Two numbers came out of the same document this morning and they point opposite ways.

The first is the one you have read: real GDP rose at a 1.5 percent annual rate in the second quarter, down from 2.1 percent, below a consensus of roughly 2.1 to 2.3 percent [1] [3].

The second is in the same release and did not make a headline: real final sales to private domestic purchasers rose 3.9 percent, up from 1.7 percent - the strongest since the first quarter of 2023 [2].

That second measure is what it sounds like. What American households and businesses actually bought, stripped of trade and inventory noise. It more than doubled.

Here is why the headline is lower.

Imports surged 11.5 percent in the quarter. Imports are subtracted in the GDP calculation - not because they represent weakness, but because they were produced somewhere else - and they took 1.0 percentage point off the total [2].

Businesses also ran down inventories by $50.8 billion, the fifth straight quarterly decline, costing another 0.7 points [2].

Add those back and the arithmetic stops being mysterious.

Now the part that is a genuine reporting error waiting to happen.

Federal nondefense spending fell 12.8 percent, which sounds like austerity and is not. BEA's own release says why: 'The pattern of nondefense spending primarily reflected sales of crude oil from the Strategic Petroleum Reserve, based on data from the Department of Energy' [1].

Sales from the Strategic Petroleum Reserve are deducted from federal consumption expenditures. It is a bookkeeping entry.

Anyone reporting that federal spending was slashed 12.8 percent has reported an oil sale as a budget cut.

The forecasters are worth a note too. Consensus sat at 2.1 to 2.3 percent depending on the poll. The Atlanta Fed's GDPNow model published 1.5 percent on July 28, two days early, and hit it exactly [4].

The composition underneath is genuinely strong in places and genuinely weak in others, and neither half should be skipped.

Consumer spending accelerated to 3.2 percent with goods up 5.2. Equipment investment rose 15.2 percent and intellectual property 8.8. Residential investment rose 1.5 percent - its first gain in six quarters [2].

Against that, structures fell 5.0 percent, a tenth consecutive contraction [2].

Michael Pearce of Oxford Economics put the summary better than the headline did. The 1.5 percent 'underplays the economy's strength as it reflects a drag from rising imports and falling inventories' [3].

None of this means the economy is fine. It means the number that says it is slowing is not measuring the thing most people think it is measuring.