Econométrie de la demande d'actifs financiers par les ménages

This thesis is structured in six chapters, and is especially related to the econometric analysis of household chapter one provides some necessary background in microeconomic of uncertainty; it starts with the introduction of the notion of expected utility maximization, risk aversion measures, dynamic programming and the induced intertemporal portfolio problem. It also presents our data set used in this framework. Chapters two and three focus on the effects of individual characteristics on the holding structure. It is found that households allocate their wealth among a very small subset of the assets available on the market. This lack of diversification may be explained by the existence of transaction, holding and monitoring costs, institutional restrictions such as minimum purchase requirement or short sales constraints, borrowing or liquidity constraints. . . Chapters four and five propose econometric specification of portfolio choice under several constraints. In chapter four, the model is based on the assumption that there are fixed costs of owning risky assets it derives demand functions for assets from an arbitrary form of the indirect utility function, which is consistent with any direct utility function. In chapter five, the econometric specification is based on a structural model of mean-variance portfolio choice under non-negativity constraints on the allocation. It leads to a multiregime model, where the associated likelihood functions allow multiple integrals, involving the use of the simulation based econometric methods. It is often suggested that the young investor should purchase more of the risky assets than old investor since these assets are long-term income-gains. The aim of the last chapter is to verify wether this assumption is true in the case of french households. We find that the share of the risky assets in the portfolio increases with the 'age' variable. We conclude giving some explanations to this phenomenon, which is clearly related to precautionary saving motives. The empirical implementations of the modellings in chapters one to five are based on microdata on security choice by french households provided by national institute of statistics and economic studies (insee) in 1992 whereas the last chapter uses data from the 'sofres' survey conducted in 1995