# CBO Baselines and Scoring Windows Every budget estimate answers the question "compared to what, over what period?" Both halves are conventions rather than facts, and both are exploitable. ## The baseline The [[Congressional Budget Office]] baseline is a projection of what happens under **current law** — not current policy, and not any economist's best guess about the future. That distinction is where the manipulation lives. Current law says a temporary provision expires on its scheduled date. If everyone involved fully intends to extend it, and has extended it every prior time, the baseline still books the expiration. Consequently: - **A permanent policy enacted as temporary looks cheaper.** A ten-year tax cut written to expire in year eight scores as eight years of revenue loss. If it is then extended, the extension is scored as a *new* cost against a baseline that assumed expiration — so the same policy is paid for twice in the accounting and never once in reality. - **Extending existing policy scores as a massive new cost**, which is why "we are just preventing a tax increase" and "this is a multi-trillion-dollar tax cut" are both accurate descriptions of the same bill. The [[Tax Cuts and Jobs Act]] is the textbook case: corporate provisions permanent, individual provisions scheduled to expire. The sunset was not a policy judgment about individuals versus corporations. It was a scoring device required to fit the bill through reconciliation's deficit constraints. ## The window Congressional scoring conventionally uses a **ten-year window**. This is arbitrary, and it creates precise, predictable incentives: **Push costs past year ten.** Any expense scheduled for year eleven is free. Phase-ins, delayed effective dates, and back-loaded provisions all exploit this. **Pull revenue into the window.** *Timing shifts* raise no additional revenue over the long run but move receipts inside the window — Roth conversion incentives are the classic example, accelerating tax collection from the future into the present while *reducing* lifetime revenue. Scored over ten years, that reads as a revenue raiser. It is the opposite. **Ignore the cliff.** A program that is fully funded for ten years and collapses in year eleven scores identically to one that is sustainable. ## Reconciliation as the forcing function These conventions matter enormously because budget reconciliation — the mechanism that allows a bill to pass the Senate with a simple majority — imposes constraints stated in terms of them. The Byrd rule bars provisions increasing the deficit beyond the budget window. That single requirement is why so much major American fiscal legislation contains sunsets nobody intends to honor. **The result is a legislative form dictated by a measurement convention.** The ten-year window was adopted as an accounting practice. It now determines the structure of tax law, and it does so invisibly to almost everyone affected. ## Why this is the branch's core pattern Baselines and windows are the purest available example of the [[Mathematics and Power Hub]] thesis. There is nothing wrong with either convention; some counterfactual and some horizon must be chosen. But the choice is: 1. **Consequential** — it changes headline numbers by trillions. 2. **Invisible** — no news report on a tax bill's cost states the baseline convention. 3. **Gameable by anyone who understands it**, which is a small professional class. The number that reaches the public is the output of a process whose most important assumptions were never presented as assumptions. ## What to ask Against what baseline. Over what window. What the ten-year figure looks like extended to twenty or thirty. Whether any provision is scheduled to expire that everyone expects to extend. And whether revenue is being genuinely raised or merely pulled forward. ## Related [[Dynamic Scoring]] · [[Rate Versus Level]] · [[Discount Rates as Policy Choices]] · [[The Laffer Curve as Existence Proof]] · [[Congressional Budget Office]] · [[Tax Cuts and Jobs Act]] · [[Mathematics and Power Hub]]