Lies and GDP Statistics

BusinessWeek's Michael Mandel has been hinting for weeks at a huge story he's been working on, and it's finally arrived. What's the scoop? Well, it seems that Mandel has a beef with US statistical methodology. And the upshot is that GDP has been overstated – we don't know by how much – due to the US importing various goods that it didn't import in the past.

Mandel's a very good journalist, but even he can't make this story exciting. What's more, he weirdly saves his best explanation of the effect he's talking about for his blog, rather than including it in the story.

My take on all this is that official government statistics, of any country, should never be considered the last word on how any given economy is doing. They're the best that we've got, and economists generally have a pretty good handle on just how good they are. But as Alan Greenspan knew full well, they don't tell you everything. That's why he would spend hours poring over all manner of obscure manufacturing statistics in order to get a more nuanced and detailed idea of how the US economy was doing.

When Republicans say that the economy is doing great, and pull out carefully-chosen statistics to back up their claims, Democrats generally cry foul and point to other statistics to show how the economy isn't doing nearly as well as the Republicans say. A similar thing happens, in reverse, during Democratic administrations. But perhaps it's the populists who have it right: never mind the statistics, they say, we know what reality is by looking at what's going on in our constituencies and using our own two eyes.

The own-two-eyes school of economics has many weaknesses, of course: it tends to overstate the effect of layoffs, and understate the impact of new jobs in the economy. (People remember being laid off from the local factory much more vividly than they remember the nail salon opening on their street.) But statisticians don't have a stranglehold on the truth.