Banks at the Chokepoint: A Cross-Country Dataset on Exposure-Conditioned Equity Losses and Recovery across Five Global Disruptions, 2019–2026
Description
This dataset provides an analysis-ready, cross-country bank–event panel designed to examine how national exposure to strategic energy and maritime chokepoints is associated with bank equity losses and post-shock recovery. It covers five disruptions: the 2019 Abqaiq–Khurais attacks, the 2021 Suez Canal blockage, Russia’s 2022 invasion of Ukraine, the 2024 Red Sea escalation, and the 2026 Strait of Hormuz shock. Event exposure is measured using pre-event bilateral trade shares: Saudi-origin HS 27 imports, a transparent Europe–Asia route proxy, Russian-origin HS 27 imports, and core or expanded Hormuz-linked HS 27 imports. The workbook contains event definitions; country–event exposure measures; country-level energy and macroeconomic controls; the bank universe; bank-level and daily event-study panels; balanced comparison and Hormuz-specificity samples; cumulative abnormal returns, maximum drawdowns, trough timing, recovery status and duration; Kaplan–Meier estimates; event-stratified Cox models; recovery-gap estimates; multiple-testing adjustments; resampling inference; sensitivity analyses; sample-flow and data-quality diagnostics; and complete source and API logs. The main balanced sample comprises 95 banks from 26 countries observed across all five events (475 bank–event observations). The broader Hormuz-specificity sample contains 665 bank–event observations from 32 countries. Daily adjusted bank and benchmark prices are from Yahoo Finance. Bilateral trade data are from UN Comtrade; renewable-energy and related indicators are from Our World in Data; and GDP per capita, inflation, and trade openness are from the World Bank’s World Development Indicators. Exposure variables and controls use only information available before each event. Expected returns are estimated over trading days −260 to −30, while event dynamics are reported from −20 to +60 and recovery sensitivity is evaluated through +120. The workbook records baseline market-model estimates alongside global-financial and augmented specifications, placebo windows, right-censoring rules, and first-passage and sustained-recovery definitions. All transformations, models, tables, figures, and diagnostic files are reproducible with the companion script “R Codes.R”. The dataset is intended for replication, robustness analysis, and research on bank resilience, geopolitical risk, energy dependence, and supply-chain chokepoints. Results should be interpreted as exposure-conditioned associations rather than causal effects.