Punitive treatment of the unemployed is usually justified in terms of the incentives it provides. So, for instance, the rationale for increasing the wait until you can claim unemployment benefits from 3 to 7 days is apparently that it  "send[s] the message from the very start that rights to benefits are conditional on the requirement to search for work".
One particular argument made is that unemployment benefits in general stop people searching for work. That's most frequently heard in the context of long-term unemployment; it is, for instance, at the heart of the myth that welfare policy needs to tackle the problem  of households with "three generations of worklessness". If welfare queens are languishing on unemployment benefit, content to be paid by the state not to work, then cutting that benefit will encourage them back into work.
But – surprise! – it seems that that plan doesn't actually work. A paper from the Boston fed  looks at the effect of the unemployment insurance on the Beveridge curve. That's the chart showing the relationship between unemployment and the number of vacancies:
The US has experienced a worrying alteration in the shape of its Beveridge curve since the recession. There are now many more people unemployed for each vacancy than there were in the years running up to 2009 (a fact easily visible in the shift between the blue and red sections of the curve in the chart above). Traditionally, that's seen as indicating a failure to match unemployed people to available jobs, perhaps through a skills shortage or a geographical dislocation. But some suggest it's due to a recent extension of unemployment insurance in the country, which allowed unemployed people to claim the benefit for 99 weeks after losing their job.
The paper's author, Rand Ghayad – the same researcher who exposed just how damaging long-term unemployment is  in April – devised a natural experiment to examine whether unemployment insurance was the cause.
(A natural experiment takes advantage of some quirk in the world at large which sorts people quasi-randomly into different groups, and then assigns different treatments to them. A classic example is to look at the fates of people who were one mark above, and one mark below, a grade boundary: their intelligence is likely equal, and so any difference in outcome can be attributed to passing the exam)
In this case, Ghayad compared long-term unemployed people who were eligible for the insurance with those who had voluntarily quit their job, those who had never worked before, and those who had left the labour market for a period, all of whom are not eligible for the extended benefits. The characteristics of the two groups are obviously different, but the comparison is revealing nonetheless. Here's the shift in the Beveridge curve for those who are eligible for unemployment insurance:
That's still an outward shift, and thus still represents a weakened labour market. But it's nothing compared to the shift in the Beveridge curve for those who are ineligible:
The unemployment rate for that group shot up in the recession – and then never dropped, even as job openings began to reappear.
In other words, unemployment benefits really don't seem to discourage people from seeking work. If anything, they appear to help: the groups which can get unemployment insurance saw their joblessness fall after the recession. It's easy to come up with reasons as to why this might be the case: perhaps not having to worry about how the bills are going to be paid in the short term gives you time to effectively look for a job in the long term? Or perhaps punitive treatment of the unemployed just pushes them into the shadow economy sooner?
Either way, the study ought to be another nail in the coffin of the idea that the way to get people back into work is with liberal application of the stick. It seems that might be the worst thing you could do.