Deteriminants of Economic Growth in Uruguay: 1955-2003 A Total Factor Productivity Analysis

During the period 1955-2003 Uruguayan economic growth showed two main characteristics. First, considerable volatility, and second a low average growth rate, even for a developing economy. The objective of this paper is to analyse the determinants of economic growth for the Uruguayan economy in the period 1955-2003, seeking to explain these two stylized facts. For this matter, the approach followed here consists of a two step procedure. First, we study the “proximate” determinants of Uruguayan growth in 1955-2003, using annual data. Based on a standard growth accounting exercise and a potential growth accounting exercise, we are able to conclude that the evolution of Total Factor Productivity is the key to explain both the volatility of Uruguayan economic growth as well as its low average growth rate. These findings are consistent with recent economic literature on the subject. Hence, the second step of this paper is to analyse the factors that determine Total Factor Productivity (TFP), which, based on the above results, are the “ultimate” determinants of Uruguayan economic growth. The analysis is divided into short run and long run TFP determinants. For the short run, an econometric model is estimated, using the cyclical component of TFP (obtained applying a Hodrick-Prescott filter to the original TFP series) as the dependent variable. Exclusion test performed to the model, and the construction of a “variable incidence indicator” allow to conclude that short run TFP (and therefore the better part of Uruguayan economic growth volatility) is mainly explained by financial capital flows and by the level of activity of Argentina. Other variables like the level activity of USA, the Terms of Trade and Real Interest Rates, provide also significant explanatory power. With respect to long run TFP, both international and domestic literature seems to highlight the importance of institutional factors, although insufficient historical data does not permit similar econometric modeling.