Why was Keynes wrong? In his 1930 essay (pdf), Economic possibilities for our grandchildren he predicted that:

the standard of life in progressive countries one hundred years hence will be between four and eight times as high as it is to-day.

This looks like being spot-on: real UK GDP per head is now 5.2 times what it was in 1930.

However, Keynes made another forecast that has been wrong - that the average working week would fall to around 15 hours. In fact, it is more than twice that. What's more, it is now falling much more slowly than it did in Keynes' time: in the last 20 years it has fallen by just 1.1 hours in the UK (3.3%) whereas in the 20 years to 1930, it fell by 8.25 hours (16%)*.

Why was Keynes so wrong about this? In a new paper, Benjamin Friedman suggests it is because of increased inequality. Since the 70s, he says, median incomes have grown much more slowly than mean incomes, so households have to work longer to get by.

Personally, I'm not so sure about this. Some of the longest working hours now are put in not by households on median incomes, but by those on high incomes, such as bankers and lawyers. This might well be due to increased inequality - younger bankers put in gruelling hours in the hope of joining the mega-rich 0.1% - but it's a different issue from the working hours of the median worker.

Instead, I suspect two other things have happened. One, as Friedman says. is that network effects have become more important. Now, more than in the 1930s, we want to spend more to keep up with the Joneses. The spread of TV might have exacerbated this process: there's evidence that this increases our material aspirations and reduces contentment with current income.

Another - which I can't quantify - is that technical progress has been more oriented towards consumption goods than Keynes anticipated. We want to work longer than he thought because there's more stuff to buy.

Both these theories are consistent with the fact that Keynes made another error. He thought that rising incomes would lead to bigger rises in savings. But in fact, the savings ratio has fallen in the last 30 years.

However, conventional economics - and indeed many forms of liberal democracy - tells us that, for policy purposes, we should take preferences as given. And many people's preference is to work longer hours. I'm thinking here not just of the unemployed and inactive but also of the 1.26 million people who are working part-time - Keynes' 15 hours a week - who want a full-time job. As Friedman points out, these preferences can best be met by more expansionary fiscal policy. It's a little paradoxical that some of the people who most strongly assert the importance of people's existing preferences - such as neoclassical economists and right-wing politicians - should be so loath to satisfy them in this regard.

* Source: Bank of England: three centuries of data, page 12.