VT elderly single 2025 29Kpwages 25Kgssi 10Kpui

[feenberg]

See http://taxsim.nber.org/out2psl/999

In PE and TA Taxable Income matches but tax before credits does not.

[PavelMakarchuk]

Re-ran on current PolicyEngine: VT income tax = $1,566, matching TAXSIM-35 ($1,566) — but the TaxAct IN-111 (line 16) shows $1,281. So PE and TAXSIM agree with each other, yet both sit ~$285 above TaxAct. The gap tracks the Vermont Social Security exclusion (elderly filer, $25K gross SS): TaxAct appears to exclude more SS from VT taxable income than either engine. Open question — is the VT SS subtraction being under-applied by both PolicyEngine and TAXSIM, or did TaxAct use a different SS inclusion?

[feenberg]

I am confused. The PE output originally posted on out2psl (now orig.txt) shows siitax of 1281, which matches TA and PE now. But I just reran PE and siitax is 1566 (shown in 999.txt).

The difference in AGI between the two PE outputs is accounted for by the Retirement Benefits Exempt From Taxation (line 12 of From IN-112, which I have just now posted to the taxsim website). The worksheet is on page 16 of:

https://taxsim.nber.org/historical_state_tax_forms/VT/2025/Income-Booklet-2025.pdf

[PavelMakarchuk]

Thanks — that pins it. With the exact inputs (AGI $60,942) current PolicyEngine gives a VT retirement/Social-Security exemption of $0 (taxable SS $20,725 fully taxed) → VT tax $1,566. TaxAct's IN-112 line 12 evidently still exempts ~$8,500 → $1,281. So this is the VT retirement/Social-Security exemption income phase-out: at AGI ~$60,942 (single) PE has fully phased it out, while the IN-112 worksheet still allows a partial exemption. Likely PE's phase-out endpoint, or the AGI base it tests against, is slightly off. We'll reconcile PE's exemption against the IN-112 worksheet (p.16 of the 2025 booklet you posted) and fix on the PolicyEngine side — this almost certainly explains the #1000 VT case too.

[PavelMakarchuk]

Confirmed live PolicyEngine-US bug, filed as pe-us #8837. (Note: "IN-111/IN-112" are Vermont form names — Form IN-111 = VT income tax return, Schedule IN-112 = adjustments — not Indiana.)

The VT SS/retirement exemption eligibility gate (vt_retirement_income_exemption_eligible) tests AGI against the CSRS reduction-end threshold for all filers, even Social-Security electors. The main formula branches correctly on use_ss, but the gate doesn't. S.51 (2025) raised the SS thresholds ($55k full / $65k eligibility single) while PE's CSRS thresholds stayed stale ($60k single end). So a single SS filer with AGI in $60k–$65k is wrongly gated out → $0 exemption.

This filer (AGI $60,942): correct partial ratio 0.41 → exemption $8,497.28 (matches the filled IN-112 Line 12 = $8,497) → VT tax $1,281.20 = TaxAct. PE and TAXSIM-35 both give $1,566 — both are outliers; TaxAct is correct per the p.16 worksheet (Section I Q2: single eligibility < $65,000). Tracking via pe-us #8837.

[PavelMakarchuk]

PR up: PolicyEngine/policyengine-us#8853 (Fixes #8837, resolves #999 + #1000). Single filer reproduces exemption $8,497.28 → VT tax $1,281.20 = TaxAct.

[MaxGhenis]

Investigated on policyengine-us 1.756.10 (git main @ a2cececb, 2026-07-05) (recomputed via the harness: policyengine + local taxsimtest binary). TAXSIM-35-side bug.

Vermont retirement/Social-Security income exemption (32 V.S.A. § 5830e / Schedule IN-112). Implied federal-AGI -> VT-AGI subtraction:

• PE: -8497 (exemption applied)
• TAXSIM: +104 (no exemption)

Because taxable income already reflects PE's § 5830e exemption but TAXSIM's does not, "tax before credits" diverges even though PE and TAXSIM agree on the pre-exemption taxable income. Single filer 2025 SS phase-out band is 55,000 -> 65,000; PE applies it, TAXSIM applies $0. PE matches the statute. Grouping with #1046 / #1000 for the NBER report.