Towards a Theory of TFP*

We begin the development of a theory of total factor productivity (TFP) by presenting a multi-sector model of general purpose technology (GPT) driven growth, which generates numerical data via simulation and confirm the theoretical predictions of Carlaw and Lipsey (2003) and Lispey and Carlaw (2004) that TFP is not a contemporaneous measure of technological change. The modeling framework enables comparison between simulated rates of technological change and measures of TFP growth calculated from the simulated data. Theoretical assumptions such as returns to scale and returns to knowledge in the production function and there effect on the TFP calculations can be tested. It is demonstrated that in cases where the generation and adoption of new technology has resource costs that are capitalized in the TFP growth calculation, TFP growth does not measure the introduction of a new GPT. Furthermore, under assumptions that the introduction of new transforming GPTs requires investment in structural adjustment, TFP growth slows while the technology diffusion rate increases and then increases while the diffusion rate declines generating a negative contemporaneous correlation between the two. Our theoretical findings are compared with the growth experience of a number of OECD economies with special focus on Australia’s and New Zealand’s experiences with ICT. TFP growth and ICT diffusion rates are calculated and compared to check the robustness of the modeling framework. Preliminary results yield limited evidence that is consistent with the model’s prediction that as a transforming new GPT such as ICT enters the economy TFP growth will slow and stay low while diffusion rates increase, then TFP growth increases as the technology matures. The results refute the