Présentation de la Maquette Retraites MARS-2003

This paper describes the model MARS, which is the shared property of the DP and the DSS. Currently, MARS aims at planning the general evolution of the first pillar pension scheme for private sector wage-earners, based on macroeconomic and demographic data. It allows to evaluate the impact of changing the parameters used when calculating one’s annuities, or changing general hypotheses. The model provides results which deal with the dynamics of the average annuity and the financial sustainability of the pension scheme. MARS relies on a macroeconomic approach and aims at forecasting the life cycle of average individuals representing their generation. In concrete terms, each generation is represented by two agents (a man and a woman), whose behaviour is supposed to match the average economic and demographic characteristics of their peers (as far as fertility, activity and employment are concerned). In addition, another source of heterogeneity is acknowledged in the model concerning the retirement age, the distribution of which is determined endogenously. Then the evolution of the representative annuity, calculated on the basis of these data, is considered to reflect the dynamics of the average annuity – and at this stage, sustainability indicators for the pension scheme can be inferred. First, the document introduces the demographic and macroeconomic framework of the model, which basically relies on the population and labour force projections of the INSEE, and takes the Conseil d’Orientation des Retraites scenario as a benchmark for the evolution of real wages and unemployment. The making of the retirement decision is a key feature of the model : it is assumed that men will retire when reaching entitlement to the full rate, and that the male and female distributions of retirement ages will match from the 1970 generation onwards. On the basis of these hypotheses, the average pension and the financing requirement of the pension scheme can be evaluated. As an illustration, the results of two variants are displayed: one deals with the impact of the lengthening of the contribution period which results from the August 2003 Act, the other one shows the sensitiveness of the results to alternative demographic scenarios. We conclude about the weaknesses of the model, which is based on the lifecycle of average individuals, and the flaws associated to a macroeconomic approach, versus a dynamic microsimulation one. Besides, MARS should encounter new limits, due to non-linear effects and deeper uncertainty associated to changes in the method used to calculate pensions after the 2003 reform. This paper is written in French.