The New York Times’ Binyamin Appelbaum wrote an excellent piece yesterday on the costs and benefits of globalization. But because I’ve thought a lot about this topic, I have some hobby-horse issues concerning how economists characterize how large the gains from trade are and how its gains and losses are distributed. Put simply, the overall net benefits of trade are much smaller than commonly advertised, but the regressive redistribution trade causes is considerable.

First, on the gains from trade policy (i.e., how much we should expect national income to rise if we sign trade agreements), Appelbaum refers to a piece from the Peterson Institute of International Economics claiming that trade liberalization added 7.3 percent of GDP to American incomes by 2005—about $9000-10,000 per American household. This is just not true. It’s a wildly inflated number that should not be in the policy debate (and if you need much smarter and better-credentialed people making the some point—here’s Dani Rodrik). This number is an effort to bully people into going along with today’s trade agreements by making them think the stakes are utterly enormous. In fact, even if it was correct (again, it’s not) this study would be irrelevant to today’s trade policy debates because the sum total of economic gains from all post-1982 trade agreements (this includes NAFTA, the completion of the General Agreement on Tariffs and Trade, the formation of the WTO, and the permanent normal trading relations with China) is estimated to be just $9 per household, meaning that  99.9 percent of the gains from trade estimated in the study happened before 1982. So even if trade liberalization really did spur mammoth gains at some point in the (distant) past, the effects were over by the early 1980s.

Second, on the distribution of gains and losses from trade, it is striking to me that so many economists who favor signing every trade agreement that comes down the pike can still feign surprise that expanded trade seems to be bad for most workers’ wages. Put simply, it is completely predicted in textbook trade economics that wages for most workers will fall and inequality will rise when the United States trades more with poorer trading partners. Yes, expanded trade is predicted to lead to higher overall national income, but it is also predicted to redistribute enough income within the United that it can (and is likely to) make most workers worse-off. This should not be a surprise to anyone familiar with the topic.

In his piece, Appelbaum quotes an economist (Gordon Hanson) who says:

“I think what we’ve learned is that U.S. labor markets aren’t as flexible and self-correcting as I think we had presumed,” said Gordon Hanson, an economist at the University of California, San Diego. “The uneasiness I have about the way we’ve handled globalization is not so much globalization itself. It’s that if you don’t have the right safety net, you’re going to impose an enormous amount of hardship.”

Hanson is an excellent economist who has been involved in some of the highest-quality academic work examining trade’s impact on wages. So I find this paragraph a real dodge on his part. “Globalization itself” really is predicted to lower most American workers’ wages and increase inequality. The empirical evidence that it does so should not puzzle people—it’s the case of evidence validating theory.

Why does this matter? Well, if theory really did predict that more trade benefited everybody all the time (have I mentioned yet that it doesn’t?), then the empirical finding that wages were lowered for the two-thirds of the workforce without a four-year college degree and inequality increased in its wake could be written off as a weird coincidence and not the predictable outcome if we moved to expand trade flows in the future. But that’s wrong. The impact of a single agreement such as the TPP will likely not be huge, but we do know the direction it will go in: lowering wages for non-college workers.

Hanson’s last sentence is key—you could in theory cut lots of the damage that expanded trade does to American workers with compensatory policies (i.e. the winners compensating the losers) that spread the gains. But these compensatory policies are not included in most trade agreements, including the TPP, at any scale large enough to mention. And anybody surprised by this hasn’t been paying enough attention to American politics.

Finally, a trade economist named David Weinstein is quoted in the piece making a common claim: we may actually worry too much about the losses from trade because losers are easy to identify but the gains from trade are hard to see because they accrue to a large and diffuse group. This corresponds with the common view that trade agreements’ costs consist only of manufacturing workers losing their jobs. But workers who have lost their jobs to imports are only the tip of the iceberg in terms of trade’s costs—the real losses from trade also include the wages of every worker in the economy who subsequently competes with the trade-displaced for remaining jobs. This essentially means all workers without a 4-year college degree. Do security guards and warehouse operators recognize that trade is dragging on their wages? Many of them probably do not; trade’s casualties do not, in fact, always know who they are. So, in fact, the losses from trade are widespread.