Ranked, sourced, bottomless
The decisions Congress could actually make
10 decisions in front of this Congress that economists — across the arguments they normally have with each other — think would raise long-run living standards. Open one and you get the levers that would deliver it, ranked by leverage. Open a lever and you get its mechanisms. There is no bottom: the first layers are written and sourced, and past them each layer is generated the moment you ask for it.
Every seeded node carries the published work it rests on, right on the node. Bookmark any depth — the route is what gets saved, so you come back to the exact provision you were reading rather than to the top of the list.
- 192
Debt held by the public is set to pass its 1946 record and keep climbing, with interest — which buys nothing — the fastest-growing thing the government spends money on. Nearly every economist who has looked at this agrees the path has to change; they disagree about which half of the ledger changes, and that disagreement is the reason nothing has.
- CBO, Budget and Economic Outlook 2026–2036Debt held by the public rises from 101% of GDP in 2026 to 120% in 2036, past the 1946 record of 106%. Net interest goes from $1.0 trillion to $2.1 trillion, 3.3% to 4.6% of GDP.
- CBO, Budget and Economic Outlook 2026–2036The FY2026 deficit is $1.9 trillion, 5.8% of GDP, against a 3.8% average over the past 50 years; it reaches 6.7% by 2036.
- 290
The retirement trust fund runs out in 2032, and the law then cuts every cheque at once — the largest automatic benefit reduction in American history, applied to current retirees with no phase-in. Economists disagree sharply about which mix of revenue and benefit changes is right; the near-unanimous point is that acting earlier makes any mix smaller, because more cohorts have time to adjust.
- Social Security Trustees, 2026 reportOASI reserves are projected to be depleted in the fourth quarter of 2032, a year earlier than the previous report. Benefits would then be cut about 22%, with incoming revenue covering roughly 83% of scheduled payments.
- Bipartisan Policy Center, 2026 Trustees Report explainedCombining the retirement and disability trust funds would push the depletion date to 2034, and requires an act of Congress.
- 387
The Supreme Court held in February 2026 that the emergency-powers statute does not authorise tariffs, which returns an unusually clean question to Congress: what, exactly, may a president tax at the border, on what finding, and for how long. Doing nothing is also an answer — it leaves the older, narrower delegations in place and the next expansion to be litigated rather than legislated.
- Learning Resources, Inc. v. Trump (U.S. Feb. 20, 2026)6–3: IEEPA does not authorise the President to impose tariffs. The Court noted that the Constitution assigns duties and imposts to Congress and that Congress has always used explicit language, with limits on scope and duration, when delegating tariff authority.
- Tax Foundation analysis of the rulingThe invalidated tariffs had raised more than $160 billion through the date of the decision and were projected to raise $1.4 trillion from 2026 through 2035 — close to three-quarters of the tariff revenue the administration had expected.
- 485
Electricity demand is rising for the first time in a generation while the queue to connect new supply keeps getting longer, and the gap is arriving as a price. Permitting reform is the rare item with support from both the industries that want to build clean power and the ones that want to build gas pipelines — which is exactly why it keeps nearly passing.
- Reported residential rates and utility filings, 2026Residential electricity rates rose 7.3% nationally between April 2025 and April 2026, and utilities requested $18.6 billion in rate increases in the first half of 2026.
- PJM capacity costsNew data-centre demand is estimated to have added about $9.3 billion to PJM capacity costs for 2025 and 2026 — roughly $18 a month on an average residential bill in western Maryland and about $16 in Ohio.
- 583
The finding economists agree on is that where housing is expensive, it is expensive because too little of it is legal to build, and that the binding rules are municipal. Congress does not write zoning codes — but it does send metropolitan areas a great deal of money, and it can decide what that money is conditional on.
- CRS, Housing Supply: Current Trends and Policy Considerations (April 2026)Supply constraints including land-use restrictions prevent housing supply from meeting demand and raise costs; restrictive zoning has been estimated to make homes substantially more expensive than the cost of building them.
- Senate passage of the 21st Century ROAD to Housing Act, 2026Passed the Senate 89–10 after clearing the Banking Committee 24–0, framed around supply. Its approach is to encourage and report rather than to condition existing formula funds on local zoning change.
- 681
The United States does not use markedly more health care than comparable countries; it pays much more per unit of it. That makes the price level the variable, and it is set by market structure and by what Medicare pays — both of which are statutes. The subsidy fight gets the attention because it has a deadline; the price fight is the larger number.
- CBO, cited in 2026 health policy coverageSite-neutral payment in Medicare is estimated to reduce federal deficits by about $157 billion between 2026 and 2034.
- Peterson-KFF and 2026 premium reportingACA marketplace insurers projected a median 18% increase in gross premiums for 2026, and the average employer family policy is approaching $30,000. Enhanced premium tax credits expired on December 31.
- 778
The employment-based categories, their numerical caps and the per-country limits were set in 1990 and have not moved since, while the economy has roughly tripled. CBO's scoring of recent inflows is unusually direct about the fiscal effect; the disagreement that matters is about composition and about wages at the bottom, not about whether the aggregate is positive.
- CBO, Effects of the Immigration Surge on the Federal Budget and the EconomyThe 2021–2026 surge lowers deficits by about $0.9 trillion over 2024–2034 and raises GDP by about $8.9 trillion over ten years.
- CBO, same analysisWage growth for workers with 12 or fewer years of education grows more slowly in the early years and the pattern reverses later as productivity and demand for more-educated work rise. CBO treats composition as mattering as much as the number.
- 876
The largest statement the economics profession has ever signed says a carbon tax is the most cost-effective instrument available, that the revenue should be returned as equal lump-sum payments, and that a border adjustment should accompany it. Almost none of the American policy built since goes through a price, which makes this the widest gap on this list between what the field recommends and what is law.
- Economists' Statement on Carbon Dividends (2019)Signed by more than 3,500 US economists — including 27 Nobel laureates, four former Federal Reserve chairs, fifteen former chairs of the Council of Economic Advisers and two former Treasury secretaries. Described as the largest statement in the history of the economics profession.
- Economists' Statement on Carbon Dividends (2019)Its four recommendations: a carbon tax as the most cost-effective instrument; revenue returned to citizens in equal lump sums; a rate that rises on a schedule until targets are met; and a border carbon adjustment to prevent leakage.
- 973
The credit is now $2,200 and indexed, and more than one child in four still gets less than the full amount or none of it, because it phases in with earnings. That design choice — not the headline amount — is what decides whether the credit reaches the children in the deepest poverty, and it is a one-paragraph amendment.
- CBPP, Child Tax Credit analysis (2026)An estimated 19 million children — more than one in four under 17 — will receive less than the full $2,200 credit or none at all in 2026 because their families' incomes are too low. The credit phases in at 15 cents per dollar of earnings above $2,500.
- CBPP, on the 2021 expansionThe temporary American Rescue Plan expansion made the full credit available regardless of earnings and coincided with child poverty falling to a record low of 5.2%; the measure rose to 12.4% the following year after it lapsed.
- 1070
The estimated return on federally funded non-defence research is among the highest of any category of public spending, and it compounds over decades — which is exactly why an annual appropriation, with its shutdowns and its stop-work orders, is the wrong instrument. The damage in 2026 has come less from the enacted numbers than from money appropriated and not moved.
- CRS, Federal Research and Development Funding: FY2026Enacted FY2026: NIH research funding up about 1% on FY2025, NSF down about 3%, NASA down about 2%, selected DOE R&D down about 0.4% — far smaller changes than the cuts of 37% at NIH and more than 50% at NSF that had been requested.
- Reported NSF award rates, FY2026NSF had funded roughly 613 grants part-way through FY2026, against the 3,000-plus it would typically have awarded by that point, with award dollars running near a third of historical averages.
- Published estimates of returns to public R&DCommonly cited estimates attribute 20–25% of US productivity growth since the Second World War to non-defence government R&D.
160 nodes open · deepest branch 4 layers · leverage scores rank a node against its siblings only, never across branches.
How the ranking works, so you can argue with it
The top layer is hand-written and deliberately not regenerated: “these are the decisions in front of Congress, in this order” is an editorial claim, and a ranking that reshuffles every time you reload is not a ranking. It orders decisions by size (what is at stake, in growth, in the budget, or in prices people pay), agreement (how close the economics is to settled, as against a distributional question economists have no special authority over) and whose decision it is — whether Congress is the body that can act, rather than an agency, a court, the Federal Reserve or a state legislature. 160 nodes are seeded this way, and the ones making a factual claim carry the source on the node.
That third factor is what separates this from a wish list, and it is also the most arguable thing here. Carbon pricing has the widest expert agreement of anything on the list — the largest statement the economics profession has ever signed — and it ranks eighth, because tractability is in the ranking. If you think tractability has no business in a ranking of what would move the ball forward, read that node as ranked first; the case for it is on the node.
Where the remaining disagreement is about who gains and who pays rather than about mechanism, the node says so instead of presenting one side as the technical answer. No node names a living legislator or attributes a position to one: a claim about what a real person believes needs evidence a reader can check.
Everything below the seeded layers is generated by a language model, on demand, and labelled AI-generated wherever it appears. A generated node can never carry a source: a model asked for citations supplies plausible ones, and a fabricated citation is worse than none because it is indistinguishable from a real one. Leverage scores rank a lever against its siblings only — a 90 under “fix the phase-in” and a 90 under “take back the tariff power” are not the same quantity, and nothing here ever adds them up. Treat a deep branch as a well-organised argument to check, not a citation.
Related: the Problem Web runs on the same instrument, pointed at the world’s biggest problems rather than at one building’s decisions, and the political spectrum visualizer maps where the traditions arguing about all of this actually sit.