What this covers

Noah Smith and Russ Roberts debate whether American workers have fallen behind, and if so, why. Smith presents evidence of a marked slowdown in wage growth since 2000 despite rising corporate profits, and proposes that rising market concentration—both in product markets and labor markets—may be suppressing worker pay. Roberts counters that wage stagnation is a misreading of the data, that total compensation has risen robustly since the 1970s, and that low-skill workers retain competitive options across multiple employers and industries. The conversation ranges across measurement pitfalls, the mechanics of monopsony power, and the economic case for co-determination, the German practice of placing workers on corporate boards.

Smith argues that the atomistic wage-taker model of employers deserves skepticism, pointing to phenomena like temp-agency wedges, persistent high markups, and evidence from locally concentrated labor markets where wages fall. He acknowledges that causation between concentration and wage suppression remains unproven, noting that economists disagree sharply—some viewing market power as only a small factor, others as substantial. The discussion of co-determination highlights a correlation vs causation trap: Germany has both lower inequality and co-determination, but attributing one to the other commits cargo cult thinking. Roberts emphasizes that rising corporate profits do not necessarily reflect worker exploitation; they may cushion employment stability during downturns, and he notes that tech dominance in product markets offers no clear mechanism for suppressing wages. Both speakers stress the importance of examining multiple wage measures over long periods rather than relying on any single metric, and they touch on how levels of analysis—macro trends versus micro industrial-organization studies—reveal different slices of the problem.

Sharpest takeaway

Smith argues that since the turn of the century there has been a real, substantial slowdown in U.S. wage growth alongside surging corporate profits, and that rising market concentration—via labor-market monopsony and markups, not just product-market dominance—is a plausible (though not yet causally proven) culprit worth investigating, while Roberts pushes back that wages have not stagnated and that low-skill workers retain competitive options across firms and industries.

  • Multiple wage measures show a marked slowdown in wage growth since 2000 even though total compensation rose since the 1970s
  • Industries that became more concentrated saw labor's share fall more, and locally concentrated labor markets show lower wages
  • The atomistic wage-taker model of employers is suspect given phenomena like temp-agency wedges and rising markups

The argument · threads7 threads · 36 claims
0.85

Corporate market concentration has risen across nearly all sectors, correlating with falling labor's share and lower local wages.

5 pointscentrality 5/5
  • Companies across many industries are charging higher markups above costs; this can't be explained by greater efficiency given the economy-wide productivity slowdown, may partly reflect unobserved quality improvements, but is also consistent with dominant companies squeezing workers, suppliers, or customers—which would explain the correlation between concentration and markups.

    “we also see markups increasing... companies are... starting to charge higher markups above whatever their costs are... if it's because all of these super-big companies are becoming much more efficient through mergers, then why do we see an economy-wide slowdown? Doesn't make sense... another possibility with these higher markups is just that companies are squeezing their workers more”

  • If concentration increases in every industry overall, then industry-specific concentration likely increases in most localities too, so a worker in chicken processing who once could choose among several local firms may now face only Tyson, and switching industries requires costly retraining—giving the dominant employer power to pay a lower wage.

    “if you have overall concentration increasing in every industry, then you have industry-specific concentrations probably increasing in most localities as well... It's Tyson that's the only game in town... and then they'll be able to pay you a lower wage because it's just so hard for you to switch.”

  • Using new local data on job postings, several papers find lower wages (and higher prices) in locally concentrated markets: a welder in an area with few welder-employing firms tends to earn less, all else equal, than a welder where many employers compete locally—because most people won't accept a two-hour commute, making local employer count matter.

    “there have been several papers that showed lower wages and higher prices in locally concentrated markets. So, for example, suppose you are like a welder and there's only a few companies in industries that employ welders within a few miles of your house: We see that your wages tend to be lower than welders... in places where there's many people who employ welders within a few miles of your house.”

  • Markets are becoming more concentrated across nearly all sectors (few large players replacing many small ones, e.g. Tyson in chicken processing, Monsanto in seeds), and in industries where the product side became more concentrated, labor's share of income fell more than in industries that became less concentrated—a correlation, though not proven causation.

    “across the board, in all sectors, markets are becoming more concentrated. You are seeing a few large players where before you saw a lot of small players... in those industries where the product side is more concentrated, you've seen labor's share of income also fall more than in industries where it's less concentrated.”

  • Over the last 17 years corporations made much more profit—a lot from overseas operations—while wages slowed, and even looking only at domestic value created (despite profit-shifting accounting issues) the ratio of corporate profits to wages rises since 2000 with a very clear inflection point.

    “if you do look at that, you still see this rise in the ratio of corporate profits to wages, since 2000, and there looks to be a very clear inflection point.”

0.79

Whether concentration causes wage slowdown is genuinely unresolved among economists despite mounting evidence of the relationship.

4 pointscentrality 5/5
  • Research on co-determination is mixed: cross-country regressions show co-determination countries tend to have lower inequality, some papers find companies with labor board representation have lower stock-market value (implying workers successfully extract value from shareholders), but other papers find no relationship—though no one finds co-determination produces worse wages for workers.

    “some papers find that companies with labor representation on their Boards do have lower stock market value. And that implies that workers are successfully extracting some of the value from shareholders and redirecting it to themselves... but then some other papers don't find that... So, it's not clear. You don't seem to find anyone who thinks that co-determination results in worse wages for workers.”

  • Whether increased corporate market power is the culprit for the wage slowdown and falling labor share is uncertain and completely unresolved—economists like the Economic Policy Institute and Jason Furman argue concentration is only a small piece, while papers by David Autor and Simcha Barkai argue declining labor's share comes substantially from concentration and rising rent extraction.

    “It's uncertain as to whether or not that is the culprit. And, it's a big question that is completely unresolved... You've got Jason Furman... he says: It's only a small piece of the wage puzzle. And then other people say... it's actually a big piece of what's going on... You've got Simcha Barkai's paper.”

  • When measured across multiple government surveys (real compensation, average hourly earnings, median real weekly earnings, the Atlanta Fed Wage Tracker) and properly deflated, total compensation rose robustly since the 1970s—so claims that wages fell are wrong—but since the turn of the century all measures show a marked, substantial slowdown in the trend rate of wage growth.

    “long time listeners know that I'm skeptical about the existence of wage stagnation. I think it's false. I think it's a misreading of the data.”

  • Local concentration only lowers wages a lot in the limiting case where a worker can move neither locations nor industries—complete immobility—which is not realistic; in reality workers retain some ability to switch industries or relocate, but frictions and barriers (cost of moving, retraining for genuinely different jobs) mean the effect operates at the margin.

    “the case where this, where local concentration really, really lowers wages a lot is where you can't move locations and you can't move industries. You are basically completely immobile... And that's the ideal, that's not real. In reality... you do have some ability to do something else.”

0.65

Worker board representation could raise productivity by capturing ground-level knowledge and may nudge companies toward longer-term thinking.

4 pointscentrality 4/5
  • Giving workers board input can raise productivity because workers know how things work on the ground (as Feynman solved the Challenger O-ring problem by asking engineers), and some papers show companies with more workers on their board have higher long-term productivity—though this is correlation, possibly because more productive companies voluntarily add more workers.

    “if you give workers more input into the running of the company, they will know what needs to be done to raise productivity, because they're the people on the ground... there have been a couple of papers showing that companies with more workers on their board do have higher productivity over the long term. But, again, that's correlation, not causation.”

  • Rather than mandating co-determination, the government could nudge companies with small monetary incentives to voluntarily try putting workers on their boards, observe whether participating companies succeed, and only then consider scaling up the program—though it may spread on its own.

    “We could nudge companies and say, 'Hey, companies: we're going to try out this thing where, try putting some workers on your board and we'll throw you a little money if you do.'... And then, if that works, if the companies who do it succeed, then we can start thinking about scaling up that program in some way.”

  • High profit margins need not mean value siphoned to shareholders; for a volatile manufacturing business like German carmakers, high profits in good years can act as a cushion to avoid going out of business and maintain high employment in bad years, so workers might accept lower wages and higher profits to protect job security.

    “you could use high profits as essentially a cushion... 'this year we might sell a lot of cars; and then next year we might not sell any cars; and so we are going to make sure that when we do sell cars, we make a bunch of profit so that we have this cushion so that we don't go out of business in the years that we don't sell many cars.'”

  • Co-determination, as practiced in Germany, gives workers status as stakeholders with representation on corporate boards (alongside workers' councils), so that employees—not just shareholders—get to vote for the board, amounting to partial worker ownership; in America only shareholders vote for the board.

    “the part that is here on the table in America, is that workers have representation on corporate boards. So basically, if you work for a company, you are considered a stake holder, and you get to vote for the Board... In Germany, basically they have partial worker ownership for companies.”

0.65

Short-term stock market pressures cause public companies to underinvest and miss long-term value creation versus private ownership.

4 pointscentrality 4/5
  • Stock prices exhibit excess volatility—bouncing around far more than the underlying dividends or earnings ever do over decades—making the market an excessively volatile forecast of future earnings, a finding for which Robert Shiller won the Nobel Prize.

    “excess volatility is this really well-established fact... you see stock prices bounce around a lot... you wait for 20 years and you look at it, and they never do. So... today, people get really optimistic. Tomorrow they get really pessimistic. And the stock price just bounces around. And then you look and you see that earnings never did the same.”

  • Public companies focus on the next quarterly result rather than long-term horizons, which can cause them to underinvest, miss key market segments, take short-term decisions, or asset-strip, and there is evidence that closely-held private companies invest more.

    “One of the main worries has been short-termisim: the idea that when you have a public company is that everyone is worried about the next quarterly result, and no one is thinking 5 years ahead... And so this could cause companies to underinvest. We have evidence that privately-, you know, held, closely-held companies invest more.”

  • For investors to price the long-term value of a stock today, the value-price link must hold through every future holder via backward induction; but because noise traders, liquidity-forced sellers, or irrational participants frequently break that chain, the long-term value fails to back-propagate to present prices, generating short-termism.

    “in order for me to care about the long-term value of the stock, I've got to assume that every single person along the way cares equally about the long-term value of that stock, so that the price back-propagates... But if that chain is broken even one time... If at some point the value/price link is broken, the whole chain is broken. And then: Why do I care?”

  • Taking a company off public markets onto private markets robustly correlates with re-investing more of its profits, and founder-run companies both invest more and are more profitable in the long term, contradicting Modigliani-Miller neutrality results which say ownership structure should not matter.

    “if you take a company off the public markets and have it one the private markets, it re-invests more of its profits. And that's a pretty robust statistical correlation... Modigliani-Miller says... 'Nothing like that should ever happen.'... We also see that founder-run companies invest more. And also are more profitable in the long term.”

0.44

Copying Germany's co-determination without understanding its broader institutional context would be cargo-cult thinking.

5 pointscentrality 2/5
  • Despite a long-term trend of moving into urban areas with more employer choices, per Jed Kolko's research the move back into cities was small and concentrated in the 2000s while the move to the suburbs has continued, and suburbs grow faster naturally because city residents have fewer children—so urbanization does not obviously offset local labor-market concentration.

    “if you look at some evidence by Jed Kolko... the move back into the city was a pretty small move, and concentrated in one decade--the 2000s. Overall, the move to the suburbs has continued; and suburbs have a much more natural rate of population growth, because people in cities don't have any kids.”

  • The current clamor over monopoly and market power comes from macro, labor, and public-finance economists examining economy-wide trends, not from industrial-organization economists who study individual mergers and specific industries; the two groups grab different parts of the elephant—macro trends contain information micro work misses, while micro work catches estimation issues macro work has—so a dialog between them is needed and currently lacking.

    “IO people have spent, especially empirical IO people, have spent a lot of their time, you know, looking very micro... Not thinking about the macro trends. So then you have... people from actual macroeconomics or labor or public finance, etc., coming in and saying, 'Okay. Let's look at the overall picture.'... they are kind of, you know, grabbing two different parts of the elephant... And they need to talk to each other more.”

  • Although Germany has less inequality than the U.S. and co-determination, attributing the lower inequality to co-determination by copying everything Germany does is cargo-cult thinking—mistaking correlation for causation—because Germany differs in size, regional inequality, education system, and a million other factors.

    “it's hard to, you know, sort of look at these aggregates and say, 'Okay, well Germany is doing better. We better just copy everything they do.' That's cargo-cult thinking.”

  • Because labor is the biggest cost—more than 50% (around 60%) of costs for retailers—claims that companies gained share by 'becoming more efficient' and 'driving down costs' almost always amount to paying workers less; while it is possible to shrink the other costs without cutting labor, it rarely happens when one factor is 60% of costs.

    “when people say, 'These companies are gaining, you know, market share because they've become more efficient; they've become more efficient because they've driven down costs'--that is almost always saying, 'You are paying your workers less.'... Because labor cost is like 60% of costs.”

  • Stock ownership is highly concentrated among the top 10% of Americans, with most middle-class people holding most of their wealth in their houses, producing a fat-tailed distribution where a little stock is owned by many and a lot of stock by few.

    “stock ownership is so concentrated among the top 10% of Americans. Most middle class people in America have most of their wealth in their houses.”

0.40

Measured wage compensation rose robustly since the 1970s but has substantially slowed since 2000, and health-care cost inflation masks real wage stagnation.

2 pointscentrality 3/5
  • The Employment Cost Index includes all forms of compensation, but if employer-based health care simply gets more expensive without improving in quality—because of an inefficient employer-based system shoveling money into health care—then workers are not made better off even though their measured compensation rises, since they receive the same services for more dollars.

    “the limitation of ECI is that if health care simply gets more expensive but yet doesn't improve in quality terms, and if the reason that health care gets more expensive is because we have an inefficient employer-based health care system that shovels extra money into this health care system... then people don't really care that they're getting the exact same health services and someone's getting more dollars”

  • Because no single wage measure is perfect and each is collected differently, looking across many measures (real compensation, average hourly earnings, median real weekly earnings, the Atlanta Fed Wage Tracker) minimizes the chance of being misled by methodological error, and one should examine decades rather than the most recent data point.

    “long time listeners know that I'm skeptical about the existence of wage stagnation. I think it's false. I think it's a misreading of the data.”

0.26

Most of the American economy lies outside tech; focusing on tech distorts policy discussion about economy-wide concentration trends.

2 pointscentrality 2/5
  • Despite spectacular market capitalization, the tech industry is a relatively small slice of American employment, revenue, and value added; most of the economy is still agriculture, food processing, construction, education, health care, real estate, business services, insurance, retail, and manufacturing, so focusing on tech distorts the picture because we pump disproportionate money into tech stocks.

    “the tech industry, despite having a spectacular market capitalization, is a relatively small slice of employment and revenue in the American economy and value added... most of the American economy is still stuff like, you know, agriculture or, you know, like food processing, or construction”

  • The traditional model treating employers as a huge number of atomistic wage-takers with no wage-setting ability should be regarded very skeptically, because the economy contains many strange industrial-organization phenomena—temp-agency wedges, huge markups, fixed real-estate commissions persisting for decades despite competition—that this model cannot explain.

    “this very traditional way of thinking of employers as these... atomistic employers which are essentially wage takers and forced to pay the going market wage and have no wage-setting ability--I think we must be very skeptical of that mental model.”