New Trade Agreements will Take Center Stage in 2015. So Will Bad Arguments Made on their Behalf.

Next year, we are going to see lots of debate over trade policy. And, like clockwork, when trade policy rises to the top of policy debate, lots of bad arguments start getting thrown around on behalf of more trade agreements. Ed Gresser submits the latest round of bad ones in a paper released last week.

Gresser goes wrong out of the gate by implying very strongly that inequality is irrelevant to the living standards of low and moderate-income households. In his own words he argues:

“But “growing apart” [editor’s note: this means the rise in inequality] appears to be a phenomenon in which wealthy people rise fastest, not one in which they rise while the middle class and poor lose ground. Americans have actually grown more affluent at all income levels.”

This implicit claim is deeply wrong—the rise of inequality over the past generation has in fact been the primary drag on living standards growth for low- and moderate-income families. Gresser arrives at his irrelevance conclusion by essentially noting that cumulative income growth for low- and moderate-income households has exceeded zero over multiple decades. Well, congratulations to us, I guess. But very few countries outside of maybe North Korea have ever posted negative income growth over decades for the majority of their population.

It’s especially ironic to get this interpretation of rising inequality wrong when discussing its with expanded trade. The standard trade theory that links falling trade costs and rising inequality in rich countries like the United States is clear that this rise in inequality is accompanied by absolute (not just relative) income declines felt by the losing group. In the United States, the losing group is generally proxied by either production and nonsupervisory labor, or workers without a college degree—in either case the majority of the workforce. And while these trade-induced losses (which I estimated to be roughly $1,800 annually for a full-time worker without a college degree) do not explain all, or even the majority, of the rise in inequality over the past generation, they’re not trivial. Gresser claims to have cast doubt on these results (which are based on off-the-shelf standard trade models), but as I’ll show below, his analysis of them is completely irrelevant.

This post will mostly focus on three arguments that Gresser makes that are clearly wrong are:

  1. Thomas Piketty doesn’t identify trade as a driver of inequality, so it can’t be one.
  2. Standard models of trade’s impact on inequality are somehow invalidated if it can be shown that manufacturing has fewer gross layoffs than other sectors.
  3. If we ignore every influence of trade except the lower prices for imports that it causes, then it helps all Americans.

Gresser does make two broad arguments along the way that are correct: the United States could benefit from faster export growth and intellectual property provisions in trade agreements could be fashioned to boost equality. But both of these insights actually argue against the desirability of the trade agreements—the Trans-Atlantic Trade and Investment Partnership (TTIP) and the Trans-Pacific Partnership (TPP)—that Gresser is pressing for.

Hiding behind Piketty

Gresser invokes the (excellent) work of Thomas Piketty as some kind of talisman to argue that trade flows cannot be affecting inequality. But he gets many things wrong here.

First, it is not the case that Piketty has shown that rising inequality in the United States has been driven entirely, or even primarily, by “investment returns…exceeding overall growth rates.” The r>g formulation that has intrigued so many about Piketty’s work is really a cautionary tale about the future, not a description of what has happened in the past.

As Piketty describes it (p. 298 in my copy): “Let me return now to the causes of rising inequality in the United States. The increase was largely the result of an unprecedented increase in wage inequality.” And it is this unprecedented increase in wage inequality that many have posited could indeed be largely driven by expanded trade (particularly with countries poorer than the United States).

Second, even if one believed that a high rate of return to investment was a prime driver of U.S. inequality, this would not constitute any grounds for dismissing the role of globalization. Textbook models of trade (literally—look in the index for “Stolper-Samuelson” in any international economics textbook) argue exactly that expanded trade can exacerbate income inequality in countries like the United States precisely by raising the rate of return to capital while reducing wages. And, in fact, plenty of research has shown exactly that the rising share of capital income across advanced countries in recent years is likely driven by trade and investment liberalization.

Standard trade models don’t account for layoff rates?

Gresser references my own work on trade and inequality and claims to debunk it by showing that layoff rates in manufacturing are lower than in other sectors. Given that my work consists of using models of trade developed by Nobelists and that are continuing to be taught today in Ph.D. international trade courses, it would be impressive indeed if this stylized fact about sectoral layoff rates constituted a devastating counter argument. But it’s completely irrelevant.

The passage of mine that Gresser takes issue with simply explains how it is that trade flows can impact even those workers who don’t work in tradeable goods sectors. The intuition is simple—workers who are displaced by imports then need to find alternative employment. They go looking in other sectors—even non-tradeable sectors. This increase in the effective labor supply available to even non-tradeable sectors stemming from import competition lowers wages. As I tried to put it simply, restaurant waitstaff and landscapers may not lose their jobs to trade flows, but their wages can suffer from having to compete with apparel workers who have been displaced by imports.

Gresser spends more than page focusing on a complete non sequitur to this argument—examining statistics on layoff rates (a gross, not net, measure of job flows) by sector. And because they’re lower in manufacturing he argues that we should abandon economic textbooks. This is wildly off base, for two reasons.

First, while manufacturing has lower layoff rates, what surely matters for assessing whether or not laid off manufacturing workers are forced to compete with workers in other sectors of the economy are net changes in sectoral employment. And on net, U.S. manufacturing has lost roughly 5 million jobs since the beginning of 2000. Retail, a sector that Gresser notes has low gross layoff rates, has added about 400,000 jobs since then. So, remind me again why we don’t think a story of former manufacturing workers seeking employment in other sectors has some relevance?

Second, you don’t actually need even a net change in employment in manufacturing to spur the Stolper-Samuelson effects I estimate—you just need a net change in demand for different types of workers. To put it far too simply, if imports of apparel displace two non-college workers and one college worker, while exports of aircraft employ two college workers but just one non-college worker, you would have no net change in manufacturing employment (three jobs displaced by imports but three jobs supported by exports), yet you would have a reduction in demand for non-college workers (two were displaced by imports but just one supported by exports) and an increase in demand for college workers (vice-versa).

So, his alleged debunking of an entire class of estimates based on textbook models simply by noting that manufacturing has lower layoff rates is just a long and irrelevant tangent.

Ignore everything but the benefits and trade is win-win

Gresser argues that the trade agreements he’s supporting in this brief (TTIP and TPP) would reduce trade costs on both imports and exports. Even assuming for a moment that’s true (it’s far from clear that it is, but we’ll run with it for now), it still does not then follow that you’re allowed to calculate the gains from trade as just the benefits consumers get from the lower cost of imported goods, for two reasons.

First, expanded trade driven by lower trade costs will indeed lower import prices, but it also raises export prices (again, I’m not breaking new trade theory ground by asserting this—it’s literally textbook international trade economics). The rough intuition is that by lowering foreign trade costs, foreign consumers will demand more U.S. output, and this will raise the price U.S. consumers have to pay for it, hence higher export prices. How these simple price impacts (lower import prices versus higher export prices) shakes out in terms of living standards for lower-income households is not as straightforward as Gresser makes out.

Second, what is straightforward is that workers intensively employed in import-competing sectors and those with similar skills throughout the economy will unambiguously lose. For them, the lower import prices will be accompanied by lower prices for import-competing production, which lowers their (real, inflation-adjusted) wages, period. The rough intuition is that the price of an import-competing good (like the price of any good) is equivalent to the costs of its production. These production costs are the wages of workers and the return to owners of capital used in its production (or, profits). For the price of import-competing goods to fall, some of these production costs have to fall as well. And, if only one of the inputs (labor or capital) sees their return fall, then their return must fall more than the price of the good. For the U.S., expanded trade (particularly with poorer countries) means that its labor that will see its return fall (since import-competing goods tend to be more labor-intensive than exported goods).

In the jargon, this is the “magnification effect” of trade, and what it means simply is that low prices for imports and imported goods are not good news for American workers harmed by trade, rather, they are the source of the wage declines they face because of trade flows.

Exchange rates and intellectual property

Gresser makes two useful points about the economics of trade and international economic policy generally, but these points are not supportive at all of efforts to pass TTIP and TPP.

First, he notes the demand stimulus that the U.S. economy could gain by increased exports. He’d be more accurate to say net exports rather than just exports (that is, even falling imports, all else equal, would provide a stimulus to the U.S. economy) but it’s true that the U.S. economy could see faster recovery if net exports rose. But trade agreements should in theory boost both exports and imports, and claims that this should be expected to lead to net export gains should be taken with boulders of salt.

Worse, the primary lever to increase net exports is policies that allow the U.S. dollar to find its competitive level—a competitive level that it is being intentionally pushed away from today by the practice of currency management on the part of several trading partners. Neither trade agreement Gresser supports has useful measures (that I know of anyway—I’m happy to be educated) that would help with this problem.

Second, he notes that intellectual property provisions in trade agreements could boost equality. He’s right in principle, but his example is very odd, in that he argues for increasing the economic distortions caused by intellectual property enforcement and claims this will benefit “lower-income intellectual property holders.” But the overwhelming majority of economic gains from intellectual property enforcement accrue to high-income households and corporations, and a genuine “free trade” agreement would aim to actually lower the costs of this enforcement. But if past practice is any guide, these treaties will surely aim to increase trade costs by forcing other countries to devote resources to become bill collectors for American holders of intellectual property (entertainment, software, and pharmaceutical companies, mostly).

Summing up

Economic theory predicts clearly that trade liberalization exacerbates income inequality in countries like the United States. No, trade and globalization don’t explain all of our increase in inequality—we still have lots of domestic discretion about how unequal a society we’d like to have. But expanded trade does push in the wrong direction, as demonstrated by ample economic evidence. And the trade agreements we sign, while habitually  referred to as “free trade” pacts, are actually a codification of protections to privileged interests, and never contain provisions for addressing the most important issue in international economics for most American workers—the persistent U.S. trade deficit. This is what policymakers should know, and which too many policy analysts are trying to obscure.