Replication package: Crises Look Like Convergence - Shock Size and Measured Business Cycle Alignment in Europe

Constructed data, estimation code, estimated artefacts and figures for the article "Crises Look Like Convergence - Shock Size and Measured Business Cycle Alignment in Europe". The package contains: the business cycle panels for 26 European Union countries extracted under seven methods (Hodrick-Prescott at three smoothing parameters, Baxter-King, Christiano-Fitzgerald, Hamilton, and an unobserved components model); the merged country-characteristic and energy panels; the full estimation chain in R, one file per stage; every estimated artefact as a CSV, one per reported table or diagnostic; and every figure. Every number reported in the article is produced by a script in 01_code and written to an artefact in 03_artefacts. No number is entered by hand. Every stage that simulates, resamples or estimates a primitive carries halting guards, and each guard is tested in both directions: it must fire on the defect it exists to catch and pass on the repair. Raw inputs are public and are not redistributed here. README.md lists them by dataset code (Eurostat namq_10_gdp, namq_10_a10, sts_inpr_q, prc_hicp_midx, nrg_bal_s, nrg_ind_ei, nrg_ind_id, ext_lt_intratrd; European Commission AMECO; Oxford COVID-19 Government Response Tracker) and gives the run order, the R version and the package requirements. MANIFEST.csv lists every file with its size and SHA-256 checksum. Software: R 4.3.3. Packages beyond base R: sandwich 3.1-0 and lmtest 0.9-40, used only for cluster-robust inference in one stage.本数据包为论文《危机看似趋同——欧洲的冲击规模与可测商业周期协同性》配套提供构建数据集、估计代码、估计生成成果文件与可视化图表。本数据包包含以下内容：基于七种方法提取的26个欧盟成员国商业周期面板数据（分别为设置三种平滑参数的霍德里克-普雷斯科特滤波（Hodrick-Prescott）、巴克斯特-金滤波（Baxter-King）、克里斯蒂亚诺-菲茨杰拉德滤波（Christiano-Fitzgerald）、汉密尔顿滤波（Hamilton）以及不可观测成分模型（unobserved components model））；合并后的国家特征与能源面板数据；完整的R语言估计流程脚本，按流程阶段拆分为单个独立文件；所有估计生成的成果文件均以CSV格式存储，对应论文中每张报告表格或诊断结果；以及所有可视化图表。论文中报告的所有数值均由01_code目录下的脚本生成，并写入03_artefacts目录下的成果文件中，无任何人工手动录入的数值。所有涉及原始数据模拟、重采样或初始估计的流程阶段均配备停机校验机制，且每种机制均完成双向测试：需能够精准触发其预设的缺陷拦截逻辑，并传递修复指令。原始输入数据为公开可得资源，本数据包未包含此类内容。README.md文件以数据集代码清单形式列出所有原始输入数据集（包括欧盟统计局（Eurostat）的namq_10_gdp、namq_10_a10、sts_inpr_q、prc_hicp_midx、nrg_bal_s、nrg_ind_ei、nrg_ind_id、ext_lt_intratrd数据集，欧盟委员会年度宏观经济数据库（European Commission AMECO），以及牛津大学新冠政府应对追踪数据库（Oxford COVID-19 Government Response Tracker）），并说明了运行顺序、所用R语言版本及包依赖要求。MANIFEST.csv文件则列出本数据包内所有文件的大小与SHA-256校验和。软件环境：R 4.3.3。基础R之外的依赖包：sandwich 3.1-0与lmtest 0.9-40，仅在某一流程阶段用于聚类稳健推断。