Are we missing a couple million jobs? These would be jobs that exist but lack workers to fill them. The notion that the recovery is being hobbled by too few skilled workers is seductive. It might explain today’s stubbornly high unemployment and why aggressive government policies to promote recovery have been so ineffective. Low interest rates and big budget deficits can’t cure bottlenecks in the job market. They can’t make construction workers into computer scientists.

There’s only one problem with this story: It’s mostly fiction.

Superficially, it seems compelling. Consider the evidence. The Labor Department’s latest estimate (February) of job vacancies was 3.9 million, up 80 percent from the latest low in July 2009. Just recently, the Wall Street Journal reported a “shortage of help hits nursing homes.” Employer complaints of scarcities abound, notes Darrell West of the Brookings Institution. Even in 2010, manufacturers said they couldn’t fill 227,000 jobs. More than half (55 percent) of state governments report difficulty hiring for IT openings. Microsoft says it struggles to fill thousands of computer science slots.

Then there’s the “Beveridge curve,” after English economist William Beveridge (1879-1963). He noted a relationship between unemployment and job vacancies. When unemployment is high, vacancies are few, because workers quickly fill them. But when unemployment falls, vacancies actually rise, as employers scramble to meet their needs. What’s puzzled economists is that there are more vacancies now than were expected at today’s high unemployment rate. This suggests job mismatches: workers lacking needed skills or living where the jobs aren’t.

On closer inspection, the logic unravels.

For starters, most vacancies are routine. They’re just-posted job openings or those reflecting workers retiring or switching employers. Some skill shortages always exist in a sophisticated economy, says Brookings economist Gary Burtless. “Are they serious enough to explain today’s high unemployment rate?” he asks. “The answer is an emphatic no.” In April, the unemployed totaled 11.7 million; another 6.3 million people wanted a job but weren’t looking. These figures dwarf the number of vacancies.

If shortages were widespread, Burtless and other economists argue, wages would be rising rapidly as employers competed for scarce skilled workers. There’s scant evidence of this. From April 2012 to April 2013, average hourly manufacturing wages rose 1 percent, reports the Labor Department. Over the same period, the gain for all private nonfarm workers was 1.9 percent. Among computer programmers, inflation-adjusted wages have remained flat for a decade, says a study by the Economic Policy Institute, a liberal think tank.

Similarly, economist Paul Osterman of the Massachusetts Institute of Technology surveyed 925 manufacturing establishments in 2012 about worker shortages. Three-quarters reported no shortages, defined as vacancies lasting three months or more. Of the rest, most shortages were less than 10 percent of their workforces. “Very few firms responded by reducing production,” says Osterman. “The most common reaction was to outsource” domestically — to send business to other American firms. Labor bottlenecks haven’t crimped recovery, he concludes. A study by economists Edward Lazear of Stanford University and James Spletzer of the Census Bureau agrees.

So, what explains more vacancies at given unemployment levels (a.k.a the shifting Beveridge curve)?

The answer almost certainly involves employers, not workers. Businesses have become more risk-averse. They’re more reluctant to hire. They’ve raised standards. For many reasons, they’ve become more demanding and discriminating. These reasons could include (a) doubts about the recovery; (b) government policies raising labor costs (example: the Affordable Care Act’s insurance mandates); (c) unwillingness to pay for training; and (d) fear of squeezed profits. In practice, motives mix.

The chief victims of this shift in business behavior seem to be the long-term unemployed (more than six months), as some fascinating research by economists William Dickens and Rand Ghayad of Northeastern University suggests. By their estimates, virtually all the reduction in hiring falls on this group, regardless of their other characteristics (age, education, industry experience). Many firms seem to have concluded that the long-term jobless are damaged goods.

To test this, Ghayad e-mailed fake resumes to hundreds of firms in response to job postings. All the fictional candidates were 2005 college graduates with identical skills; they differed only in their length of unemployment (0-12 months) and experience in the hiring industry. The long-term unemployed received few responses. In many cases, software filters apparently eliminated their applications automatically. Similarly, six months of joblessness erased the value of industry experience. Employers preferred candidates with less joblessness over those who had worked in their industry.

No doubt the economy’s future would be brighter if workers had more skills. But we shouldn’t mistake a long-term goal for a short-term problem. The idea of widespread labor shortages in an era of high unemployment seems absurd — and is. Today’s crucial scarcity is not skills. It’s confidence.

Read more from Robert Samuelson’s archive.