Does taxsimtest's fica skip the Social Security wage base for the employer share?

[MaxGhenis]

@feenberg — found while diagnosing the Additional Medicare Tax double count (#1225). fica is not part of the federal income-tax score, so this does not move the benchmark headline. It does make fica comparisons meaningless above the Social Security wage base.

Summary

Since the mid-2025 builds (2025Jul5 onward), taxsimtest's fica exceeds employee plus employer OASDI/HI (plus Additional Medicare Tax, in builds from May 2026) by exactly 6.2% × (wages − $147,000) for 2022 wages above the base. That is what you get if the employer's 6.2% OASDI share is applied to all wages instead of stopping at the contribution and benefit base. tfica, the employee share, is capped correctly. The October 2024 build and the emulator both give the capped total.

Evidence

Probe: 2022, state 0, single or joint, all wages earned by the primary, idtl 2.

Wages	taxsimtest cd2026081819 fica	capped total (emulator output where shown)	difference	6.2% × (wages − 147,000)
$150,000	22,764	22,578 (emulator)	186	186
$300,000 (single)	37,314	27,828 (emulator; also the Oct-2024 build)	9,486	9,486
$400,000 (joint, primary earner)	46,864	31,178 (emulator; also the Oct-2024 build)	15,686	15,686

For $300,000 the capped total is: employee OASDI 9,114 + HI 4,350, doubled for the employer, plus Additional Medicare Tax 900, which is 27,828. The taxsim35 2025Aug23 build gives 36,414, which is also 27,828 − 900 + 9,486. So the uncapped employer share dates from the mid-2025 builds.

26 U.S.C. §3111(a) (Cornell LII, fetched 2026-09-24): "... there is hereby imposed on every employer an excise tax, with respect to having individuals in his employ, equal to 6.2 percent of the wages (as defined in section 3121(a)) paid by the employer with respect to employment (as defined in section 3121(b))."

26 U.S.C. §3121(a)(1): "the term 'wages' means all remuneration for employment ... except that such term shall not include— (1) in the case of the taxes imposed by sections 3101(a) and 3111(a) that part of the remuneration which, after remuneration ... equal to the contribution and benefit base (as determined under section 230 of the Social Security Act) with respect to employment has been paid to an individual by an employer during the calendar year ..., is paid to such individual by such employer during such calendar year."

2022 Instructions for Form 1040, Schedule 3, line 11, "Excess Social Security and Tier 1 RRTA Tax Withheld": "If you, or your spouse if filing a joint return, had more than one employer for 2022 and total wages of more than $147,000, too much social security or tier 1 railroad retirement (RRTA) tax may have been withheld. You can take a credit on this line for the amount withheld in excess of $9,114."

Minimal reproducer

printf "taxsimid year state mstat page sage depx pwages idtl\n1 2022 0 1 40 0 0 300000 2\n" | resources/taxsimtest/taxsimtest-osx.exe

The binary gives fica 37314.00, tfica 14364.00. The emulator (PolicyEngineRunner, benchmark configuration) gives fica 27828.00, tfica 14364.00.

Question: is the uncapped employer OASDI in fica intended, for example as a deliberate employer-cost concept? If not, the capped total (27,828 here) is what §3111(a) and §3121(a)(1) give.

Related: #1225 (Additional Medicare Tax double count); #21 (employer FICA in fica); #712 (OASDI on SE income above the wage base).