Definitions and Measures of ICT Impact on Growth: What is Really at Stake?
Many innovations have been introduced in national accounts in order to better gauge the information and communication technologies (ICT) diffusion impact: new ICT definitions; recognition of business and government software expenditures as fixed investment; hedonic price index. Nevertheless, there still does not exist any clear consensus about the magnitude of the ICT impact on growth. Our aim is to propose some explanations of this relative failure and also show that the debate should not be exclusively centered on quantitative methods. To this end, we take a close look at the two main questions concerning the debate surrounding the measure of the ICT impact: 1) Are there any substantial total factor productivity (TFP) gains generated by ICT diffusion or is it only a classic story of capital deepening increase ? 2) If there are indeed TFP gains, are they limited to ICT producers, as Robert J.Gordon claims, or is there any diffusion to ICT users ? The answer to the first question is really important only if it determines the length and the extent of an eventual growth cycle impulsed by ICT. The possibility that productivity gains mainly due to capital deepening generate strong and durable growth has been theoritically demonstrated by Greenwood and Jovanovic (1998), thanks to a vintage capital model. We precise the conditions under which this result can be obtained and discuss their empirical relevance. According to this approach, the true debate concerns the durability of the present technological shock, instead of its capacity to generate an autonomous technical progress. The answer to the second question is crucial because it could guide industrial policy choices. If TFP gains are limited to ICT producers, should a country always be an ICT producer, or will it anyway grow at a strong pace thanks to the fall of ICT prices ? The relevance of this economic debate is unfortunately poised by the shortcomings of available statistical tools. On one hand, the distinction between ICT users and producers is purely discretionary. On the other hand, TFP measure is completely distorted by the method used to evaluate the value of capital (cost-based prices against adjusted-quality prices). That is why we argue that the international diffusion of growth gains due to ICT essentially depends on the capacity of ICT producers' countries to stay in a rent keeping situation. The text is divided into two parts. The first one first makes a quick assessment of the adaptation of american national accounts to the " new economy ", and then underlines the limits of these changes. The second one shows that the economic debate on the importance of TFP gains acceleration and where they occur, although more complex because of these limits, can quite ignore them thanks to the implications of some endogeneous growth and international trade models.