The results suggest that, had these policies not been implemented, real gross domestic product (GDP) would have been nearly 1 percent higher relative to our alternative baseline both last year and this year. That’s about $300 billion in lost output per year.
The next line—per capita GDP—tells an interesting and revealing story. Because economic growth is the sum of productivity growth and the growth in hours worked, deportations—which led to reduced aggregate hours worked—show up as a drag on GDP growth in the model. But because both real GDP and population fell by similar percentages, the per-capita GDP measure shows less of an impact.
That’s also why unemployment is largely unchanged in the model (down slightly last year; up slightly this year). Labor demand, proxied by job growth, suffered greatly from the Trump administration’s policies, down about 1.2 million jobs last year and 2 million jobs this year. But because labor supply was also significantly reduced by the administration’s policies, the unemployment rate—which is affected by both supply and demand—hasn’t gone up very much. Another way to put this is that, had the Trump administration not implemented its agenda, the U.S. economy would have a lot more labor demand (jobs) and a lot more labor supply, such that the simulated unemployment rate would not be that different from the actual rate.
This doesn’t mean that those job losses are harmless. An unemployment rate in the neighborhood of 4 percent feels a lot better to people if achieved through strong labor demand and supply versus weaker versions of those key variables. True, layoffs have remained low so far, but the very low hiring rate means the job market is uniquely unwelcoming to new entrants and has little buffer against layoffs should demand falter further—a distinct possibility as the war’s inflation impacts erode real incomes.
Inflation and interest rates are considerably higher due to the Trump administration’s policies, a point that’s also often been made (regarding inflation) by Federal Reserve Chair Jerome Powell. Last year, the tariffs were the main factor behind the just-under-a-point higher inflation than would otherwise have occurred, and in 2026, the war’s energy price impacts are on top of that. Empirically, the 30-year mortgage rate tends to track the 10-year Treasury rate, and both would be about 60 basis points lower in both 2025 and 2026 in the counterfactual scenario. Higher interest rates translate into higher mortgage payments (and fewer refinancings), more expensive debt service, and dampened investment.
All told, this simulation suggests that the Trump administration’s agenda has generated stagflation: slower growth and faster inflation. It is not 1970s level stagflation, to be sure, and even with the growth-reducing agenda, real GDP growth has been around 2 percent and neither unemployment nor layoffs have spiked. But the counterfactual shows that the administration’s policies have been costly—and this is only over the short-term. Longer-term costs from reduced investment in both people and public goods will also take a toll on future growth.
The author thanks Ernie Tedeschi for macroeconomic modeling.
Methodology
This analysis measures the economic effects of 2025 and 2026 policies using MA/US, a large structural macroeconomic model of the United States originally developed by Macroeconomic Advisers and currently maintained by S&P Global. The analysis uses the October 2024 baseline for MA/US, the final baseline before the 2024 election, which largely reflects pre-Trump administration trends and policies. While all counterfactuals are necessarily uncertain, this baseline reflects consensus expectations as of late 2024.
Table 2 below summarizes baseline assumptions and adjustments across the variables that the CAP analysis shocked. In most cases, “2025 policy” was measured as the difference between actual shock outcomes and their October 2024 baseline expectations. For example, in October 2024 S&P assumed that the average effective tariff rate (as a percentage of goods imports) would end 2025 at 3 percent. The actual rate in 2025 Q4 was 12.5 percent. The difference between actual and baseline (9.5 percentage points in this case) constituted the 2025 Q4 policy shock in CAP’s simulation. Measuring “2026 policy” is more uncertain, since only the first quarter of the year has elapsed. In these cases, the analysis assumed simple forward projections, either maintaining end-2025 levels—as in the case of the tariff rate or federal employment—or using credible forecasts, such as the Goldman Sachs forecast of quarterly Brent and West Texas Intermediate (WTI) oil prices.
One note in particular: The analysis assumed $50 billion higher nominal federal defense consumption and investment in GDP as a result of the war in Iran in calendar year 2026. This assumption is not an estimate of the conflict’s gross cost, but rather the portion that would manifest as higher net GDP in 2026. The $50 billion assumption represents a middle ground between two extreme possibilities: the war has little net effect on GDP in 2026 (because it primarily draws from preexisting munitions and weapons inventories in 2026, and supplemental defense funding does not replenish these resources until after 2026) or the effect is front-loaded to 2026 by more than $50 billion (because a defense supplemental passes quickly and funds are outlaid quickly).
Most versions of the simulations for this analysis assumed static or unchanged Federal Reserve nominal policy rates from the October 2024 baseline, so headline results should be interpreted as the underlying “gross effect” without a Fed policy response. An alternative assumption swapped in actual policy rates and the latest policy forecasts from the Fed’s March 2026 Summary of Economic Projections (SEP). This latter assumption treats the change in Federal Reserve policy itself as a separate shock to be incorporated. Over the short term, these two assumptions only drove modest differences in model outcomes.