Fitting parsimonious household-portfolio models to data

US data and new stockholding data from fifteen European countries and China exhibit a common pattern: stockholding shares increase in household income and wealth. Yet, there is a multitude of numbers to match through models. Using a single utility function across households (parsimony), we suggest a strategy for fitting stockholding numbers, while replicating that saving rates increase in wealth, too. The key is introducing subsistence consumption to an Epstein-Zin-Weil utility function, creating endogenous risk-aversion differences across rich and poor. A closed-form solution for the model with insurable labor-income risk serves as calibration guide for numerical simulations with uninsurable labor-income risk.