This project uses a panel data set of Swedish households' wealth portfolios and socio-demographic variables to document households' investment behavior. The investment behavior is compared to that predicted with state-of-the-art life-cycle portfolio choice models. The purpose is to improve understanding of the driving forces in households' investment behavior and of the different investment styles in the cross-section, and over the life cycle. The project also contributes to the asset pricing literature.
The project utilizes a unique dataset on households' wealth portfolios. It contains both households' disaggregated financial portfolios, down to the level of individual stocks and mutual funds, and information about real estate ownership.
The first stage of the project consists of an evaluation of different features of life-cycle portfolio models. This stage is a natural continuation of existing research on optimal household portfolio choice. Among the features that have been proposed to improve the predictive ability of the class of life-cycle portfolio choice models with risky labor income are: (i) costs associated with stock-market participation; (ii) preference heterogeneity; (iii) the realization of disastrous economic shocks; and (iv) housing as an investment opportunity in the decision to own or rent.
The data period is 1999-2006. Unlike other datasets with detailed information on households' wealth portfolios, the same households are sampled each year. Due to the panel dimension of the data, research in a second stage of the project is able to document facts about households' financial behavior over their life-time, conditional on the realization of shocks to health status and labor income. Examples of such responses are portfolio rebalancing and self-insurance. For welfare purposes, the ability (propensity) to consume out of real-estate wealth or other forms of wealth is of importance for non-elderly households that suffer major economic shocks and for the elderly. There are broader impacts of the results particularly for investment strategy and for the choice of policies for retirement savings and social security.
The project also contributes to two strands of the asset-pricing literature. In recent research, heterogeneity across agents in combination with incomplete financial markets have been used to explain financial price relations such as the size of the equity premium. Access to high-quality micro data facilitates taking account of such matters and deriving appropriate model calibrations. The panel dimension also facilitates contributions to the literature on the estimation of asset pricing models (Euler equations) using Generalized Method of Moments (GMM). Combined with heterogeneous financial portfolios at the household level, the Euler equations for a given household produce powerful identifying restrictions.
