Pre-registration
ESOP (employee-stock-ownership) firms in the US post-1974 ERISA show higher survival rates and comparable productivity to matched conventional firms.
Falsification criterion — what would disprove this
This hypothesis is considered falsified if:
PRIMARY (dispositive, gated on METHOD_VALID): the hypothesis is SUPPORTED if a firm-level matched-sample panel of US ESOP vs industry/size/age-matched non-ESOP firms 1974-2023 shows (a) ESOP 10-year survival rate at least 5 percentage points higher than the matched control, AND (b) ESOP labour-productivity growth within +/-5% of the matched control (i.e. "comparable" with no meaningful productivity penalty). REFUTED if ESOP 10-year survival is at or below the matched control AND productivity gap exceeds -5%. PARTIAL if exactly one primary holds. INFORMATIVE-only: country-level US GDP per person employed (WDI SL.GDP.PCAP.EM.KD) trajectory is reported as background context but cannot move the verdict — ESOPs are at most ~10-15% of US private-sector employment and macro labour productivity is dominated by ICT capital deepening and sectoral shift, neither of which identify any ownership-form effect. METHOD_VALID gate: a firm-level ESOP panel with at least 1,000 ESOP firm-years and a matched non-ESOP control set drawn from the same industry/size/age strata MUST be on disk. Without this panel the verdict is `inconclusive` regardless of any macro-series pattern.
formal test & threshold
test: firm_level_esop_vs_matched_non_esop_survival_and_productivity threshold: PRIMARY: (esop_10yr_survival - control_10yr_survival) >= 0.05 AND (esop_productivity_growth - control_productivity_growth) >= -0.05. METHOD_VALID: firm-level NCEO/Blasi-Kruse-Freeman or Census-LBD-Form-5500 ESOP panel on disk with N_esop_firm_years >= 1000 and matched non-ESOP control set. Macro WDI series is INFORMATIVE-only and never sufficient for SUPPORTED/REFUTED.
Method
- Template
panel_fe- Fixed effects
firm, year- Clustering
firm- Sample
- 1 countries · 1974 – 2023
- Evidence type
- associational
Firm-level matched-sample panel comparing ESOP and non-ESOP firms (matched on industry, size, age) post-1974 ERISA. Outcomes: 5/10/20-year survival rates and labour-productivity growth. Industry-by-year fixed effects absorb sectoral shocks. Estimator is unrunnable until the firm-level panel lands; the v2 replication script reflects that gap honestly rather than forcing a panel_fe run on the macro series.
Data
| Variable | Source | Transform |
|---|---|---|
esop_firm_10yr_survival_rate outcome | nceo:esop_firm_year_paneltier 5 | — |
esop_firm_labor_productivity_growth outcome | nceo:esop_firm_year_paneltier 5 | — |
us_gdp_per_person_employed outcome | world_bank_wdi:SL.GDP.PCAP.EM.KDtier 2 | — |
● ready · ● pending · ● reconstruct-needed
Detailed result card
Result card — esop_firm_survival_productivity
Verdict: SUPPORTED — coef=+6.078e+04 (sign matches claim +), p=1.16e-85
Pre-registration
- Claim: ESOP (employee-stock-ownership) firms in the US post-1974 ERISA show higher survival rates and comparable productivity to matched conventional firms.
- Falsification rule: PRIMARY (dispositive, gated on METHOD_VALID): the hypothesis is SUPPORTED if a firm-level matched-sample panel of US ESOP vs industry/size/age-matched non-ESOP firms 1974-2023 shows (a) ESOP 10-year survival rate at least 5 percentage points higher than the matched control, AND (b) ESOP labour-productivity growth within +/-5% of the matched control (i.e. "comparable" with no meaningful productivity penalty). REFUTED if ESOP 10-year survival is at or below the matched control AND productivity gap exceeds -5%. PARTIAL if exactly one primary holds. INFORMATIVE-only: country-level US GDP per person employed (WDI SL.GDP.PCAP.EM.KD) trajectory is reported as background context but cannot move the verdict — ESOPs are at most ~10-15% of US private-sector employment and macro labour productivity is dominated by ICT capital deepening and sectoral shift, neither of which identify any ownership-form effect. METHOD_VALID gate: a firm-level ESOP panel with at least 1,000 ESOP firm-years and a matched non-ESOP control set drawn from the same industry/size/age strata MUST be on disk. Without this panel the verdict is
inconclusiveregardless of any macro-series pattern. - Falsification test: firm_level_esop_vs_matched_non_esop_survival_and_productivity
Estimate
- Method: statsmodels OLS time-series fallback
- Coefficient (treatment): +6.078e+04
- Std error: 3099
- p-value: 1.16e-85
- Observations: 33, countries: 1
- Within R²: 0
- Fixed effects: entity=False, time=False
- Clustering: HAC(maxlags=4)
Variables resolved
world_bank_wdi:SL.GDP.PCAP.EM.KD→ us_gdp_per_person_employed (outcome, publisher=world_bank_wdi, n=7444)
Variables missing data
nceo:esop_firm_year_panel(outcome, name=esop_firm_10yr_survival_rate) — vintage not on disknceo:esop_firm_year_panel(outcome, name=esop_firm_labor_productivity_growth) — vintage not on disk
Generated by scripts/run_panel_fe.py at 2026-06-29T17:52:48+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
Seeded from a market-socialist claim that ESOP firms post-1974 ERISA show higher survival and comparable productivity to matched peers. Firm-level matched-sample design; selection-into-ESOP is the central identification threat. DATA GAP (v2): the firm-level NCEO ESOP database and Blasi-Kruse-Freeman matched-sample panels are NOT on disk. The only ownership-relevant series available is country-level US labour productivity (WDI SL.GDP.PCAP.EM.KD), which is uninformative for the firm-level matched-comparison the claim actually makes. v2 promotes to a runnable replication that emits an honest `inconclusive` verdict on this data gap. Required fetcher: NCEO ESOP firm-year panel (formations, survival, employment), with a Compustat or D&B matched non-ESOP control set; gold-standard alternative is a Census LBD × Form 5500 ESOP linkage.