It is often said that developing countries will be much more vulnerable to climate change than developed countries, that are much more able to adapt to new climate conditions (e.g. Smith et al. (2001)). Economic impact assessments that have been carried out up to now lead to small economic consequences of climate change in developed countries (see Tol (1996), Tol (2002a), Tol (2002b), Mendelsohn and Neumann (1999), Nordhaus (1998) or Frankhauser and Tol (2002)). However, all these authors mentionned some limitations of their assessments, and particularly the fact that indirect impacts are ignored ("for simplicity, we assume that there is no impact of different policies or of climate change on relative prices, so impacts can simply be summed across different sectors", in Nordhaus (1994), p.51).
But developed economies are always based on heavy infrastructures. Among them can be mentioned the energy production and distribution infrastructures ; the transportation infrastructures ; the housing, leisure and working buildings and many others (these economic sectors constitute the social overhead capital of Hirschman (1958)). The very long time scales involved by such sectors make them more vulnerable to climate change than other sectors, in which the productive capital is replaced more frequently and can thus be modified to adapt to climate change without earlier capital withdrawing. Moreover, these vulnerable sectors are often the most directly dependent on environment and climate conditions. Even if these sectors represent only a small fraction of the total economic production in developed countries, most of the other economic and non-economic activities are dependent on the services provided by the vulnerable sectors. In consequence, it is very possible that the greatest part of the climate change economic damages in developed countries would come from propagation effects from the vulnerable sector to the rest of the economy.
This paper aims at assessing the seriousness of these potential effects and at identifying the parameters, which are the most pertinent for this assessment. It is based on a very crude model, it has only two sectors and a lot of major processes are disregarded in order to focus on the propagation effects.
This paper emphasizes a few important results : (i) Accounting for dynamic and economic propagation may dramatically modifiy the climate change damage assessment. (ii) It also changes significantly the temporal repartition of the damages. (iii) The model suggests that some parameters are of the greatest importance in the assessment : investment flexibility (particularly in the infrastructure sector, which may require public support if private investment time horizon is shorter than one century) ; substitution between infrastructure services and productive capital in the rest of the economy ; damage non-linearity that delays and enhances the medium term damages. (iv) Accounting for natural variability may change the damages and the model suggests that the future damages may depend on the particular realization of a random process. Finally, rigorous definitions of the direct and the indirect economic impacts are proposed in order to clarify the problem.
Full paper available soon.