The Federal Reserve Bank of New York's dynamic stochastic general equilibrium model, updated on September 18, 2026, delivered a forecast that is notable less for a sharp change in direction than for a subtle shift in the composition of risk. The model kept its 2026 GDP growth projection at 1.2 percent, unchanged from June, but now expects core inflation to revert toward the FOMC's long-run goal of 2 percent more slowly than it did three months earlier. The short-run real natural rate of interest, or r*, was revised down slightly for 2026 to 1.9 percent annualized from 2.0 percent, while r* forecasts for 2027 through 2029 were revised modestly higher. The update matters because it shows a staff modeling tool becoming more cautious about the persistence of inflation even as it acknowledges that the economy has repeatedly outperformed its own growth expectations.
The mechanics behind the forecast are explicit. The model uses data released through 2026:Q2, augmented for 2026:Q3 with median forecasts for real GDP growth and core PCE inflation from the August release of the Philadelphia Fed Survey of Professional Forecasters, short-run inflation expectations from that same survey, expectations of the future federal funds rate from the July release of the New York Fed Survey of Market Expectations, and yields on 10-year Treasury securities and Baa-rated corporate bonds. The model remains pessimistic on growth and, by its own account, keeps being surprised when the economy turns out stronger than expected. The SPF projects GDP growth in 2026:Q3 to be more than 1 percent higher in annualized terms than the DSGE model expected in June. The model attributes that likely forecast miss to more buoyant financial conditions than expected and to the positive growth effect of the AI-related boom in investment, captured through marginal efficiency of investment shocks.
The forecast revisions are concentrated in the outer years. GDP growth projections for 2027, 2028, and 2029 are now 0.1, 0.5, and 1.2 percent, compared to 0.2, 0.7, and 1.5 percent in June. The model explains the downward revision in part by an expectation that monetary policy will be more restrictive than previously assumed, and in part by a downward revision to expected total factor productivity growth. On inflation, the projections for 2026, 2027, 2028, and 2029 are 3.3, 2.1, 1.8, and 1.8 percent, compared to 3.1, 1.8, 1.6, and 1.7 percent in June. The slower reversion toward 2 percent is attributed in part to lower TFP growth, which the model associates with higher inflation. The r* path shows a similar divergence: 1.8, 1.4, and 1.2 percent for 2027, 2028, and 2029, compared to 1.7, 1.3, and 1.1 percent in June.
The market and policy implications are indirect but meaningful. A model that sees more persistent inflation and a slightly higher medium-term r* is consistent with a less accommodative policy path than the June forecast implied, even though the 2026 r* estimate was trimmed. The attribution of the growth surprise to AI-related investment and buoyant financial conditions suggests the model is incorporating a supply-side or investment-demand channel that has not yet fully translated into its longer-run productivity assumptions. Because the DSGE forecast is explicitly not an official New York Fed forecast but only an input to the Research staff's overall forecasting process, the update should be read as one structured view among several, not as a policy signal.
The limitations are stated clearly in the source. The model's own narrative acknowledges repeated forecast misses on growth, and the September update relies on survey-based conditioning data for the current quarter rather than realized outcomes. The downward revision to TFP growth is a modeling judgment that could reverse if the AI-related investment boom produces measurable productivity gains. What to watch next is whether the model's pessimism on 2027 growth and its slower inflation reversion persist in subsequent updates, and whether the higher r* path for 2027 through 2029 begins to influence the broader staff forecast or market pricing of the longer-run neutral rate.