Hanging over the $800 billion Paycheck Protection Program, one of the government’s most expensive pandemic relief efforts, is a simple question.

Did it work?

New research, drawing on millions of wage and payroll records, suggests a complicated answer: Yes, but at an extraordinarily high cost.

One new analysis found that only about a quarter of the money spent by the program paid wages that would have otherwise been lost, partly because the government steadily loosened the rules for how businesses could use the money as the pandemic dragged on. And because many businesses remained healthy enough to survive without the program, another analysis found, the looser rules meant the Paycheck Protection Program ended up subsidizing business owners more than their workers.

“Jobs and businesses are two separate things,” said David Autor, an economics professor at the Massachusetts Institute of Technology who led a 10-member team that studied the program. “We tried to figure out, ‘Where did the money go?’ — and it turns out it didn’t primarily go to workers who would have lost jobs. It went to business owners and their shareholders and their creditors.”

Questions about the success of the program have gained urgency as the Omicron variant of the coronavirus disrupts the country’s economic upswing, intensifying calls from hard-hit industries like restaurants for a new round of federal aid.

Congress rushed to create the Paycheck Protection Program in the pandemic’s early days, trying to prevent struggling small companies from gutting their work forces and adding to the staggering unemployment rate. The program offered business owners low-interest loans of up to $10 million to cover roughly two months of payroll and a few additional expenses. The loans would be forgiven as long as the money went to permitted costs.

Nearly every company in America with 500 or fewer workers (and some larger ones) qualified: law firms, construction companies and restaurant chains as well as Uber drivers, freelancers and the bars, boutiques, grocery stores and hair salons that are the backbone of many Main Streets.

Early studies of the program — which generally focused on the largest small companies — were not flattering, finding it had little effect on preserving jobs. But Michael Dalton, a research economist for the Bureau of Labor Statistics who drew on extensive wage records collected by the government that other researchers did not have access to, said it had performed better than he expected.

Within one month of being approved, companies that got loans had an average head count 8 percent higher than comparable businesses that didn’t. After seven months, their work forces were still 4 percent larger, maintaining a lead even as hiring nationwide began to bounce back.

And some ventures that would have been forced out of business stayed alive. Businesses that received a loan from the program were 5.8 percent less likely to be closed one month after receiving the money, and 3.5 percent less likely to be shut down after seven months, Dr. Dalton found.

The effects were strongest for the smallest businesses and for those in areas with higher poverty ra