Can technology set off a new boom in job creation? The question is a fundamental one for the American economy given that policy makers in Washington often look to the technology sector to pick up the slack in the employment market. Meanwhile, the fortunes of Silicon Valley start-ups continue to be closely followed, in part because of the spectacular wealth they can generate for their founders, but also because of the assumption that these new companies are a significant source of new employment.

So it will likely disappoint many people that four prominent economists assembled for a recent panel discussion to explore the link between technology and job creation were, in large part, bearish in their outlook. Some went so far as to suggest that technology actually increases unemployment and adds to other problems in the U.S. economy, notably the growing wage disparities between an extremely elite group of earners and everyone else.

The discussion, titled “Can Tech Power the Next Jobs Boom?” took place at Wharton’s San Francisco campus and was co-sponsored by the Churchill Club, a Silicon Valley business and technology forum.

Several troubling data points emerged during the evening, including one offered by Erik Brynjolfsson, a professor at MIT’s Sloan School of Management and director of the MIT Center for Digital Business.

Working with fellow MIT professor Andrew McAfee, Brynjolfsson compared the market capitalization and payrolls of four of the biggest tech companies. His conclusion: While the companies had astronomical values on Wall Street, their job production was minimal.

The four — Apple, Amazon, Facebook and Google — at the time had a market cap in the neighborhood of $1 trillion, which is roughly 6.25% of the combined market cap of all U.S. companies. But the four employ about 190,000 people, fewer than the number of jobs the U.S. economy needs to add approximately every six weeks to just keep pace with population growth. The implication, said Brynjolfsson, is that even hugely successful tech companies cannot be counted on to create the kinds of jobs the economy needs.

Brynjolfsson also described what he called a “superstar” effect in technology-related wealth distribution, a trend that has become pronounced in the last decade. In recent years, he said, the majority of GDP growth has benefited a very small part of the population, less than 1%. In many cases, even college-educated workers are not sharing in the growing pie. “It’s becoming a winner-take-all situation,” he said.

“Technology doesn’t automatically lift the fortunes of all people,” Brynjolfsson noted. “It’s something of a paradox. Profits have never been higher, innovation is roaring along, GDP is high, but job creation is lagging terribly, and the share of profits going to labor is at a 60-year low. This is one of the most important issues facing our society.”

Citing the work of economist Joseph Schumpeter, Brynjolfsson noted that technology has historically provided “creative destruction” for an economy, causing some jobs to disappear while bringing others into existence. “But the last 10 years have been different. Technology simply hasn’t been creating jobs as it did before.”

It’s a double-barreled effect, Brynjolfsson added. Not only are today’s technology companies creating fewer jobs, but the products they make, notably computerized automation equipment, often lead to further job losses in other parts of the economy. These second-effect job losses are further encouraged by off-shoring and by the declining power of labor unions.

“Improvements in technology can improve productivity,” he said. “For most of the 20th century, those productivity increases were associated with job growth and growing wages. But there is no economic law saying that always has to be the case. It’s quite possible to make the pie bigger, but with most people having a smaller slice. That is what has been happening recently.”

Flint, Michigan, vs. Austin, Texas

Somewhat similar themes were echoed in the remarks of Enrico Moretti, an economics professor at the University of California, Berkeley, whose recent book, The New Geography of Jobs, was widely praised for its insights into the changing nature of the American workforce.

According to Moretti, there was not one single labor market in the United States, but hundreds, corresponding to metropolitan areas. Overall, he said, these markets fell into three different groups: those doing well in the new technology economy, those doing poorly and those hanging in the balance.

The differences between the emerging job market “winners” and “losers” are striking. In 1980, Moretti noted, both high school and college graduates in Austin, Tex., made half as much as their counterparts in Flint, Mich. Now, however, those ratios are reversed, and the gap between wages in Flint’s rust belt and the booming tech sector in Austin continues to grow. “So when people ask, ‘Can technology create the next job boom?’ my answer is, ‘It depends,'” said Moretti.

While technology may not be making as many jobs as it once did, the jobs it does create are among the most economically valuable. Moretti noted that the average tech position creates five additional jobs in various support industries, from doctors to hairdressers to dog walkers. However, the “multiplier effect” for manufacturing jobs is much lower: 1.6 instead of 5. Much of that, he added, was simply the result of the higher wages generally paid by tech jobs.

Because of that high multiplier, the majority of people will never be employed directly by technology, even in thriving tech hubs like Silicon Valley. “Technology jobs will be the minority, [about] 30%,” Moretti said. “What’s important is to build that foundation of 30%….”

Moretti and several other panelists suggested that the jobs being lost in the traditional manufacturing sector are never coming back. Or, if they do, there will be far fewer of them, as is the case with the heavily automated manufacturing facilities that Apple has discussed opening in the United States.

Michael Chui, who studies job creation for the McKinsey Global Institute, said that “employment transparency” has become a crucial issue for college students attempting to pick a field of study. They need to know where the jobs of tomorrow are likely to be, but the data is not available to them during the period of their lives when they are making decisions that will weigh most heavily on their career options.

He also said the United States needs to increase the number of college graduates studying science, technology, engineering and mathematics, the so-called “STEM” curriculum. More than 40% of China’s college population are in a STEM field, and the figure in Germany is 28%. But in the United States, said Chui, it is only 15%.

Even within STEM, he noted, priorities may need to be readjusted. For example, a traditional elite technical education typically includes a healthy dose of calculus. But perhaps statistics should receive more attention, Chui said, because of the need for future managers to be able to more intelligently use the huge mountains of data now being routinely collected by businesses.

Education Is Key

The fourth member of the panel, Hal Varian, chief economist at Google and an emeritus professor at the University of California, Berkeley, told the audience that the “secret” of guaranteeing oneself employment in an increasingly technology-oriented society is “to make yourself an expensive complement to something that is becoming cheaper and more ubiquitous.” As an example, he echoed Chui’s reference to the growing demand for “data scientists” who can work with businesses’ ever-larger databases.

Varian also urged greater appreciation for the “supporting” jobs created by technology, saying many of them, like doctors and lawyers, require sophisticated educations and usually provide excellent salaries.

On a more positive note, Varian offered a global, long-term perspective on the U.S. employment problem. In the last 30 years, he noted, more than a billion people have been lifted out of poverty. Economists of a generation ago might have assumed that the global process would be a zero-sum game; that for there to be a billion big “winners” in the developing world, there would need to be an equal number of losers elsewhere, notably in parts of the world, like the U.S., that are already prosperous. While Varian noted that those better-off regions are indeed experiencing a number of employment-related challenges, “in some senses, it’s amazing that we’ve done as well as we have in this country, considering these massive global changes.”

In terms of specific policy recommendations, Moretti said he favored a substantial and permanent investment tax credit, pointing to the importance of federal support for technology research. He described the many technology successes of the Pentagon’s DARPA (Defense Advanced Research Projects Agency), the agency whose early research money made possible everything from the Internet itself to the self-driving car that Google has recently popularized.

Proper education was also frequently stressed by the panel. “There is no better time than right now to be an entrepreneur who can make use of all these new technologies,” said Brynjolfsson. “But there is no worse time to be a worker with no special skills, because all of those jobs are being automated.”