Tim Worstall points to the likely high earnings of successful Olympians such as Victoria Pendleton and says: "Stop subsidising the training.It's just stealing money from the poor so that a gilded few can get ever richer." But there's a different way of looking at this.

It's that such subsidies are a form of investment in human capital. In subsidizing training, the government helps to raise athletes' earnings and it gets a return on this in the form of higher tax revenues.

Granted, the additional tax alone doesn't recoup the spending on elite sport. But it is only a fraction of the financial payback; for example, there's also lower health spending, to the extent that lardies are inspired by Olympians to get fit.

What's true for sport funding is more true for welfare spending generally; it's a form of investment which helps raise peoples' earnings and tax revenue. To this extent, the welfare state pays for itself. For example:

- Health spending patches people up and gets them back to work

- Education spending raises our earnings; if I hadn't had a state education, I'd be struggling to hold down a minimum wage job. As it is, my higher earnings have repaid the cost of my schooling many times over.

- Welfare benefits don't just keep people alive whilst they wait to get a job. They can give people time to find the right job, which raises their future earnings. And they can subsidize writers and musicians whilst they hone their craft; if it hadn't been for the dole (and an Arts Council grant), we might not have the Harry Potter books and JK Rowling's large tax payments.

You might ask why, if such spending pays for itself, the private sector does not do it, as Hopi suggests? One reason is that there are large costs of identifying the profitable investments such as those children who would benefit most from schooling. Venture capitalists find it hard enough to spot profitable companies, let alone profitable people. Another, bigger, reason is that private contracts to hand over 40%+ of one's future income are unenforceable. Only the state has the power to extract returns on investments in education and health.

So much for intuition. What of the evidence? Here are three things:

1. Between 1830 and 1870, when the UK's welfare state was minuscule, real GDP per head grew by just 1.4% a year. Between 1945 and 1985, when it grew rapidly, real GDP per head rose 1.9% per year.

2. Nordic states have for years combined large tax-funded welfare states with good economic outcomes, which suggests that big welfare states in themselves are not bad for economies.

3. Peter Lindert's research suggests that the welfare state is a free lunch. He says (pdf): "There is no clear net GDP cost of high tax-based social spending."

I say all this the welfare state is often seen as a cost. It's not. It's also an investment and perhaps a profitable one.