Key points are not available for this paper at this time.
Purpose The purpose of this paper is to uncover the relevant environmental (macroeconomic, technological and demographic) predictors of crypto adoption at the country level through the lens of institutional and entrepreneurship theories. Design/methodology/approach The study utilizes three different empirical methods to examine the determinants of crypto adoption (1) automatic machine learning (AML), (2) SHAP analysis and (3) cluster analysis, using 20 independent variables, cross-sectional data of 137 nations. The different methods allow for the identification of the most important determinants, direction of influence, nonlinear relationships and clustering of the different countries. We also performed some robustness tests using OLS and Random Forest analyses. Findings The AML analysis identified governance stability, informal entrepreneurship and network readiness as the most critical predictors of national cryptocurrency adoption. Partial dependence plot (PDP) and SHAP analyses revealed a strong negative correlation for governance stability, while informal entrepreneurship and the network readiness index exhibit strong positive impacts on adoption. The models further indicate that cryptocurrency diffusion is positively driven by favorable crypto legality, higher inflation rates, mobile phone penetration and moderate unemployment, but is negatively impacted by strong fiscal balances, high Moody's credit ratings and high formal new business density. Finally, cluster analysis corroborates these complex dynamics, demonstrating that emerging economies characterized by lower governance stability achieve higher adoption rankings than highly stable, tech-ready developed economies. Originality/value We checked two leading theories – institutional and entrepreneurship – of crypto adoption and found that the Kirznerian entrepreneurship explanation is better suited. The leading variables show that worse institutions and informal entrepreneurship promote adoption, rather than discourage it. In addition, the cluster analysis reveals different typologies of adoption based on economic development.
Szrajber et al. (Tue,) studied this question.