農業總要素生產力成長因素之跨國研究--以中南美洲國家為例

Abstract

ABSTRACT This study comprehends two phases for the analysis of agricultural Total Factor Productivity (TFP), its sources and determinants in Central America. In the first phase, DEA Malmquist TFP Index measures are calculated using FAO panel data to build a time trend that will help us to assess the behavior of agricultural TFP and its components during 1961 to 2005. The second phase uses econometric analysis to characterize the agricultural TFP time trends built in the first phase, by regressing the Malmquist index and its technical change component with determinants of agricultural productivity. The explanatory variables include fixed capital formation, human capital, international spillovers, irrigation infrastructure, environmental shocks and foreign aid. The efficiency change component of TFP change was not included since most of the results did not report much technical efficiency change for the exception of Honduras. Agricultural productivity studies have been justified by associating agricultural development with poverty reduction since the majority of the poor population inhabit rural areas and rely on agricultural activities as their main source of livelihoods. It is clear that production increases if more inputs are used or if productivity is improved, the higher conversion of forest land to agricultural uses and the increased needs to secure food. This makes an imperative for a sustainable increase of production by means of productivity growth rather than increasing land area. In the first phase we find four distinctive growth patters of TFP in CA. A steep increasing TFP growth pattern with non detrimental technical efficiency change (Belize and Costa Rica), a gradual TFP growth with non detrimental technical efficiency change (El Salvador and Guatemala), a humble TFP growth with detrimental technical efficiency change (Honduras and Panama) and detrimental TFP growth with non detrimental technical efficiency change (Nicaragua). An important aspect to add is that the main engine that drives agricultural TFP growth is technological change, therefore, if a country fails to update its technology compared to the other countries would not exhibit any improvements in TFP as seen in the case of Nicaragua. The possible error caused by multicollinearity and heteroskedasticity during the econometric phase obliged us to modify Luh’s et al (2008) model to accommodate a system equation using panel data. The Seemingly Unrelated Regressions (SUR) finds positive impacts of human capital, fixed capital formation, international spillovers and irrigation infrastructure on TFP growth and technical change. Results from the SUR estimation do not give any support to negative impacts from environmental shocks over agricultural TFP growth. Actual Official Development Assistance (ODA) in current dollars, ODA in capital stock form, ratio of ODA to GDP, and ratio of ODA to agricultural GDP were used to represent different variables of foreign aid in the model. For the four specifications, foreign aid was found to have a negative impact, whereas a positive impact in the quadratic form, although this behavior can be specific in countries that compound a heavy external debt like those in Central America. The interaction term used to test for changes of foreign aid effectiveness before and after 1990 was significant for ODA in current dollars and ODA capital stock form. This is indicative of positive changes after the reforms occurred during the 1990’s also giving insights of the contribution of “good” policies. When studying foreign aid ratios to GDP and agricultural GDP, these two variables was not significant giving evidence that foreign aid dependence has a negative impact on agricultural development that does not depend on “good” policies. When calculating the elasticities of TFP to foreign aid, it was found that when ODA takes the form of capital stock, there is a better chance for foreign aid to have a significant positive impact on the development of agriculture.