Integration design: dynamics → Populace panel → full tax-benefit incidence
The end state (per the program's design and the maintainer's direction): a Social Security reform scored here should report household-level tax-benefit incidence — benefits, taxation of benefits, SNAP/SSI/Medicaid interactions, state taxes, marginal rates — not benefit-side deltas alone. Three workstreams, ordered by dependency:
W1 — Histories onto the Populace panel (the funded build)
Impute gate-certified earnings histories (gate-1-passing generator) and demographic histories (gate-2 transition models, pending that gate's first pass) onto the certified Populace CPS file: CPS↔PSID covariate transport, then calibration against SSA administrative targets (Statistical Supplement award/beneficiary tables, already staged with provenance in this program). Deliverable: Populace gains a longitudinal dimension with per-record benefit-relevant state (AIME/PIA, marital/claiming history). Each stage inherits the pre-registration discipline (transport validation gates before scored use).
W2 — Benefits into policyengine-us (the one-variable seam)
Finding from the survivor-plumbing build (#80): pe-us takes social_security as an uprated survey input — no benefit formula exists upstream. Integration is therefore replacing a survey input with the modeled variable; all downstream incidence (taxation of benefits, program interactions, MTRs) flows through existing policyengine machinery unchanged. Interim product available before W1 completes: dynamics-computed benefit deltas for a reform, fed as inputs to a standard policyengine simulation on the current-year certified Populace file → full current-year tax-benefit incidence of an SS reform. (The taxation-of-benefits work already exercises half this pattern.)
W3 — Axiom as the determinism layer (upstream the auxiliary rules)
The §415 chain here is cross-validated against the Axiom engine 240/240 exact-to-cent. Next: encode 402(b)/(c)/(e)/(f) (spousal, survivor incl. RIB-LIM, the remarriage-at-60 predicate — implemented here and validated against SSA worked examples in #80, but absent from upstream pe-us) as rulespec encodings; cross-validate against this repo's implementation; upstream real formulas to pe-us. Outcome: pe-us gains auxiliary-benefit computation, the encodings gain a validated Title II slice, and this program's statutory layer becomes independently verified per the estimation/determinism split.
Sequencing: W3 and W2-interim are startable now; W1 is the funded build and gates the panel-based projection products. Every scored output continues through the pre-registered evaluation protocol of this repository.
Integration design: dynamics → Populace panel → full tax-benefit incidence
The end state (per the program's design and the maintainer's direction): a Social Security reform scored here should report household-level tax-benefit incidence — benefits, taxation of benefits, SNAP/SSI/Medicaid interactions, state taxes, marginal rates — not benefit-side deltas alone. Three workstreams, ordered by dependency:
W1 — Histories onto the Populace panel (the funded build)
Impute gate-certified earnings histories (gate-1-passing generator) and demographic histories (gate-2 transition models, pending that gate's first pass) onto the certified Populace CPS file: CPS↔PSID covariate transport, then calibration against SSA administrative targets (Statistical Supplement award/beneficiary tables, already staged with provenance in this program). Deliverable: Populace gains a longitudinal dimension with per-record benefit-relevant state (AIME/PIA, marital/claiming history). Each stage inherits the pre-registration discipline (transport validation gates before scored use).
W2 — Benefits into policyengine-us (the one-variable seam)
Finding from the survivor-plumbing build (#80): pe-us takes
social_securityas an uprated survey input — no benefit formula exists upstream. Integration is therefore replacing a survey input with the modeled variable; all downstream incidence (taxation of benefits, program interactions, MTRs) flows through existing policyengine machinery unchanged. Interim product available before W1 completes: dynamics-computed benefit deltas for a reform, fed as inputs to a standard policyengine simulation on the current-year certified Populace file → full current-year tax-benefit incidence of an SS reform. (The taxation-of-benefits work already exercises half this pattern.)W3 — Axiom as the determinism layer (upstream the auxiliary rules)
The §415 chain here is cross-validated against the Axiom engine 240/240 exact-to-cent. Next: encode 402(b)/(c)/(e)/(f) (spousal, survivor incl. RIB-LIM, the remarriage-at-60 predicate — implemented here and validated against SSA worked examples in #80, but absent from upstream pe-us) as rulespec encodings; cross-validate against this repo's implementation; upstream real formulas to pe-us. Outcome: pe-us gains auxiliary-benefit computation, the encodings gain a validated Title II slice, and this program's statutory layer becomes independently verified per the estimation/determinism split.
Sequencing: W3 and W2-interim are startable now; W1 is the funded build and gates the panel-based projection products. Every scored output continues through the pre-registered evaluation protocol of this repository.