Here are two interesting new papers on inequality. First, LSE researchers say:

Household characteristics account for only part of the cross-country variation in household wealth and its distribution. The largest share of the differences remains unexplained pointing towards the importance of country specific effects.

If this seems unsurprising, it shouldn't be. It undermines the rightist idea that inequality is due merely to differences in individuals' skills or savings propensities. In this sense, it corroborates other evidence which shows that there is no correlation across countries or over time between inequality of income and inequality of human capital.

Instead, the causes of inequality lie in institutions and ideology. On the one hand, the US has institutions and ideologies which engender "winner take all" markets in which CEOs and "superstars" - of either an Adler (pdf) or Rosen (pdf) type - can get very rich. On the other hand, Scandinavians have strong welfare states which reduce income inequality but - by deterring saving amongst the middlingly poor - increase wealth inequality.

But could it be that the Americans have got it right? This is the possibility posed by our second paper. Che-Yuan Liang asks: how much inequality would agents choose behind a veil of ignorance? Rawls' answer - derived from the maximin principle - is: very little. But what if instead people in the original position used prospect (pdf) theory? Then, says Mr Liang, they'd have two reasons to favour inequality:

- People overweight the small chance of a big gain. "Very low probabilities are generally overweighted" said Kahneman and Tversky. "People prefer what is in effect a lottery ticket over the expected value of that ticket."

- Big losses have diminishing marginal utility; this is why people who have made small losses often take bets that have poor expected value.

These two aspects of prospect theory could, says Mr Liang under some conditions (such as leaving aside loss aversion) lead people in the original position to choose a US-style economy in which many are relatively poor, but a few are extraordinarily rich.

But is this a rational preference? It's not, if the overweighting of the low probability is due to overconfidence or a misunderstanding of probabilities. But it could be rational if there are non-linearities in the marginal utility of income. For example, a £50,000pa salary gives you a life of dull drudgery in which you can't afford a house in London, whereas a £1m income gives you freedom, power and positional goods. You might, therefore prefer a 1% (or lower) chance of the £1m to a 20% chance of the £50,000.

If you want an argument for inequality, this might be a stronger one than claims that such inequalities are the "deserved" result of differences in people's abilities and aptitudes.