A loyal reader has chastised my defence of welfare benefits for ignoring incentive effects. This deserves a reply.

The issue here is not about Osborne's welfare cuts, simply because these haven't, net, made much difference to incentives. On the one hand, the rise in the "living wage" relative to Universal Credit will increase incentives to work. But on the other hand, this won't affect under-25s (who aren't covered by the wage floor); cuts to in-work tax credits reduce work incentives; and the higher taper rate reduces incentives to work longer. Overall, says the Resolution Foundation's David Finch:

These changes will do very little to improve the incentives for low paid families to find a path into work and then to progress.

Instead, the issue is a fundamental trade-off facing any benefit system - that of incentives versus risk reduction. Low out-of-work benefits sharpen incentives for the unemployed to find work. But they also mean that people losing their jobs face a bigger cut in income, which could deepen any recession.

I'll concede that the incentives argument has some merit. Some of the unemployed do reject job offers because they'd prefer to stay on benefits. And Barbara Petrongolo has shown that a tougher benefits regime does incentivize job-finding. The question is: how strong is this argument?

Let's do a back-of-the envelope estimate. If we could move 300,000 people - almost all those who have been unemployed for more than two years - from unemployment to full-time minimum wage work GDP would rise by around £5bn: this is £3.8bn of wages plus around £1.2bn of extra profits from employing them. This is around 0.3% of GDP. However, a serious recession could easily cost 5% of GDP. A generous welfare state, being a strong automatic stabilizer, would save a big fraction of this.

Which should we prefer? It depends on many issues:

1. How sensitive is job-finding to unemployment benefits? I doubt if many people think: "my benefits have been cut, so I might as well stop watching Jeremy Kyle and take up that offer of a professor of maths." I suspect many of the voluntarily unemployed are borderline unemployable and so not very amenable to incentives. This is consistent with evidence that the 2010-15 welfare reforms, such as the benefit cap, did not greatly increase job-finding.

2. What's the mechanism whereby the demand for labour increases to meet the increased supply? One possibility is that wages get bid down. But this channel is silted up by a rising minimum wage. Another possibility is that vacancies get filled faster, which makes firms more efficient. But Ms Petrongolo shows that those who are incentivized to find work by benefit cuts are less likely to stay in work - which suggests that low benefits lead to worse job matches, which is bad for firms.

3. Can macro policy be used to stablize the economy? If the answer's yes, then there's less need for a generous welfare state as a stabilizing device. My view, though, is that recessions are unpredictable and so policy cannot prevent them.

4. Do recessions have permanent effects on GDP? In my example, the benefits of getting the unemployed into work - subject to the above caveats - are long-lasting, whereas the benefits of stabilizing recessions come only once every few years. However, if recessions have long-lasting adverse effects upon future growth, then it becomes more important to cushion ourselves from them.

5. How important is it to punish those who violate the norm of reciprocity? The issue here is not merely one of economics, but ethics. Many people hate the idea that some of the unemployed are getting something for nothing at the expense of the rest of us. How much weight does this preference have?

6. How should we weigh false positives against false negatives? A tough welfare regime punishes "scroungers", but also the "deserving poor" - those unlucky to lose their jobs, whilst a more generous welfare state is kinder to both.

Many people's attitudes to these issues are based in part upon ignorance - an overestimation of benefit spending. Nevertheless, in the fact-based world, reasonable people will disagree here. But let's be clear: I am not ignoring incentives, but merely doubting how much weight we should put upon them.