This study aims to document cross-country configurations of procurement centralisation and digitalisation, focusing on the unrecorded combination of a central purchasing body (CPB) without national e-procurement.
The analysis combines Global Public Procurement Database indicators with public-finance and macroeconomic data whose source fields span 2018–2023. The estimation sample contains 203 country-year outcome observations representing 79 countries. Country-level OLS and pooled fractional-logit models provide supporting, non-causal expenditure comparisons.
Under the main coding rule, which codes missing CPB entries as zero, three configurations are recorded: No CPB/No eProc, No CPB/eProc and CPB/eProc. No recorded case combines CPB with no e-procurement. Supporting expenditure and governance interaction estimates are imprecise and not statistically significant.
CPB reporting is incomplete, the architecture indicators capture formal presence rather than implementation and the design cannot establish a reform sequence. The missing cell is therefore a recorded-data pattern, not proof that the architecture does not exist.
Countries considering central purchasing should assess whether shared digital infrastructure, interoperable reporting and implementation capacity are already in place or must be developed alongside it.
The study treats configuration support and missingness as substantive features. It reframes an unsupported interaction test as a research agenda on sequencing, reporting capacity, measurement and institutional architecture.
