Friday, April 29, 2005

Deconstructing GDP

Yesterday’s big economic news was that U.S. GDP grew 3.1% in real, annualized terms, while forecasters had pegged growth at 3.5% (see the full report here). Yet, when one takes a look at the coverage of this story, nearly all reports do a very poor job of explaining why the economy did worse than expected.

The problem is that people need a “story” to make sense of raw data. Needless to say, there’s always a strong temptation (conscious or unconscious) to twist facts so that they fit the narrative, particularly if it’s simple and compelling.

In this case, the story goes like this: high oil prices forced consumers and firms to cut spending on other items in the first quarter, resulting in weak growth. This piece by the AP offers the clearest example.

But is it true? Not really.

Personal consumption expenditures grew at a 3.5% annual clip, down from 4.2% in the 4th quarter. Meanwhile, private investment spending rose at a 12.5% rate, slightly below the 13.3% increase in the previous quarter. Actually, if you add both, they grew 5.1% in total.

Obviously, consumers and firms were just as eager to spend as ever. And let’s not forget that expenditures on gasoline and oil represent just 3% of household expenditures.

The real issue is how the money is spent. Basically, growth in the first quarter was held down by spending on imports, which rose a very impressive 14.7% and, yes, that includes oil. Even after adding exports, the trade balance subtracted 2.2 percentage points of growth, versus 1.7 percentage points in the previous quarter. This alone explains much of the slowdown between the fourth quarter and the first.

Runaway import growth was not the only troubling aspect of the GDP report. Around half of the growth in private investment took the form of inventory accumulation. If consumer spending doesn’t grow at a faster clip this quarter, high inventories may lead to production cutbacks and employment growth –the key element in sustaining consumer expenditures—will be minimal.