When we model economic behavior with micro-foundations and are interested in unemployment, we need to take a stand about what the unemployed worker is doing with his time. Remember that work gets remunerated, and that is because work is not leisure, work is something would prefer not doing if there were no income from it. Thus we need to factor in that an unemployed person has more leisure than an employed one. But how much?
Indeed, when unemployed one may still do some work, like looking for a job. Luckily, the American Time Use Survey (still not saved from budget cuts) comes to the rescue. Alan Krueger and Andreas Mueller use it to draw some interesting conclusions. The average US unemployed worker spends 41 minutes a day looking for a job, 9 minutes a day on week-ends. While this seems to be very low, it is still much higher than in Europe. Note that these average numbers are drawn down by the large number of those who do not spend time searching at all. Conditional on search, the average time is 167 minutes a day which still leaves plenty of added leisure.
Taking characteristics of state unemployment insurance into account, Krueger and Mueller find that incentives work like they should: a more generous system leads to less search time. As documented abundantly for unemployment-work transitions, they also find a spike in searching during weeks 15 to 26 of an unemployment spell. Somewhat surprisingly, those eligible for unemployment insurance search a little more than those that ran out, even after controlling for demographic characteristics.
What I take from all this is that unemployed workers on average search very little, respond properly to incentives, but there is a puzzle about why they search less once they run out of unemployment insurance.