where 4yijt is the change in an outcome of interest for household i residing in county j at year
t, and 4hijt is the health shock for household i (to be dened in Section 4). The rst-dierence
setup is equivalent to the inclusion of household xed eects in a levels regression; it already
eliminates unobserved household heterogeneity (such as preferences and health endowment) which
may determine the outcome variable and illness at the same time. Xijt includes a set of demographic
variables, including change in log household size, change in the share of children in the household,
7
change in the share of working-age adults in the household, age and age-squared of the household
head, indicators for the education level of the household head, and whether the household head
is single. The inclusion of Xijt controls for changes in the outcome variables related to household
demographics.
j and
t are county and year xed eects, respectively.
j
t are county-by-year
xed eects, which control for the average change in outcomes across all households or the change
in aggregate resources within a county in a given year.