Based on Cimoli, Porcile and Sossdorf "Crisis, Secular Stagnation and “Creative Destruction”: The Creative Path and Industrial Policy in Latin America", ECLAC, manuscript, October 2014.
Complex, diversified structures, that comprise a large set of capabilities and knowledge-intensive sectors, are less vulnerable (or more resilient) to changes in the rate of growth of the international economy. In other words, the structure “filters” fluctuations in the international economy in such a way that the diversified economy cushions or moderates the impact of shocks in global growth.
This can be tested through a simple econometric model based on a rather conventional growth model, which includes on the right hand side the initial level of GDP per capita (as in models of conditional convergence), human capital, openness, the rate of growth of world GDP and the country’s investment rate. In addition, it is included an interaction term between world growth and an index of knowledge intensity of the production structure. This index of knowledge intensity combines patents per million population plus the share of high-tech exports in total exports.
Table 1
Panel Estimation of the vulnerability of GDP growth to changes in the international rate of growth, 1975-2011
Source: World Bank; COMTRADE;Barro-Lee Database; USTPO
Notes: A panel data and generalized least squares (GLS) estimation method corrected by heteroskedasticity and autocorrelation is estimated. All variables are defined as average for each period, excepting Initial GDP and Education which correspond to the initial year of each period. Remaining variables are defined as follows: Education as completion rates at the secondary level as % of population aged 15 and over; Investment as average of fixed capital investment rate; World GDP Growth; GDP per capita growth; Openness as exports plus imports over total GDP and Structure as an index of patent applications per million population and exports of high technology over total exports. t-statistics with robust standard errors in parenthesis, * Significant at 10%; ** Significant at 5%, *** Significant at 1%.
All the coefficients of the explanatory variables of the model are significant (except for openness) and their signs are broadly in line with what the literature reports. While GDP growth affects domestic growth positively, as expected, the interaction variable is negative. This suggests that while world demand favours growth in all countries, some of them are more vulnerable or dependent on world growth. The variable that makes the difference in terms of vulnerability with respect to global growth is the knowledge-intensity of the production structure.
Notes: A panel data and generalized least squares (GLS) estimation method corrected by heteroskedasticity and autocorrelation is estimated. All variables are defined as average for each period, excepting Initial GDP and Education which correspond to the initial year of each period. Remaining variables are defined as follows: Education as completion rates at the secondary level as % of population aged 15 and over; Investment as average of fixed capital investment rate; World GDP Growth; GDP per capita growth; Openness as exports plus imports over total GDP and Structure as an index of patent applications per million population and exports of high technology over total exports. t-statistics with robust standard errors in parenthesis, * Significant at 10%; ** Significant at 5%, *** Significant at 1%.
All the coefficients of the explanatory variables of the model are significant (except for openness) and their signs are broadly in line with what the literature reports. While GDP growth affects domestic growth positively, as expected, the interaction variable is negative. This suggests that while world demand favours growth in all countries, some of them are more vulnerable or dependent on world growth. The variable that makes the difference in terms of vulnerability with respect to global growth is the knowledge-intensity of the production structure.