IDA Subsidy Accountability

Industrial Development Agencies and similar subsidy programs operate across more than 30 states, granting billions in property tax exemptions with minimal public accountability. Our analysis cross-references subsidy beneficiaries against campaign finance records, lobbying disclosures, and federal enforcement actions.

What we found

Across the 14 states with complete donor-subsidy crossref analysis, 22-28% of subsidy beneficiaries also donated to political campaigns affiliated with the officials who control subsidy approvals. How that compares to the baseline business donation rate varies by state: in most states with a computed baseline, beneficiary donation rates are at or below the all-business rate (Ohio statewide 0.9×, Texas 0.8×, Pennsylvania 0.6×). Elevated ratios appear in the largest-employer size class — 2–3× among 500+ employee firms in Ohio and Illinois. All baseline ratios are derived from Census County Business Patterns and are upper bounds.

Formal regression analysis using 438,714 New York organization donors as controls estimates that subsidy recipients donate approximately 130% more than non-recipient organizations in the same county and era (OLS, p < 0.001). A logit model finds recipients are about 24 percentage points more likely to be substantive donors ($1,000+) — among organizations already appearing in state campaign-finance records. These findings document a correlation; they do not establish that donations caused approvals.

Status and caveats

This investigation is in beta. Findings are sourced from public records and reproducible, but the project has not completed full peer review across all states. Specific caveats:

• All control group ratios are upper bounds — Census CBP covers employer establishments only; the true baseline donation rate is likely higher
• State-level variance exists — FL (36.1%) and CA (42.3%) are outliers above the 22-28% core range; CA is inflated by labor unions in CalAccess data
• Donation timing: post-approval collapse confirmed — Callaway-Sant’Anna doubly-robust estimator finds donations decline −3.11 log points after IDA approval (p<0.001), with a pre-approval surge in the 1-2 years before the vote. Prior TWFE event study was superseded after Goodman-Bacon decomposition showed 95.7% forbidden comparisons. (Updated 2026-04-16.)
• Spike methodology is transitioning — some states use z-score detection, others have been rerun with nonparametric tests (see methodology)

Flagship analysis: New York

New York has the deepest data: 104 IDAs, $12.19B in tracked exemptions, 2,753 county-level scorecards.