Demographic Transition and International Flows of Capital: What Can an OLG Model Tell Us?

DEMOGRAPHIC TRANSITION AND INTERNATIONAL FLOWS OF CAPITAL: WHAT CAN AN OLG MODEL TELL US? INGENUE Team (Aglietta CEPII , Breton MODEM , Chateau CEPII , Fayolle OFCE , Juillard CEPREMAP , LeCacheux OFCE , Touze OFCE ) Demographic transition happens and will continue to unfold at different rythms in different region of the world. That means that the concentration of high saving groups will move from region to region and so will the location of the largest groups in age of working which in turn determines the location of most profitable investment. Important flows of capital between regions would result. In order to study this question, we built INGENUE, a world--wide, multiregional, general equilibrium, overlaping generation model. Its general architecture has been presented at last conference. This year, we want to report on first results and methodological difficulties that one encounters in such endeavour. The basic scenario takes the historical situation at the end of the XXth century as initial conditions and simulates households consumption and saving decisions in a transition toward a very far away balanced growth path. The demography is taken as given and is based on UN population forecasts. Technological change and the speed of convergence of productivity between regions is as well taken as given. Many OLG models have been calibrated to national economies. However, at a world level, there is a much greater degree of uncertainty concerning the value of key parameters. This makes only more necessary a detailed sensitivity analysis which we present in the paper. Among others, we present different scenarios concerning the speed of convergence and future evolution of life expectancy. As one can expect in such a model, the institutions of social security (in particular the relative importance of the pay--as--you--go system) play a crucial role in the determination of savings. So different policy experiments are conducted, varying the age of retirement, the amount of social security benefits or the rate of social security tax. In conclusion, we present the main areas needing improvement and outline the main ideas for a second stage in our modeling effort.