Abstract:

This paper aims to study the economic consequences of cognitive variation among elderly in a family context with two waves China Health and Retirement Longitudinal Study (CHARLS) data in 2011 and 2015. I construct various indicators of household assets and liabilities and two measures of cognition for 3015 married-couple households. Using panel data methods, I find that (1) There is little evidence of correlation between changes in household assets and cognition level of either spouse after ruling out variation between households. (2) To explain the variation of assets within each household, difference of cognition within the spouses matters. Reducing difference of memory scores within spouses by one standard deviation leads to increase in total assets of the family by 8.7 percent, which provides evidence of the gain from positive assortive mating.  (3) Poor health of either spouse leads to decrease in liquid asset by more than one quarter, which suggests that medical expenditure is a great burden for Chinese households.