Americans spend more time at work than citizens of other wealthy countries. But many of us are not actually working when we’re on the clock.

According to a study published on Monday by the National Bureau of Economic Research—which cites the American Time Use Survey—the average American spends about 34 minutes of their time at work not actually working. Half of this time is spent eating, though it’s up to the individual respondent whether or not he considers a lunch break taken outside of the office part of the official workday. When you remove from the sample the folks who claim to not do any slacking off at work whatsoever, that figure rises to just over 50 minutes of loafing per day. Individual results vary, however, with “a tiny fraction of respondents even reporting not working the entire time on the job.”

The economists who published the paper (Michael Burda, Kaie Genadek, and Daniel Hamermersh) weren’t interested in determining how much time Americans spent at work not working. Instead, they wanted to determine which outside factors might lead workers to slack off more, or less, on the job.

They found that the longer a person’s work week is, the greater the chance that the given worker will spend time at the office not working, up until a 43-hour work week. Beyond 43 hours, the incidence of slacking off actually decreases. This makes sense if you believe there are a lot of undemanding jobs that still require workers to be present in the office 40 hours per week. Jobs that require more than 43 hours of work per week are probably actually demanding in terms of hours needed to get the job done, and the long hours are about more than just giving the appearance that you are working hard.

The broader state of the economy also affects how much we slack off at work. During times of high unemployment, the economists found that workers who may otherwise be major loafers slack off less at work, suggesting that these particular workers are more afraid of losing their job when the economy is weak.

But when you look at the overall population, the data shows that non-work time actually increases along with the unemployment rate. This underscores a fact that’s often ignored by the financial press: the first response of employers to poor economic conditions is not to start firing people. As the authors write:

Firms face variable and imperfectly forecastable demand for their products…. In an economic downturn which is perceived as temporary, it is easy to show that a layoff, even if temporary, is an inferior choice to maintaining employment, possibly even at standard hours. This behavior is often referred to as labor hoarding. In this case, firms assign workers to “unproductive” tasks such as cleaning, maintenance, painting, etc. or even tolerate more worker-initiated non-work.

In other words, you may be more comfortable slacking off at work today, when the unemployment rate is at 5% versus in 2009, when it was almost double that. But there is probably more work for you to do too.