| Abstract: | The paper reports on the statistical aspects of work done on the National Investigation into the Economics of Milk Production in 1965/66. It argues that conclusions draw from the tabulated results of random samples invariably need checking by more rigorous methods. The method of step-wise regression with the introduction of zero-one variables is the principal technique for examining feed input/output relationships and their reflection in economic terms. Five basic dependent variables are considered: yield per cow, stocking density, output per forage acre, utilised starch equivalent per forage acre and margin over feed cost per forage acre. The discussion of the results argues the overwhelming importance of stocking rate in the profitability of milk production and that on many farms concentrates are being used at uneconomic levels. The latter part of the paper considers variation in labour productivity and argues the existence of economies of scale in the use of labour regardless of the system of herd operation. |
