03Projects / Econometrics
Variance decomposition of cross-country crypto adoption: 28 determinants, 144 countries, relative-importance routines implemented in MATLAB.
Cryptocurrencies, blockchain-based digital assets, are rapidly shifting from niche instruments to globally significant financial tools. However, understanding the diverse factors behind their adoption remains challenging. This study explores global cryptocurrency adoption between 2020 and 2024 across 144 countries, analyzing macroeconomic, socio-cultural, infrastructural, financial, institutional, and risk-related determinants. Employing variance-decomposition metrics (LMG and PMVD) to a cross-sectional, linear model’s estimates, it quantifies each factor’s relative contribution to adoption, also introducing a custom Monte-Carlo method for computational efficiency. Results highlight financial determinants, especially bank concentration, as major drivers, while traditional factors like GDP per capita and education show minimal influence. Attained findings underscore cryptocurrency adoption as complex and context-dependent, shaped by relationships rather than isolated factors, thus providing nuanced insights for any stakeholder within this context.
A special thanks to Prof. Vicente Ríos Ibáñez, the best and most impactful professor I have ever had in my entire academic career.
The most interesting parts
A bank distrust index
No dataset measured distrust in banks country by country, so I built one: 13 terms of banking stress in 106 languages, searched on Google country by country, per internet user.
OpenA faster LMG and PMVD
My own Monte Carlo routine in MATLAB: one Cholesky factorisation gives every sequential R² of an ordering, and from 23 predictors up it is cheaper than the exact method.
OpenThe thesis
Loading the document
The preview could not be loaded. Download the file
Want to know more about LMG and PMVD, bank concentration, the distrust index, Cholesky, crypto adoption, relaimpo, the MATLAB code?