The preliminary GDP release yesterday provided a number of surprises. The first surprise was not that GDP was higher than the advance release (given the June trade figures reported earlier this month), but rather that at 3.3% it exceeded the 2.8% [SAAR] of the consensus . The second surprise is that the reduction in imports comprises an even larger proportion of the overall growth.
Let's turn first to the surprise. The change in the contribution due to net exports was anticipated, given the release of the monthly trade figures for June, which were unavailable at the time of the advance release. However, one big change was in the contribution of inventories. Apparently, they decreased by a smaller amount (-1.44 ppts vs. originally estimated -1.92 ppts). While this increases GDP in 08Q2, this might suggest a bigger reduction in output in the current quarter, as pro.ducers seek to match inventory stocks to anticipated output.
But let's return to trade. Figure 1 illustrates real GDP growth in ppts SAAR (blue bars), and the contributions from Net Exports (red line) and Imports (green line).
Figure 1: GDP growth q/q at annual rates (blue bars), contribution from net exports (red line), and contribution from imports of goods and services (green line). Source: BEA GDP advance release of 28 August 2008.
Interestingly, the change in trade sectors accounts for almost all of growth (in the advance, it accounted for more than all GDP growth). And while the proportions coming from export growth and import shrinkage are about the same, here I think absolute percentage point contributions are important. Reductions in real imports now are calculated to contribute 1.45 ppts, as opposed to the earlier estimate of 1.26 ppts.
Of course, some of this reduction is due to the weaker dollar, as journalistic accounts have stressed. But unlike exports, imports are -- according to macroeconomic estimates -- highly insensitive with respect to exchange rates, and much more elastic with respect to income . So I suspect that real income is either actually stagnants or falling, or perceived to fall in the near future (I'm not so Keynesian as to think only current income matters). Figure 2 depicts the log real dollar exchange rate (calculated against a broad basket of currencies (blue), lagged two years, and both net exports to GDP and net exports ex oil to GDP ratios (red, teal, respectively).
Figure 2: Log real dollar exchange rate (Fed broad index) (blue, left axis), net exports to GDP ratio (red, right axis), net exports ex-oil to GDP ratio (teal, right axis). Source: BEA GDP release of 28 August 2008, Federal Reserve Board, and author's calculations.
The steep dropoff in the non-oil trade balance to GDP ratio suggests to me a big income (and perhaps wealth) effect, but I admit to not having conducted a formal decomposition into income and price effects.
Now we come to the issue that absorbed so much interest at the time of the advance GDP release. The divergence between real GDP and real Gross Domestic Purchases remains. Indeed it has expanded, from 2.4 ppts to 3.1 ppts. In Figure 3, I plot real GDP and Gross Domestic Purchases growth, and in Figure 4, the respective growth rates of the deflators.
Figure 3: GDP growth q/q annual rates (calculated using log differences) (blue line) and Gross Domestic Purchases growth (red line). Source: BEA GDP release of 28 August 2008, and author's calculations.
Figure 4: GDP deflator growth q/q annual rates (calculated using log differences) (blue line) and Gross Domestic Purchases deflator growth (red line). Source: BEA GDP release of 28 August 2008, and author's calculations.
Recalling the following definitions:
GDP = C + I + G + EX - IM
Gross domestic purchases = C + I + G
It's clear that the amount that we're expending on (from consumers, businesses, and government) is barely growing; given q/q annualized growth of 0.3 ppts (log terms), it's essentially zero. So the factories may be humming, but that's because exports are up, thereby illustrating how much continued growth in GDP depends upon the trends in the rest of the world.
Figure 4 illustrates why -- as noted earlier -- the GDP deflator does not jibe with what we as consumers see. It's because the prices for what we produce have diverged in a significant way from the prices for what we consume.
To me, this last outcome is not a surprise  (pdf). When a country consumes much more than it produces, then at some point, it has to repay some of the debt incurred. Repaying involves producing more than consuming. We're not even at that point yet -- we're merely moving toward producing more than we're consuming. How much longer the process will continue, and how far (i.e., will we actually run a trade surplus in the foreseeable future?) depends on a lot of things, including the desirability of American assets, and hence how much foreigners want to lend to us. So, as I say, two surprises, and -- for me -- one non-surprise.
What are the prospects abroad? For Europe, I'll just refer to Jim's post. It remains to be seen what happens in East Asia. But while GDP is not yet shrinking (at least not according to the current data), there's still a good chance in the future, as the slowdown spreads across borders.
A lot more coverage at Economists View.
Update: Earlier remarks of a similar nature in WSJ RealTime Economics