Most people think the Congressional Budget Office produces numbers. That’s true, but it’s also incomplete. What the CBO actually produces is a story—a narrative about a bill’s future, rendered in tables and confidence intervals. That story, more than any floor speech or press release, determines whether a piece of legislation reaches a vote. If you don’t understand how the CBO constructs that narrative, you can’t understand why so many bills die before they ever see the light of the House or Senate floor.
Consider a hypothetical but entirely realistic major authorization bill: the National Infrastructure Resilience and Modernization Act (NIRMA). The bill authorizes $85 billion over ten years for grants to states for flood mitigation, grid hardening, and drought resilience. It includes a new formula allocation, a competitive grant program, and a set of regulatory streamlining provisions. The sponsors believe it will save money in the long run by reducing disaster recovery costs. The CBO, however, will not score those savings unless they are mandatory and clearly attributable to the bill’s provisions. That gap—between what the sponsors believe and what the CBO can count—is where the story begins.
The Baseline Is the First Draft
Every CBO score starts with a baseline. The baseline is not a prediction of what will happen. It is a projection of what would happen if current law remained unchanged, adjusted for inflation and economic trends. For NIRMA, the baseline assumes that existing disaster recovery spending—primarily through the Federal Emergency Management Agency’s Disaster Relief Fund—will continue at its historical average, adjusted for inflation. The CBO’s baseline for disaster spending over the next decade is roughly $200 billion. NIRMA’s $85 billion in new authorizations would appear, in the CBO’s ledger, as an additional cost on top of that baseline, not as a substitute for it.
This is the first narrative choice. The CBO’s baseline methodology, governed by the Balanced Budget and Emergency Deficit Control Act of 1985 and subsequent scorekeeping guidelines, treats discretionary spending as a fixed path. If a bill authorizes new spending, the CBO scores it as an increase relative to that path, even if the bill’s proponents argue it will reduce future emergency appropriations. The CBO cannot assume that future Congresses will appropriate less for disaster relief just because NIRMA builds a more resilient grid. That assumption would require predicting legislative behavior, which the CBO explicitly avoids. The result: NIRMA’s score shows a cost of $85 billion, with no offsetting savings.
Behavioral Responses and the Art of What Gets Counted
The second narrative layer involves behavioral responses. The CBO’s models attempt to account for how individuals, firms, and state governments will react to a policy change. For NIRMA, the key behavioral question is whether the availability of federal grants will cause states to reduce their own infrastructure spending—a phenomenon known as the “crowding-out” effect. The CBO’s analysts, drawing on academic literature and historical data from similar grant programs, estimate that for every dollar of federal grant money, state and local spending on resilience projects will decline by roughly 20 cents. That reduces the net national investment, and the CBO’s score reflects it: the effective impact of the $85 billion authorization is closer to $68 billion in new resilience spending, with the rest simply replacing state dollars.
This behavioral adjustment is not a political judgment. It is a modeling choice, grounded in peer-reviewed research. But it becomes a political weapon the moment the score is released. Opponents of the bill will cite the crowding-out estimate as evidence that NIRMA is inefficient. Supporters will argue that the CBO’s model underestimates the catalytic effect of federal investment—that the grants will spur additional state and private spending, not replace it. The CBO’s analysts, bound by their mandate to provide objective, impartial analysis, will not engage in that debate. They will publish their methodology, answer technical questions from staff, and let the numbers speak. But numbers do not speak; people speak for them.
Time Horizons and the Discount Rate Dilemma
The third narrative choice is the time horizon. The CBO typically scores legislation over a ten-year window, a convention established by the Congressional Budget Act of 1974. For NIRMA, the ten-year window captures the full cost of the authorization but only a fraction of the benefits. Flood mitigation projects, for example, have useful lives of 30 to 50 years. The avoided disaster recovery costs—the savings that NIRMA’s sponsors tout—will accrue over decades, not years. The CBO’s ten-year score shows $85 billion in costs and, because the savings are not mandatory and not clearly attributable, zero dollars in benefits. The story the score tells is one of pure fiscal burden.
This is not a flaw in the CBO’s methodology. It is a constraint imposed by the budget process. The ten-year window exists because longer-term projections become increasingly uncertain, and because the budget resolution that governs congressional action typically covers a decade. But the constraint has consequences. It systematically disadvantages legislation with long-term payoffs, like infrastructure resilience, preventive health care, and early childhood education. It advantages legislation with immediate, visible costs and benefits, like tax cuts or direct transfers. The CBO is not making a value judgment; it is following the rules. But the rules themselves shape the story.
Uncertainty Disclosure and the Weaponization of Ranges
The CBO’s scores are not single numbers. They are ranges, accompanied by descriptions of uncertainty. For NIRMA, the CBO might estimate that the bill will increase direct spending by $85 billion over ten years, with a 90 percent confidence interval of $70 billion to $100 billion. The report will note that the estimate is “highly uncertain” because it depends on assumptions about state participation rates, construction costs, and the frequency of extreme weather events. That language is a professional necessity. It is also a political gift.
In markup, a committee member opposed to the bill will seize on the upper bound: “The CBO says this could cost $100 billion.” A supporter will cite the lower bound: “The CBO’s best estimate is $70 billion, and that doesn’t count the savings.” Both statements are technically true, and both are misleading. The CBO’s uncertainty disclosure, designed to promote transparency, becomes a menu of talking points. The analysts who wrote the report have no control over how their ranges are used. Their job is to produce the estimate; the politics belong to the members.
This dynamic is not unique to NIRMA. It is a feature of every major CBO score. The Brookings Institution has documented how CBO cost estimates shape legislative strategy, noting that “the CBO’s modeling choices and baseline assumptions influence which legislation advances or stalls” (Brookings – Quality. Independence. Impact.). The score is not just a number; it is a framing device. It defines the terms of debate before the debate begins.
How Staffers and Policy Professionals Respond
Experienced legislative staffers do not wait for a CBO score to land and then react. They anticipate it. The process begins during bill drafting, when the legislative counsel’s office works with committee staff to structure provisions in ways that minimize scorable costs. For NIRMA, that might mean designing the grant program as a capped authorization rather than an entitlement, or including a sunset clause that limits the scoring window. It might mean adding a provision that requires states to maintain their own spending levels as a condition of receiving federal funds, directly addressing the crowding-out concern.
But drafting is only half the battle. The other half is narrative preparation. Before the CBO releases its score, staffers prepare a counter-narrative: a memo that explains what the score does and does not say, highlights the limitations of the ten-year window, and provides alternative estimates from outside analysts. This memo is not for public consumption. It is for the members, the leadership, and the relevant committee chairs. Its purpose is to inoculate the bill against the most damaging interpretations of the score before those interpretations take hold.
This is where institutional research tools become valuable. Policy professionals increasingly rely on Congressional Research Service memorandums and CBO’s own preliminary estimates to model how different assumptions will affect a bill’s score, test alternative provisions, and generate the kind of narrative memos that can keep a bill alive. The goal is not to replace the CBO’s analysis but to contextualize it—to show that the score is one story among several, and that the bill’s merits cannot be reduced to a single number.
That same discipline applies to long-form organization: before publishing, editors need a way to test whether a complicated body of material has a coherent beginning, middle, and end, which is where an AI book generator that fits the project can function as a planning aid rather than a substitute for domain evidence.
The Score as a Veto Point
The most important thing to understand about CBO scores is that they function as veto points. A bill that receives a high cost estimate, particularly one that exceeds the allocation in the budget resolution, cannot proceed to the floor without a waiver of the relevant budget point of order. In the Senate, that waiver requires 60 votes. In the House, it requires a special rule from the Rules Committee. Either way, the score creates a procedural hurdle that can be fatal.
For NIRMA, assume the Senate Budget Committee has allocated $50 billion in new budget authority for the relevant function. The CBO scores the bill at $85 billion. The bill is now $35 billion over the allocation. The chairman of the Budget Committee can raise a point of order against the bill, and unless the majority leader can secure 60 votes to waive it, the bill is dead. The score has become a story about fiscal irresponsibility, and that story has become a procedural barrier.
This is not an accident. The budget process was designed to enforce fiscal discipline, and the CBO’s scores are the enforcement mechanism. But the process also creates perverse incentives. It encourages sponsors to underfund programs, to rely on unrealistic assumptions, or to structure bills in ways that game the scoring rules. It discourages long-term investment. And it concentrates power in the hands of the few members and staff who understand how the scoring process works.
What the Public Misses
The public debate about a bill like NIRMA will focus on the top-line number: $85 billion. Opponents will call it wasteful; supporters will call it essential. Almost no one outside the committee rooms will discuss the baseline assumptions, the behavioral models, or the time horizon. The CBO’s methodology will remain a black box, even though the agency publishes detailed documentation of its models and invites public comment on its methods.
This is a failure of policy reporting, but it is also a failure of institutional communication. The CBO is not designed to explain itself to the public. Its audience is Congress. Its reports are written in the language of professional economics, not public discourse. The result is a gap between what the score actually says and what the public thinks it says. That gap is filled by partisans, lobbyists, and journalists who often lack the technical background to interpret the score correctly.
Pew Research Center surveys consistently show that public understanding of federal budget processes is low, and that “CBO scores become political weapons that shape markup and floor debate” (Pew Research Center | Nonpartisan, nonadvocacy, public opinion polling and data-driven social science research). The public hears the number, not the narrative. And the number, stripped of context, is almost always misleading.
How to Read a CBO Score
If you want to understand what a CBO score actually means, start with the baseline. Ask what current law assumes and whether those assumptions are realistic. Then look at the time horizon. Is the bill’s impact measured over ten years, and if so, what happens in year eleven? Check the behavioral assumptions. Does the model account for substitution effects, crowding out, or induced demand? Read the uncertainty section. How wide is the confidence interval, and what drives the uncertainty? Finally, look at the mandatory versus discretionary classification. Is the spending scored as direct spending, subject to PAYGO rules, or as discretionary, subject to appropriations?
These questions will not give you a definitive answer about whether a bill is good or bad. They will give you something more valuable: an understanding of what the score is actually saying, and what it is leaving out. That understanding is the difference between being manipulated by a number and using it to make a judgment.
The Institutional Lesson
The CBO is one of the most respected institutions in Washington, and for good reason. Its analysts are rigorous, its methods are transparent, and its leadership is committed to nonpartisan analysis. But the CBO operates within a set of rules that were written decades ago, for a different fiscal environment. Those rules—the ten-year window, the baseline conventions, the treatment of behavioral responses—shape the stories the CBO tells. And those stories, in turn, shape the legislation Congress considers.
If we want better policy outcomes, we need to understand not just the numbers but the narratives they create. We need to ask why certain costs are counted and certain benefits are not. We need to recognize that a CBO score is not a fact; it is an estimate, built on assumptions, constrained by rules, and interpreted by political actors. The score is a story. The question is whether we are reading it critically or just repeating the headline.