This week’s episode with David Autor did not go as I planned. I thought we would spend most of the time on the empirical estimates he and his co-authors have made of the impact of US trade with China. Eventually, we got to those estimates along with how reliable they are and the possible implications for public policy. But along we way, we had a lengthy discussion of the fundamentals of trade, trade deficits, and the complexity of labor markets. Because this was such a lively conversation and because it is such a timely issue, I invited four blogging economists–Don Boudreaux, Arnold Kling, Adam Ozimek, and Noah Smith to respond to the conversation. You’ll find their reactions below the fold. And you have a chance to respond to those reactions in the comment section.

Quick navigation:

Noah Smith

[originally published March 16, 2016 at “Autor on EconTalk” at Noahpinion.]

David Autor recently went on Russ Roberts’ EconTalk podcast to talk about his new paper about China trade. The Autor paper, co-authored with Dorn and Hanson, has been a real bombshell in the econ world, since it seems to challenge the “consensus” on free trade. Economists have long asserted confidently – in public, though not always in private – that free trade is always good, and now some top economists have shown that maybe it’s not always good. That’s a big deal.

Roberts and Autor do a very good job of zeroing in on why free trade might not be good, and why it might not have been good in the case of China. The reason is distribution – trade can hurt some people, especially if the government is not perfectly efficient in using redistribution to cancel out the distributional effects of trade (which of course it never is). This has been known for a long time, but economists often wave their hands and ignore distribution.

Autor et al. have basically made a splash by failing to ignore this very important thing. They treat distribution as important by assumption, and then they check that assumption by showing that govt efforts to cancel out the distributionary effects of trade have been woefully inadequate. Roberts, to his great credit, does not do the “annoying online libertarian” thing, and simply wave away distributionary concerns (as, say, Bryan Caplan might do if he were doing podcasts). In fact, Russ once again comes off as a very pragmatic, situationalist, intuition-driven sort of fellow – the kind of thinker I believe we need more of (and that Brad DeLong says have made America great).

One very interesting point in the discussion is when Autor says that trade between rich countries is basically all upside. I think that this is a hugely important point that gets completely ignored in today’s trade policy discussions, including by some of my favorite writers like Paul Krugman. Unfortunately, Russ doesn’t follow up on this point, but it matters a lot, since the trade agreements we’re now considering – the TPP and TTIP – are almost entirely agreements with rich countries. Some people seem to be treating the Autor paper like it applies to TPP and TTIP, but it just doesn’t, and I wish this were talked about more.

Autor also mentions a couple of other potential downsides of trade. He mentions trade diversion from multilateral agreements – unfortunately this doesn’t get followed up on. He also mentions trade deficits, correctly pointing out that in real terms, trade deficits are a loan, not a gift. This means that running trade deficits today is, at the country level, an impatient thing to do – it sacrifices the consumption of future generations to allow our current generation to consume more. Sadly, this point is not pursued much, but I guess podcasts have time constraints, and this one was already a bit long.

Anyway, a very good episode. I think I will do more reviews of EconTalk episodes! One thing I think Russ needs to get better at is critiquing empirical papers. In this episode, he brings up the possibility of multiple comparisons (which of course Autor accounted for, Autor being Autor). A better critique might have been to question the structural assumptions of the linkage model that allowed Autor et al. to conclude that the China trade shock reduced employment in aggregate. Now that econ is becoming more empirical, being able to get into the empirical weeds like that is going to be more and more important.

Arnold Kling:

[also published March 17, 2016 at “The Plunge in Manufacturing Jobs in the U.S.” at askblog.]

David Autor’s work on the adjustment problems for American workers caused by China’s rapid export growth is very interesting. I believe that these sorts of adjustment problems, rather than “aggregate demand,” are the main source of unemployment in the economy. I have discussed this way of thinking about unemployment as the problem of discovering patterns of sustainable specialization and trade.

I was intrigued that the podcast with David Autor and Russ Roberts discussed structural unemployment and corkball. One of my late father’s favorite expressions was, “We are going to have to pay people to play corkball.” More than thirty years ago, he foresaw a future in which society would have to deal with widespread unemployment by finding it socially acceptable to pay people to engage in leisure. Some additional points:

  • 1. The rules of corkball in St. Louis, where my I grew up, evidently differed from those in Memphis, where Roberts grew up.
  • 2. My father thought in terms of workers displaced by technology, and Autor talks about workers being displaced by trade, and in particular by rapid growth in manufacturing in China.
  • 3. My father thought that once low-skilled workers were displaced, it would not be long before what we would today call non-STEM college majors would find themselves displaced as well.
  • 4. My father was not an economist. In my view, he seemed to be committing the “lump of output” fallacy. That is, imagine that the world is one big factory, producing a fixed amount of output. If the factory “foreman” (meaning the impersonal forces of the market) finds some Chinese welders who can perform welding at less cost than American welders, then American welders will be out of a job. The fallacy is in treating the amount of output as fixed. In a dynamic economy, the factory “foreman” should be able to find another use for the American workers that serves to increase output. Since we teach that the economic problem consists of unlimited wants and limited resources, the unemployment of American workers should be of limited duration.
  • 5. In the podcast, it sometimes sounds as though Roberts wants to defend what we teach against what he sees as “lump of output” thinking. However, one way to characterize the dispute between Roberts and Autor is that what they really disagree about is the meaning of “limited duration” for structural unemployment. For Roberts, it might mean a few months or at most a couple of years. For Autor it might mean a decade, or perhaps from now until the worker reaches the legal retirement age.
  • 6. I believe that this is the most useful way to think about unemployment in general. That is, I do not think that very much unemployment is “cylical,” meaning that factories are shut down until unwanted inventories have been worked off, at which point they are re-opened and the same workers are rehired. Rather, I think that most unemployment is structural. The factory “foreman” finds millions of new jobs for workers each month (see the JOLTS survey), but the “foreman” cannot find a fit for everyone. Even a small increase in the rate of displacement and/or a small decline in the rate of fit-finding gives rise to what we call a recession.
  • 7. Both trade and technological innovation constitute progress. Displaced workers are the victims of that progress. The policy issue raised at the end of the podcast is how best to deal with the victims. I would lean against trying to slow the pace of progress by restricting trade or by making it difficult to introduce technological innovation. I would lean instead in the direction of cutting payroll taxes and decoupling employment from health insurance. Those steps would reduce the cost of labor and thereby make it more likely that the “foreman” will find a use for displaced workers.
  • Adam Ozimek:

    [also published March 17, 2016 at “What to Do About Losses From Trade” at The Dismal Scientist Blog.]

    The work of David Autor and his co-authors on trade and labor markets has caused many economists, including me, to adjust their beliefs. I would like to address some issues that have been raised by the research and by Autor’s interview with Russ, and also to focus on policy implications.

    The first point I’d like to make is about the robustness of these results. Scott Sumner has raised a skeptical eyebrow and claimed this is “just one paper,” but that vastly understates the strength of the research. There have been several papers from Autor and coauthors that take broadly different approaches: looking at industries, looking at local labor markets, and looking at individuals. The last approach, used in Autor, Dorn, Hanson, and Song is notable in its granularity as it tracks hundreds of thousands of individual workers from 1992 to 2007 using Social Security Administration data. And other studies found similar effects for European firms, for Norway, for Denmark, and for U.S. industries using a different instrument. It’s also true that there are many benefits of trade, but it seems clear that the many people and places have faced large and long-lasting adjustment costs.

    So if we believe the results—and I think we should—an important question is why trade with China appears to generate more job losses than trade has in the past? One likely factor is that the recent losses from trade are concentrated in low-skilled workers who have a harder time adjusting. This was a point brought out in Autor’s interview but it bears repeating: trade shocks did not harm the top third of manufacturing workers by pay.

    Why have low-skilled workers fared worse? One reason is they are less likely to pack up and move someplace with better job opportunities. Indeed, other research has shown that mobility for low-skilled workers is below that of skilled workers and has also fallen over time. There is no single explanation for this, but Ganong and Shoag found that rising house prices have reduced the incentives for moving to more productive parts of the country for low-skilled workers.

    Another factor that likely contributes to the large job losses is that labor market dynamism has fallen in general, which makes finding a new job harder. This helps explain why low-skilled workers have been hit hardest by trade, as Davis and Haltiwanger found they disproportionately hurt by falling dynamism.

    One issue that remains under-discussed is what the appropriate policy response is to the losses from trade. A common suggestion is to subsidize manufacturers, either directly or with higher tariffs. But given the ongoing mechanization in manufacturing, these subsidies could end up simply supporting factories filled with robots and few jobs. In addition, trade increases productivity and innovation, and propping up unproductive manufacturers risks reducing those positive effects.

    A better policy response would address the core problem, which is that low-skilled workers are not finding good jobs to replace the ones they lost. The Earned Income Tax Credit remains the best policy to address this by making low-skilled employment a more profitable option for workers and firms.

    One criticism of this approach is that the task of increasing employment is hopeless given the weak demand for labor in the U.S. economy. But high-skilled workers who have been hit by trade shocks have managed to find other jobs that pay as well.

    A benefit of the EITC over subsidizing manufacturing is that it boosts labor demand for firms that have the most profitable use for low-skilled workers rather than just manufacturers. This is important given that many high-skilled workers have survived trade shocks by moving to entirely new sectors and out of manufacturing. Subsidizing manufacturing does nothing to improve this channel of adjustment for low-skilled workers.

    In short, the work of Autor and his co-authors is extremely important, but economists are still digesting it. On the policy front, what is needed is more demand for low-skilled labor, not more demand for manufacturing output.

    Don Boudreaux

    [originally published March 17, 2016 at “Russ Talks with David Autor” at Cafe Hayek.]

    I just listened to Russ’s most-recent excellent podcast with M.I.T. economist David Autor. The subject is the effects of increased U.S. trade with China on the U.S. economy – in particular, on certain American manufacturing workers. Autor’s new paper on this topic, written with David Dorn and Gordon Hanson, is getting a lot of attention, mostly quite favorable.

    I have not yet read this paper, so all that I say in this post is based upon the podcast.

    In (very) short summary, Autor argues that Americans’ increased trade with the Chinese over the past few decades, while it unquestionably increased the size of America’s and China’s economic ‘pie,’ has left some American workers worse off over the long-run. Further, the extent of this harm to these U.S. workers is larger than what Autor believes economists’ priors would have predicted. The implication seems to be that freer trade is no longer as overwhelmingly positive as economists once believed: the costs are larger both in size and duration; economists’ theoretical case for a policy of free(r) trade is, therefore, weaker than even economist once believed.

    Here are some reactions; I’ll likely have more later, especially after I read the paper carefully.

    First (on the reasonable assumption that the podcast supplies a sound description of the main points of the paper) I agree with Scott Sumner that (a) one piece of empirical research in economics does not justify rejection of a long-standing conclusion among economists, and (2) much more importantly, it’s unclear that the findings of this paper reveal anything new or that should have been unexpected by those of us who accept the standard economic case for free trade.

    The economic case for free trade

    (1) has never denied that some workers (and other input suppliers), as well as some business owners and creditors, suffer reductions in their economic well-being as a result of changes in the patterns of trade – reductions, that is, from what that well-being would have been had trade patterns not changed;

    (2) says nothing about how much time it takes those who are ‘harmed’ by changes in trade patterns to optimally adjust to those changes;


    (3) was never premised on the argument or belief that the particular flesh-and-blood people who are ‘harmed’ by changes in trade patterns will themselves eventually adjust to these changes in ways that ensure that none of these people remain worse off than they were before trade patterns changed.

    Autor would not dispute (1), but he seems to dispute (2) and (3).

    He says at least twice in the podcast that during the “Bretton Woods era” – roughly mid-1940s through early 1970s – we in America saw no significant or long-lasting harm caused by changing trade patterns. That might be – and, if it be, it was indeed good. But I am unaware of any theoretical case for a policy of free trade that rests on the prediction that adjustment costs will be small, that adjustment times will be short, and that, in the end, literally everyone will be better off than he or she would have been had the particular change in trade patterns not occurred.

    On this point, see Russ’s truly spectacular book The Choice.

    Second, I disagree with Autor’s quick dismissal of Russ’s point that the greater availability today (compared to in the Bretton Woods era) of government cash assistance for displaced or otherwise unemployed workers might explain the lengthening of time that Autor et al. find that it takes workers today to adjust to changes in the patterns of trade. It’s true that no one wants to be unemployed, but incentives operate on the margin. With tens of millions of workers affected by changes in trade patterns, relatively small responses to even relatively small increases in government assistance – and, also, to increased wealth in society – can plausibly have detectable effects on the ways that workers respond to losing their jobs to changes in trade patterns.

    Third, Autor is mistaken to equate the trade deficit with debt. A U.S. trade deficit might become American debt (as when, for example, non-Americans lend some dollars to Uncle Sam), but a trade deficit is not necessarily debt (as when, for example, non-Americans use their dollars to buy shares of stock on Wall Street). Russ rightly pressed Autor on this point, but I don’t think Autor adequately responded, for he (Autor) kept talking about the trade deficit as if it necessarily represents a drain of demand today from the American economy.