Modelling the composition of household portfolios: A latent class approach

Abstract

We explore portfolio allocation in Great Britain by introducing a latent class modelling approach using household panel data based on a nationally representative sample of the population, namely the Wealth and Assets Survey. The latent class aspect of the model splits households into four groups, from lowest-wealth and least-diversified through to highest-wealth and most-diversified, which serves to unveil a more detailed picture of the determinants of portfolio diversification than existing econometric approaches. A pattern of class heterogeneity is revealed that conventional econometric models are unable to identify because the statistical significance and the direction of the effect of some explanatory variables vary across the groups. For example, the effect of labour income on the number of financial assets held influences the level of diversification for the two middle classes, whereas no effect is found for households with the lowest or the highest levels of diversification. Noticeable differences in the magnitude of the effects of pension wealth and occupation are also revealed across the four classes. Such findings demonstrate the importance of accounting for latent heterogeneity when modelling financial behaviour. Ultimately, treating the population as a single homogeneous group may lead to biased parameter estimates, whereby policy based on such models could be inappropriate or erroneous.