Abstract only

In 1999, the total science and technology budget in Argentina reached US$ 1,517 million. This corresponds to 0.54% of the national PBI and represents US$ 50,800 of annual investment per capita for the 20,911 research staff. Agricultural, forestry and fisheries research receives 10.6% of the budget. The focus of the present study is to evaluate the productivity of this sector by analyzing the efficiency of agricultural research scientists in the National Research Council of Science and Technology of Argentina (CONICET). The data used in the study was obtained from CONICET databases and covers 1996-2000. The specific discipline of Agricultural Sciences has 107 researchers representing 3% of a total of 3,685 individuals. There are also a significant number of researchers who belong to other disciplines but with research competences in agricultural subjects and who are not included in this sample. To evaluate efficiency, we use the approach proposed initially by Farrell in 1957. It consists of the estimation of a production function that allows the calculation of the maximum output (y *) that can be obtained by each production unit for an input combination. The level of technical efficiency (TE) of each production unit can be defined as the relationship observed among the actual product (y) and this maximum (y *): O t: ET = y /y* E 1. Therefore, to evaluate TE, it is necessary to know the underlying production function; i.e., the production function that could be estimated from the available data. Then, TE can be calculated from the best practices observed. The proposed model is: yi = f (xi, b) + ei where yi is the output of unit i; xi is the input vector, b is a vector of parameters and ei is a composed random error: ei = vi + ui .vi is a symmetrical error that represents the aleatory variations in production, taken as i.i.d. N(O, s2). The ui is an asymmetric term that represents technical inefficiency and it is assumed that it is distributed independently of vi and that it satisfies the condition ui f O. The stochastic frontiers models used in the study are those proposed by Battese and Coelli (1996). These models allow the use of panel data (a sample of N individuals in T periods), and have different specifications about the composition of uit (technical inefficiency effects). To assess the determinants of CONICET Agricultural Sciences Researchers' efficiency and productivity, research output will be measured in quantity and quality of publications, patents, theses, and advice given. Different input measures will be considered such as project budgets, actual expenses, and wages. Specific variables will be included to assess efficiency effects and to identify scale effects and size economies. These variables are namely: type of research institution (university, public research institute, private institution); geographical location; use of technological and scientific networks; individual characteristics (age, gender); and environmental aspects.

English

0310|R01CIMPU|AGRIS 0301|AL-Economics Program

Juan Carlos Mendieta

CIMMYT Staff Publications Collection