An Empirical Inquiry into the Development of Financial Structures and Economic Growth

This thesis evaluates indicators that measure financial intermediaries, capital markets, international flows and money supply for redundancy as well as their empirical relationship with economic growth in 210 countries from 1960-2009. The dimensionality of the data is reduced, retaining indicators that explain maximum variation using Principal Components Analysis. The empirical relationship is investigated using cross-country methodology with extreme bounds for sensitivity check. The thesis also employs Partial Least Squares to serve a dual mandate of reducing dimensionality while explaining economic growth. Principal Components Analysis reveals that private credit, bank deposits, other financial assets, liquid liabilities and non-life insurance explained the maximum variation. Cross-country analysis with extreme bounds supports, in the most restrictive of all models that bank deposits, bank assets, other financial assets, financial system deposits, liquid liabilities, private credit and liquidity were the positive robust indicators. Partial Least Squares conclude that liquid liabilities, private credit, bank deposits, bank assets, public bond markets, other financial assets, life insurance premiums and international debt were considered important predictors of economic growth. Among the other explanatory determinants of growth, population growth rate, investments (as mapped by both investment’s share in real GDP per capita and growth rate of new capital accumulation) and trade openness were estimated to be very important variables. Inflation rate was considered important only with imputed estimates and once again, government’s share in real GDP per capita and taxes were not considered important. Though there are some indicators that perform better than others, no single model or indicator alone was able to explain the growth variations in all countries of varying characteristics suggesting that a tailored country specific approach is optimal for policy design.