Budget Scoring

A budget score is an official estimate of a bill’s fiscal impact—how much it will cost (or save) the government over a ten-year period. In the United States, the Congressional Budget Office (CBO) and the Senate and House budget committees produce these estimates; other legislatures employ similar processes. Scoring rules are arcane but consequential: a pessimistic score can doom a bill; an optimistic score can get it passed, even if both rest on identical underlying assumptions.

Why ten years?

Budget scorekeeping traditionally uses a ten-year window. This is not rooted in economic theory; it is a political choice. A ten-year horizon is long enough to capture the main effects of a policy (a tax cut affects revenues for the full decade; a new programme affects spending for ten years) but short enough that extreme uncertainty does not render the score useless.

If the CBO scored only one year ahead, bills that phase in slowly would appear cheap; Congress could hide long-term costs. If the window were thirty years, compound uncertainty would dominate, and the score would be little more than a guess. Ten years is a pragmatic compromise.

That said, the ten-year window creates perverse incentives. Policies that back-load costs (paying large sums in year eleven) or front-load benefits appear cheaper than they are. A tax increase that starts in year twelve does not affect the ten-year score, so a bill can appear deficit-reducing even if it genuinely pushes costs to the future.

The CBO’s scoring process

The CBO is an independent legislative agency, formally nonpartisan, that produces budget estimates for Congress. When a bill is introduced, one or both chambers may request a score. The CBO’s analysts examine the bill’s text, consult with agencies, and model its fiscal impact.

The baseline is “current law”—the assumption that existing tax and spending rules remain unchanged. Against this baseline, the CBO estimates the bill’s effect. If a bill cuts Medicare payments, the CBO estimates how much Medicare spending will drop. If a bill creates a new tax credit, the CBO estimates the revenue cost and behavioural responses.

The CBO’s estimates are not the only views in the room. The bill’s sponsors often request alternative analyses, sometimes from the “Joint Committee on Taxation” (JCT) or the CBO’s scoring team under different assumptions. The most charitable interpretation of the bill’s sponsors’ assumptions gets published alongside the CBO’s estimate, allowing Congress to choose which view to credit.

This is not a bug; it reflects the genuine uncertainty inherent in projection. Reasonable analysts can disagree about elasticities (how sensitive behaviour is to price), feedback loops, and macroeconomic effects. The CBO generally publishes the central case; alternatives show the range of plausible outcomes.

Baseline and current law

The “baseline” against which a bill is scored is crucial because it determines the apparent cost. If the baseline assumes that the Bush tax cuts expire (as law dictated until they were extended), a bill that extended them appears to cost far more than a bill that extended them under a baseline that already assumed continuation.

Congress sets the baseline methodology. It has changed over time. In the 1990s, the baseline assumed tax laws would be extended indefinitely (the “current-policy baseline”), making tax cuts appear costless. In 2010, Congress shifted to a baseline where temporary laws expire on schedule (the “current-law baseline”), making extensions appear expensive.

These are not subtle matters. They can swing the apparent cost of a bill by hundreds of billions of pounds. When Congress debated tax cuts, the dispute over the baseline was often more consequential than the substantive policy debate.

Behavioural assumptions and feedback

A scored bill typically assumes current law continues, but it does incorporate some behavioural responses. If a bill taxes the wealthy more heavily, the CBO will estimate that some earning will shift to lower-tax jurisdictions and some will be deferred. If a bill expands a welfare programme, the CBO estimates take-up rates and account for the fact that not every eligible person will sign up.

However, the CBO does not score dynamic macroeconomic effects—the idea that a tax cut might boost GDP growth so much that revenues rise. Most economists agree that very large tax cuts have small positive macroeconomic effects, but estimating them is controversial. Congress’ rules forbid the CBO from incorporating macroeconomic feedback unless explicitly instructed. This means a tax cut that reduces the deficit in year one, according to the CBO, might actually increase it once growth effects are included.

Conversely, bills’ sponsors often argue that the CBO is too static—that it ignores the growth effects of tax cuts or spending programmes. This is a perennial dispute, and scorekeeping rules have shifted to allow (in some circumstances) dynamic scoring, though the CBO remains cautious about assuming large feedback loops.

The political incentive structure

Because budget scores carry weight—a bill that worsens the deficit struggles to pass under fiscal consolidation rules or pay-as-you-go restrictions—there is enormous political pressure to craft bills so they score well. This creates opportunities for what critics call “gimmicks.”

One common gimmick is the sunset clause: a programme is funded only for a limited time within the ten-year window, making it appear cheaper. When year eleven arrives, the programme must be re-authorised and funded, but that cost sits outside the score. Congress then faces a choice: cut the programme, raise revenue, or break its fiscal rule.

Another gimmick is the timing trick: a tax is delayed to start in year six, so the ten-year cost is smaller. Or spending is front-loaded, so the early years’ costs eat the full score while later years are cheap. These stratagems are legal but hollow.

A third is to pair a costly bill with offsetting savings elsewhere. If a bill costs £200 billion but Congress attaches a provision to cut spending by £200 billion (even if the cut is never actually implemented), the bill scores as paid-for. This lets Congress pass substantive changes whilst appearing fiscally responsible.

Accuracy and limitations

The CBO’s ten-year scores are forecasts, and forecasts have errors. Econometric models miss structural shifts. Policymakers respond in ways the baseline did not anticipate. Unforeseen events (recessions, wars, pandemics) alter the fiscal landscape.

Historical analysis shows that CBO’s out-year estimates (forecasts for years five to ten) have substantial uncertainty bands around them. A bill scored as costing £100 billion could plausibly cost £80 billion or £120 billion once the future unfolds. This is not a failing of the CBO; it is inherent in long-term forecasting.

Nonetheless, the score carries political weight as if it were hard fact. A bill that scores as adding £50 billion to the deficit is politically poisoned, even if the estimate’s confidence interval is huge. The precision of the number (£50 billion, not “roughly half a trillion over ten years”) invites people to treat it as certain.

Contested assumptions: Medicare and dynamic scoring

Two perennial disputes over budget scoring are the treatment of Medicare and the use of dynamic scoring.

Medicare is scored by assuming payment rates to providers remain as law specifies, even though Congress has repeatedly overridden these rates (the “Sustainable Growth Rate” fix). If the CBO assumes rates fall sharply, programmes that reduce Medicare spending appear to save money that Congress will never actually realise. This asymmetry—assuming deep provider cuts are real but assuming Congress will reverse them—has plagued scoring for decades.

Dynamic scoring—incorporating macroeconomic feedback effects—has become more contentious in recent years. Conservative economists argue that the CBO’s static scores understate the growth benefits of tax cuts and thus overestimate their costs. Liberal economists counter that dynamic effects are small and that assuming large feedback loops is wishful thinking. Congress has authorised dynamic scoring in some contexts, but the CBO remains circumspect.

International comparisons

Most legislatures employ budget-scoring systems. The UK’s Office for Budget Responsibility, Canada’s Parliamentary Budget Officer, and Australia’s budget-costing systems all perform analogous functions. However, the specific rules differ—some score ten years, others five; some incorporate macroeconomic feedback, others do not; some use current-policy baselines, others use current-law.

The CBO is widely regarded as particularly rigorous and nonpartisan, though it is not immune to political pressure or assumption disputes. Its credibility rests on independence, transparency, and long institutional history.

See also