Thomas Piketty, author of a best-selling book on the widening gap between rich and poor, relied on faulty data that skewed his conclusions, The Financial Times reported on its website.

The mistakes undercut Piketty’s argument that wealth inequalities are heading back up to levels last seen before World War I, Chris Giles, the FT’s economics editor, wrote.

Data underpinning Picketty’s “Capital in the Twenty-First Century” includes apparent transcription errors, unexplained statistical modifications and “cherry picking” of sources, he wrote.

Piketty defended his findings in a separate posting on the newspaper’s website, saying that he used a very diverse set of data and so needed to make adjustments.

“I have no doubt that my historical data series can be improved and will be improved in the future,” wrote Piketty, who is a professor of economics at the Paris School of Economics.

“But I would be very surprised if any of the substantive conclusion about the long-run evolution of wealth distributions was much affected by these improvements.”