When we model unemployment, we like to think that unemployed workers enjoy more leisure, but still spend a significant time searching for jobs. This modeling choice has important implications as it means that it is normal for the start of an unemployment spell to lead to lower consumption, as market consumption is substituted for home consumption and leisure. Do these choices by theorists make sense?
This is one question that Alan Krueger and Andreas Mueller can answer in a pair of papers that use the American Time Use Survey (which still needs saving) and other surveys. Overall, unemployed people behave in expected ways: they sleep more, they spend more time in home production, including care of children, and have more leisure, in particular TV watching. They also spend more time shopping and, of course, looking for a job.
This is where it becomes interesting. On a weekday, an American spends 40 minutes looking for a job. While this is surprisingly low, consider that Scandinavians bring it down to 5 minutes. These numbers hide considerable within country variation as, for example, only about 20% unemployed Americans actually spend time looking for a job. For those that actively search, the time spent is determined by the "urgency" of the job search: being married or female leads to less job search time. Educated people search more. But again, this is conditional on searching at all.
Unemployment insurance benefits do not seem to explain the cross-country variation in time spent searching, except maybe the change in benefits as the unemployment spell last over six months. However, in the US, cross-state variation can be explained by UI benefits. And for those eligible, search time increases as the end of eligibility approaches.
All in all, it looks like unemployed workers are behaving as theorists would expect. Except that the length of time spent looking for a job is an order of magnitude shorter that what is generally assumed.