Inlemational Diffusion of Technology: A Case Study of Sugarcane Varieties 55 



investment in experiment stations during the long gestation 

 period before they began producing stage 4 varieties. 



The second computation is the value of stage 4 varietal output. 

 Research cost data for 1950 can be used to compute the average 

 cost of research per percent of varietal turnover. These costs were 

 $2 1 ,000 in South Africa, $25 ,000 in Australia, and $7,000 in India. 

 The estimates in table 3.7 indicate that the varietal change pro- 

 duced by a one-dollar increase in research investment is worth 

 roughly $15 in South Africa, $25 in Australia and $35 in In- 

 dia.^^ These estimates compare current research with current out- 

 put value. The actual research that results in varietal change re- 

 quires a long period of time. In the case of sugarcane research, the 

 average lag is probably eight years or so. Assuming no yield 

 deterioration and an eight-year average lag between investment 

 and benefit realization, the 'MnternaP' rates of return are approxi- 

 mately 40 percent in South Africa, 50 percent in Australia, and 60 

 percent in India. The South African data allow an adjustment for 

 age bias based on a comparison of the table 3.7 estimates and the 

 comparative data by variety. A downward adjustment of 30 per- 

 cent is indicated. This adjustment reduces the internal rates of 

 return only slightly. 



Sugarcane as a Prototypical Case 



The relationship between production and research is examined 

 with more sophisticated models in later chapters. The crude 

 measures utilized here indicate that investment in sugarcane 

 research yielded a very high return. The sugarcane experience is 

 also valuable in another context. The historical sequences 

 revealed by the varietal censuses and the pattern of experiment 

 station success show a pattern that has some prototypical aspects. 



Sugarcane development history reveals a pattern of in- 

 creasingly complex technological discovery through time. In ad- 

 average age would result in an increase in average yields of about 0.2 tons per 

 acre. This is the maximum adjustment that might be made. It would represent 

 about one-third of the Australian yield effect and all of the Caribbean and South 

 African effect. The South African coefficient is approximately what it should be 

 according to the yield comparison data. 



13. These calculations are based on adjusted coefficients from table 3.7 and a 

 sugarcane price of $5.50 per ton. 



