Printable page E-mail this

IN EUROPE, immigrants are blamed for rising crime and for stealing jobs which would otherwise go to native workers. So it is to America that defenders of immigration turn for ammunition. There, opposition to immigration is more muted. Immigration is widely perceived to bring many benefits, and not just in the form of people willing to do the dirtiest jobs that natives no longer want to do. They also come in the form of entrepreneurialism and innovation: imagine Silicon Valley without immigrants. Certainly, immigration entails some costs, the perceived view goes. Immigrants take some jobs that locals could otherwise have had. Yet on balance, the effect is positive, if not always easily measurable.

At first blush, the new findings of two professors at Columbia University, Donald Davis and David Weinstein, appear to challenge this view. Their analysis*, which relates uniquely to America, claims that immigration into the United States costs native workers around $72 billion a year, equivalent to nearly 1% of GDP. That is a much bigger figure than economists previously reckoned. As it happens, it is about the same scale of losses which America makes from pursuing protectionist trade policies.

The starting point for the new analysis is America's highly productive economy. Thanks in large part to its technological lead, the United States attracts flows not just of unskilled labour, but of skilled labour and capital too: all kinds of workers and capital flow to the world's biggest economy.

Using census data, Mr Davis and Mr Weinstein calculate that 11.7% of the American labour force was born abroad. This is a high proportion. It reflects a large rise in immigration in recent years—up from about 250,000 people a year in the 1950s to 1m a year now. Immigrants have a higher participation rate in the workforce than do native workers, which also helps boost their proportion of the labour force. And they account for an increasing share of population growth. As for their level of education, other survey data suggest that, of those workers who emigrate to America, the proportion with college degrees is, at 29%, only slightly less than that among native workers, at 32%. At the bottom of the labour pile, however, the proportion of immigrants with only the scantiest education is much higher than among indigenous workers.

The enormous capital flows which America has recently enjoyed have also had a big cumulative impact. For the past two decades, foreign capital flows have accounted in any one year for 5-21% of gross capital formation. Using previous work as the basis for their calculations, the authors conclude that in 1998—the latest year for which they have run the numbers—the foreign-financed capital stock accounted for 11.8% of the total. That is almost exactly the same proportion as the foreign-born share of the labour force, which, say Messrs Davis and Weinstein, reinforces the view that all factor inflows, capital and labour, should be looked at together. Hitherto, they have been studied largely in isolation.

It is not hard to see that such big factor inflows will increase America's output of goods. But one consequence of this is to lead to the conclusion that native American workers pay a cost. For as the production of American goods rises, so their price on world markets—that is, America's terms of trade—will tend to fall. In 1998, the authors calculate that the deterioration in the terms of trade was worth 0.7% of GDP, or a little over $58 billion.

The flows of labour and capital into America also lead to a lower output in the rest of the world than would otherwise have been the case. Foreign output therefore becomes scarcer in relation to American output—and so pushes American relative prices (or America's terms of trade) down further. That takes the loss to the terms of trade up to 0.8% of GDP.

When the authors refine their calculations still further, by making an adjustment to take account of the different skills composition of immigrant labour (something that, in theory at least, could have improved America's terms of trade) the total losses to the terms of trade rise again, to 0.9% of GDP, or $80 billion. And nine-tenths of these losses, or $72 billion, would, they calculate, be carried by American-born workers.

Opponents of immigration everywhere will seize on the new calculations with glee. Yet the authors urge caution. For a start, their work applies specifically to America, not to every country with a high rate of immigration. What is more, even in the American case, the findings do not lead to the conclusion that immigration is bad. There may be costs for native workers in America, yet the gains to immigrant workers far outweigh them. World income is higher as a result of the immigration process: it is on the distribution of benefits that Messrs Davis and Weinstein have thrown new light.

The authors leave open the possibility that immigrants may bring with them important benefits not reflected in their analysis. They propose more work to discover what these might be. At bottom, the authors make the case for a new analytical framework.

That one is needed is a sad reflection on much of modern economic research. Too often, for instance, economists preoccupied with labour flows have little (if anything) to do with international-trade theorists—a vital omission. The authors' call for economists in different fields to communicate more effectively—even, they might have said, to show some interest in what other researchers are working on—is surely right.