| Forecaster | 2026Q3, SAAR | Vintage |
|---|---|---|
| Macro Edge Research | +2.2% | 2026-10-28, data through October 28 |
| Atlanta Fed GDPNow | +3.3% | final update, October 28 |
| Phila Fed SPF median | +2.5% | August 2026 survey (released August 14) |
| NY Fed Staff Nowcast | +2.7% | October 23 update |
Steady domestic demand, a cooling labour market, and a trade surge. Real consumer spending grew at a 2.9% annualised pace through July and August, supported by a recovery in real disposable income, and core capital-goods orders rose at a 15% annualised pace, while housing starts fell about 15% annualised. Payrolls rose by about 50k a month on average in the quarter and unemployment held at 4.1% to 4.2%. Headline inflation cooled sharply from Q2, but crude rebounded from about $80 a barrel in July to a $107 peak in mid-September on renewed Gulf escalation. The Fed raised rates 25 bp to 3.75% to 4.00% on September 16 and held there on October 28, with two dissents in favour of a further increase.
Trade and inventories carry the quarter, and largely cancel. The goods and services deficit averaged $99 bn a month in July and August against $67 bn in Q2, and the advance goods-only deficit widened further in September, to $135 bn from $132 bn in August, consistent with importers front-running threatened tariffs. We attribute −2.7 pp to net exports and +2.0 pp to inventories, judging that most of the inflow was stocked rather than sold. That split is the largest judgement in this call: if inventories absorbed the whole net-export drag, the arithmetic of our attribution would put the quarter near +2.9%.
Where we sit. +2.2% is the lowest of the four forecasts on this page, 1.1 pp below Atlanta Fed GDPNow. We see the balance of risk tilted to the upside, mainly because more of the import surge may have been stocked than we assume.
What we cannot see. September personal income and spending publish the same morning as GDP; September services trade and the detailed inventory books follow in November. Net exports and inventories, estimated from incomplete data at the advance stage, are typically the widest error channels in nowcasts of the first print.
| Forecaster | 2026Q2 forecast | BEA advance | |Miss| (pp) |
|---|---|---|---|
| Atlanta Fed GDPNow (final update, T-2) | +1.54% | +1.50% | 0.04 |
| Macro Edge Research NowCast (T-2) | +1.10% | +1.50% | 0.40 |
| Phila Fed SPF median (May survey, T-76) | +2.10% | +1.50% | 0.60 |
| NY Fed Staff Nowcast (backcast, T-6) | +2.80% | +1.50% | 1.30 |
| Forecaster | n | MAE (pp) | RMSE (pp) | Median |err| | Sign hit |
|---|---|---|---|---|---|
| Macro Edge Research NowCast | 12 | 0.75 | 1.04 | 0.40 | 100% |
| Atlanta Fed GDPNow | 12 | 0.81 | 1.21 | 0.63 | 100% |
| Phila Fed SPF median | 12 | 1.34 | 1.70 | 0.79 | 92% |
| NY Fed Staff Nowcast | 12 | 1.19 | 1.45 | 1.10 | 92% |