Finman Identification

Identification Strategy (finman-identification)

When to trigger

• The headline claim rests on a regression of an outcome on an endogenous corporate-policy choice plus controls
• A difference-in-differences exploits a regulation or shock with staggered adoption and uses plain TWFE
• An instrument is invoked but the exclusion restriction and first-stage strength are not defended
• You are asserting a channel or mechanism where the design only supports a correlation

The FM identification bar

FM is an applied-finance journal that prizes practical relevance, but practical relevance does not buy a pass on identification — the editors list rigor alongside relevance among the five criteria. The realistic standard is credible, well-defended causal or economic identification appropriate to corporate-finance data, not the absolute frontier bar of JF/JFE/RFS. What FM referees reward: a clearly named source of exogenous variation, a design matched to it, and an honest statement of what is and is not identified — paired with an economically meaningful magnitude. What they punish: endogeneity hand-waved away with firm fixed effects, a "shock" that is anticipated or confounded, and a mechanism claim the design cannot reach.

Branch paths

Branch A: Corporate-finance event / regulation designs (DiD, event study)

• With staggered adoption, move beyond TWFE — use Callaway–Sant'Anna, Sun–Abraham, or de Chaisemartin–D'Haultfœuille; report a Goodman-Bacon decomposition to show what TWFE was averaging.
• Show a clean event study with pre-period leads flat around zero; FM referees read the pre-trend plot before the table.
• Defend that the shock is unanticipated and not bundled with a confounding policy; address anticipation/reversal.
• Cluster inference at the level of treatment assignment (firm, state, industry); flag few-cluster problems.

Branch B: Endogenous corporate-policy choices (capital structure, payout, governance, M&A)

• Treat the policy as a choice, not an exogenous regressor; name the omitted variable / reverse-causality story explicitly and show how the design breaks it.
• Where an IV is used, defend the exclusion restriction in institutional and economic terms, report first-stage strength, and use weak-IV-robust inference (Anderson–Rubin) when F is modest.
• Matching / entropy balancing must show covariate balance and acknowledge selection on unobservables (Oster-style sensitivity is persuasive and cheap).

Branch C: Asset pricing / return predictability

• Identification here is about separating signal from data-snooping: control for known factors, address multiple testing, and show the result survives reasonable transaction costs and out-of-sample.
• Distinguish a risk explanation from a mispricing one explicitly rather than leaving the channel ambiguous.

Branch D: Mechanism / channel claims

• A correlation between treatment and outcome is not a channel. To claim a mechanism, show the intermediate variable moves and that shutting it down attenuates the effect (mediation, heterogeneity by mechanism intensity).

Execution bridge (StatsPAI / Stata MCP)

Estimate and audit the design, don't only describe it. Full map: execution-with-mcp. Financial Management is empirical corporate finance + asset pricing; corporate-causal chain (DiD/IV/RDD) plus the factor-zoo haircut for cross-sectional pricing.

• detect_design → recommend → fit with as_handle=true → audit_result.
• Observational causal claims: staggered DiD (callaway_santanna / sun_abraham + bacon_decomposition + honest_did_from_result); IV (effective_f_test + anderson_rubin_ci); RDD (rdrobust + mccrary_test).
• Experiments: randomization-based inference + romano_wolf for many-outcome control.
• Sensitivity: oster_delta / sensemakr for observational claims.

Report the magnitude in interpretable units; route the full battery to the appendix. A run end-to-end (synthetic data, real returns) is in the JF execution walkthrough.

Checklist

• The source of exogenous (or quasi-exogenous) variation is named in one sentence
• Staggered designs use a modern estimator; event-study leads shown and flat
• Endogenous policy choices have the OVB/reverse-causality story named and addressed
• IVs defend exclusion in economic/institutional terms; weak-IV inference where F is modest
• Inference clustered at the assignment level; few-cluster issues flagged
• Channel claims backed by intermediate-variable evidence, not just the reduced form
• The economic magnitude is stated and plausible — not just statistical significance