Bessent and Senate Republicans Are Looting Four Trillion Dollars from Working Families

Bessent and Senate Republicans are looting four trillion dollars from working families. The four trillion funds a tax cut for the top tenth of one percent. The 2025 reconciliation bill is the instrument. The current-policy baseline is the cover. The dynamic-scoring claims used to sell the bill repeat a methodology the documentary record has falsified four times since 1981.

Here are the numbers. The Congressional Budget Office scored the tax-cut title of the reconciliation legislation at approximately $4.6 trillion in conventional revenue loss over the ten-year budget window, against a current-law baseline as required by §257 of the Balanced Budget and Emergency Deficit Control Act of 1985. The Joint Committee on Taxation produced the distributional table. The top one percent’s share of the tax-cut benefit exceeds the combined share of the bottom four quintiles. The §199A passthrough deduction expansion — the carve-out that taxes income from pass-through businesses (partnerships, S-corps, sole proprietorships) at lower rates than wage income — concentrates in returns above one million dollars, alongside the corporate-rate reductions and the capital-gains preferences. On the program side, the offsetting cuts in SNAP, Medicaid, ACA premium subsidies, the refundable portion of the Child Tax Credit, federal-employee benefits, and the federal-match formula for state-administered programs total approximately one trillion dollars to one-point-two trillion. Net deficit effect on a conventional basis: roughly three-and-a-half trillion dollars in additional federal borrowing over the window, before debt-service. The bottom quintile loses net after the benefit cuts. The top quintile keeps the money. The math is in the JCT table.

The procedural device that made the bill scorable as cheaper than $4.6 trillion is the current-policy baseline. The standing baseline convention since the 1974 Budget Act, codified at BBEDCA §257, assumes that scheduled expirations occur — provisions written with a sunset are scored as if the sunset arrives. The 2001, 2003, and 2017 tax cuts fit inside their reconciliation windows on this convention: the architects accepted Byrd-Rule sunsets because the alternative was scoring multi-trillion-dollar permanent deficit effects the procedure would not allow. The 2025 reconciliation moved the goalposts. The Senate Budget Committee chair invoked statutory authority to declare a current-policy baseline — which assumes the expiring 2017 provisions continue indefinitely — and scored the extension as costless against that fictional baseline. The Committee for a Responsible Federal Budget and the Peter G. Peterson Foundation reached the same documentary finding from independent shops: the current-policy baseline is not a baseline convention as the term has been used in Congressional budget practice since 1974. It is a press release. The cost of extending the 2017 individual provisions on a current-law baseline is approximately $4.6 trillion. The cost on a current-policy baseline is zero. The bill became scorable as costless by the mechanism of describing it as costless. That is fraud against the budget process the country has used to govern itself for fifty-two years.

The dynamic-scoring claims attached to the bill are the second piece. Treasury Secretary Bessent’s office produced an analysis claiming the cuts would pay for themselves through macroeconomic feedback — the assertion that lower rates produce enough growth to generate offsetting revenue. The one-page Treasury “analysis” in 2017 made the same claim about the Tax Cuts and Jobs Act and contradicted JCT’s contemporaneous $385 billion dynamic estimate (JCX-67-17, November 2017) by an order of magnitude. JCT scored the 2017 bill’s net dynamic feedback at $451 billion of revenue offset less $66 billion of higher debt-service, against a $1.46 trillion static cost — a 26 percent offset, not the 100-percent-plus offset advocates predicted before the vote. The actual outturn aligned closer to JCT’s number than to Treasury’s: corporate investment effects were modest; share repurchases rose approximately 88 percent in 2018; the wage gains the Council of Economic Advisers projected for non-supervisory workers did not materialize. The 2025 dynamic claims rest on the same modeling choice that produced the 2017 overshoot: a small-open-economy assumption — treating the United States as if foreign capital flows in elastically enough to finance any tax-cut-driven investment without bidding up domestic interest rates — paired with capital-supply elasticities the empirical literature does not support for a country of approximately one-quarter of global output. The corporate-tax-incidence split that produces the headline wage-gain finding — labor bears 70 percent of the corporate tax — is the Tax Foundation outlier. Treasury’s Office of Tax Analysis uses 18 to 20 percent on normal returns and zero on super-normal. JCT uses 25 percent. The Tax Policy Center uses 20 percent. The 70 percent figure is reachable only under assumptions the empirical literature does not support. That is not a technical accident; it is the rhetorical operation that makes the wealth transfer politically marketable.

The historical analytical record is what makes the 2025 claims unrecognizable on the merits. The 1981 Economic Recovery Tax Act was sold with the claim that cuts would pay for themselves; David Stockman, who was Reagan’s OMB director and is the architect of record, wrote in The Triumph of Politics (1986) that he “out-and-out cooked the books, inventing $15 billion per year of utterly phony cuts” to fit the headline number. The deficit subsequently exploded; TEFRA partially reversed the loss in 1982. The 2001 and 2003 cuts were sold the same way; the CBO retrospective put the combined cost at approximately $1.6 trillion in actual revenue loss, and the Heritage Foundation’s contemporaneous projection that the cuts would eliminate the publicly held debt by 2010 missed the actual debt path by approximately eleven trillion dollars. The 2017 TCJA was sold the same way; JCT’s contemporaneous dynamic score offset 26 percent of the static cost, and the outturn looks closer to JCT’s number than to Treasury’s.

Bruce Bartlett, who helped draft the original 1981 cuts, wrote in the Washington Post on September 28, 2017 that the Republican claim that tax cuts pay for themselves is “hogwash” and “a lie.” This is the inside witness. The 2025 dynamic claims are the same claim being made a fifth time, by some of the same operators, with the same modeling choices, against the same documentary record showing the claim has been falsified each prior time. The pattern is forty-five years old.

The credentialing operation around the bill is the third piece. Douglas Holtz-Eakin chaired the Congressional Budget Office from 2003 to 2005. Under his directorship, CBO produced a study finding that the 2001 tax cuts’ new revenue did not offset their cost. He was honest as director. He left CBO in 2005 and in 2010 founded the American Action Forum, an advocacy organization that supports the coalition writing the 2025 reconciliation bill. AAF’s 2025 brief describes the bill’s tax-cut title as “growth-financed” — a characterization the CBO he ran would not have produced and that the JCT distributional table available to him does not support. The press cites him as the former CBO director. The credential is operative because of the documentary record CBO produced under his leadership; the advocacy operates against that record. Dr. Holtz-Eakin knows what CBO scored and how. He chooses to write something else. Kevin Hassett, who chaired the Council of Economic Advisers from 2017 to 2019 and made the wage-gain projections JCT’s models did not support, is again at Treasury producing the same forecasts that did not match the prior outturn. Larry Kudlow’s 2025 defenses repeat his 2017 defenses. Stephen Moore’s most recent brief carries the same projection error the 1981, 2001, and 2017 briefs carried. The credentialing of former federal-fiscal-office personnel by advocacy organizations is the operational pattern. None of this makes the analysis wrong on its face. It does mean the analysis is advocacy, and the methodological choices are the ones that produce the conclusion the advocates wanted.

The symmetric application is honest and is the point. The same methodology discipline applies to the analytic shops on the other side of the room; the volume of cases is smaller but the temperature is identical. The Center on Budget and Policy Priorities, whose program-effectiveness work on the 2025 cuts is documentary and cited above, also overstated the revenue offset from the 2022 IRA’s $80 billion in IRS funding; CBO’s initial $204 billion estimate was revised to approximately $180 billion after Treasury’s $400,000-floor enforcement guidance, and CBPP’s contemporaneous projections did not survive the revision. The Roosevelt Institute’s published multipliers for federal-investment dynamic effects on occasion exceed what the empirical multiplier literature supports — multipliers in the 1.5-to-2.0 range are defensible during recessions and slack labor markets, not at full employment, and the briefs deploying them sometimes do not condition. The Center for American Progress’s distributional briefs on prior tax legislation have occasionally used aggressive refundable-credit assumptions on the upside that the data did not support. The point is not coalitional balance. Asymmetric output produced by symmetric application of consistent standards to an asymmetric world is fairness working; asymmetric application is fairness violated. The receipts are densest where the receipts are densest. The standard is the same.

The 2025 bill does the opposite of what federal fiscal policy is supposed to do, in three specifications the documentary record makes legible. Fiscal policy is supposed to project distributional impact accurately; the 70/30 corporate-incidence assumption marketed for this bill projects a distribution the empirical literature does not support, while the JCT distributional table available to the drafters shows the actual concentration in the top one percent. Fiscal policy is supposed to match revenue with expenditure over the long term; the current-policy baseline removes that match by accounting fiction, and the long-term debt-to-GDP trajectory CBO published in the January 2026 Budget and Economic Outlook crosses the post-1946 record under current law and crosses sooner under the bill. Fiscal policy is supposed to use scoring honestly rather than as a rhetorical instrument; honest scoring of the tax-cut title at $4.6 trillion against a current-law baseline, with the JCT distributional table appended, would price the bill differently in the Congressional Record and the bill would not pass under that pricing. The current-policy gimmick is what makes the bill politically marketable. The dynamic-scoring claims are what make the price palatable to legislators. The methodology choices and the procedural innovations are the rhetorical operation. They were chosen to produce the conclusion.

Here is what the receipts say, and what they have said for forty-five years. The 1981 cuts did not pay for themselves; their architect later wrote that the numbers were cooked. The 2001 and 2003 cuts did not pay for themselves; the Heritage projection missed the debt path by eleven trillion dollars. The 2017 cuts did not pay for themselves; JCT scored 26 percent feedback against a 100-percent-plus advocate claim. The 2025 extension is being sold on identical projections by some of the same operators, against the documentary record showing the projections were wrong each prior time. Secretary Bessent’s Treasury, Senator Thune’s Senate, Chairman Smith’s Ways and Means Committee, the Heritage Foundation’s tax-policy operation, the American Action Forum’s brief carrying the former CBO director’s name, the Wall Street Journal editorial page and the Hassett-Kudlow-Moore commentariat cycle — all of these institutional actors had the JCT distributional table available, the CBO conventional score available, the §257 baseline construction available, and the historical analytical record showing each prior round of these claims falsified. They wrote the bill anyway. They scored the bill anyway. They are selling the bill anyway. The honest reading is on the receipts: allow the 2017 individual provisions to sunset as the Budget Act sunset them; restore the §257 baseline construction by Congressional action; broaden the corporate-tax base by repealing the §199A passthrough deduction, the carried-interest preference, and the expanded capital-expensing whose incidence concentrates at the top; preserve the SNAP, Medicaid, ACA, and refundable-CTC architectures the empirical literature supports as cost-effective. None of those positions is exotic. The Budget Act mechanisms exist to support them. The 2025 reconciliation bill is the active choice to use those mechanisms for the opposite purpose. This was decided in advance and the methodology was retrofitted to the conclusion. The score is the score. The author of the bill does not get to grade it.