Information, Technology, Embodiment And Growth

The sector of information technologies has been recently invoked to crucially matter in the recent trends and performances of national economies (see for example Gordon, 1999, and Greenwood and Jovanovic, 1999). In this paper, we propose a computable general equilibrium framework in which this sector is explicitly modeled. Indeed, we model this sector to meet the following characteristics. First at all, technological progress is mainly embodied. The new softwares are more and more efficient since they incorporate the most advanced R&D. Additionally, these new softwares can only be run on the most recent vintages of hardwares for technical compatibility requirements. Secondly, the innovators are rewarded by a market power (patents) so as to stimulate innovation and growth. This is a typical specification in research-based growth models (see Romer, 1990, for example). In our model, innovating corresponds to expanding the varieties of available softwares (vertical differentiation). Since the problem of market power and imperfect competition is specially crucial in the information technologies sector (Cf. Microsoft trial), a particular attention will be paid to these issues within our computable general equilibrium set-up. Finally, while the new softwares are patented during a fixed period of time, they become public knowledge at a certain point in time, which generates positive externalities in the rest of the economy. In this sense, the information technology may a be a powerful engine of growth.Here come some details on the constructed model. The model is written in discrete time with an infinite time horizon. Two production sectors and one research sector are explicitly modeled. The final good sector uses a vintage capital technology and employs raw as well as skilled labor. Only the most recent generations of capital (hardware) are compatible with the new varieties of intermediate goods (softwares). The softwares are produced in a second sector, the intermediate goods sector, by the means of a simple linear technology with a single input, unskilled labor. The producers of patented softwares charge a markup while those who produce softwares already in the public domain use marginal-cost pricing. The research sector is designed to expand the varieties of softwares according to a linear technology with a unique input, skilled labor. To model externalities, we assume that the cost of research decreases with the level of public knowledge, which in turn depends on the number of the existing (produced) softwares. Research is financed by households through a perfect financial market. The households block is standard. In particular, skilled and unskilled labor supplies are taken inelastic.After deriving the optimality conditions, we show that the model reduces into a system of eight dynamic (difference) equations. No analytical resolution is allowed even for the steady state equilibrium. So we use the Dynare package to run simulations. These experiments allow us to address several crucial issues. In particular, we are able to study how the dynamics of innovation (through the expanding varieties of softwares) interact with the accumulation of physical capital in the growth process, which helps to assess to which extent growth can be sustained by R&D in the information technology sector. Several other interesting questions are also raised within our computable general equilibrium setting (skill premium, growth-optimizing lifetime of patents, speculative crisis in the stock market...).