Global CBAM Exposure, Green Compliance Capability, and Latent Trade-Readiness Regimes: Replication Data and Code, 2007–2024

This dataset and replication package supports a global longitudinal study of carbon-border exposure, green-compliance capability, and trade-readiness mobility under the European Union Carbon Border Adjustment Mechanism (CBAM). The construction panel contains 3,566 country-year observations for 201 non-EU exporting economies and territories over 2007–2024. EU-facing HS07 exports are classified into direct CBAM products, near-CBAM value-chain products, and broader green-compliance products. The baseline weighted exposure measure assigns regulatory-proximity weights of 0.60, 0.25, and 0.15 to these three channels, respectively. The package also contains the Green Compliance Capability Core, constructed from renewable-energy, electricity-access, and green-transition-capacity indicators, together with macroeconomic and structural variables used to examine regime mobility. The primary latent-state sample contains 3,010 observations for 178 economies over 2007–2023. The covariate-dependent transition analysis contains 1,831 observations and 1,695 consecutive one-year transitions for 135 economies. The state analysis ends in 2023 because the capability components required for the baseline index are not sufficiently populated for 2024. The replication workflow estimates two- to six-state hidden Markov models, identifies latent carbon-compliance regimes, calculates posterior state probabilities, transition matrices, expected regime durations, and one-, three-, and five-year mobility probabilities. It also evaluates state-dependent transition effects and performs robustness analyses using alternative exposure weights, capability definitions, state structures, and sample exclusions. The deposited materials include processed and analysis-ready data, product-exposure variables, posterior country-year classifications, model-selection results, transition estimates, robustness outputs, a data dictionary, source-provenance documentation, checksums, Python estimation scripts, R verification and figure-generation scripts, and publication-ready monochrome figures. The package supports numerical verification, figure reproduction, and full model re-estimation. The data should be interpreted as evidence on historical exposure-readiness structures and mobility preceding the definitive financial implementation of CBAM in 2026. They do not constitute a causal estimate of realized post-2026 CBAM effects.