1 1. Causal Inference in Financial Event Studies
Citation: Goldsmith-Pinkham and Lyu [2025]
Research question: Standard financial event studies measure stock market reactions to corporate announcements via cumulative abnormal returns (CARS). When does a non-zero CAR identify a causal effect, and when is it confounded by anticipation or endogenous event timing?
Identification strategy: Goldsmith-Pinkham and Lyu [2025] formalise the potential outcomes framework for event studies, clarifying the conditions under which CARs recover causal effects. They show that CAR estimates are contaminated when events are partially anticipated (information was already priced in) or when event timing correlates with pre-announcement returns. They propose tests for exogeneity of event timing and IV corrections using regulatory announcement calendars.
Key result: In applications to earnings announcements, a substantial share of the measured CAR reflects partially anticipated information, biasing event study estimates. IV-corrected estimates are smaller for firms with high analyst coverage, where market learning about earnings is faster.
Takeaway: The credibility standards of the empirical economics literature are now reaching financial economics; even apparently clean event studies require careful examination of whether information is truly new.
2 2. Securitisation, Screening, and Mortgage Default: A Regression Discontinuity
Citation: Keys et al. [2010]
Research question: Did the securitisation of mortgage loans lead to reduced screening effort by originators, thereby increasing defaults? The moral hazard hypothesis predicts that loans sold to securitisers are riskier than retained loans.
Identification strategy: A rule-of-thumb in the 2000s sub-prime mortgage market caused loans to borrowers with FICO scores ≥ 620 to be securitised at substantially higher rates than loans just below 620 creating a sharp regression discontinuity in securitisation rates at the 620 threshold. Since FICO scores are measured in discrete units and mortgage originators knew this threshold, borrowers just above and below 620 are otherwise comparable, but loans just above are more likely to be securitised.
Key result: Loans with FICO scores just above 620 defaulted at a rate approximately 10-25% higher (relative to similar loans just below 620) within 2 years of origination, despite similar borrower observables. The gap is largest for soft-information loans those where originator due diligence would be most valuable consistent with securitisation reducing screening incentives.
Takeaway: The originate-to-distribute model of securitisation causally reduced lender screening effort, contributing to the elevated default rates that triggered the 2008 financial crisis. This is one of the clearest quasi-experimental demonstrations of moral hazard in financial markets.
3 3. Consumer Debt Relief and Labour Market Outcomes
Citation: Dobbie and Song [2015]
Research question: Does receiving Chapter 13 bankruptcy protection which allows households to discharge unsecured debt via a court-supervised repayment plan improve financial and labour market outcomes for distressed households?
Identification strategy: Household bankruptcy cases are quasi-randomly assigned to bankruptcy judges who differ in their propensity to approve Chapter 13 repayment plans. Judge approval propensity serves as an instrument for whether the household successfully completes the bankruptcy process and obtains debt discharge. The first stage is strong and confirmed with pre-determined covariate balance tests.
Key result: Receiving Chapter 13 protection increases employment probability by 3.7 percentage points and annual earnings by approximately $6,000 over a 7-year follow-up, relative to filing households whose plans are denied. The employment effects are driven by households with the most severe financial distress, consistent with debt overhang constraining labour supply.
Takeaway: Debt relief has substantial causal effects on labour market participation and earnings, beyond any income transfer effect suggesting that financial distress acts as a binding constraint on economic productivity.
4 4. Tracking the Credibility Revolution Across Economics and Finance
Citation: Goldsmith-Pinkham et al. [2025]
Research question: How has the credibility revolution in empirical methods the shift toward quasi-experimental designs diffused across fields of economics and finance over the past four decades?
Identification strategy: Text analysis of approximately 44,000 NBER working papers (1982-2025) and articles from eleven top economics and finance journals (2011-2024), using keyword-based classification of research designs (RCT, IV, RD, DiD, synthetic control) to track method adoption over time and across subfields.
Key result: The use of quasi-experimental methods in economics rose from approximately 5% of papers in the mid-1980s to over 30% by the mid-2020s. Adoption has been fastest in labour economics, health economics, and development economics. Finance journals show substantially slower adoption, with approximately 15-20% of empirical corporate finance papers using quasi-experimental methods by 2024. Asset pricing and macrofinance remain predominantly structural or reduced-form without quasi-experimental identification.
Takeaway: The credibility revolution has transformed empirical economics unevenly: certain subfields have been fully reshaped by quasi-experimental norms, while others particularly finance retain methods that would be viewed sceptically in top economics journals. Cross-field contamination (finance papers published in economics journals) appears to be slowly spreading quasi-experimental norms into finance.
References
- Dobbie, W. and Song, J. (2015). Debt relief and debtor outcomes: Measuring the effects of consumer bankruptcy protection. American Economic Review, 105(3):1272-1311.
- Goldsmith-Pinkham, P. and Lyu, T. (2025). Causal inference in financial event studies. Working paper.
- Goldsmith-Pinkham, P., Sokolova, A., and Yagan, D. (2025). Tracking the credibility revolution across fields. NBER Working Paper 35051.
- Keys, B. J., Mukherjee, T., Seru, A., and Vig, V. (2010). Did securitization lead to lax screening? Evidence from subprime loans. Quarterly Journal of Economics, 125(1):307-362.