In this paper, we test the Rodrik et al (2004) framework to explain
differences in development levels across countries by using a broader
set of definitions for institutions, geography and economic variables.
We use a multi-faceted database to measure institutions in an attempt to
go beyond the single-dimension measures that are often employed. We find
that institutions trump other factors (geography and trade) when we use
GDP per capita as an independent variable. When we expand the dependent
variable to include other aspects of development, such as growth and
investment, we find that institutions, growth and geography are all
important variables. In this case, institutions no longer trump the
other factors. In this case, we also find that the same institutions
variable that was positively associated to GDP per capita is now
negatively correlated with the more dynamic development variable.
Keywords: institutions, geography, openness, governance, economic development
JEL Classification: O1, O16, O17