AMONG the most controversial of Thomas Piketty's arguments in his bestselling analysis of inequality, "Capital in the Twenty-First Century", is that wealth is increasingly concentrated in the hands of the very rich.
Rising wealth inequality could presage the return of an 18th century inheritance society, in which marrying an heir is a surer route to riches than starting a company.
Critics question the premise: Chris Giles, the economics editor of the Financial Times, argued earlier this year that Mr Piketty's data were both thin and faulty. Yet a new paper suggests that, in America at least, inequality in wealth is approaching record levels.*
Earlier studies of American wealth have tended to show only small increases in inequality in recent decades. A 2004 study of estate-tax data by Wojciech Kopczuk of Columbia University and Emmanuel Saez of the University of California, Berkeley, found an almost imperceptible rise in the share of wealth held by the top 1% of families, from about 19% in 1976 to 21% in 2000.
A more recent investigation of the Federal Reserve's data on consumer finances, by Edward Wolff of New York University showed a continued but gentle increase in inequality into the 2000s. Mr Piketty's book, which drew on this previous work, showed similarly modest rises in wealth inequality in America.
A new paper by Mr Saez and Gabriel Zucman of the London School of Economics reckons past estimates badly underestimated the share of wealth belonging to the very rich. It uses a richer variety of sources than prior studies, including detailed data on personal income taxes (which the authors mine for figures on capital income) and property tax, which they check against Fed data on aggregate wealth.
The authors note that not every potential source of error can be accounted for; tax avoidance strategies, for instance, could cause either an overestimation of the wealth share of the rich (if they classify labour income as capital income in order to take advantage of lower rates) or an underestimation (if they intentionally seek out lower yielding investments for their tax advantages). Yet they believe their estimates represent an improvement over past attempts.
The results are enough to make Mr Piketty blush. The authors examine the share of total wealth held by the bottom 90% of families relative to those at the very top. Because the bottom half of all families almost always has no net wealth, the share of wealth held by the bottom 90% is an effective measure of "middle class" wealth, or that held by those from the 50th to the 90th percentile. In the late 1920s the bottom 90% held just 16% of America's wealth--considerably less than that held by the top 0.1%, which controlled a quarter of total wealth just before the crash of 1929.
From the beginning of the Depression until the end of the second world war, the middle class's share of total wealth rose steadily, thanks largely to collapsing wealth among richer households. Thereafter the middle class's share grew along with national wealth thanks to broader equity ownership, middle-class income growth and rising rates of home-ownership. The expansion of tax breaks for retirement savings also helped. By the early 1980s the share of household wealth held by the middle class rose to 36%--roughly four times the share controlled by the top 0.1%.
From the early 1980s, however, these trends have reversed. The ratio of household wealth to national income has risen back toward the level of the 1920s, but the share in the hands of middle-class families has tumbled (see chart).
Screenshot 2014 11 09 12.22.34The Economist
Tepid growth in middle-class incomes is partly to blame; real incomes for the top 1% of families grew 3.4% a year from 1986-2012 while those for the bottom 90% grew 0.7%. But Messrs Saez and Zucman reckon the main cause of falling middle-class net worth is soaring debt. Rising home values did little to raise middle-class wealth since mortgage debt also soared. The recession battered home prices but left the debt untouched, further squeezing middle-class wealth.

The really, really rich get much, much richer

On the other side of the spectrum, the fortunes of the wealthy have grown, especially at the very top. The 16,000 families making up the richest 0.01%, with an average net worth of $371m, now control 11.2% of total wealth--back to the 1916 share, which is the highest on record.
Those down the distribution have not done quite so well: the top 0.1% (consisting of 160,000 families worth $73m on average) hold 22% of America's wealth, just shy of the 1929 peak--and exactly the same share as the bottom 90% of the population.
Meanwhile the share of wealth held by families from the 90th to the 99th percentile has actually fallen over the last decade, though not by as much as the net worth of the bottom 90%.
The outsize fortunes of the few would not be too worrying were they largely the product of entrepreneurial activity: riches amassed by hardworking billionaires who are as likely as not to give their bounty away through philanthropy. Messrs Saez and Zucman find some evidence for this dynamic. Wealthy families are younger than they were a generation or two ago, and they earn a larger share of the country's income from labour: 3.1% in 2012 versus less than 0.5% prior to 1970.
Yet one should not yet rule out the return of Mr Piketty's "patrimonial capitalism". The club of young rich includes not only Mark Zuckerbergs, the authors argue, but also Paris Hiltons: young heirs to previously accumulated fortunes.
What's more, the share of labour income earned by the top 0.1% appears to have peaked in 2000. In recent years the proportion of the wealth of the very rich held in the form of shares has levelled off, while that held in bonds has risen.
Since the fortunes of most entrepreneurs are tied up in the stock of the firms that they found, these shifts hint that America's biggest fortunes may be starting to have less to do with building businesses, just as Mr Piketty warned.
Click here to subscribe to The Economist