Sunday, March 10, 2013

The Left and wages



Matt Yglesias has a post about the left and wages that I think makes a number of errors. Here's the intro:
The left is rightly concerned about the real wages and incomes of ordinary working people. But the left fallaciously analogizes from a bargaining dynamic at a single firm to the entire economy.
Well, I think many people do make errors of this type ("The government is like a household" being the most common). But I don't think that the Left is doing this when it comes to wages. Let's see what Matt thinks the Left is doing. He continues:
The way a given worker or class of workers improves his real wages is by persuading his boss to give him a nominal raise that outpaces the growth in the cost of living. But the way the economy as a whole works is that my income is your cost of living. If Slate doubled my salary, I'd be thrilled. But if everyone's boss doubled everyone's salary starting on Monday, we'd just have one-off inflation. As it happens, a little inflation would do the macroeconomy some good at the moment. But the point still holds that while "you get a raise" is the way to raise your living standards, "everyone gets a raise" is not the way to raise everyone's living standards.
The first error I see here is the idea that wages = incomes. Only about 55% of Americans' income comes from "labor". The rest comes from "capital" - land rents, corporate profits, and interest income. It is true that the sum of all expenditures across an economy equals the sum of all incomes. It is not true that the sum of all expenditures across an economy equals the sum of all wages. Thus, it is possible to raise every wage-earner's wage at the same time. This would simply result in an increase in labor's share of national real income, and a decrease in the share allocated to capital. In other words, it would be the exact reverse of what has happened to our economy since (at least) 2001:

Graph of Business Sector: Labor Share

The second error I think Matt makes is that he talks about the mean, while the Left generally cares about the median (or some lower percentile). It is possible for median real wages to increase even while mean real wages stay flat. This would mean a more equal distribution of income...in other words, the reverse of what has happened over the past three decades:



(Source: The Atlantic)

Matt goes on:
To raise real wages across the economy...what you need to do is make things cheaper. And that generally entails an accumulated series of events that make selected groups of people's nominal incomes lower. If everyone could be chauffered around cheaply by autonomous cars, then the people who currently earn a living driving cabs and buses and trucks will lose out. If computers take over routine diagnostic work, then doctors will lose out. If online courses that serve tens of thousands of students displace community colleges, then community college instructors will lose out.
This is just not right. The process that Matt describes need not reduce nominal incomes. Suppose, first of all, that everyone had to stay in the same line of work forever (which is not true, but which is a necessary assumption if you want to use phrases like "cab drivers will lose out"). Now an industry-specific productivity shock comes along - someone invents a latte machine that makes better-tasting lattes, twice as fast, while using only half the milk. The real price (i.e. the relative price) of lattes will fall. The quantity of lattes sold will rise. The income of latte-makers (the # of whom, remember, is fixed!) equals the price of lattes sold multiplied by the quantity of lattes sold.

Will this income go up or down? Well, you can see for yourself, with a bit of Econ 101. Draw a supply-demand graph for the latte market. Model the productivity improvement as a rightward shift in the supply curve. Take the derivative of equilibrium income with respect to some level parameter of the supply curve. You will find that whether latte-makers' incomes go up or down depends on the elasticities of demand and supply for lattes. In particular, if demand for lattes is very elastic, then the incomes of latte-makers will go up. (This result will also hold in general equilibrium.)

Now, those are real incomes, and Matt did say "nominal". But if the latte market is small compared to the overall market - in other words, if the change in the price of lattes is not enough to change the aggregate price level much - then real and nominal are approximately equal. So it's easily possible for latte-makers' nominal incomes to go up when their productivity improves, even if latte-makers are unable to shift to other jobs.

(Note that "latte-makers' incomes" includes the wages of latte-making workers and the capital income of latte-making firms. Who pockets more of the increase depends on other stuff, including workers' bargaining power.)

So I see Matt making some econ mistakes here. But anyway, what is the takeaway? Well, it's not clear what Matt thinks the Left is trying to do at the national level. If he sees the Left as a bunch of Luddites, trying to block the implementation of new technology, then his case would be strengthened by the points I made above.

But I do not see the Left as a bunch of Luddites. I see them as trying to increase the bargaining power of workers relative to capital owners, by any means available. That rarely involves Luddism; in fact, I can't really recall seeing any American Leftist types trying to block the spread of new technology. On the contrary, labor advocates like Dean Baker routinely trumpet advances in productivity; rather than trying to halt these advances, they decry the fact that too much of the surplus from the advances is flowing to owners rather than workers.

So I guess I don't see what specific mistaken thing Matt thinks the Left is actually doing.

Now this is a bit beside the point, but I personally do see the Left making some mistakes. I think the Left probably tends to overstate the effect of politics (e.g. union-friendly laws) relative to structural changes (especially globalization) in eroding workers' power. I think they often overestimate the degree to which American trade policies can affect the process of globalization. And I think they see immigration as more of a threat to native-born workers than it really is.

But I don't think any of these things has to do with a mistaken analogy between a single firm and the entire nation. The Left wants to increase labor's share of income, and decrease the skewness of the wage distribution. How to do that is an open question, but those goals seem perfectly reasonable to me.

35 comments:

  1. Anonymous5:05 PM

    I agree completely, I had a similar post and the conclusion I came to was:

    All this isn’t to say that the Left hasn’t made a “big mistake” regarding wages. The real problem among the Left is an aversion to capital and technology. Where Matt and I see lower prices, many liberals see structural unemployment. The real challenge of governance in this century will be the transition from wage-intensive income to capital-intensive income, as the above figure painfully demonstrates.

    In the postwar era, America embraced the idea of a “property-owning democracy” through various schemes to promote homeownership. We need to embrace the idea of broad capital ownership. Some radical ideas might be a lump-sum provision of stock at the age of majority: at least then the median worker can take some pleasure as capital incomes rise.



    ----

    So I guess my point is labor share of income isn't going to start increasing, it may level out if we're smart (increasing distribution of human capital etc.) We just need broad capital ownership, somehow.

    I disagree with a lot of what Matt said, but I guess the conclusion is something you can take away from Econ 101 – liberals should stop fighting for demand-side inflation to raise wages, but supply-side cost decreases.

    Sounds in many ways similar to the tussle between Sachs and Krugman.

    What do you think about lumpsum capital transfers?

    ReplyDelete
    Replies
    1. Anonymous5:10 PM

      Or, even more, some sort of "right to capital" campaign. I don't know how it would look, or how it would be implemented without screwing incentives, but simply redistributing income won't work.

      Matt's point is becoming less true by the day.

      Delete
    2. "Some radical ideas might be a lump-sum provision of stock at the age of majority: at least then the median worker can take some pleasure as capital incomes rise."

      I think Noah mentioned this before, and I mentioned why I don't like it.

      1. Are you sure you want to give 18 year olds (of all people) a big hunk of assets all at once to do them for the rest of their lives. Sorry, but the story of the prodigal son can be interpreted in a number of ways. But one thing that sure comes out, is that people learn.
      2. Circumstances still make this grossly unfair (because asset prices are highly unstable). And how will you phase such a thing in? It is not something you can do for all 18 year olds, but not at all for 19 year olds is it (jealousy - you ain't seen nothing yet)!

      If you want to do this - make it cumulative!

      Delete
    3. This comment has been removed by the author.

      Delete
    4. "but simply redistributing income won't work."

      Why not? Is the government having a share in the economy and then distributing the government's capital income as a national dividend somehow different? People can then either spend or invest as suits their circumstances.

      Delete
  2. Anonymous5:11 PM

    I think your argument on labor vs. capital shares misses the point.

    Is it possible for real wages to go up? Yes, it's called economic growth. Does this imply a fall in the capital share? Not necessarily, see Kaldor's growth facts (which motivated the widespread use of the Cobb-Douglas production function).

    ReplyDelete
    Replies
    1. Well, Kaldor's facts are not laws of the Universe. The fall in labor's share of income in the U.S. from just under 2/3 to about 55% certainly contradicts a major Kaldor fact. A Cobb-Douglas function may well approximate U.S. production in the short run, but in the long run the labor share parameter might be subject to change, and in fact does appear to have changed.

      Delete
    2. Anonymous5:31 PM

      Sure.

      In the past, the main cause of real wage growth was productivity growth. So if people are doubtful of future wage growth, is it because they don't believe in more productivity growth, or do they think future productivity advances will mainly benefit capital? That seems to be the real question.

      Delete
    3. Yep.

      Also note that growth rates != levels. It may be that the Left wants not just a resumption of past rates of growth of wages, but faster-than-historical growth, in order to reestablish labor's historical 2/3 share of income.

      Whether that's possible, and how it might be done, are of course open questions.

      Delete
  3. Anonymous5:24 PM

    Oh, and I don't know if this is confirmed by theory/empirics (will check it out), but in the short-run because humans aren't rational and don't discount for inflation, an expected increase in nominal wage will probably result in greater consumption, which by itself can decrease our output gap.

    ReplyDelete
  4. Pretty girl in the photograph.

    ReplyDelete
    Replies
    1. John S3:51 AM

      +1, the photo is all I remember from this post.

      cute girl pic is not the best lead in to economics discussion, imo.

      Delete
    2. Really? Guess she's not my type.

      Delete
    3. cute girl pic is not the best lead in to economics discussion, imo.

      Never under estimate the economic importance of people's sex drives - plus a pretty young woman is a suitable decoration for almost any occasion.

      Noah - tastes differ.

      Delete
  5. I think the Left probably tends to overstate the effect of politics (e.g. union-friendly laws) relative to structural changes (especially globalization) in eroding workers' power.

    You are making the error here in assuming this matters. There is a tendency for some in the inequality debate to engage in root cause fallacy, where first it is argued that inequality is the natural result of powerful economic forces either beyond our control or that we do not wish to reverse (often technology or globalization), and then the logical leap is made that if that is the cause then nothing other than reversing it will reverse inequality, and so the conclusion follows that in the face of rising inequality we should throw up our hands, understanding that whatever inequality we get is the inevitable result of the march of history. But this is silly; increasing worker bargaining power, progressive taxation, and more generous public services can alleviate the problem no matter what's causing it. What you observe is the left focusing on what can and should be changed to address the problem, rather than what you see as the real root cause. So what? Unless your concern is that the left needs to focus more on how to stop globalization?

    ReplyDelete
  6. Noah,I'm not sure if I miss your point, but it seems to me that in your latte model, you are assuming that technology or capital is complementary to labor (i.e. those latte machines need labor to operate), hence labor will at least capture some of the income in latte making (which depends on bargaining power, which is your point)

    I think Yglesias's point is that technology can be labor-substituting rather than complementary. New technology can completely replace labor in some sector/occupation. For example, if a new latte machine is invented such that it does not need people to operate at all, then there will be no need for latte-making workers at all and whatever change in income from this technological shock will be captured or borne by latte making firms. (Yglesias's automatic car & diagnosing machines are similar examples)

    This paper by Acemoglu & Autor, for example, makes an explicit distinction between factor-augmenting technology and -substituting technology to explain the disappearing of the medium wage jobs: http://www.nber.org/papers/w16082

    ReplyDelete
    Replies
    1. Of course such machines already exist - they are called automats. They still need labour to operate, but the labour comes from the consumer (this is happening everywhere in our society - cash machines, self-service checkouts etc). I don't see this consumer labour anywhere in our production functions. (And of course it raises the question - if the labour is coming from consumers, where will their income come from?)

      Delete
  7. I think the Left probably tends to overstate the effect of politics (e.g. union-friendly laws) relative to structural changes (especially globalization) in eroding workers' power.

    I'd say those are both politics. It is the political creation and regulation of corporations that enables globalization and other structural changes. And it is the political enactment free-market attitudes of winner-take-all, no social safety net, etc. that makes workers bear the brunt of those changes.

    ReplyDelete
    Replies
    1. Anonymous12:00 PM

      I disagree that globalization is primarily due to political/regulatory changes. What you're missing is the enormous changes in technology (primarily communications, but also in shipping logistics with containerization, etc) that have made it possible to manage an enterprise on the other side of the world. Forty years ago a transatlantic phone call was still kind of a big deal... now you can email hundreds of megabytes of data from Cape Town to Anchorage, instantly, for essentially free.

      Delete
    2. I'd say those are both politics.

      But, crucially, someone else's politics.

      Corporations are - often - a very potent system of organization, in terms of producing stuff. If Japan forms corporations and we don't, their exports will outcompete ours in markets all over the world, and no matter what we do, we will lose out. Disestablishing corporations, shutting down trade, etc. would only shoot ourselves in the foot, because either way we'd lose out to foreign rivals.

      Delete
    3. Germans and Japanese have corporations without having such egregious side effects because the political enactment of corporate law there protects the populace from some of the toxic side effects. We could do that in the US if we had the political nerve.

      I don't know how you got "Disestablishing corporations, shutting down trade, etc" from what I wrote.

      Delete
  8. Anonymous11:49 AM

    I actually have a question about the share of income captured by wages.

    I know that in the UK for example, most of the change you observe is really the sum of 2 effects: the increase in taxation and the increase in self employment. Especially the second one seems very similar to wage income to me. Even a large part of capital income is really the result of mislabelling wage income for capital (see hedge fund bosses).

    Are you totally sure that that chart actually reflect reality?

    ReplyDelete
  9. Anonymous7:05 PM

    "The Left wants to increase labor's share of income, and decrease the skewness of the wage distribution. How to do that is an open question, but those goals seem perfectly reasonable to me."

    I'm curious how you would suggest going about doing that, if that happened to be a goal you held, in addition to one you found reasonable.

    ReplyDelete
    Replies
    1. I can think of two avenues, legislative and cultural.

      Use the tax code to inhibit inherited dynasties, take higher % of capital gains, use that wealth for health care and safety net.

      Culture, the things we value as a society, is influenced by the speech and example of prominent people, starting with politicians. When the president talks about our obligations to each other it has an effect. When powerful people use their wealth in manifestly generous ways it has an effect. When we collectively ease off our worship of money and one-upmanship the world changes in ways most would praise as just.

      Delete
  10. Matt Yglesias said:

    "To raise real wages across the economy...what you need to do is make things cheaper."

    This is absolutely true, if "cheaper" is meant in real terms, i.e. "requiring fewer resources to produce" (or equivalently, producing better things with the same resources, which makes the things cheaper if you measure quantity in terms of a hedonic index). It is also approximately true in nominal terms given the way central banks behave: when productivity is rising rapidly, things tend to get cheaper in nominal terms compared to what would happen if productivity is not rising so rapidly.

    Continuing:

    "And that generally entails an accumulated series of events that make selected groups of people's nominal incomes lower."

    Also true, provided that one interprets "generally" to mean "usually in real life" In real life, there are usually winners and losers from technological change (although you can always imagine a model in which there are no losers from a particular change). The losers are those "selected groups of people" whose nominal incomes get lower. However, the longer the time frame you're looking at, the more likely that there will be no net losers.

    Now of course you can model a situation where a particular technical change produces only winners even in the short run. But I don't think such situations are typical.

    But how is it that what Matt says is "just not right"? It may be imprecise, as blogging usually is, but it seems basically right to me.

    As for what the Left wants:

    "But I do not see the Left as a bunch of Luddites. I see them as trying to increase the bargaining power of workers relative to capital owners, by any means available. That rarely involves Luddism"

    That depends on your model. If your model has the characteristic (as many models do) that capital income is more likely to be invested than labor income, then there is a tradeoff between current labor income and growth, and provided the economy is dynamically efficient, increasing the bargaining power of labor will reduce aggregate income in the long run.

    But I think Matt's overall point is that trying to get everyone's real wages up in the long run by trying to get each individual's nominal wages up in the short run isn't a very effective strategy. And I agree, at least in normal times. (Of course, if you're in a liquidity trap and there's a Eggertsson-Krugman "paradox of toil," then it is an effective strategy, but I'm sure Matt realizes this, and he either doesn't believe in the Paradox of Toil or he's anticipating leaving the liquidity trap.)

    ReplyDelete
    Replies
    1. Andy is tend to respect your opinion enormously, but I wonder a bit if this:
      "That depends on your model. If your model has the characteristic (as many models do) that capital income is more likely to be invested than labor income, then there is a tradeoff between current labor income and growth, and provided the economy is dynamically efficient, increasing the bargaining power of labor will reduce aggregate income in the long run."
      isn't empirically challenged. We are currently running an experiment where we are decreasing labours share, and an investment boom is nowhere in sight.

      Delete
    2. Anonymous10:25 AM

      Thank you Andy Harless for confirming what I've already suspected: Noah's post is nothing more than an exercise in nitpicking.

      Delete
    3. So Noah Smith on this blog, Seth Ackerman here - http://jacobinmag.com/2013/03/wage-and-productivity-effects-of-internet-trolling-by-matt-yglesias/ , and JW Mason (linked to in this very comment section) slam Matt Yglesias' for his post. Andy Harless provides a charitable reading of Matt's post, defending him against the sharp and adroit Smith, Ackerman, and Mason.

      Mason points out that changes in wages can merely change the distribution of income and they don't always have to lead to price changes. Ackerman complains about the assumption on the part of Yglesias that the conversation should be about everyone's wages, when Yglesias is talking about raising the real wages of "ordinary working people," namely, janitors, cooks, nurses, etc. Similarly Smith points out that wages are the incomes of "labor" but not of "capital" so raising nominal wages would change the income distribution. Yglesias should take note - not too shabby.

      Then again, when Yglesias says **selected** groups of people's nominal incomes often go down in response to technological innovation, Smith responds that nominal incomes in general don't go down through the process Ygelsias describes, so one wonders if Smith is talking past Yglesias here. And after all, Yglesias's talk of "everyone's" wages seems consistent with Harless's charitable reading of Ygelsias (that Yglesias was talking about the long run, in which case Matt's belief could be consistent with a belief in diverging policy prescriptions under differing circumstances, like for example in a liquidity trap).

      Left over disagreement will remain even if the charitable reading of Yglesias goes through (e.g. the Twitter exchange between Smith and Harless on same topic). But Matt's post would seem to rise to the level of plausibility if the charitable interpretation of the offending post is accepted by all parties.

      Geez, if only Matt Yglesias had a forum, you know, where he could tell us what he really had in mind... Then we could find out whether he meant

      "We can NEVER raise the real wages 'for every janitor, short-order cook, yoga instructor, and nurse in America' by increasing their nominal wages,"

      or whether he meant

      "as a general historical matter, what really matters is productivity growth, but you know of course we can increase real wages by increasing nominal wages, in particular circumstances."

      I mean, a leftover disagreement over how often we're in which kind of circumstance, or over whether Matt and Andy's point even applies in the long run, (as, I suppose it's possible that the income distribution could matter even when we're not in a liquidity trap) seems less intense than the current beef. The current beef seems to be, at least according to Smith, Ackerman, and Mason, that Yglesias denied that real wages and incomes of ordinary workers can actually be increased with nominal wage increases, like, ever.

      Delete
  11. Poor Matt once had a fine mind but he took it to Washington where all but the finest things are tarnished by proximity to power (I'm pretty sure it was Ezra Klein wot did it to Matt, so that makes Matt not all that fine to start with, but I had hoped for better...).

    The idea that things happen in economics unaffected by the power in the political world around them is the core fallacy New Keynsians and other classical revivalists were created to perpetuate. If you can just keep money out of your models you don't have to deal with all its corrosive and coercive affects.

    If you want labor to have bargaining power, pursue full employment and maximization of the labor participation rate. There are policies that will do that. Or actually enforce anti-trust in the spirit it was intended: to end industry consolidation rather than to encourage low every day prices (which have historically coincided with low every day wages).

    If you want bankers to behave better, threaten their access to the Fed until they behave. Better yet actually threaten to press charges for money laundering and the other white collar crimes they currently build their business around. It worked with Hank Greenberg at AIG, I bet you could get some political miles out of that if you, as a politician (not you Noah, but, you know "you") were inclined to look to the public good rather than your next paying gig once chased from office.

    If you want to control globalization (and you are the issuer of the world reserve currency) quit issuing government bonds and spend fiat dollars directly into the national economy: you're the bloody fucking government, you can buy local if you choose: if your goal is to employ your resources, including your human stock of capital, you are uniquely well positioned to do so. These are all political decisions that have huge impacts on economics.

    We get the economics we choose politically. When we believe the bull shit artists who say economics occupies some technocratic orbit outside the world of politics and must be kept in that valence we double down daily on a status quo that is increasingly exhausted. It's just like asking the Chinese to observe our copyright law and our internet rules so they can't beat the Pentagon in an electronic war as if they were docile American citizens.

    These are delusions and can only be believed by those insulated from reality by money and the power it can buy when everyone else is convinced that economics is some god given organic equilibrium seeking ideal independent of the seven billion people it actually organizes, or disorganizes as the case may be. But it is all political, money does not exist in nature and despite centuries of economic models with out money in them, the division of labor, investment, innovation and growth in living standards just about can't happen without money, but with money only happen where there is the political will to make it so.

    ReplyDelete
  12. I've notice that whenever Matt writes about anything I know really well, he is almost always completely wrong. It makes me think that he hand-waves a lot of arguments in a way that sound appealing if you aren't an expert, but really dumb when you are.

    ReplyDelete
    Replies
    1. A fairly common trick. That's how The Economist stays so popular.

      Delete
    2. I stopped reading the economists due the insanity of Lexiton in the early Bush years. But Ryan Advent seems mostly sensible.

      Delete
  13. I claim dibs :) : http://theredbanker.blogspot.com/2013/03/disagreeing-with-matt-yglesias.html

    That being said, Noah makes a very good point about the median vs. average salary thingy. A good way to represent wage inequality. I'll steal the idea in the future... :)

    ReplyDelete
  14. This is a good post.

    My own response to Matt is at http://slackwire.blogspot.se/2013/03/recession-is-time-of-harvest.html.

    ReplyDelete
  15. My impression of Matt's blog was that it primarily concerned manufacturing debates and international trade. That is, the left focuses too heavily on "keeping jobs at home" and not enough on the extent to which imports increase real wages. I don't think any of your points counter that thesis.

    ReplyDelete