-
Notifications
You must be signed in to change notification settings - Fork 206
ny_inflation_refund_credit assigned to wrong tax year (2025 instead of 2023) #7902
Copy link
Copy link
Open
Description
Summary
ny_inflation_refund_credit is modeled as a 2025 refundable credit using 2025 NY AGI, but the actual program is based on 2023 tax year returns and paid as direct mailed checks — not as a credit on the IT-201.
This causes PE to overstate the 2025 NY refund by $150–$400 for every eligible filer, and understate the 2023 refund by the same amount.
Related TAXSIM issue: PolicyEngine/policyengine-taxsim#788
Evidence
From the official NY Tax Department page:
- Eligibility: "for tax year 2023, you: filed Form IT-201"
- Income used: "2023 New York Adjusted Gross Income (Form IT-201 line 33)"
- Delivered as mailed checks, not claimed on the IT-201
Root Cause
formula_2025inny_inflation_refund_credit.pycomputes the credit usingny_agifor period 2025- The credit is included in
refundable.yamlfor 2025-01-01 - Per PE convention, one-time rebates should be assigned to the tax year whose liability determines the amount — that's 2023, not 2025
Suggested Fix
- Change
formula_2025→formula_2023andformula_2026returns 0 →formula_2024returns 0 - Move
ny_inflation_refund_creditfrom the 2025 entry to the 2023 entry inrefundable.yaml - Update existing tests to use 2023
Integration Test
- name: NY inflation refund credit - HoH 20K wages - should be in 2023
period: 2023
input:
people:
person:
age: 54
employment_income: 20_000
child:
age: 2
tax_units:
tax_unit:
members: [person, child]
households:
household:
members: [person, child]
state_fips: 36
output:
ny_inflation_refund_credit: 200
- name: NY inflation refund credit - should be zero in 2025
period: 2025
input:
people:
person:
age: 54
employment_income: 20_000
child:
age: 4
tax_units:
tax_unit:
members: [person, child]
households:
household:
members: [person, child]
state_fips: 36
output:
ny_inflation_refund_credit: 0Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
No labels