Institutions, geography, and terms of trade in Latin America : a longitudinal econometric analysis

Recent World Bank reports indicate that the world has seen a severe reduction in poverty, representing not just a cutback in the relative incidence of poverty but a significant decline in the total number of poorʺ people. However in Latin America, the incidence of poverty has remained approximately constant, while the number of poor people has increased. Because of this it is important to understand why Latin America is not making the same strides that the rest of the world is in terms of poverty reduction. It is also necessary to understand how Latin American countries can achieve a long-term economic development to both reduce poverty and improve income inequality. Since long-term economic development is a complex phenomenon, the paper focuses on three deepʺ determinants: geography, integration, and institutions. Traditionally, authors study the impact of institutions, integration and geography on per capita income using worldwide cross-section data. Conversely, this paper employs the Hausman and Taylor (1981) estimator to examine the influence of these three determinants on per capita income in Latin American countries. That longitudinal econometric method allows us to consider unobserved heterogeneity across countries and to obtain direct parameter estimates of the time invariant independent variables, like geography or some institutional measures. Our results demonstrate that not just the quality of institutions, as much of the previous literature has claimed, but also the terms of trade both have strong impacts on per capita income. Once institutions are controlled for, measures of geography have relatively weak direct effects on incomes. Similarly, once institutions are controlled for, we find that both openness to trade and appreciated real exchange rates are detrimental to growth. -- Institutions ; Development ; Openness ; Geography ; Panel Data ; Latin America