Repository navigation
Fourteen SOI families bind a stale Historic Table 2 US row beside a newer Table 1.x national row on national_state #1243
Description
On the pinned feed (51aa40f, compiled at 2024 as the release does), the national_state surface binds two different national all-AGI targets for the same model quantity (same variable, measure mode, filing status, return universe and filters) in 14 SOI families. One is the stale Historic Table 2 US row (TY2022); the other is a newer Pub 1304 Table 1.x row (TY2023). No weight vector can meet both.
| quantity | national rows | max/min | 51 HT2 states / smaller national |
|---|---|---|---|
taxable_social_security sum |
HT2 TY2022 vs Table 1.4 TY2023 | 1.1219 | 0.998 |
taxable_social_security indicator_sum |
same | 1.0507 | 0.997 |
itemized_taxable_income_deductions sum |
HT2 vs Table 1.2 | 1.0306 | 0.997 |
ira_distributions sum |
HT2 vs Table 1.4 | 1.0223 | 1.019 |
ira_distributions indicator_sum |
same | 1.0161 | 0.997 |
taxable_pension_income indicator_sum |
same | 1.0139 | 1.009 |
unemployment_compensation indicator_sum |
same | 1.0107 | 0.998 |
taxable_income sum |
HT2 vs Table 1.2 | 1.0076 | 0.999 |
employment_income indicator_sum |
HT2 vs Table 1.4 | 1.0069 | 0.995 |
taxable_pension_income sum |
same | 1.0066 | 0.995 |
count indicator_sum |
Table 1.1, Table 1.2, HT2 US, state_2022 US |
1.0060 | 0.994 |
adjusted_gross_income sum |
Table 1.1, Table 1.2, HT2 | 1.0035 | 0.993 |
employment_income sum |
HT2 vs Table 1.4 | 1.0031 | 0.992 |
unemployment_compensation sum |
same | 1.0028 | 0.997 |
For ira_distributions sum and taxable_pension_income indicator_sum, the 51 unrebased HT2 state rows already sum past the Table 1.4 national target.
The count family has a fourth national row: irs_soi.ty2023.state_2022.us.return_count, from 22in54us.xlsx, "Historic Table 2 state data, United States total". It is the HT2 US total (159,651,330, TY2022 data) stamped ty2023.
Vintage, not concept. For every one of the 14 columns, the HT2 US row agrees with Table 1.4's same-year "All returns, total" within 2.7% in each of TY2020-TY2023 (HT2/Table 1.4 from 0.974 to 1.009) (20in14ar.xls-23in14ar.xls vs 20in55cmcsv.csv-23in55cmcsv.csv), and the HT2 doc guides define each column at the same Form 1040 line as Table 1.4. So each gap is vintage times a small same-year HT2 / Pub 1304 difference, not a concept difference (the cause of that difference is not established; both cover returns processed in the same calendar year). HT2 amounts age 2022 to 2023 by realized AGI growth (wages by realized wage growth), while each column's own change differed: taxable Social Security grew 15.0% and taxable IRA distributions 0.1%. For the AGI and wage amounts, the gap is the same-year difference alone (0.3%).
The taxable-interest (#489/#518) and capital-gains (#1035/#1036) families already rebase stale HT2 rows as shares onto the latest concept-matched national control and retire the HT2 national all-AGI rows. These 14 families need the same treatment, with each concept match verified.
Evidence: ~/PolicyEngine/_reviews/soi-national-controls-20261012/ (identities.json, doc guides, compile dumps).
Activity
Test, pin and document the Historic Table 2 national controls (#1243)
Rebase stale Historic Table 2 rows onto their national controls (#1243)
Test, pin and document the Historic Table 2 national controls (#1243)