60 Agriculniral Research and Productivity 



varieties are grown successfully from North Africa to India, while 

 specific maize varieties have to be bred for the various regions of 

 the same countries.'* 



Formulation and Specifications 



To analyze the relationships between agricultural research and 

 yields, we start from a simple summary framework. Yields are a 

 function of soil, climate, and technology. Soil and climate deter- 

 mine yield potential (with a given technology). Weather causes 

 year-to-year variations in yield. Shifting cultivation between 

 different soils also causes yield variations:^ 



y{t)=f[S{t),T{t)] +u{t) (4.1) 



y{t) Yield in year t 



S(t) Soil and climate; since the observations are 



on countries, this variable can be taken as 



"country-specific conditions" 

 T{t) Technology 

 u(t) Random weather effect. 



Technology is the form in which knowledge is revealed in pro- 

 duction. It is a function of indigenously created and borrowed 

 stocks of knowledge K, B, respectively: 



nt) = T[Kit),B(t)] (4.2) 



K(t)= \ p{s)ds (4.3) 



Jo 



B^^'>= )^b{s)ds (4.4) 



4. Myren (1969) attributes much of the relative success and failure of the 

 Rockefeller project in Mexico in wheal and maize, respectively, to these charac- 

 teristics. 



5. For notational convenience, in the general discussion t is treated as a con- 

 tinuous variable. It appears as a subscript in regression equations. Similarly, inte- 

 grals in the general discussion are represented by sums in the regressions. 



