The noisy agitation over offshore "outsourcing" died down a bit. But assertions repeated with such stridency often remain believed long after they are proven false. The original story focused on a supposedly permanent loss of high-paid jobs among computer specialists. And these jobs were said to have been lost to India, rather than to cyclical business failures.

 "What are the unemployed computer engineers and information technology workers supposed to retrain for?" asked economist Paul Craig Roberts earlier this year. "I know young software engineers who are substitute teachers in middle schools."

 At the same time, however, the Bureau of Labor Statistics (BLS) released projections of the occupations expected to see the most rapid job growth from 2002 to 2012. The following sample shows the average salary in 2002 (in parentheses) followed by the estimated increase in jobs through 2012:

 Computer and information scientists ($80,510), up 29.9 percent. Computer software engineers-systems software ($75,840), up 45.5 percent. Computer systems analysts ($64,890), up 39.4 percent

 These and other computer specialists account for more than 96 percent of the nearly 3 million jobs in the BLS category called "Computer and Mathematical Science Occupations" -- such as software engineers, programmers, systems analysts and database administrators. These jobs grew at an astonishingly rapid pace in the late '90s and then declined relatively modestly in 2001 and 2002. By 2003, when sensational news stories appeared about a supposedly horrific loss of these jobs, the news was far behind the reality. The number of computer specialist jobs rose by 54,390 in 2003, or 2 percent. From 2002 to 2012 the BLS expects there to be more than a million additional jobs for computer specialists -- an impressive increase of 36 percent.

 What accounts for the widespread misapprehension that many of these jobs vanished last year and that the future will only be worse? One problem is the chronic inability to distinguish between cyclical recessions and secular trends. To take an extreme example, payroll employment in "computer systems design" doubled between April 1996 and April 2001, rising from 679,300 to 1,339,200, before settling back to 1,098,900 by April 2004. People focus on the bust and forget the boom, expecting to see every occupation immediately restored to its previous cyclical peak.

 Another reason the press exaggerated the loss of high-tech jobs is that the extreme cycle in Silicon Valley has been wrongly generalized into a nationwide phenomenon. The unemployment rate in San Jose fell from 6.7 percent in July 1994 to 1.3 percent by December 2000. But local unemployment subsequently soared to 9.1 percent by January 2003 -- a two-year shakeout as frightening as the corresponding collapse of Silicon Valley stocks. By this March, San Jose unemployment was still 6.8 percent, nearly double the 3.6 percent rate in Orange County. Yet even in 2000-2002, when computer specialist jobs were declining nationwide, Jacob Kirkegaard at the Institute for International Economics (iee.com) showed such jobs were rising substantially in Virginia, Maryland and New Jersey.

 The notion that huge numbers of computer specialist jobs are being outsourced to India implies the experts at our own Bureau of Labor Statistics must be horribly mistaken. This insult to the BLS is not based on any economic facts but on estimates from Forrester Research -- former cheerleaders for the dot.com craze, with no expertise in economics. In an overly publicized 11-page $249 report, "Forrester has increased its estimate of how many U.S. services jobs will go offshore in the near term. Long term, we believe that our previous projection of 3.3 million by 2015 is still accurate."

 To put such a blue sky "projection" in perspective, The Economist noted the United States routinely loses about 30 million jobs every year (millions even quit!) but gains even more. The gullible business press failed to notice that Forrester is not talking about a net loss of jobs at all. To do so would require serious economics, of the sort used in BLS projections. Forrester doesn't do economics, or at least nothing recognizable as economics. Continued...