The Limits of GDP Statistics

Zubin has the quarter-by-quarter GDP chart; he, like many others, is particularly taken that in the fourth quarter of 2007, the GDP figure fell below zero. But while the Q4 revision from +0.6% to -0.2% was big, the Q2 revision from +3.8% to +4.8% was even bigger.

In a world where GDP revisions made a full year after the time period in question can amount to a whole percentage point up or down, it's a bit silly to get overexcited about whether any one reading manged to squeak below 0. GDP figures, it seems, are much less exact than the market seems to give them credit for, and today's +1.9% is probably well within the margin of error vis-a-vis the +2.3% expectations.

Here's Mike Mandel:

Most people--and most journalists--take the economic statistics as fact, and report them that way. In fact, they are approximations, estimates, and in some cases educated guesses. Much better than nothing--but not reality.

Mike adds that the BEA releases "better and better estimates over time" - which means, I guess, that the 4.8% figure for Q2 2007 is significantly more reliable than the 3.8% figure was. My feeling is that the best you can really do with GDP statistics is lump them into three vague buckets: "good", "mediocre", and "bad". GDP growth was good in mid-2007, and it was mediocre in the first half of 2008. Was it bad for one quarter at the end of last year? I dunno: what with all the seasonal adjustments surrounding the holiday shopping season, I think it's hard to get that specific. I'd be happier just saying that we've gone from good to mediocre, and leaving it at that.