Keep extending jobless benefits, and long-term unemployed people will be less motivated to find work.

That’s the conventional conservative wisdom, and one of the primary arguments GOP lawmakers have used to hold up extending federal emergency unemployment benefits for 2.3 million Americans since the beginning of 2014.

But new research suggests that separating jobless people from long-term unemployment benefits isn’t putting more of those people back onto the payrolls. “Losing benefits isn’t prodding unemployed back to work,” says a headline this month on the numbers-crunching website Five Thirty Eight.com, which measures and predicts all manner of social, political, sports and economic trends.

On April 7, the U.S. Senate approved a jobless benefit extension retroactive five months to when benefits expired in December . The bill was adopted in a bipartisan 59-39 vote in the Senate, but it is so far dead on arrival in the Republican-led House, where Speaker John Boehner (R-Ohio) said that any bill must include incentives to create jobs. “I told the president I would consider this as long as it was paid for and as long as there were provisions attached that would actually help the economy and help people get back to work,” he said.

On Dec. 8, 2013, Kentucky Sen. Rand Paul said on Fox News, “When you allow people to be on unemployment insurance for 99 weeks, you’re causing them to become part of this perpetual unemployed group in our economy. “While it seems good, it actually does a disservice to the people you’re trying to help.”

Is it a disservice? No, says Ben Casselman, writing for Five Thirty Eight . “The evidence doesn’t seem to support that theory,” he wrote on April 14. Acknowledging that they only used three months of job market data tracking, the group found that short-term unemployed “are seeing a steady improvement” in job prospects, “but long-term jobless are not.” He added: “The cutoff of federal unemployment benefits doesn’t seem to be helping the long-term unemployed get back to work.”

That conclusion is shared, surprisingly, by the conservative American Enterprise Institute. Writer  James Pethokoukis looked at the state of North Carolina, where conservative lawmakers last July dropped the state’s average weekly unemployment benefit of $298 by one third, while reducing the number of benefit weeks that would be available.

“So how’s that worked out in the Tar Heel State?” Pethokoukis asked. “Well, if you listen to Republicans, it’s worked out pretty well. The state’s unemployment rate has dropped to 8.0 percent in October from 8.8 percent in June. (And then to 6.4 percent in February). So clearly cutting jobless benefits creates jobs and gets residents back in the workforce, right? When you dig a bit deeper, things look less bright.”

In June, the month before benefits were cut in North Carolina, the American Enterprise Institute writer found that there were 416,314 residents classified as unemployed. By October, there were 371,756 – a decline in unemployment of 44,558. But it wasn’t because they all suddenly found work: over that same period, the number of employed residents only rose by 1,902. So what happened, the AEI asked, to those 42,656 residents who left unemployment but did not move to employment?

“It looks like the cut in unemployment benefits moved people out of the labor force rather than into employment,” the writer concluded. “Did reducing the number of North Carolina residents eligible for federal extended unemployment benefits boost employment? These data suggest it did not, a reality Washington policymakers might want to consider.” If those jobless workers had not been moved off of the unemployment rolls, Pethokoukis said the state’s jobless rate would have actually increased to 9.1 percent.

Politifact, which examines the truthfulness of statements by politicians, said Senator Paul’s claim that long-term jobless are less likely to find work is backed up by research – to an extent. “But if Paul means to suggest that ending benefits actually increases employment, the support for that proposition is more mixed,” Politifact says. “In particular, there’s evidence that unemployed people may drop out of the labor market rather than find a job when they stop getting benefit checks.”

There’s an ongoing problem with government jobless numbers, Politifact pointed out, that never seems to get resolved. The unemployment rate can drop either because jobless people find work, or because they move out of the labor force and stop collecting benefits, and are thus out of the official equation used to calculate the unemployment rate.

So what happens to those uncounted unemployed people? Do they go live in mom and dad’s basement? Do they go live off the grid under and overpass or in the woods? No one really knows – it’s one of the few things for which the federal government doesn’t keep statistics.