Trade in the GDP accounts
Trade was a significant factor in the weak GDP report today and as usual when this happens you see many comments that do not understand why imports are a negative in calculating GDP.
We do not directly calculate GDP. Rather, we calculate consumption and adjust that for trade and inventories to obtain GDP indirectly. To go from consumption to production in the US we have to subtract imports because they were not produced in the US. However, imports show up twice in the GDP accounts. They show up once in final demand –consumer spending, government spending or investments. So for example if you buy a Volvo, which are not yet built in the US, it will be recorded in both personal consumption expenditures as a positive and in trade as an import. So when you subtract imports all it does is offset the positive contribution recorded in final demand. So your buying a Volvo will have a zero impact on GDP, which is the way it should be because GDP is a measure of what is made in the US. Most people, like Larry Kudlow on CNBC do not seem to understand this and keep saying this it is a mistake to subtract imports.
Their is another big difference in how trade is treated in the GDP accounts. Final demand –personal consumption expenditure, government spending and investments –is calculated as the average of the three months data that is reported monthly. But trade is calculated as the difference between what it was in the final month of the previous quarter and the final month of the current quarter– data that is not yet reported when the first estimate of GDP is released. In the GDP accounts trade and inventories are reported as the change over the quarter rather than the average during the quarter. This is an adjustment that is necessary to go from the estimate of final demand or consumption to a measure of production which is what GDP measures. It is also normally the major reason why the first and second revisions to GDP or so significant.
i did not realize that “trade is calculated as the difference between what it was in the final month of the previous quarter and the final month of the current quarter” and if so, i can’t see how that makes sense…
you’re saying if, for example, November’s trade deficit was $70 billion vs a hypothetical deficit of $48 billion in September, October and December, that November’s trade wouldn’t count in GDP? i’ll have to look at the data, but if that’s true i can’t see it…if you imported and consumed cars in November, you’d want those November car imports subtracted from GDP…
i can understand end of the quarter totals for inventories, but not for trade…
Spencer, here’s the “Key source data and assumptions” for today’s advance GDP… https://www.bea.gov/newsreleases/national/gdp/2018/xls/gdpkeysource_4q17.xlsx
obviously, there are estimates for some December data, but scroll on down to lines 160 to 186 for the trade data and you’ll see that the 4th quarter figures given in column S is an average or sum of October (column M), November (column O) and December (column Q)
“Rather, we calculate consumption and adjust that for trade and inventories to obtain GDP indirectly. To go from consumption to production in the US we have to subtract imports because they were not produced in the US. ”
Quite so and well put. Note a corollary as well. We have to add exports because they are produced in the US but not consumed there. So, if we start with consumption then ……
Further, GDP is, by definition, all production, all consumption and all incomes. The three must, again by definition, be equal to each other. The income from exports accrues to US persons (OK, maybe profits, rent, interest if not direct wages) but the consumption doesn’t. Equally, the income of imports accrues to foreigners but the consumption doesn’t.
Thus (X-M).