Paul Samuelson’s outsourcing “bombshell”

Steve Lohr breathlessly reports in the New York Times that Nobel prize winner and undisputed godfather of modern economic theory Paul Samuelson is coming out with an article in the Journal of Economic Perspectives on outsourcing that contradicts the mainstream economic take:

At 89, Paul A. Samuelson, the Nobel Prize-winning economist and professor emeritus at the Massachusetts Institute of Technology, still seems to have plenty of intellectual edge and the ability to antagonize and amuse. His dissent from the mainstream economic consensus about outsourcing and globalization will appear later this month in a distinguished journal, cloaked in clever phrases and theoretical equations, but clearly aimed at the orthodoxy within his profession: Alan Greenspan, chairman of the Federal Reserve; N. Gregory Mankiw, chairman of the White House Council of Economic Advisers; and Jagdish N. Bhagwati, a leading international economist and professor at Columbia University. These heavyweights, among others, are perpetrators of what Mr. Samuelson terms “the popular polemical untruth.” Popular among economists, that is. That untruth, Mr. Samuelson asserts in an article for the Journal of Economic Perspectives, is the assumption that the laws of economics dictate that the American economy will benefit in the long run from all forms of international trade, including the outsourcing abroad of call-center and software programming jobs. Sure, Mr. Samuelson writes, the mainstream economists acknowledge that some people will gain and others will suffer in the short term, but they quickly add that “the gains of the American winners are big enough to more than compensate for the losers.” That assumption, so widely shared by economists, is “only an innuendo,” Mr. Samuelson writes. “For it is dead wrong about necessary surplus of winnings over losings.”

Sounds like a radical break — oh wait, let’s get into the details:

Mr. Samuelson, who calls himself a “centrist Democrat,” said his analysis did not come with a recipe of policy steps, and he emphasized that it was not meant as a justification for protectionist measures…. According to Mr. Samuelson, a low-wage nation that is rapidly improving its technology, like India or China, has the potential to change the terms of trade with America in fields like call-center services or computer programming in ways that reduce per-capita income in the United States. “The new labor-market-clearing real wage has been lowered by this version of dynamic fair free trade,” Mr. Samuelson writes…. For his part, Mr. Bhagwati does not dispute the model that Mr. Samuelson presents in his article. “Paul is a great economist and a terrific theorist,” he said. “And in markets like information technology services, where America has a big advantage, it is true that if skills build up abroad, that narrows our competitive advantage and our exports will be hit.” But Mr. Bhagwati, the author of “In Defense of Globalization” (Oxford University Press, 2004), says he doubts whether the Samuelson model applies broadly to the economy. “Paul and I disagree only on the realistic aspects of this,” he said. The magnified concern, Mr. Bhagwati said, is that China will take away most of American manufacturing and India will take away the high-technology services business. Looking at the small number of jobs actually sent abroad, and based on his own knowledge of developing nations, he concludes that outsourcing worries are greatly exaggerated…. The Samuelson model, Mr. Bhagwati said, yields net economic losses only when foreign nations are closing the innovation gap with the United States. “But we can change the terms of trade by moving up the technology ladder,” he said. “The U.S. is a reasonably flexible, dynamic, innovative society. That’s why I’m optimistic.” The policy implications, he added, include increased investment in science, research and education. And Mr. Samuelson and Mr. Bhagwati agree that the way to buffer the adjustment for the workers who lose in the global competition is with wage insurance programs. “You need more temporary protection for the losers,” Mr. Samuelson said. “My belief is that every good cause is worth some inefficiency.”

Before I throw my two cents in, let me just add the following caveats:

1) I haven’t seen Samuelson’s essay (anyone who’s got a copy of it, e-mail it to my brand-new gmail address listed on the right); 2) I’m not an economist; 3) Paul Samuelson is way, way, way, way, way, way, way smarter than I am.

That said, this dispute boils down to a few empirical questions:

1) Just how many well-educated workers are there in China and India? 2) Will U.S. firms have an incentive to offshore sophisticated value-added work in areas where the United States currently has a comparative advantage? 3) Will the United States continue to be a locus for value-added innovations? 4) To what extent are wages and employment in the affected industries declining because of outsourcing as compared to technological innovation standardizing and commodifying what used to be highly complex (and highly paid) tasks?

In the past, my answers to these questions have been a) not as many as you think; b) no, c) yes, and d) not a lot. [On (d), see Tyler Cowen’s and Arnold Kling]. Which is why I side with Bhagwati on the outsourcing question. Furthermore, Samuelson appears to partially fall into the Douglas Irwin trap of firing a warning shot on outsourcing but providing little in the way of a solution that departs from those who believe outsourcing is not a problem. Indeed, Samuelson explicitly rejects the solution most favored by those who oppose outsourcing — higher trade barriers. So, in the end, I’m not convinced that Samuelson’s dissent changes the substantive issues of debate. But as a political scientist, it is impossible to deny the extent to which Samuelson’s article will alter the rhetorical balance of power in this policy debate. Samuelson will succeed in reigniting debate on this topic, as well as provide aid and comfort to those who wish oppose the practice of offshore outsourcing. So let the debate be joined. UPDATE: Arnold Kling links to a draft version of the response paper by Jagdish Bhagwati, Arvind Panagariya, and T.N. Srinivasan alluded to in Lohr’s Times story. Kling’s summary:

The authors point out that some of the concern is not about trade per se but about the accumulation of capital and know-how in China and India. They suggest that this could harm the U.S. if it reduces trade by eliminating the division of labor. That is, suppose that the U.S. stays stagnant, but China and India learn how to do everything that we know how to do. Then they will no longer export cheap goods to us, and we will lose. This, they claim, is what Samuelson’s theoretical paper describes. If so, then it does not really describe outsourcing.

LAST UPDATE: Douglas Irwin – who’s read the paper – is underwhelmed. This is from an e-mail he sent to me:

[Samuelson’s paper] doesn’t have much to do with outsourcing. If a foreign country experiences technological progress in a home country’s export industry, it can deteriorate the terms of trade of the home country and make it worse off (not vis a vis autarky, but its previous trade situation). We’ve know this since the U of C’s great Harry Johnson pointed it out in the 1950s…. Pretty thin stuff.

VERY LAST UPDATE: One of the commenters linked to Joe Stiglitz’s outsourcing essay in the Singapore Straits-Times from May of this year. That essay contains the following:

Many of globalisation’s advocates continue to claim that the number of jobs outsourced is relatively small. There is controversy, of course, about the eventual size, with some claiming that as many as one job in two might eventually be outsourced, others contending that the potential is much more limited.

Sounds dispassionate, except for one thing — I have not seen any estimate even remotely suggesting that “one job in two might eventually be outsourced.” That’s way higher than any of the upper bound numbers I’ve seen (the highest I’ve seen is 30%). Readers are invited to post a link to any study that suggests otherwise.