“…In the experiment, we find random forest performs better than MLP, and that the financial development index [23] is particularly the most important feature for our model, followed by the GDP per capita and the voice and accountability, when explaining the CBDCPI drivers for August 2020. This corresponds [3], which concluded that more developed CBDC projects can be found in countries with higher financial development index, digital infrastructure, GDP, and institutional characteristics. As a robustness check, we have performed the same analysis with full and aggre-gated data and with December 2020 CBDCPI.…”