1. How much of a role did trade play in the long-term decline in the manufacturing share of total employment, which fell from around a quarter of the work force in 1970 to 9 percent in 2015? The answer is, something, but not much.
2. How much of a role did trade play in the absolute decline in manufacturing employment, down about 5 million since 2000? Here the role is bigger, basically because you’re comparing the same effect with a much smaller denominator; even so, trade is less than half the story, but by no means trivial.
But what about the now-famous Autor-Dorn-Hanson paper on the “China shock”? http://www.ddorn.net/papers/Autor-Dorn-Hanson-ChinaShock.pdf It’s actually consistent with these numbers. Autor et al only estimate the effects of the, um, China shock, which they suggest led to the loss of 985,000 manufacturing jobs between 1999 and 2011. That’s less than a fifth of the absolute loss of manufacturing jobs over that period, and a quite small share of the long-term manufacturing decline.
I’m not saying that the effects were trivial: Autor and co-Autors show that the adverse effects on regional economies were large and long-lasting. But there’s no contradiction between that result and the general assertion that America’s shift away from manufacturing doesn’t have much to do with trade, and even less to do with trade policy.

  1. 全雇用に占める製造業の比率は、1970年には労働力人口のおよそ4分の1だったが、2015年には9%となった。この長期的な低下について、貿易はどの程度の役割を果たしたのだろうか? その回答は、少しは果たしたが、大きなものではない、というものだ。
  2. 製造業の雇用者数は、2000年以降に約500万人減少した。この絶対的な減少幅について、貿易はどの程度の役割を果たしたのだろうか? これについては、基本的に同じ効果をかなり小さな分母について比較することになるため、もっと大きい、ということになる。それでも貿易の影響度は全体の半分以下に留まるが、とは言え、些末とは言えない。

では、今や有名になった「中国ショック」に関するAutor-Dorn-Hanson論文はどうか? http://www.ddorn.net/papers/Autor-Dorn-Hanson-ChinaShock.pdf *1 実際のところ、この論文は以上の数字と整合的である。Autorらは論文のタイトル通り中国ショックの影響を推計したに過ぎないが、1999年から2011年の間に98万5000の製造業の職の喪失がもたらされた、という結果を示している。これは同期間の製造業雇用者数の絶対的な減少の5分の1以下であり、長期的な製造業の低下に比べると僅かな比率に過ぎない。

これに自ブログでTim Duyが噛み付いたEconomist's Viewエントリ)。

Nothing is wrong with the analysis here. But I think Krugman is downplaying the transition costs, especially regional impacts. Politically, that is the important part. Economists tend to just pay lip-service to the negative effects as we seek what is perceived to be the bigger prize, the aggregate effects. Fundamentally. Krugman is looking for what we got right in trade theory, and he finds it in Autor et al.
For me, Autor et al is not about what we got right in trade theory, but what we got wrong. Spectacularly wrong:
The speed of regional labor market adjustment to shocks is agonizingly slow in any area that lacks a critical mass of population. Rural and semi-rural areas remain impacted by negative shocks for at least a decade, but often longer. Relative to life spans, in many cases the shocks might as well be permanent.
And note that this is not just about negative trade shocks. Trade is an easy punching bag for Trump, but his message carries wider because we are really talking about structural shocks in general. ...
We don’t have answers for these communities. Rural and semi-rural economic development is hard. Those regions have received only negative shocks for decades; the positive shocks have accrued to the urban regions. Of course, Trump doesn’t have any answers either. But he at least pretends to care.
Just pretending to care is important. At a minimum, the electoral map makes it important.
These issues apply to more than rural and semi-rural areas. Trump’s message – that firms need to consider something more than bottom line – resonates in middle and upper-middle class households as well. They know that their grip on their economic life is tenuous, that they are the future “low-skilled” workers. And they know they will be thrown under the bus for the greater good just like “low-skilled” workers before them.
The dry statistics on trade aren’t working to counter Trump. They make for good policy at one level and terrible policy (and politics) at another. The aggregate gains are irrelevant to someone suffering a personal loss. Critics need to find an effective response to Trump. I don’t think we have it yet. And here is the hardest part: My sense is that Democrats will respond by offering a bigger safety net. But people don’t want a welfare check. They want a job. And this is what Trump, wrongly or rightly, offers.

それに対しクルーグマンが表題のエントリ(原題は「Trade, Facts, and Politics」)で以下のように反応した

...This wasn’t a political manifesto, and never claimed to be. Nor was it a defense of conventional views on trade. It was about what the data say about a particular question. Are we not allowed to do such things in the age of Trump?
Actually, maybe not. Part of the whole Trump phenomenon involves white working class voters rallying around a candidate who promised to bring back the coal and industrial jobs of the past, and lashing out at anyone who refuses to make similar promises. Yet the promise was and is fraudulent. If trying to get the analysis right is elitist, we’re in very big trouble — and perhaps we are.
So what would a political manifesto aimed at winning over these voters look like? ...
And I have to admit to a strong suspicion that proposals for regional policies that aim to induce service industries to relocate to the Rust Belt would not be well received, would in fact be attacked as elitist. People want those manufacturing jobs back, not something different. And it’s snooty and disrespectful to say that this can’t be done, even though it’s the truth.
So I really don’t know the answer. But back to the starting point: when I analyze the effects of trade on manufacturing employment, the goal is to understand the effects of trade on manufacturing employment — not to win over voters. No, dry statistics aren’t good for political campaigns; but that’s no reason to ban statistics.

*1:cf. ここここ