23For a fuller survey, see Carroll, Slacalek, and Tokuoka (2015), which documents that the income process described in section 3.1 fits cross-sectional variance in the data much better than alternative processes which do not include a permanent, or at least a highly persistent, component.