Introduced in Senate · Hassan / Young

Stronger Start for Working Families Act (S. 3596)

Bill text ↗Axiom bills ↗analyzed 2026-07-31
Conditional forecasts

What does S. 3596's $1 earned-income threshold do to the number of TY2027 Additional Child Tax Credit claims?

Baseline · current law holds
17.6M
80% interval 5.1M – 30.1M
Full cell + reasoning trace →
Conditional · bill enacted
27.2M
80% interval 14.7M – 39.7M
Full cell + reasoning trace →
Both arms on one scale — 80% intervals
Baseline
17.6M
Enacted
27.2M
The gap is the forecasted effect of the bill
27.2M − 17.6M = 9.6M The 9.6-million-return gap between arms is the forecasted effect of dropping the earned-income threshold from $2,500 to $1 — published before Congress decides. Each arm resolves against the IRS Table 3.3 first print when its registered condition holds; an intermediate enactment (say, a $500 threshold) resolves neither.

Computed impact — PolicyEngine

policyengine-us@1.764.6populace buildpmodaluncertified pairing
2026 federal impact
−$1.83B
Ten-year federal (2026-2035)
−$17.1B
Child poverty
−1.2%
People gaining
6.5%

Gains concentrate in the lowest income deciles (decile 1 averages +$32/yr vs +$0.30 for decile 10); Gini falls 0.02%. The annual cost declines from $1.86B to $1.53B across the window as earnings growth lifts families past the old $2,500 floor. State income-tax revenue falls about $0.35M per year — itemized separately and excluded from the federal figures above.

Reform parameters
{
  "gov.irs.credits.ctc.refundable.phase_in.threshold": {
    "2026-01-01.2100-12-31": 0
  }
}

Certification pending (prb4 re-review, 2026-07-31): the committed evidence validates against policyengine-us@1.784.3 while the Modal artifact records pe_us_version 1.764.6, and the validation path has fail-open branches — 'certified' is not yet proven at exact runtime. Figure retained as claimed-tier evidence with this caveat. Static microsim, one evidence stream with its own error bars. Anchors: PolicyEngine -$1.6B/yr on an older data build; Tax Policy Center ~$1.0B/yr.

Provisions

Section 2

Earned income threshold for the refundable child tax credit: $3,000 → $1

Under current law the refundable portion of the child tax credit (the additional child tax credit) phases in at 15 percent of earned income above a threshold — $3,000 in the permanent statute at §24(d)(1)(B)(i), overridden to $2,500 by §24(h)(6). The bill lowers the permanent threshold to $1 and strikes the $2,500 override, so the phase-in starts from the first dollar of earned income. The 15 percent phase-in rate and the refundable cap are unchanged; the change takes effect for taxable years beginning after December 31, 2025.

Quoted from the bill ▸
Section 24(d)(1)(B)(i) of the Internal Revenue Code of 1986 is amended by striking `$3,000'' and inserting `$1''. … Section 24(h) of such Code is amended by striking paragraph (6). … The amendments made by this section shall apply to taxable years beginning after December 31, 2025.
Full section text ▸
SEC. 2. EARNED INCOME THRESHOLD FOR REFUNDABLE CHILD TAX CREDIT.

(a) In General.--Section 24(d)(1)(B)(i) of the Internal Revenue Code of 1986 is amended by striking ``$3,000'' and inserting ``$1''.

(b) Conforming Amendment.--Section 24(h) of such Code is amended by striking paragraph (6).

(c) Effective Date.--The amendments made by this section shall apply to taxable years beginning after December 31, 2025. <all>

Countersignable goals

Extend the refundable child tax credit to working families with the lowest earnings by starting the phase-in at the first dollar earned.
Raise refund amounts for families whose earnings currently fall at or below the phase-in threshold.

Likely effects — shown regardless of the goals

Phase-in floor

Families with earned income below the current $2,500 threshold become newly eligible for a refundable credit; every family whose refundable credit is limited by the earnings formula gains up to 15 percent of $2,499 (about $375), holding the rate and cap fixed.

Filing requirement

The newly eligible population is concentrated among very-low-income households, many of whom are not currently required to file; claiming the credit requires filing a return, so realized take-up depends on filing behavior, not only on eligibility.

Unchanged parameters

The $2,200 credit amount, the refundable cap, and the 15 percent phase-in rate are untouched — the provision changes who reaches a given refundable amount at what earnings level, not the credit's ceiling.

Implementation barriers

IRS

The threshold change must land in TY2026 forms, instructions, and e-file schema for the 2027 filing season; the conforming strike of §24(h)(6) also touches guidance that currently cites the $2,500 figure.

Outreach / VITA

Reaching newly eligible non-filers is an outreach problem: the households gaining eligibility are those least connected to the filing system.

Candidate outcome metrics

In Thesis registry2 servesDirect implementation/outcome

IRS Statistics of Income, Individual Income Tax Returns (Publication 1304 line-item tables): number of returns claiming the additional child tax credit and total ACTC dollars, TY2027 versus prior years.

Live forecast 16.2M
Why this metric ▸

The provision operates entirely through the ACTC formula, so ACTC returns and dollars are the most direct administrative readout: if the threshold change does anything, it must appear here first. Official, first-print resolvable from a fixed IRS publication. The registered conditional pair targets TY2027, matching its policy premise. Known weakness: SOI line-item tables publish with roughly a two-year lag, so TY2027 resolves in 2029; considered the weekly Filing Season Statistics as a faster proxy and rejected it because it does not break out the ACTC line.

In Thesis registry1 serves · 1 orthogonalDownstream outcome

Census Bureau Supplemental Poverty Measure, child poverty rate — the SPM accounts for refundable credits, so a phase-in change for the lowest earners shows up here if realized take-up is material.

Live forecast 13.1%
Why this metric ▸

Selected because it is the outcome the goal actually claims (resources of the lowest-earning families) and because the SPM is already a live Thesis registry series with existing forecast cells. The conditional pair targets CY2027 so its outcome can reflect the provision's TY2027 application. Thesis commits the conservative 2028-08-01 through 2028-12-31 release window and 2028-12-31 resolve-by deadline; the registered Census announcement authenticates revised-methodology identity only and establishes neither timing value. The pair is non-exhaustive: an intermediate enacted threshold satisfies neither arm. Known weakness: attribution is noisy — a sub-$400-per-family change sits inside the SPM's year-to-year movement, so this metric is informative in the enacted-versus-unchanged-current-law pair, not as a standalone level.

Unmapped2 orthogonalHonest gap

No official series separates ACTC claims by earnings band at the threshold margin in a timely publication; SOI tabulations by AGI band are the closest proxy and lag by about two years.

Why this metric ▸

Recorded as a gap rather than omitted: the sharpest test of this provision would be claims among families earning under $2,500, and no official publication isolates that band on a useful timeline. Stating the missing series is part of the analysis — it is the demand signal a registry would need to close.

Conditional forecast sketches

P(IRS SOI TY2027 ACTC returns and dollars | §24(d)(1)(B)(i) threshold ≤ $1 for TY2027 vs unchanged $2,500 current-law threshold for TY2027; an intermediate enacted threshold is outside both arms)
P(Census SPM child poverty rate for CY2027 | §24(d)(1)(B)(i) threshold ≤ $1 for TY2027 vs unchanged $2,500 current-law threshold for TY2027; an intermediate enacted threshold is outside both arms)