Sarah hated having to fire 18 employees at once, but she simply could not afford their salaries anymore. Her small business had received a $1.4 million grant under the American Recovery and Reinvestment Act (ARRA) to run a job training program in Arizona, a project that required her to hire those 18 workers up front to do the actual training. Once the stimulus project was completed, however, she had to let those workers go. Their jobs weren’t sustainable past the grant, ironically putting her former employees in a position similar to that of the people they’d just finished training.

Stories like these are depressingly common, as we discovered in a recent Reason Foundation survey of businesses that received ARRA grants and contracts. Only 23 percent of the 8,381 companies we were able to contact in our sample hired new workers to complete their stimulus project and kept all of them once the project was done. In other words, more than seven out of 10 companies did not hire workers at all or had to lay off the workers they did hire.

The survey covered businesses that had completed stimulus projects in areas such as construction work, medicine, and education. The companies that responded were scattered across the country and ranged in size from a handful of workers to more than 500 employees, and were fully representative. We asked the stimulus recipients about their employment levels during different phases of the stimulus and whether they had to lay off workers once the stimulus money was spent.

Most estimates of “total jobs created or saved” are dubious, derived from models built on estimates of estimates. But estimating overall levels of employment is not a good way to understand how the stimulus affected unemployment. A company could have hired 10 workers for one day and claimed it created 10 jobs, even though the labor market was essentially unchanged.

This survey was the first of its kind: a broad collection of data from stimulus recipients about their employment practices in a standard format using telephone and email questionnaires. It allows us to view the stimulus from the perspective of the companies and thus gain a better understanding of how the stimulus interacted with the labor market.

We found that only 41 percent of the companies that received stimulus grants or contracts actually hired workers for the relevant project. The other 59 percent either hired no workers for the stimulus project or reduced their payrolls because they couldn’t afford to retain all their staff even with the stimulus money coming in.

Within the 41 percent of companies that did hire workers, the survey found that 55 percent kept all their workers once the project was done. But 30 percent were, like Sarah’s job training company, forced to lay off workers once the government money stopped flowing. Another 15 percent of companies hired workers but couldn’t retain all of them. These findings suggest that ARRA did not create much lasting employment among firms that received grants or contracts for stimulus projects.