Banks at the Chokepoint: A Cross-Country Dataset on Exposure-Conditioned Equity Losses and Recovery across Five Global Disruptions, 2019–2026

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.本数据集为一份可直接用于分析的跨国银行-事件面板数据，旨在探究各国对战略性能源与海运咽喉要道的暴露程度，如何与银行股权损失及冲击后恢复情况相关联。该数据集涵盖五起冲击事件：2019年阿布盖克-胡赖斯袭击事件、2021年苏伊士运河堵塞事件、2022年俄罗斯入侵乌克兰事件、2024年红海局势升级事件，以及2026年霍尔木兹海峡冲击事件。事件暴露程度通过事件发生前的双边贸易份额衡量，包括：源自沙特阿拉伯的HS编码（Harmonized System）27类进口商品（作为欧亚透明航线的代理变量）、源自俄罗斯的HS编码27类进口商品，以及核心或扩展型与霍尔木兹海峡相关的HS编码27类进口商品。 本数据集配套工作簿包含以下内容：事件定义、国家-事件暴露程度测算指标、国家级能源与宏观经济控制变量、银行样本池、银行层面的每日事件研究面板、平衡对照样本与霍尔木兹海峡专项样本、累计异常收益、最大回撤、谷底时点、恢复状态与恢复时长、Kaplan–Meier估计量、事件分层Cox比例风险模型、恢复缺口测算、多重检验校正、重采样推断、敏感性分析、样本流与数据质量诊断，以及完整的源数据与API日志。核心平衡样本包含来自26个国家的95家银行，覆盖全部五起事件，共计475组银行-事件观测值；更广泛的霍尔木兹海峡专项样本则包含来自32个国家的665组银行-事件观测值。 经每日调整的银行股价与基准价格数据来自雅虎财经（Yahoo Finance）；双边贸易数据来自联合国商品贸易统计数据库（UN Comtrade）；可再生能源及相关指标来自《我们的世界在数据》（Our World in Data）；人均GDP、通货膨胀率与贸易开放度数据来自世界银行世界发展指标（World Bank’s World Development Indicators）。暴露程度变量与控制变量仅使用各事件发生前已公开的信息。预期收益率基于交易日窗口[-260, -30]进行估计，事件动态分析窗口设定为[-20, +60]，恢复敏感性评估则延伸至交易日+120。本工作簿还记录了基准市场模型估计结果，以及全球金融拓展模型设定、安慰剂窗口、右删失规则、首次通过与持续恢复的定义。 所有数据转换、模型、表格、图表与诊断文件均可通过配套脚本"R Codes.R"复现。本数据集旨在用于复现研究、稳健性检验，以及围绕银行韧性、地缘政治风险、能源依赖与供应链咽喉要道展开的相关研究。研究结果应被解读为基于暴露程度的关联关系，而非因果效应。