Tri-State Corporate-Law Reform Event Study · v3 (working draft)

Did the 2025 corporate-law reforms move firm value?

Delaware S.B. 21, Texas S.B. 29, and Nevada A.B. 239 — measured firm by firm, day by day, on real market data, and subjected to a locked analysis plan specified ex ante, with a battery of robustness and falsification tests.
Headline finding. With the treatment variable measured as one-third single-holder voting power, Delaware shows no detectable controlled-firm reaction: the controlled-minus-dispersed estimate is , statistically indistinguishable from zero. Texas is a power-limited null; Nevada is an underpowered null; the movers are reported separately and remain provisional.
1 · The questions

Two questions, three states

Track A — incumbents. When a state changed its corporate law, did it move the firms already incorporated there? Track B — movers. Did reincorporating to Texas or Nevada move the stock? The grid below is the whole study; green = executed on real data, amber = pending.

How to read it: each cell is one state × one question. Delaware and Texas incumbents are estimated; Nevada incumbents are estimated but underpowered; the mover track is provisional pending the full as-of-event treatment lock.
2 · How an event study works

The method, explained three ways

Full methodology — data, factor model, test statistics, long-horizon & identification battery, every method cited to its peer-reviewed source →

An event study asks: on the day news hit, did the stock move more than it should have? The part of the move that the market and known risk factors cannot explain is the abnormal return — the market's verdict on the news.1

Plain English

Every stock normally moves with the market. We first learn each firm's "normal" relationship to the market over a long quiet period before the event. Then on event day we compare what the stock actually did to what that normal relationship predicted. The gap is the reaction to the news.

Worked example — the student test

A student usually scores about 5 points above the class average. This week the class averaged 80, so we predict ~85. She scores 92. The +7 is her "abnormal" score — the part not explained by how the class did. An event study does exactly this for a stock instead of a student.

The math

Market model fit on the estimation window:
\[ R_{i,t}=\alpha_i+\beta_i' F_t+\varepsilon_{i,t} \]
Abnormal return on event day \(t\):
\[ AR_{i,t}=R_{i,t}-\big(\hat\alpha_i+\hat\beta_i' F_t\big) \]
Cumulated over a window and averaged across \(N\) firms:
\[ CAAR=\tfrac1N\sum_{i}\sum_{t\in W} AR_{i,t} \]

Our benchmark \(F_t\) is the Fama–French five-factor model plus momentum (FF5+UMD)2 — it strips out market, size, value, profitability, investment, and momentum, so what's left is closer to the news effect and not a style tilt.

Plain English

Small firms, cheap firms, and profitable firms tend to move together for reasons that have nothing to do with corporate law. The factor model subtracts those common patterns so we don't mistake "this is a small-cap day" for "this is the law's effect."

Worked example — the medical baseline

Before saying a drug raised someone's heart rate, a doctor adjusts for age, caffeine, and exercise. FF5+UMD is that adjustment for a stock: subtract the known drivers, then see if anything is left.

The math

\[ R_{i,t}-R_{f,t}=\alpha_i+\beta_{mkt}\,\mathrm{MKT}_t+\beta_{smb}\mathrm{SMB}_t \] \[ +\beta_{hml}\mathrm{HML}_t+\beta_{rmw}\mathrm{RMW}_t+\beta_{cma}\mathrm{CMA}_t+\beta_{umd}\mathrm{UMD}_t+\varepsilon_{i,t} \]

We report the test statistic robustly. Because event-day returns are fat-tailed and firms share a single event date, standard t-tests overstate significance; we report standardized and cross-sectionally adjusted statistics alongside the parametric tests.3

StatisticWhat it guards against
Patell Z (1976)standardizes each firm by its own estimation-window noise
Boehmer–Musumeci–Poulsen (1991)event-induced volatility (variance jumps on news days)
Kolari–Pynnönen (2010)cross-correlation — firms sharing one event date are not independent
Corrado rank / sign (1989)non-normal, outlier-heavy return distributions

Walk through the windows

The interactive panel below shows the estimation window (used to learn each firm's normal behaviour) and the event window (the days the news effect is read), with no event-day data leaking into the baseline.

Figure 7. Event-study walkthrough — estimation and event windows
Source: tristate_results_canonical.json (build v3, as of 2026-06-21) · BUILD_MANUAL §152.149–§152.164. Hover for values; drag/scroll where enabled; ⤢ for fullscreen inside the panel.
3 · The data & the cohort

What we built it on

Returns are CRSP CIZ daily (delisting-adjusted) through 2025-12-31, extended to 2026-06-18 on the reviewer-built canonical panel; fundamentals from WRDS wave-2 (130 parquet files, 1.68 GB); factors from Fama–French.4 Every excluded firm carries a named structural reason — there is no black-box dropout.5 Of the firm universe, are retained (%) after a named-reason exclusion ledger of ; the % figure reported elsewhere is the share of the reincorporation cohort linked to CRSP, a different population.

The treatment variable — and the error that mattered

A controller-protective statute should bite hardest on controlled firms — those with a founder or insider holding a large voting block. The Delaware reform's bright line is ≥ 33⅓% of the voting power.6

Controlled-firm treatment variable. A Delaware firm is coded controlled when a single largest holder holds at least one-third (33⅓%) of the voting power, read from DEF 14A proxy voting-power tables. On this single-holder measure of the Delaware incumbents are controlled. A firm-by-firm rebuild from DEF 14A Item 403 tables is still owed.
4 · Delaware S.B. 21 — no detectable controlled-firm reaction

The within-Delaware controlled-firm result

Two distinct questions sit side by side here, and we report both as co-equal findings.

(a) The aggregate Delaware comparison. Following Khoo & Tallarita's design — all Delaware firms versus all non-Delaware firms, covariate-adjusted — the Delaware indicator carries a statistically significant discount at their headline [−1,+5] window: −0.85 pp (t −2.56, p .011; winsorized 1/99 per day).7 This is a genuine, significant aggregate result, not a null.

(b) The within-Delaware controlled-vs-dispersed test. The separate question this page set out to answer — do controlled Delaware firms move differently from dispersed ones — is null: −0.84% (p .50) once the treatment variable is built correctly (single-largest-holder voting power, not all-insider ownership).

The aggregate discount is not window-robust. At the symmetric [−1,+1] announcement window the Delaware coefficient flips to +0.25 pp (p .37); the discount appears only once the window is widened to absorb a post-event drift over days +2…+5 (window grid below). It is therefore a concentrated post-event drift, not a clean announcement-core reaction.

The controlled-minus-dispersed estimate is −0.84% (p .50) — statistically indistinguishable from zero.7 It remains insignificant under alternative event windows, alternative factor models, and winsorization of the cumulative abnormal returns at the 1st/99th percentiles. No controlled-firm reaction is detectable.

The forest plot: the single-holder controlled-firm estimate (grey, indistinguishable from zero) against the dispersed baseline. Hover any marker for the underlying values. The single-holder ≥33⅓% estimate is built on controlled Delaware firms.

Symmetric-window robustness — the aggregate Delaware coefficient across event windows

Figure 1. Robustness decay — the candidate Delaware coefficient under successive inference layers
Source: tristate_results_canonical.json (build v3, as of 2026-06-21) · BUILD_MANUAL §152.149–§152.164. Interactive: hover for values, drag/scroll where enabled, ⤢ for fullscreen inside the panel.
Figure 2. Cumulative calendar-time portfolio paths — controlled-minus-dispersed and incumbent hedges
Source: tristate_results_canonical.json (build v3, as of 2026-06-21) · BUILD_MANUAL §152.149–§152.164. Interactive: hover for values, drag/scroll where enabled, ⤢ for fullscreen inside the panel.
Figure 3. Buy-and-hold abnormal returns by cohort and horizon (1m / 3m / 6m / 12m)
Source: tristate_results_canonical.json (build v3, as of 2026-06-21) · BUILD_MANUAL §152.149–§152.164. Interactive: hover for values, drag/scroll where enabled, ⤢ for fullscreen inside the panel.
Figure 4. Texas S.B. 29 — sector decomposition of the April-3, 2025 tariff window
Source: tristate_results_canonical.json (build v3, as of 2026-06-21) · BUILD_MANUAL §152.149–§152.164. Interactive: hover for values, drag/scroll where enabled, ⤢ for fullscreen inside the panel.
Figure 5. Controller density by state — the conflict channel
Source: tristate_results_canonical.json (build v3, as of 2026-06-21) · BUILD_MANUAL §152.149–§152.164. Interactive: hover for values, drag/scroll where enabled, ⤢ for fullscreen inside the panel.
Figure 6. Reincorporation flow to result status
Source: tristate_results_canonical.json (build v3, as of 2026-06-21) · BUILD_MANUAL §152.149–§152.164. Interactive: hover for values, drag/scroll where enabled, ⤢ for fullscreen inside the panel.
Figure 7. Event-study walkthrough — estimation and event windows
Source: tristate_results_canonical.json (build v3, as of 2026-06-21) · BUILD_MANUAL §152.149–§152.164. Interactive: hover for values, drag/scroll where enabled, ⤢ for fullscreen inside the panel.

The controlled-minus-dispersed estimate is statistically indistinguishable from zero. It stays insignificant under alternative event windows, under alternative factor models (CAPM, FF3, FF5, FF5+UMD), and when the cumulative abnormal returns are winsorized at the 1st/99th percentiles. The full endogeneity battery is reported below for completeness; none of it produces a detectable controlled-firm effect.10

MethodWhat it correctsResult
OLS baseline (SIC FE)covariate adjustment only
Propensity-score matching (1:3 NN)observable confounding
Entropy / IPW reweightingcovariate imbalance (no exclusion needed)
Heckman two-stepself-selection into "controlled"
Instrumental variables (2SLS)endogenous incorporation choice
Endogeneity, in one line: firms choose their state and ownership jointly with everything that drives returns. The within-state differential mitigates this; the Heckman and IV rows are sensitivity checks only — the Heckman selection equation has no valid exclusion restriction (no point estimate is reported) and the 2SLS instrument is weak (first-stage F well below 10) and inadmissible.

Did it persist over the next year?

If the controlled-minus-dispersed difference (or K-T's −1.4% aggregate) were a real revaluation, controlled firms should have kept underperforming. With the panel extended through 2026-06-18 we held the portfolio over the full ~14–16 months after the event and tested two ways, with daily returns winsorized 1/99.11

Figure 2. Cumulative calendar-time portfolio paths — controlled-minus-dispersed and incumbent hedges
Source: tristate_results_canonical.json (build v3, as of 2026-06-21) · BUILD_MANUAL §152.149–§152.164. Hover for values; drag/scroll where enabled; ⤢ for fullscreen inside the panel.

Plain English

Over the full year-plus, did the controlled firms stay down, recover, or bounce? They reverted — the gap closed. To the data's edge (June 18, 2026, ~16 months) the cumulative gap is , not significant.

Worked example — the medical follow-up

A patient dips after a treatment, then recovers to baseline over the year. The dip was transient, not a permanent decline.

The math

Calendar-time hedge \(\alpha\): \[ R^{hedge}_t=\alpha+\beta'F_t+\varepsilon_t \] Through the full window to 2026-06-18 (334 td): unconditional \( \bar R^{hedge}=+0.005\%/\text{day} \), \(t_{NW}=0.14\), \(p=.89\), cum +0.9%; CAPM \(\hat\alpha=+0.003\%/\text{day}\), \(p=.93\). FF5+UMD \(\hat\alpha\) (\(p=.82\), cum +2.8%) is factor-bounded at 2026-04-30. BHAR = +3.1% (median −17.7%, 95% CI [−21.8, +34.8]). Returns winsorized 1/99.

Delaware bottom line. No detectable controlled-firm effect — at announcement, under alternative windows, factor models and winsorization, and over the full ~14–16 months that followed (the cumulative gap reverts and is not significant).
5 · Texas S.B. 29 — a power-limited null (MDE ≈ 2%)

The one negative window is the April-2 tariff shock, not the law

On the dates S.B. 29 actually moved through the legislature, nothing happened. The only negative window (April 3) sits on the "Liberation Day" tariff crash and is concentrated in nine oil & gas names — remove them and it vanishes.12

Figure 4. Texas S.B. 29 — sector decomposition of the April-3, 2025 tariff window
Source: tristate_results_canonical.json (build v3, as of 2026-06-21) · BUILD_MANUAL §152.149–§152.164. Hover for values; drag/scroll where enabled; ⤢ for fullscreen inside the panel.
Drop the 9 energy firms and the collapses to −0.25%. A separate non-Texas control diff-in-CAAR is −2.10% (p=.093, n.s.) — a distinct diagnostic, not fused.

Plain English — power

A "no detectable reaction" finding only means something if the test could have found a reaction. With Texas firms, the smallest effect we could reliably detect is ~2%. So we can say "no large reaction," not "the law is proven irrelevant."

Worked example — the home test

A negative result is meaningful only if the test is sensitive enough at the stage you are testing. A weak test giving a negative tells you little. The Texas leg is a weak test.

The math — MDE & equivalence

Minimum detectable effect: \[ MDE=(z_{1-\alpha/2}+z_{power})\cdot \sigma_{CAR}/\sqrt N \approx 2\% \] Equivalence (TOST) against ±50 bp does not clear — the band we can support is ~±2%, not ±0.5%.13

Reading the Texas sample sizes. The cohort is TX incumbents; have estimable market-model betas (the announcement-window sample); the long-horizon BHAR cells are horizon-specific ( firms at 3 months, at 12 months, as later horizons drop firms past the data edge); and the “clean industrials” robustness cut is 15 firms (cohort with energy, materials, and financials removed).

Did anything show up over the following year? PROVISIONAL

We held the Texas-incumbent cohort from the S.B. 29 signing forward and asked whether any abnormal return accumulated over 1, 3, 6, and 12 months — a calendar-time portfolio (factor-model alpha) and buy-and-hold abnormal returns vs the market.13

Texas bottom line. No detectable repricing on the legislative dates; the lone window is a sector-specific tariff reaction; and no abnormal return shows up over the following year either. It is a power-limited null (MDE ≈2%), not a bounded "the law is proven irrelevant" null.
6 · Nevada A.B. 239 & the movers

Nevada estimated but underpowered; the movers measured from their move date

Did moving repay the shareholders? PROVISIONAL

For the firms whose reincorporation we have locked to an as-of-event date, we measure buy-and-hold abnormal returns vs the market at 1, 3, 6, and 12 months from each firm's own move date — the move is the event, not the legislative date. The mover track is reported separately from incumbents and is small, so read it as directional.15

Figure 3. Buy-and-hold abnormal returns by cohort and horizon (1m / 3m / 6m / 12m)
Source: tristate_results_canonical.json (build v3, as of 2026-06-21) · BUILD_MANUAL §152.149–§152.164. Hover for values; drag/scroll where enabled; ⤢ for fullscreen inside the panel.

Where the firms went

Destination by firm count: the DExit wave is overwhelmingly to Nevada, then Texas, with a small set going to other states and a counterflow back to Delaware. The flow diagram below maps each destination cohort to its current result status.

Figure 6. Reincorporation flow to result status
Source: tristate_results_canonical.json (build v3, as of 2026-06-21) · BUILD_MANUAL §152.149–§152.164. Hover for values; drag/scroll where enabled; ⤢ for fullscreen inside the panel.
Nevada's MDE is ≈2.7% (σ≈8.5%), so any "null" there is near-vacuous; the leg is reported as underpowered rather than as evidence of no reaction. The mover sample is the as-of-event-locked subset; the full move-effect study still needs the complete EDGAR incorporation lock.
7 · The complete test battery

Every test, its status, and what it found

Thirty-five tests across seven families. Status is reported as executed: nothing is marked run before it was.

8 · Why the states differ

It is the firms, not only the statutes

Setting aside the retracted number, one structural fact stands: on the corrected single-holder ≥33⅓% measure, controller density runs Nevada (≈13.8%) > Delaware (≈2.6%) > Texas (0%). Texas's cohort is dominated by dispersed banks and royalty trusts, with no controlled firm at the bright line; Delaware sits in the middle; Nevada — whose NRS 78.138 extends business-judgment protection to directors and officers and whose NRS 78.240 limits controller fiduciary duties — is the densest.14

Figure 5. Controller density by state — the conflict channel
Source: tristate_results_canonical.json (build v3, as of 2026-06-21) · BUILD_MANUAL §152.149–§152.164. Hover for values; drag/scroll where enabled; ⤢ for fullscreen inside the panel.
Share of incumbents that are controlled. The primary bars use the corrected single-holder ≥33⅓% measure (ranking Nevada > Delaware > Texas); the ghosted reference bars show the older, over-inclusive all-insider measure. The DGCL §144 safe harbor for controlling-stockholder transactions only bites where there are controllers to protect, and where the default rule is the demanding "entire fairness" standard. Texas defaults to the deferential business-judgment rule, and the Texas cohort contains no firm at ≥33⅓% single-holder voting power. For Nevada the corrected single-holder and legacy all-insider bars coincide (both 13 of 94 ≈ 13.8%): Nevada’s controlled firms are its dual-class firms, so the correction that pruned Delaware’s over-inclusive flag leaves Nevada’s count unchanged.
8b · Cross-state summary

Forest plot and governance matrix

The forest plot collects each leg's headline estimate with its status badge; the governance matrix records, per cohort, the legislative event and the per-leg result. Every value binds to the canonical dataset.

Legislative dates and statutes are the canonical record; result colors track the status badges above.
8c · Khoo–Tallarita replication & long-horizon performance

Replicating the published Delaware estimate — and tracking it over a year

Khoo & Tallarita (“KT”) study the Delaware S.B. 21 episode by comparing the top 1,000 Delaware-incorporated firms to the top 1,000 non-Delaware firms around the February 18, 2025 event date.7 We re-ran that design on our as-of-event incorporation panel ( usable Delaware firms, non-Delaware), then carried the two cohorts forward through to test whether any short-window difference persists. This section reports the replication, the size-decomposition that drives it, an identification battery, and the long-horizon path — including the windows where the difference is null.

Reading the counts (they are nested, not contradictory). The Delaware cohort is the top 1,000 firms by 2024 market cap (≈1,000); of these, ≈988–993 have an estimable event-window CAR (those with sufficient estimation-window returns), which is the basis for the per-window CAAR and the buy-and-hold horizons; the cross-sectional regression pools both legs and so runs on n = 1,904 firms (Delaware plus non-Delaware) that also carry the full Compustat control set.

Short-window replication, all five event windows

Cumulative average abnormal returns (CAAR), winsorized 1%/99% per day, against the Khoo–Tallarita Table 5 targets. The Delaware leg tracks their estimates closely; the non-Delaware control leg falls rather than rises, so the published positive spread does not reproduce in our panel.

Khoo–Tallarita's voting-block cut (paper-faithful)

Khoo & Tallarita's own voting-block analysis uses three groups keyed to the S.B. 21 33⅓% bright line, over the [-1,+5] window. We reproduce their published values here for a paper-faithful comparison. Any finer five-bin split (15–20, 20–25, 25–30, 30–35, >35%) shown elsewhere on this page is our finer refinement of K–T's voting-block cut, low-power, and not part of their published result.

Size decomposition (Gelbach)

A sequential Gelbach decomposition of the regression coefficient on the Delaware indicator, window [-1,+5]. The raw coefficient is positive; it turns negative once firm size (log assets) enters, and settles near the published magnitude under the full control set — locating the difference in the size gap between the two cohorts rather than in the law.

FINAL Covariate-adjusted Delaware coefficient, [-1,+5]: (vs Khoo–Tallarita target ).

Identification battery on CAR[-1,+5]

Five estimators recover the Delaware–minus–non-Delaware effect under progressively stronger balance. The pre-match imbalance is large (max standardized mean difference ), driven by size; how aggressively that imbalance is removed determines the answer.

SMD = standardized mean difference (balance target < 0.10). Synthetic-control inference is an Abadie RMSPE-ratio in-space placebo.
MIXED

Long-horizon performance (one year and beyond)

Carrying the two cohorts forward, the chart below traces cumulative calendar-time portfolio paths for the Delaware–minus–non-Delaware hedge; the second chart shows buy-and-hold abnormal returns at fixed horizons. Factor-alpha regressions are capped at the Fama–French factor availability date (); the window beyond is shown as a market-adjusted path only.

Figure 8. Cumulative calendar-time portfolio paths — Delaware-minus-non-Delaware hedge (KT cohort)
Source: stage5_longhorizon_results.json (Stage 5, run 2026-06-21). Equal- and value-weighted self-financing spreads; FF3 and FF5+UMD calendar-time alphas with Newey–West HAC. Interactive: hover for values.
Figure 9. Buy-and-hold abnormal returns by horizon and incorporation group (KT cohort)
Source: stage5_longhorizon_results.json (Stage 5, run 2026-06-21). BHAR vs panel value-weighted market; Lyon–Barber–Tsai skewness-adjusted t. Interactive: hover for values.
NULL
9 · Honest limits & open data

What this does not claim, and what we still need

10 · Methods, sources & references

References

  1. Event-study design: A. Craig MacKinlay, Event Studies in Economics and Finance, 35 J. Econ. Literature 13 (1997); Stephen J. Brown & Jerold B. Warner, Using Daily Stock Returns: The Case of Event Studies, 14 J. Fin. Econ. 3 (1985).
  2. Eugene F. Fama & Kenneth R. French, A Five-Factor Asset Pricing Model, 116 J. Fin. Econ. 1 (2015); Mark M. Carhart, On Persistence in Mutual Fund Performance, 52 J. Fin. 57 (1997) (momentum, UMD). Factors: Kenneth R. French Data Library; our pull ff5_umd_daily_2020_2026.csv.
  3. James M. Patell, Corporate Forecasts of Earnings Per Share and Stock Price Behavior: Empirical Tests, 14 J. Acct. Rsch. 246 (1976); Ekkehart Boehmer, Jim Musumeci & Annette B. Poulsen, Event-Study Methodology Under Conditions of Event-Induced Variance, 30 J. Fin. Econ. 253 (1991); Charles J. Corrado, A Nonparametric Test for Abnormal Security-Price Performance in Event Studies, 23 J. Fin. Econ. 385 (1989); James W. Kolari & Seppo Pynnönen, Event Study Testing with Cross-Sectional Correlation of Abnormal Returns, 23 Rev. Fin. Stud. 3996 (2010).
  4. CRSP CIZ daily security file (delisting-adjusted), through 2025-12-31, extended on the reviewer-built canonical panel to 2026-06-18; WRDS wave-2 (v3plus_wrds_data_2026-06-19, 130 parquet, 1.68 GB). BUILD_MANUAL §152.144, §152.154–155.
  5. CCM-missing exclusion ledger: 5,542 firms, 12 named structural reason codes (private, pre-CRSP, no-equity subsidiary, foreign, SPAC/trust, …). BUILD_MANUAL §152.159.
  6. S.B. 21, 153d Gen. Assemb., Reg. Sess. (Del. 2025), amending Del. Code Ann. tit. 8, § 144; enrolled text at LegiScan. The § 144(e)(2) "controlling stockholder" definition requires both ≥ one-third (33⅓%) of the voting power and the power to exercise managerial authority over the business and affairs of the corporation (a holder below one-third can still qualify via de facto managerial control). Our treatment flag uses the ≥33⅓% single-holder voting prong. Single-cleansing safe harbor (committee or majority-of-minority); squeeze-outs still need both. § 220 books-and-records narrowed.
  7. Kenneth Khoo & Roberto Tallarita, The Price of Delaware Corporate Law Reform (2025), SSRN No. 5318203 (posted June 24, 2025), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5318203 — top-1,000 DE vs top-1,000 non-DE, market model (CRSP-VW, est. −250..−20), event window [−1,+5], event date Feb 18 2025: DE coefficient ≈ −1.4% (controls + SIC3 FE); blockholder >15% −1.6%; dual-class −2.4%. Our replication (full-WRDS run, §8c): following their method on the as-of-event panel, our covariate-adjusted Delaware coefficient over [−1,+5] is −0.75% (t −1.96, p .050, n 1,904) — a size-sensitive partial replication: about half Khoo & Tallarita's −1.4%, reaching the −1.0 to −1.2% range only under balance-credible estimators (entropy balancing −1.21%, AIPW −1.05%). An earlier partial-sample run estimated −1.11% (t −2.5, p .013, n 1,598); it is superseded by the full-WRDS −0.75% reported throughout §8c. This is distinct from the retracted −2.17% (a within-DE controlled-vs-dispersed cut, not K-T's DE-vs-non-DE), and the announcement effect reverts over the following ~16 months (fn 11). BUILD_MANUAL §152.174.
  8. Joseph P. Romano & Michael Wolf, Exact and Approximate Stepdown Methods for Multiple Hypothesis Testing, 100 J. Am. Stat. Ass'n 94 (2005); Yoav Benjamini & Yosef Hochberg, Controlling the False Discovery Rate, 57 J. Royal Stat. Soc'y B 289 (1995). Re-run: 07_REDTEAM/redteam_grid.py.
  9. Winsorization sensitivity: 06_OUTPUTS/winsorization_sensitivity_2026-06-19.md; BUILD_MANUAL §152.166. Convention: 1/99 cross-sectional, disclosed ex ante, robustness reported across levels.
  10. James J. Heckman, Sample Selection Bias as a Specification Error, 47 Econometrica 153 (1979); Jens Hainmueller, Entropy Balancing for Causal Effects, 20 Pol. Analysis 25 (2012); Kai Li & Nagpurnanand R. Prabhala, Self-Selection Models in Corporate Finance, in 1 Handbook of Corporate Finance 37 (B. Espen Eckbo ed., 2007) (on weak exclusion). Re-run: 07_REDTEAM/extrev_b_selection.py; BUILD_MANUAL §152.164.
  11. Long-horizon: calendar-time portfolio (Eugene F. Fama, Market Efficiency, Long-Term Returns, and Behavioral Finance, 49 J. Fin. Econ. 283 (1998); Mark L. Mitchell & Erik Stafford, Managerial Decisions and Long-Term Stock Price Performance, 73 J. Bus. 287 (2000)) + buy-and-hold abnormal returns (Brad M. Barber & John D. Lyon, Detecting Long-Run Abnormal Stock Returns, 43 J. Fin. Econ. 341 (1997)). Re-run on the full reviewer panel daily_returns_panel_2026-06-20_v3.parquet (CRSP-canonical →2025-12-31, Compustat-canonical 2026-01..06-18), keyed on gvkey, returns winsorized 1/99. The FF5+UMD factor-model alpha is bounded at 2026-04-30 by Ken-French daily factor availability; the unconditional long-short hedge and a panel-EW-market CAPM alpha extend to the data edge 2026-06-18 (cum +0.9%, p=.89/.93), as does the BHAR. BUILD_MANUAL §152.171, §152.173 (supersedes the ~10.4-month CRSP-only result, §152.167).
  12. S.B. 29, 89th Leg., Reg. Sess. (Tex. 2025), signed by Gov. Abbott and effective immediately on May 14, 2025 (the September 1, 2025 effective date belongs to companion bills S.B. 1057 / S.B. 2411, not S.B. 29); enrolled text at capitol.texas.gov. Codified at Tex. Bus. Orgs. Code § 21.419 (business-judgment-rule codification) and the paired § 21.418(f) (BJR shielding section), with § 21.552 (3% derivative-standing threshold). E1 = Senate passage 2025-04-03, inside the April-2 "Liberation Day" tariff window (market −5.1%/−5.9% on Apr 3–4). Energy decomposition robust to 300k random 9-firm exclusions (99.9th pct). BUILD_MANUAL §152.149–158.
  13. Minimum detectable effect & TOST equivalence: Donald J. Schuirmann, A Comparison of the Two One-Sided Tests Procedure and the Power Approach for Assessing the Equivalence of Average Bioavailability, 15 J. Pharmacokinetics & Biopharmaceutics 657 (1987); Brown & Warner, supra note 1. TX MDE ≈2% clean / 4–6% clustered. BUILD_MANUAL §152.162, §152.164.
  14. Controller exposure from Capital IQ insider + dual-class screens; BUILD_MANUAL §152.152, §152.157. Primary bars use the corrected single-holder ≥33⅓% measure (DE 52/1999≈2.6%, TX 0/43=0%, NV 13/94≈13.8%; ranking Nevada > Delaware > Texas); the all-insider levels are over-inclusive and shown only as a ghosted reference.
  15. A.B. 239, 83d Leg. (Nev. 2025), amending Nev. Rev. Stat. § 78.138 (director/officer shield) & § 78.240; Gov. Lombardo signing 2025-05-30, effective immediately. Statute at Justia (NRS 78.138). Long-horizon Texas-incumbent & mover tracks (2026-06-20): calendar-time portfolio (FF3/FF5 alpha, Newey-West) + buy-and-hold abnormal returns vs a panel-equal-weighted market benchmark, at 1/3/6/12 months. TX anchor = S.B. 29 signing 2025-05-14; movers anchored at each firm's as-of-event reincorporation date (LOCKED subset, N≈15). Returns and per-firm BHAR winsorized 1/99. Both PROVISIONAL and underpowered. BUILD_MANUAL §152.179. See also Kenneth Khoo & Roberto Tallarita, The Price of Delaware Corporate Law Reform, SSRN No. 5318203 (2025), link.
Figures bind to tristate_results_canonical.json (single source). Provenance: BUILD_MANUAL §152.136–§152.167. SMU Cox · Corporate Governance Initiative · Cox & Dedman Law. Empirical research only — not legal, investment, or tax advice. All interpretations are the author's own and do not represent SMU. Working draft v3 — not for external distribution until the voting-power rebuild + as-of-event lock land.