This minimum wage business is tricky. On its face, raising the wage seems an easy way to fight poverty. Just pay low-wage workers more. After all, some scholarly research finds that, within reasonable limits, there’s no job penalty. A higher minimum doesn’t reduce employment much, if at all. By and large, that’s the position of the Obama administration, congressional Democrats and liberal groups. Unfortunately, it may not be that simple.

Democrats propose raising the present federal minimum of $7.25 an hour to $8.20 this year, $9.15 in 2015 and $10.10 in 2016. Assuming no job losses, almost 28 million workers would benefit by 2016, estimates the Economic Policy Institute (EPI), a liberal think tank. That’s about 17 million who now make less than the proposed minimums, plus 11 million slightly better-paid workers who would get increases to keep them above the minimum.

Someone working 40 hours a week at the minimum would see annual wages go from $15,080 now to $21,008 in 2016. Today’s annual wage is about 20 percent below the federal poverty line for a family of three, while the 2016 wage would slightly be above the line for a family of three (though not of four), says EPI. Not all workers would receive big increases, because many work part-time (46 percent), don’t stay for a full year or already are above the minimum. Still, wage gains could be sizable.

Economist John Schmitt of the Center for Economic and Policy Research, another left-leaning think tank, says that recent minimum-wage studies find that “modest increases” have “little or no employment effect.” Businesses turn to other ways of absorbing the added costs rather than reducing payrolls or workers’ hours, he says. Better-paid workers mean less turnover. This cuts firms’ recruitment and training costs; it also raises workers’ productivity, because they’re more familiar with their jobs. Finally, firms adopt “small price increases.”

All this sounds plausible; it may also be incomplete.

For starters, the minimum wage is a blunt instrument to aid the poor because it covers many workers from families that are well above the federal poverty line. By the administration’s figures, 53 percent of workers who would benefit from a higher minimum come from families with incomes above $35,000, including 22 percent with incomes exceeding $75,000.

Next, economists still disagree on the job effect. In studies — and their review of other studies — economists David Neumark and J.M. Ian Salas of the University of California at Irvine and William Wascher of the Federal Reserve conclude that higher minimums do weaken low-wage employment. Under plausible assumptions, even a small effect (say, a 1 percent job loss for each 10 percent increase in the minimum) implies nearly a million fewer jobs over three years.

But scholarly research, regardless of conclusions, may be beside the point. Businesses don’t consult studies to decide what to do. They respond based on their own economic outlook. They may not react to a higher minimum wage now as they did in the past. Two realities suggest this.

First, the proposed increase is huge. By 2016, it’s almost 40 percent. Similar gains usually have occurred when high inflation advanced all wages rapidly. The minimum mainly kept pace. That’s not true today. Compared to average wages, the proposed hike in the minimum appears to be the largest since the 1960s.

Second, businesses have been reluctant job creators. They curb hiring at the least pretext. They seem obsessed with cost control. The Great Recession and the 2008-09 financial crisis spawned so much fear that they changed, at least temporarily, behavior. Firms are more cautious.

Would employers take the minimum’s steep costs in stride — or react by cutting hiring and automating more low-paid jobs (example: supermarket checkouts)? That’s the crucial question. Which matters more for low-income workers: added jobs or higher incomes? There’s a powerful symbolism to raising the minimum, but the notion that it can be boosted sharply without any job penalty may be a mirage.

Read more from Robert Samuelson’s archive.