Research and Piodiictivily in Wheat and Maize: An Inlernalional Analysis 71 



mulation replaced the flow by the stock, Kj(t), \mp\y\ng that it is 

 not the work done but the amount of knowledge a country 

 possesses that determines its borrowing ability. The flow formula- 

 tion proved superior. 



4. A ''world stock of knowledge," defined similarly to the 

 definition of the regional stocks was constructed and tried in the 

 regressions. Two hypotheses were tested with this variable: (a) 

 that countries borrow from a world pool over and above the 

 regional borrowing; and (b) that borrowing takes place from the 

 world at large with no regard to regional pattern. These two hy- 

 potheses had to be rejected. 



5. Several ways to introduce depreciation of knowledge or lags 

 in its effect were attempted, but with no improvements in the 

 results. Similarly, experiments in constructing stocks of 

 knowledge by accumulating flow prior to 1948 did not yield better 

 results.^ There can be two explanations for the outcome of these 

 experiments, (1) the "noise" component in the data is too large to 

 permit such fine distinctions; (2) knowledge accumulated before 

 1948, mostly before World War II, was either obsolete or totally 

 disseminated by 1948. 



Economic Implications 



The economic contribution of knowledge varies considerably 

 with the model presumed. First consider the direct contribution 

 of indigenous research in the linear model. The marginal con- 

 tribution of a publication calculated from equation 4.13 is 



^ = ll|^ (4.14) 



dp d ^P 



and, by construction, ^ = 1. Equation 4.14 is calculated in 

 terms of yield per unit area. Total contribution in the country is 



dp a 



9. There are, however, built-in lags in the data— the lag from the completion of 

 a research work to its publication, and the lag from publication to the appearance 

 of the abstract. 



