TCS Daily

Taking Advantage

By Arnold Kling - January 20, 2004 12:00 AM

"England may be so circumstanced, that to produce the cloth may require the labour of 100 men for one year; and if she attempted to make the wine, it might require the labour of 120 men for the same time...

To produce the wine in Portugal, might require only the labour of 80 men for one year, and to produce the cloth in the same country, might require the labour of 90 men for the same time. It would therefore be advantageous for her to export wine in exchange for cloth. This exchange might even take place, notwithstanding that the commodity imported by Portugal could be produced there with less labour than in England."
-- David Ricardo

Is the American middle class in jeopardy because modern communications technology enables U.S. firms to use workers in India for tasks such as call-center staffing and software development? Pundits appear to be divided on this issue. However, if you look closely, you will see that professional economists, regardless of ideology, all disagree with the claim that the American middle class will be impoverished by trade with India. We remain loyal to the analysis first propounded by David Ricardo, who would spin in his grave if he could see the contrarian views of outsourcing recently espoused by policy wonk Michael Lind or columnist Paul Craig Roberts and Senator Charles Schumer.

What accounts for the persistent belief that trade with poor countries will make us worse off? Recently, it occurred to me that evolutionary psychology might provide the answer. Anthropologist Alan Fiske has pointed out that there are four ways in which humans transact: on the basis of authority; on the basis of communal sharing; on the basis of equality matching; and on the basis of market pricing. In the era of small hunter-gatherer tribes in which our brains evolved, only the first three were needed. Market pricing is required once you start to interact with strangers.

My hypothesis is that people are not "hard-wired" to understand market pricing, so that they often fall back on the models of authority ranking, communal sharing, or equality matching to guide them. Thus, people interpret trade with India as if it were communal sharing with India. It certainly is true that if we share with India, we will be poorer. However, it is not true that trading with India at market prices will lower our well-being.

Is Math the Issue?

In his book The Blank Slate, Steven Pinker suggested that market pricing is counterintuitive because it involves mathematical calculations. Indeed, Paul Krugman has written that the gulf between economists and noneconomists is due to math phobia on the part of the latter. In 1994, journalist Michael Lind made the claim that American workers were being set back because of international trade. In response, Krugman wrote,

"The question here is not why Lind got these numbers wrong. It takes considerable experience to know where to look and what to worry about in economic statistics, and one should not expect someone who does not work in the field to be able to get it right without some guidance. The question is, instead, why Mr. Lind felt that it was a good idea to make sweeping pronouncements about this subject, when he clearly was unwilling to invest time and energy in actually understanding it."

Krugman's words seem to have had little effect. Lately, in the January-February 2004 issue of the Atlantic Monthly, Lind writes,

"the productivity gains in heavily automated, capital-intensive sectors such as manufacturing, agriculture, banking, and other routine services have gone almost entirely to the investors who own the machines and the software, not to the workers who remain, and certainly not to the workers displaced into other sectors."

This is exactly the same pronouncement that Lind made in 1994. If it were true, then we would observe a fall in the share of national income going to labor. However, the labor share remains as stable today as it was when Krugman pointed out that this data refuted Lind's claim ten years ago. Stability of the share of income means that as the economy grows, workers are gaining as much in percentage terms as capitalists.

Krugman's essay implies that if noneconomists cannot be taught the theory of international trade then they should learn to shut up. This may sound harsh, but advising an unqualified pundit to lay off international trade theory is not unlike advising a massage therapist to refrain from attempting open heart surgery.

My sympathy for Krugman's position increased earlier this month, when I attended The Real State of the Union, an event sponsored by The New America Foundation (a think thank) and The Atlantic Monthly. There, Michael Lind was given an equal spot on a panel with Martin Neil Baily, the former Chairman of the Council of Economic Advisers under President Clinton. Baily is a real economist. His remarks reinforced what I wrote last year for TCS, where I argued that the key to increasing net exports is to increase net national saving. Meanwhile, Lind's comments, such as "we need to create comparative advantage," or "low wages are India's comparative advantage," betray a fundamental misunderstanding of economic concepts.

Ultimately, Krugman despaired of explaining Ricardo to math-phobic non-economists. It is true that in the process of attempting to teach students the Ricardian model of international trade, economic textbooks have settled on using equations and graphs. In spite of this 200 years of experience, I think that there is hope for the math-phobics -- and the rest of this article will attempt to get at the economics of outsourcing without resorting to such devices.

Comparative Advantage

Economists use the term "comparative advantage" to describe Ricardo's analysis of trade. It describes relative productivity, and it is to be distinguished from absolute advantage.

Why do doctors use assistants to type up reports, rather than type them up themselves? Even a doctor who can type faster than a secretary will still use a secretary. The doctor uses the secretary because the doctor's time is more profitably spent practicing medicine than doing typing. Even though the doctor may have an absolute advantage over the secretary in typing, the doctor has a much bigger absolute advantage in practicing medicine. Therefore, we say that the doctor has a comparative advantage in practicing medicine, and the secretary has a comparative advantage in typing.

This example -- which makes sense to most people -- illustrates why there is no need to "create" comparative advantage. As long as two people have different skill sets, there will be comparative advantage. Even if one person has better skills in all areas, there is a pattern of specialization that is more efficient than simply randomly assigning tasks to each person. In baseball, even if Derek Jeter could play first base better than Jason Giambi, we would still say that Giambi's comparative advantage is playing first base.

Comparative advantage is a characteristic that determines the most efficient way to allocate resources. Comparative advantage would exist even on a commune, with no markets and no wages. If the members of a commune rotate all tasks evenly among themselves, there is still comparative advantage -- it's just that such a commune fails to exploit the opportunities for efficiency that would come from specialization.

Equilibrium Conditions

Although we may be able to do away with equations, we cannot do away with what the equations keep track of: equilibrium conditions. An equilibrium condition is a stable situation. In physics, a rock that is lying on the ground can be in equilibrium. A rock that is suspended in midair with nothing holding it up is not in equilibrium. As economist Daniel Davies pointed out in a blog post, the mistakes that can arise from forgetting an equilibrium condition are "fearfully easy to make."

One mistake is to posit an economy in which there are overlooked profit opportunities. For example, it is wrong to assume an economy in which workers are paid less than their marginal product, meaning the worker's contribution to the value of what she produces. Paul Craig Roberts made such a mistake quite explicitly at a recent Brookings Institution event devoted to a discussion of the Schumer-Roberts op-ed piece. Roberts claimed that Indian software workers are paid less than their marginal product. (The transcription reads "margin of product" instead of "marginal product")

It violates an equilibrium condition to suppose that Indian software programmers are paid less than the value of what they produce. Such an assumption raises questions of what is being overlooked. Why don't firms hire more programmers, since they are getting more in value than what they are paying? If the reason that firms do not increase their hiring is that all of the qualified programmers are already working, then why do the workers not demand higher wages? Until wages come in line with productivity, there are unexploited profit opportunities.

General equilibrium does not require that Indian software programmers be paid the same salary as American programmers. However, it does require that they be paid the same per unit of output. If Indian software programmers have one-fourth the productivity of American developers, then the equilibrium wage for Indian software developers is one-fourth the wage of workers in the United States. General equilibrium analysis would say that low wages in India are not a competitive tool; instead, they are a consequence of the productivity differential.

General equilibrium implies that you can have low wages with low productivity or high wages with high productivity, but never low wages with high productivity. Thus, the non-economist's fear of low-wage countries is misguided. In the essay quoted above, Krugman writes, "In fact, one never teaches the Ricardian model without emphasizing precisely the way that model refutes the claim that competition from low-wage countries is necessarily a bad thing, that it shows how trade can be mutually beneficial regardless of differences in wage rates."

Other equilibrium conditions are trade balance and full employment. To see why trade balance is an equilibrium condition, imagine a barter economy. In a barter situation, in order to pay for Indian software programmers, we would have to provide goods or services to India in exchange. Thinking in terms of barter helps to clarify the symmetry in trade that is an equilibrium condition.

Because of symmetry, as American software programmers are displaced the demand for labor in export industries increases. Wages go up in those industries, inducing more Americans to work there. Some of these new workers come from the software industry, and some come from other industries which in turn must hire from the software industry to fill in. The most logical outcome with symmetry and comparative advantage is that American workers shift to where they are relatively more productive, which means that their wages go up.

Full employment is another equilibrium condition. In the long run, wages adjust upward in some sectors and down in others, and workers seek the best opportunities relative to the skills that they have or are able to acquire. In fact, that process plays out over time and gets caught up in issues of macroeconomics, which is beyond where I want to go here. These days, an economist might invoke monetary policy rather than equilibrium as the guarantor of full employment.

When people claim that the overall economy has "lost" jobs because of trade, they are ignoring an equilibrium condition and speaking as if there were no symmetry (or no Federal Reserve). That is a fundamental error -- the equivalent of ignoring gravity in physics or evolution in biology.

What is true is that the demand for labor can fall in some easily-identified industries while it increases in other industries in ways that are less visible. Baily is one of a number of economists who advocate programs to ease the plight of workers displaced by trade. I myself tend to find the case for government intervention less persuasive. The way I see it, in a dynamic economy there are plenty of ways that new jobs are created and old jobs are made obsolete. I trust individual initiative more than government programs to deal with those circumstances. Still, at least Baily and I (and Krugman, for that matter) can discuss our ideological differences within a reasonable economic framework. With the noneconomists like Lind or Roberts, you cannot get to policy until you first clean up their conceptual mess.

Violating Equilibrium Conditions

In his Atlantic article, Michael Lind's thesis is that unless the government steps in, the middle class will disappear as businesses increase productivity through outsourcing and the use of technology. As he portrays it, today's office workers are going to wind up working cash registers at fast food restaurants.

This sort of downward mobility has been forecast repeatedly by noneconomists over the past several decades, and it has yet to occur. Economists see upward mobility as more likely, because it is extremely difficult to obtain Lind's prediction without violating equilibrium conditions. If wages and productivity are linked, and Indian office workers become as productive as American office workers, then Indian office workers will earn American wages, and everyone will be better off. The only way that Indian office workers can continue to earn less than American office workers is if Indian office workers are less productive than American office workers.

The only way that Indian office workers can displace American office workers without being as productive as American office workers is if there is some other industry in which America's productivity edge over India is even greater. In that case, American labor will tend to shift to that relatively more productive industry. That is more likely to lead to higher wages in America than to lower wages.

I cannot go so far as to claim that it is impossible to construct a theoretical example that satisfies equilibrium conditions in which American workers earn lower wages as a result of globalization. I believe that I could construct such an example, but in order to work it requires a very large difference in wage rates between two types of Indian workers -- the average wage differential between India and the United States is not the driver. I also do not think that the example would have any real-world relevance, given actual data.

Moreover, I can state with complete confidence that Michael Lind, Paul Craig Roberts, Senator Charles Schumer and other math-challenged pundits could never in a million years show how their thinking about trade could be made consistent with equilibrium conditions. Giving such commentators credence in discussions of economic policy is like trusting your open-heart surgery to someone who has never so much as dissected a cat.


TCS Daily Archives