THE RATE OF RETURN ON R&D IN THE SOUTH AFRICAN SUGAR INDUSTRY, 1925-2001

The rate of return (ROR) on R&D in the South African Sugar Industry is estimated from a Ridge Regression of a production function of time series data for the period 1925 to 2001. The Industry has kept records on R&D expenditure, yields, rainfall and related factors over a 75-year period. Sugar cane yield was measured in tons sucrose to account for quality improvement. In this function, R&D expenditure lagged three years was significant (t = 6.5) in explaining increased sucrose production per ha. Other highly significant variables in this model were rainfall (t = 5.2) and real cost of production (t = 8.4). A dummy interaction with R&D was significant (t = 2.9) implying a greater impact for R&D technology during the period 1959 to 1975 than either before or after this period. The standardised regression model indicated that the R&D variable was one of the most important variables in explaining yield. Using the elasticity of production estimate for the R&D variable of the un-standardised model, a Benefit/Cost ratio for this variable of 1.41 was estimated, if benefit of millers is excluded and 1.59, if the gain to millers is included. In the latter estimates, the exports realisation price of sugar was used as th