Thursday, 5 May 2011

Bad Data In...

The IMF seems to be having a bad year. First it was accused of 'groupthink'. Then it was rumoured (again) to be considering Gordon Brown as its leader.

Now it's The Economist's turn to sink the slipper. It points out that the IMF's forecast of China's current-account surplus assumes a steadily depreciating yuan, and a widening surplus. That provides ammunition for protectionists to impose tariffs on Chinese goods. Yet China's current-account surplus has actually declined since 2006.

But let's not just pick on the IMF. In "Botox and Beancounting", The Economist also points out the cosmetic effect of US official measures of government debt, productivity and economic growth compared to European measures. "The snag comes if investors fail to grasp that official national figures can show the American economy in an overly flattering light."

Of course, none of these is an isolated incident. There are also fundamental problems in comparing corporate financial data, for instance, given differing accounting standards.

And we live in a world where auditors are still trying to figure out what "scepticism" means...

But we don't have a problem detecting bad data. Plenty of people warned others about what Madoff's fraud, for example, and investment analysts routinely uncover issues such as The Economist has reported. Short-sellers make this their business. No, as David Einhorn elucidated in "Fooling Some of the People All of the Time", the problem is how to give the same weight of publicity to the prudent interpretation of the data as is given to the release of the data itself.

The media and social media clearly play a significant role. But even if we create an Office of the Devil's Advocate, ultimately each of us must accept responsibility for thinking critically about the data we're given if we're to avoid making some big mistakes.


Image from TraceyNolte.

No comments:

Post a Comment