This paper investigates the mechanisms behind wealth inequality by extending the standard income fluctuation model to incorporate households’ decisions to acquire information about future economic states. Motivated by empirical evidence from the Survey of Consumer Expectations (SCE), which shows that wealthier households tend to possess better information, leading to more accurate expectations, we develop a model that closely replicates the observed wealth distribution, particularly among the top percentiles. Our model demonstrates that wealthier households, facing lower relative costs of acquiring information, make more informed financial decisions, leading to accelerated wealth accumulation and a concentration of wealth among the top percentiles. Additionally, a counterfactual analysis reveals that reducing the cost of acquiring information significantly shifts the wealth distribution, with full cost elimination promoting a more equitable distribution of wealth. These findings suggest that information acquisition plays a role in driving wealth inequality and that policy interventions aimed at reducing information access costs could help mitigate disparities in wealth distribution.

This paper proposes a numerical solution method for Dynamic Programming problems that uses the technique of quantization. The basic idea is to use the best possible discrete approximations to the continuous variables appearing in the problem. The numerical properties of our method are illustrated in the context of a simple consumption-investment “cake eating” model for which a closed form solution is available.

In this paper, we investigate the effect of expectations on house prices, which have long-lasting impacts on households. Household decisions are strongly influenced by their expectations about the future, making these expectations crucial determinants of house prices. However, estimating the effect of expectations is challenging because they are not directly observable. To address this challenge, we generate indices for expectations using dynamic factor models, which offer several advantages. This approach mitigates the issue of selecting a specific proxy, as the indices are estimated using all available survey-based proxies. It also accounts for measurement errors present in the surveys when estimating the indices for expectations. Using data from the Survey of Consumer Expectations (SCE), we demonstrate that models incorporating these estimated indices outperform models that rely on direct survey-based measures in terms of in-sample fit.