Why Q3 U.S. GDP Declined
If we look at the different components, it seems increasingly clear after yesterday's weak report that personal consumption fell in the third quarter for the first time since the 1990-91 recession. Average real (2000 dollars) consumer spending for July and August was $8293.2 billion, down from the Q2 number of $8341.3 billion. What this means that unless real consumer spending soared more than 1.74% in September (which is extremely unlikely), then Q3 consumer spending will be lower than in the second quarter. Assuming the number is unchanged instead (which might be over optimistic), then we're talking about a decline of 2.3%, something which will subtract roughly 1.6%:points from GDP.
Residential investments will also likely to continue to subtract from GDP. Exactly how much is very uncertain due to insufficient data data, but we're talking about at least 0.5%:points. After last week's weak durable goods report, it seems clear that equipment and software investments will likely decline too. As only July data is available for non-residential construction, it is more uncertain whether that will be positive or not, but it is unlikely that it will rise enough to push business investments as a whole on the plus side.
Trade also suffer from the same uncertainty factor (only July data available yet), but it seems likely that the trade deficit will fall in August and September due to cheaper oil. However, because the July deficit was so big, the Q3 deficit might still not fall. And given that the decline is driven by falling oil prices, this means that trade will more likely contribute negatively in the volume number (and perhaps even in the proper number).
The two factors which will likely add to GDP are inventories and government spending. Inventories will probably continue to fall during the third quarter, but as it will fall less than in the second quarter, it will still add to GDP. Just how much these two factors will contribute is uncertain, but is unlikely to come even close to compensating for the other factors.
So, while I will not at this point give a more precise Q3 forecast due to the insufficient data data available, I can already say that GDP (both the headline and the proper measure) will be below zero.