This paper applies cointegration techniques to a model of induced innovation based on the two-stage constant elasticity of substitution (CES) production function. This approach results in direct tests of the inducement hypothesis, which are applied to data for South African commercial agriculture for the period 1947-1991. South African data is used because the policy changes have been substantial enough that the factor and price ratios have turning-points, rather than being monotonic. The time series properties of the variables are checked, cointegration is established, and an error correction model (ECM) constructed, allowing factor substitution to be separated from technological change. Finally, the ECM formulation is subjected to causality tests, which show that both the factor price ratios and R&D and extension expenditures are Granger-prior to the factor-saving biases of technological change. Thus, each stage of the analysis corroborates the inducement hypothesis. However, straightforward price-inducement is only part of the explanation of changes in factor ratios. Policy-induced innovation, in response to tax concessions and subsidised credit, is also present. © 1998 Elsevier Science B.V. All rights reserved.