Pre-registration

registration ordering unverified

first-spec commit 4c8ce8e · 2026-07-18T22:11:21Z

run generated · 2026-05-15T20:30:42Z

Run timestamp predates this path's first git-add commit (rebase, rename, or pre-git local run). Spec hash is still the path's first-add commit — not repository HEAD — but ordering is not a clean pre-registration proof.

The 2021 expansion of the US Child Tax Credit under the American Rescue Plan (full refundability + monthly payments + raised maximum) reduced the official + Supplemental Poverty Measure child poverty rate by at least 3 percentage points within the six-month payment window (July- December 2021), with a sharp reversion after expiration in 2022Q1. The discriminating test is the symmetric pre/expansion/post pattern in SPM child-poverty rates.

Falsification criterion — what would disprove this

set before the run · honoured after

This hypothesis is considered falsified if:

SUPPORTED if SPM child-poverty rate falls by at least 3 percentage points 2020-2021 AND rises by at least 2 percentage points 2021-2022 after expiration. REFUTED if either move is below 1.5 percentage points at p<0.10.

formal test & threshold

test:      US descriptive time-series 2018-2023 SPM and OPM child-poverty rates around 2021Q3-2021Q4 expansion window with explicit expiration break at 2022Q1.

Method

Template: event_study
Fixed effects: year
Clustering: year
Sample: 1 countries · 2018 – 2023
Evidence type: descriptive

Event year: 2021. Archived run uses a bespoke replication for the exact registered descriptive test: 2020 to 2021 SPM child-poverty drop during the expanded CTC year, followed by 2021 to 2022 rebound after expiration. The generic event-study template is retained as the broader design family, but the result-card verdict is governed by the explicit Census Table B-2 pre/expansion/post rule.

Data

Variable	Source	Transform
child_poverty_rate outcome	us_census:spm_child_poverty_ratetier 1	level
all_people_spm_poverty_rate outcome	us_census:spm_child_poverty_ratetier 1	level
ctc_expansion_window_2021h2 treatment	constructed:indicator = 1 for 2021Q3-2021Q4tier 5	indicator
post_ctc_expiration_2022_indicator treatment	constructed:indicator = 1 for year >= 2022tier 5	indicator
unemployment_rate control	fred:UNRATEtier 1	level
cpi_inflation control	fred:CPIAUCSLtier 1	yoy_pct_change
log_real_gdp control	fred:GDPC1tier 1	log

● ready · ● pending · ● reconstruct-needed

Detailed result card

Result card - tax_inequality_biden_ctc_2021_child_poverty

Verdict: SUPPORTED - SPM child poverty fell 4.5pp in 2020-2021 and rebounded 7.2pp in 2021-2022; both clear the registered thresholds and p<0.10 MOE check

Registered Test

Rule: SUPPORTED if SPM child-poverty rate falls by at least 3 percentage points 2020-2021 AND rises by at least 2 percentage points 2021-2022 after expiration. REFUTED if either move is below 1.5 percentage points at p<0.10.
Source: US Census Bureau, P60-283 Table B-2, published child SPM poverty rates.
Caveat: This is a descriptive accounting test. Refundable tax credits enter SPM resources, so the mechanism is partly mechanical rather than a clean randomized causal design.

Estimate

{
  "shape": "registered_pre_expansion_post_descriptive_check",
  "primary_drop_2020_2021": {
    "label": "2020->2021 drop",
    "start_year": 2020,
    "end_year": 2021,
    "start_rate": 9.7,
    "end_rate": 5.2,
    "diff_pp": 4.499999999999999,
    "se_diff_pp": 0.3438817172943696,
    "z": 13.085894869333542,
    "p_two_sided_normal_approx": 3.9646328312482584e-39
  },
  "expiration_rebound_2021_2022": {
    "label": "2021->2022 rebound",
    "start_year": 2021,
    "end_year": 2022,
    "start_rate": 5.2,
    "end_rate": 12.4,
    "diff_pp": 7.2,
    "se_diff_pp": 0.3892476740080759,
    "z": 18.4972203580865,
    "p_two_sided_normal_approx": 2.1740093560585953e-76
  },
  "rates": [
    {
      "year": 2018,
      "value": 13.7,
      "under18_spm_poverty_rate_moe_90_pctpt": 0.5,
      "all_people_spm_poverty_rate_pct": 12.8
    },
    {
      "year": 2019,
      "value": 12.6,
      "under18_spm_poverty_rate_moe_90_pctpt": 0.5,
      "all_people_spm_poverty_rate_pct": 11.8
    },
    {
      "year": 2020,
      "value": 9.7,
      "under18_spm_poverty_rate_moe_90_pctpt": 0.4,
      "all_people_spm_poverty_rate_pct": 9.2
    },
    {
      "year": 2021,
      "value": 5.2,
      "under18_spm_poverty_rate_moe_90_pctpt": 0.4,
      "all_people_spm_poverty_rate_pct": 7.8
    },
    {
      "year": 2022,
      "value": 12.4,
      "under18_spm_poverty_rate_moe_90_pctpt": 0.5,
      "all_people_spm_poverty_rate_pct": 12.4
    },
    {
      "year": 2023,
      "value": 13.7,
      "under18_spm_poverty_rate_moe_90_pctpt": 0.6,
      "all_people_spm_poverty_rate_pct": 12.9
    }
  ],
  "method_note": "Census Table B-2 reports 90 percent margins of error. The run uses a normal approximation for differences in annual published rates; the CTC effect is partly mechanical because refundable tax credits enter SPM resources."
}

Key Rates

| year | under-18 SPM poverty rate | 90% MOE | all-people SPM poverty rate | | --- | ---: | ---: | ---: | | 2018 | 13.7 | 0.5 | 12.8 | | 2019 | 12.6 | 0.5 | 11.8 | | 2020 | 9.7 | 0.4 | 9.2 | | 2021 | 5.2 | 0.4 | 7.8 | | 2022 | 12.4 | 0.5 | 12.4 | | 2023 | 13.7 | 0.6 | 12.9 |

Generated by engine/runs/tax_inequality_biden_ctc_2021_child_poverty/replication.py at 2026-05-15T20:30:42+00:00.

Strongest opposing argument

Every hypothesis ships with its charitable opposing argument. The framework earns credibility by handling objections at their strongest, not weakest.

Notes

Tax-inequality candidate, swarm-S6 batch 2. Pairs with IRA tax provisions separately.

In ordinary language

Results

Who has skin in the game — schools predicting on this