This paper aims to explain cross-national differences in central bank digital currency (CBDC) development by distinguishing broad digital-state capacity from observed digital-payment use. It treats CBDCs as public digital infrastructures and asks whether advancement is associated separately with administrative capability and with the salience of digital-payment use.
The study reconstructs a reproducible complete-case dataset of 109 jurisdictions by merging the Bank for International Settlements (BIS) CBDC project database as of 1 March 2024, the World Bank Global Findex 2021 indicators and the United Nations E-Government Development Index (EGDI) 2024. Ordered logit models estimate four CBDC stages. Diagnostics include a Brant-style proportional-odds test, correlation and variance-inflation analysis, a retail-CBDC outcome, an alternative account-reach measure, Firth penalized logit and simulated predicted probabilities.
In the preferred model, EGDI is positively associated with higher CBDC project stages, while digital-payment use remains positive but is marginal at the 5% threshold (p = 0.054); high-income status is negatively associated after conditioning on both digital variables. The proportional-odds restriction is not rejected and preferred-model VIFs remain at or below 3.11. The retail-CBDC ordered-logit model yields positive and significant coefficients for both digital variables, whereas digital-payment use remains positive but is not independently significant in the Firth advanced-stage robustness model. The EGDI-by-payment-use interaction is not significant.
This paper does not claim that cross-national CBDC estimation is new. It extends Auer, Cornelli and Frost by combining their updated public project-stage data with independently published digital-government and payment-use indicators, making all country-level observations and diagnostics visible. Its contribution is a transparent governance-capacity account with explicit nonclaims, temporal boundaries and reproducible country-level measurement.
