Despite the paralysis in Washington, D.C., there is a consensus that job creation is a prime policy objective. A new analysis of American employment from the McKinsey Global Institute, the nonprofit research arm of the big consultancy firm, sheds welcome light on a murky subject.

The current “jobless recovery” is a new phenomenon in post–World War II history, an ugly hallmark of the 2000s. A common recovery index is the number of months it takes for total employment to return to its prerecession peak.

In all postwar recoveries until the 1990s, the trough-to-peak employment recovery took about six months. With the anemic growth after the early-2000s recession, however, it took thirty-nine months to reach prerecession employment levels. And at the current “recovery” growth rate—a near-recessionary 1.4 percent or so—it will take a full five years to claw back the jobs lost in the 2008 crash.

What has changed according to the McKinsey analysts is the behavior of employers. Until the 1970s, employers responded to a 1 percent recessionary fall in GDP with job cuts equivalent to about a third of 1 percent of GDP, while the rest came from “productivity losses”—that is, lower profits. Since then, employers have been shifting more and more of the burden of a recession to their workers. In the recent crash, labor bore essentially all of the costs of the downturn.

Employers interviewed by McKinsey blamed the shift on global competition. But that’s nonsense: customers don’t care whether a price cut is financed with layoffs or lower margins. But CEOs and shareholders certainly do. What’s changed is the loyalty employers once felt toward their workers, and the relentless drive of executives and the investing classes to achieve mythic levels of wealth.

The McKinsey researchers calculate that at the current rate of “recovery” growth, the economy will barely absorb new workers entering the labor force, and that unemployment rates will continue at the 9 to 10 percent level essentially forever. To return to “full employment” by 2020, defined as the 5 percent rate we had in 2007, we will have to add some 22 million new jobs. That would require increasing the current rate of job creation by about two and a half times and sustaining it for the rest of the decade—a daunting prescription.

Three industry sectors will be key: health care, business services, and manufacturing. Manufacturing is a top priority not because it’s so large but because its employment base is shrinking so fast. Depending on how you count, the United States is still the world’s largest manufacturer and U.S. output is growing strongly. But manufacturing productivity is growing so fast that companies can raise output with fewer and fewer workers.

The answer—easier stated than accomplished—is to increase America’s global manufacturing share. There are some favorable straws in the wind. Wage increases in China and India, and the complexities of long-distance supervision, are causing many companies to rethink their love affair with “offshoring” work. A falling dollar would help as well, by lowering the foreign prices of U.S. exports. So would a big infrastructure construction program with “Buy American” strings attached.

The sectors that are now the country’s strong suits are health care and business services, both among the largest employers. Although both comprise large numbers of lower-skill positions (orderlies and janitors), each has a strong tilt toward well-paying jobs that require substantial education—physician’s assistants and radiation therapists are hot job categories in health care, while business services employs armies of computer programmers, accountants, and industrial engineers.

Alarmingly, there is a growing mismatch between the skills of the American work force and the new well-paying jobs being created. Throughout history, American educational attainment has always grown faster than the skill demands of the job market. That positive mismatch undoubtedly has much to do with the high level of American innovation. But over the past three decades, that education/jobs mismatch has flipped, and American education levels are no longer keeping up with the market.

Currently, fewer than a third of workers have four-year college degrees; that rate is still growing but very slowly compared to projected demand. Just as important, the majority of graduates are weak in the sciences, mathematics, and technical skills that are most in demand.

To sum up, the McKinsey report shows (barely) plausible paths toward a robust, job-rich economy. But to get there will require a national commitment, determined leadership, and large-scale investment. It’s not impossible, but you can’t like the odds.