Two weeks ago, we published the Congressional Budget Office's baseline projections of what the federal budget and the economy would look like in 2026 and coming years if laws governing taxes and spending generally remained unchanged. Under that assumption—which is required by the statute that governs our baseline projections—various factors could cause economic and budgetary outcomes to differ from our projections.
First, some key points from those projections:
- The budget deficit widens to $3.1 trillion, or 6.7 percent of gross domestic product (GDP), in 2036.
- The government's net outlays for interest rise to $2.1 trillion in 2036, or 4.6 percent of GDP.
- The sustained large deficits in our projections are historically unusual, given that the unemployment rate is projected to remain below 5 percent.
- Federal debt held by the public grows from 99 percent of GDP at the end of 2025 to 120 percent of GDP in 2036 and 175 percent of GDP in 2056.
- The balance of Social Security's Old-Age and Survivors Insurance Trust Fund is exhausted in 2032, one year earlier than we projected in January 2025.
Our projections continue to indicate that the trajectory for budget deficits is not sustainable. The projections do not reflect last week's decision by the Supreme Court about tariffs. We are currently analyzing the effects of that decision and subsequent administrative actions.
The baseline projections we published two weeks ago in The Budget and Economic Outlook were our central estimates. At the time those estimates were made, we assessed that under laws and administrative policies then in place, the chances were roughly equal that deficits and debt would be higher or lower over the next 10 years than in our projections. Since those projections were released, several Members of Congress have asked me what might cause deficits and debt to be lower than we have projected. I will briefly recap my answers to their questions about economic growth, revenues, and spending.
Could Economic Growth Be Faster Than It Is in CBO's Forecast?
Yes, it could. In the Outlook, we published a range of possible paths for nominal GDP over the next decade. Those paths result from simulations that are based on the variability of economic outcomes in the past. They include some paths that have much faster growth (see the figure). By 2036, the cumulative growth of nominal GDP in our simulations varied widely: About 5 percent of the simulations had growth of more than 100 percent (more than doubling) since the end of 2024, and about 5 percent had growth of less than 25 percent.
Uncertainty of CBO's Projections of the Cumulative Growth of Nominal GDP Since the End of 2024
Percent
In the Outlook, we identified various factors that could contribute to faster-than-expected economic growth. For example, if improvements in generative artificial intelligence (AI) continued rapidly, and if businesses integrated those improvements effectively into their production processes, the pace of AI diffusion could be faster than anticipated. That faster diffusion would accelerate the growth of real GDP (that is, GDP adjusted to remove the effects of changes in prices) over the long run. In our central estimate, AI-related increases in productivity boost the rate of economic growth by 0.1 percentage point per year, on average. If the adoption of generative AI proved more costly than expected, gains in productivity could be smaller.
One of the main ways we evaluate whether economic outcomes are about equally likely to be above or below CBO's central estimate is to compare that estimate with projections from the private sector. In the two months since we completed our current economic forecast, private-sector forecasters have generally boosted their short-term economic projections. Among the forecasters surveyed by the Blue Chip Economic Indicators this month, 26 projected faster rates of real GDP growth for 2026 than our estimate of 2.4 percent, and 20 projected slower rates (Wolters Kluwer, February 10, 2026). Our projection is equal to the average projection of all of the forecasters surveyed.
Our forecast for real GDP growth in the rest of the coming decade averages 1.8 percent a year. That growth is slower than the nation has experienced over recent decades, largely because of slower expected growth in the number of workers. If the growth of the labor force and its effect on real GDP growth over the next decade were as large as they were over the past 75 years, on average, economic growth would average 2.8 percent a year—a full percentage point faster than we project.
Among the private-sector forecasters most recently surveyed by the Blue Chip Economic Indicators about their projections of economic growth from 2032 to 2036, the 10 forecasts with the fastest growth averaged 2.1 percent a year, and the 10 with the slowest growth averaged 1.7 percent a year (Wolters Kluwer, October 10, 2025). Our projection for that period falls near the Blue Chip consensus: It is 0.1 percentage point lower than the average projection of all of the forecasters surveyed.
If the cumulative growth of real GDP over the next decade equaled the rate at the 95th percentile of our simulations, it would average 2.6 percent a year. If that rate of real GDP growth resulted from faster productivity growth, greater labor force participation, and more investment, our preliminary analysis is that deficits would be smaller than in our projections by a total of about $1.3 trillion over the next decade. That difference in deficits would occur because revenues would be about $2.5 trillion greater over 10 years than in our central estimate, while net outlays for interest would be about $1.0 trillion greater and noninterest spending would be about $0.1 trillion greater. Those effects on spending partly reflect the projection that stronger productivity growth would increase interest rates. In 2036, the deficit would equal 5.5 percent of GDP (instead of 6.7 percent in our central estimate), and debt held by the public would equal 109 percent of GDP (instead of 120 percent).
Could Federal Revenues Be Greater Than They Are in CBO's Projections?
Yes, they could. Since 1982, our revenue projections for the next year have differed from actual revenues by an average of 6 percent. (That average ignores whether the differences were positive or negative.) With our current revenue projections, a difference of that size would equal about $330 billion (or 1.0 percent of GDP) in 2027, compared with a projected primary deficit in that year of $779 billion (which excludes net outlays for interest).
One of the factors discussed in the Outlook that could cause revenues to exceed our central estimate is the 2025 reconciliation act. For instance, if individuals and businesses responded more strongly to that law's reductions in tax rates on labor and capital income than they have to similar changes in the past, GDP growth over the next few years would be stronger than we project, leading to higher revenues.
As part of our analysis of the 2025 reconciliation act, we looked back at our projections of the effects of the 2017 tax act. Our April 2018 baseline projections incorporated the anticipated effects of the 2017 tax act in boosting investment and economic growth. The macroeconomic effects incorporated in those projections have been broadly consistent with the findings of recent research on the effects of specific provisions of the 2017 tax act.
In April 2018, we projected that real GDP growth in 2018 and 2019 would average 2.8 percent annually. Actual growth in those two years averaged 2.7 percent. In 2018, actual revenues totaled $3.330 trillion, or 99.8 percent of our projection of $3.338 trillion. In 2019, actual revenues totaled $3.463 trillion, or 99.2 percent of our projection of $3.490 trillion. In other words, our April 2018 projections of GDP growth and revenues in 2018 and 2019 were slightly too high but very close to the actual values.
To analyze how the 2025 reconciliation act's effects on the economy would in turn affect the budget, we used methods similar to those used to analyze the 2017 tax act. Those findings support our assessment that our central estimate of the 2025 reconciliation act's budgetary feedback has a roughly equal chance of being too large or too small.
Could Federal Spending Be Less Than It Is in CBO's Projections?
Again, yes, it could. Since 1985, our spending projections for the next year have differed from actual outlays by an average of 2.5 percent (ignoring whether the differences were positive or negative). With our current spending projections, a difference of that size would equal about $200 billion (or 0.6 percent of GDP) in 2027. For major programs such as Medicare and Medicaid, our spending projections for the next year have historically been about 2 percent more than actual spending.
One of the factors discussed in the Outlook that could cause spending to be less than in our central estimate is our forecast for interest rates. If, for example, the labor force grew more slowly than expected—because rates of labor force participation or net immigration were lower than projected—interest rates would be lower than we project. The opposite could also be the case.
The growth of federal spending in the long term stems in part from spending on programs that support older people, such as Social Security and Medicare. Growth in the number of beneficiaries of those programs is a major contributor to spending growth. Given the aging of the population, our projections of the number of Social Security and Medicare beneficiaries are among the most certain of our projections.
Director's note: I would like to recognize the many people at CBO who contributed to the analyses discussed here. In our collective assessment, the central estimates I have described are in the middle of the likely range of outcomes under current law. We will continue to work hard to provide Congress with the best information we can to help inform their decisions, and we will incorporate the effects of any changes in law in our future baseline projections.
Phillip L. Swagel is CBO's Director.