few extreme cases are worth addressing with anti-trust law. I’d say the same about monopsonists in the labor market.
michael pettengill
Nov 3 2018 at 3:03am
Economies are zero sum. Cost minus benefit must equal zero.
Workers can not consume more than they are paid in the long run.
Arguing workers should be paid as little as possible argues for limiting production to the least possible.
Lower labor costs as virtue means low, no, or negative GDP growth is a virtue.
A worker is worth more than his production because his (paid) consumption is just as critical as his production.
This is why robots will never replace workers. A self replicating robot will drive GDP to zero, unless robots get paid, and become consumers of GDP.
Scott Sumner
Nov 3 2018 at 11:08am
You lost me at “Economies are zero sum.”
Mark Bahner
Nov 4 2018 at 10:45pm
GDP is the total value of goods and services. How is the total value driven to zero?
P.S. I think you’re missing the fact that simply because a good or service is provided at less cost, doesn’t mean it’s of less value.
Philo
Nov 3 2018 at 10:29am
“A sound public policy will involve a set of labor market regulations, and income redistribution, that maximizes the well-being of society as a whole.” You forgot to add: “in the long run.” And, of course, it is anybody’s guess how the regulations and laws put into effect *now* will evolve over time, so—given this inevitable but unknowable evolution–it is anybody’s guess which set of regulations and laws would maximize long-run well-being *if put into effect now*. Anyway, the issue is not worth my attention, since I have virtually no power to change the set of regulations and laws that actually exist. I’ll leave it to you to “fight the battles.”
Scott Sumner
Nov 3 2018 at 11:09am
Philo, You said:
“I’ll leave it to you to “fight the battles.”
That seems like a good idea.
Benjamin Cole
Nov 3 2018 at 8:14pm
My only advice is if a democratic nation wishes to remain mostly a free-market economy, then keep labor markets as tight as possible.
If voters think that income and wealth stratification is tearing apart the social fabric, then they will not vote for free markets.
If globalism coincides with lower perceived living standards and job scarcity, then voters will not embrace free trade and open borders for immigration .
nobody.really
Nov 5 2018 at 3:08pm
I think Cole has a point: We can strive for “open markets.” But if this results in a populist backlash, we may end up with less free markets than if we initially constrained the markets to bias results in favor of labor. We let the perfect be the enemy of the good.
Yet analyses of constrained markets often find that society losses something like $300,000 for each $50,000 job preserved. Yet this outcome seems to be politically stable. So here’s the next question: Should we strive for…
A. A free market that maximizes GDP and drives down wages–but proves to be politically unstable and is replaced with a protectionist market with lower GDP and higher wages?
B. A moderately constrained market that results in an intermediate level of GDP (less than free market, but more than populist market) and wages, but is politically stable?
C. A free market with low wages subsidized by a mechanism for sharing wealth, resulting in an intermediate level of GDP and political stability?
Protectionism leaves so much money on the table. But if alternative economic systems will forever be vulnerable to populist attack, protectionism may be the best we can hope for. I yearn to make free markets popular–but to do that, I think we need to give more voters a bigger slice of the pie. That’s option C.
art andreassen
Nov 4 2018 at 6:20pm
Scott: My comment concerns one small part of this discussion which is the actual values that are used to measure productivity and compensation.
I have read this and the BLS paper and see no discussion in either of how the appropriate deflators are created and whether this has any impact on the resulting productivity and compensation discrepancy. This is an area I have had a morbid interest in.
As a general observation concerning employment and productivity, I don’t feel there has been enough interest in the increase in intermediate inputs due to specialization, offshoring, outsourcing and imports at the expense of internal production . The more a firm substitutes outside production for internal production the more will be the appearance that the productivity of the remaining employees has increased. This is separate from the deflation problem.
The BLS has for generations been responsible for product and industry deflators. It has always included estimates of the value of quality changes in the products and industries resulting in a change in the real dollar output. As stated these BLS deflators are the basis for the productivity indexes. The BLS makes these estimates in house and base them on no market transaction. These deflators are applied to the output side of the National Accounts only and have no matching impact on the input side. Is it any wonder that there is a discrepancy between the growth of industry productivity and compensation? The BLS paper blandly states that deflating compensation by the industry deflator rather than the CPI deflator removes a lot of the difference. DUH. They seem oblivious that their deflation procedure is the reason.
What is so surprising is that there has never been a suggestion that the BLS publish the impact that its quality adjustment has had by product and industry and in total since this data are created in the derivation of the indexes. The BEA has for generations published an Inventory Valuation Adjustment that balances an increase in demand ( due to price changes) with no commensurate increase on the income side, a similar problem to what the BLS is doing. I bet the value of the quality measurement is greater than that of the IVA.
The BLS is in the Department of Labor not Commerce.
Andrew_FL
Nov 5 2018 at 10:19am
It is consumer sovereignty, not marginal product theory, that establishes that only a laissez faire distribution of income is just and right.
Plato’s Revenge
Nov 5 2018 at 3:25pm
I agree that MP arguments don’t show what people may assume they show, but I disagree on your complementary claim that regulation (with the effect of evening out compensation) is still utility increasing. I’d say it undermines social mobility. Today, it’s much more possible to rise to middle and even upper class through sophisticated skills (and hard work) — that’s much more widespread than a few decades ago, no?
You might argue that a class society (in this sense) is adverse to utility, but the irony is that lack of social mobility cements class society
John Fembup
Nov 6 2018 at 1:08pm
As a lifetime wage earner, I never expected that my wages, or any increase to my wages, would reflect the full value of my contribution. If it had, the company would make nothing. I was willing to exchange some of the value I briught to the company for some intangible values I obtained in return – mainly job security and the discipline of forcing me to save (both explicitly in my bank, and implicitly in reduced wages) for post-retirement benefits; but also professional advancement from association with my co-workers and perhaps even a certain amount of prestige for having this or that title in a respected company.
And I always reminded my children as they were growing up, if you really want to be paid what you are worth, start your own company. I’m very proud of all my children, one of whom does have her own business.
Was I all wrong about this?
Hazel Meade
Nov 7 2018 at 2:47pm
I think what you are trying to say here is that while it’s true that workers are generally paid their marginal product (approximately), that doesn’t address the question of fairness. Many people think it is “unfair” for workers to be paid less than what they need to live on, regardless of whether that is their marginal product or not, for instance.
The thing is that MP theory is needed not so much to make arguments about the fairness of wages, but to rebut bad economic theory about what determines wages – for instance old Marxist ideas that capitalists derive profits from exploiting the “surplus labor value” of workers – paying workers less than what they are actually worth. A lot of the argument for things like minimum wages is derived from that idea – that minimum wages and unions and so forth are needed so that labor can capture a “greater share” of what their labor is worth. That so long as there are “profits” that means someone is being exploited. MP theory tells us that is incorrect – workers are already getting paid their marginal product. Whether what their marginal product is worth is a “fair” state of existence is a separate question, but if we want to get to that discussion we need to have an accurate description of how the economy works and that means getting past old defunct ideas like “surplus labor value”.
Larry Stevens
Nov 9 2018 at 7:43am
Old joke about an automated factory. There is one employee and a dog. The dog’s job is to keep the writer from messing with the robots. The worker’s job is to feed the dog. What is the worker’s MP? The dog’s?
and if the factory gets rid of them both, what share of its output should go to labor?
Comments are closed.