Abstract
This paper exploits the properties of the third order approximation of the translog cost function which allows nested tests regarding the nature of technical change and specifically a direct test for price-induced technical change. Using data from 1949 to 1990 the input demand elasticities and factor saving biases were estimated for South African agriculture. Large machinery-using biases of technical change were evident with no labour-using biases. This, together with substitutability between machinery and labour, has-not contributed to alleviating the unemployment problem currently faced in South Africa. The hypothesis of no price induced bias was rejected suggesting that changing pricing policies may be a consideration for alleviating the current technology biases in South African agriculture.