Since January 2001, a three-year period during which the economy has experienced one year of recession and two years of recovery, the U.S. economy has lost 2.6 percent of its private-sector jobs. These losses are not evenly distributed. Construction employment has declined by only 0.1 percent, and employment in oil and gas extraction by 0.7 percent.

Employment declines in manufacturing and knowledge jobs, however, have been dramatic.

Tables prepared by Charles McMillion of MBG Information Services from government data show employment in primary metals down 24 percent; machinery 21.6 percent; computer and peripheral equipment 28 percent; communications equipment 38.8 percent; semiconductors and electronic components 37 percent; electrical equipment and appliances 22.8 percent; textile mills 34.1 percent; apparel 37.3 percent; chemicals 8.3 percent; plastics and rubber products 13.8 percent; Internet publishing and broadcast 40 percent; telecommunications 19.4 percent; ISPs, search portals, data processing 22.6 percent; securities, commodity, investments 6.8 percent; computer systems design and related 17 percent.

Where has employment grown? Private service sector jobs have declined by 0.1 percent. Growth in state education jobs and local government jobs has boosted overall service employment by 0.6 percent.

During the past 12 months, the second year of economic recovery, the U.S. economy eked out 57,000 net new jobs in nontradable low-pay services, leaving the economy millions of jobs short of normal performance.

This is not a picture of an economy that is doing well. Low income jobs in nontradable services are the only sources of employment, while high value-added jobs that pay good incomes continue to disappear.

This job record is not one of a powerful U.S. economy dominating world markets and building consumer incomes for sustained recovery. It is not a record that promises jobs for university graduates. It is not a record that promises a future.

Economists have apologies, but no real explanations, for the loss of jobs in tradable goods and services. They are careful not to blame outsourcing of manufacturing and service jobs, which they claim creates as many new jobs as it loses.

Outsourcing platforms, especially the knowledge jobs platforms in India, are commissioning U.S. think tanks and consultants to do "independent" studies that prove "outsourcing is good for the U.S."

Certainly, the people who are benefiting from outsourcing want us to think it is good for us.

For years as U.S. multinationals moved manufacturing offshore, Americans were told that their future was in "knowledge jobs." Today, knowledge jobs are being moved offshore more rapidly than manufacturing jobs were.

What are the unemployed computer engineers and information technology workers supposed to retrain for? What high value-added job can't be outsourced?

Only those in the nontradable sector, such as dentists and surgeons. If everyone becomes a dentist or a surgeon, those incomes will be driven down.

Many young engineering graduates have discovered that they invested in acquiring skills for which there are no jobs and are headed to law schools in an effort to retrieve their future. I know young software engineers who are substitute teachers in middle schools, and others who are trying to organize rock bands to play the club and bar circuits.

They have no idea what to retrain for, and neither do the economists who tell them retraining is the answer.

What is happening is easy to discern from the daily announcements of the multinationals themselves. Cheap foreign labor is being substituted for U.S. labor over a wide range of goods and services produced for U.S. markets. Americans are losing the incomes associated with the production of the goods and services that they consume. Because of extraordinary differences in domestic prices and living standards, foreign labor can offer its services to U.S. capital and technology for far lower wages than can Americans. Capitalists maximize profits, not employment in their home countries.

This is a new development. Until the collapse of world socialism and the rise of the Internet, first world capital stayed in the first world, and offshore production was not the motive of foreign investment.

As offshore production takes hold and spreads, the United States will lose more high value-added jobs.

A rise in Asian currency values could dampen and eventually end the exodus of jobs from America. The question is: How long does the exodus last before there is a new equilibrium?

In an important new work in trade theory (Global Trade and Conflicting National Interests, MIT Press), Ralph E. Gomory and William J. Baumol show that it very much matters which industries and occupations countries retain. They explode the free trade assumption that free trade always produces mutual gains. Gomory and Baumol show that in many cases, perhaps a majority, gains for countries come at the expense of other countries. The authors explain why the "man in the street," so derided by economists, is right in his understanding that free trade produces winners and losers.

University of Maryland economist Herman Daly has been making this point for 15 years and Senator Charles Schumer and I more recently. Now, Gomory and Baumol have provided powerful demonstration that trade has winners and losers. Right now, the United States is losing.

COPYRIGHT CREATORS SYNDICATE, INC.

Paul Craig Roberts was Associate Editor of the WSJ editorial page, 1978-80, and columnist for "Political Economy." During 1981-82 he was Assistant Secretary of the Treasury for Economic Policy. He is the author of Supply-Side Revolution: An Insider's Account of Policymaking in Washington.