Data Provenance and the Tax Institution Growth Nexus in Afghanistan (1996–2024): A Reproducible WDI Official Source Comparison Protocol

Abstract

This methodological article specifies a reproducible comparison of World Development Indicators (WDI) and Afghan official macro-fiscal series in analyses of taxation, institutions and growth during 1996–2024. The author-supplied harmonized archive contains 29 annual observations, while the previously estimated dynamic benchmark is a BIC-selected ARDL(1,0,0,1,0) for 1997–2020 (N = 24). The baseline reports a bounds F statistic of 6.306, a statistically imprecise long-run tax coefficient of 0.9963 (p = .450), and an adjustment parameter of −1.0762 (p < .001). These are archived model results, not outputs from alternative-source regressions. Four proposed configurations (R0–R3) would compare WDI and Afghan official GDP-growth and tax-revenue source packages using a matched annual sample, fixed covariates, and a common lag structure. Their coefficient contrasts have not been executed or verified in the supplied evidence; no data-source-driven sign reversal is claimed. The article provides a measurement crosswalk, provenance-flow diagram, year-level ledger schema, evidence classification and statistical reporting protocol. It argues that source choice should be evaluated as a transparent measurement-sensitivity decision rather than a causal instrument. The protocol establishes a transparent basis for subsequent matched-source sensitivity analysis, conditional on access to the required source records.