Notes: Standard errors are reported in parentheses and are clustered by state-immigration status groups. In the “all
persons” sample, the regressions reported in rows 1 and 2 have 784 observations, while the regressions reported in
row3 have 600 observations. In the children sample, the regressions reported in rows 1 and 2 have 377 observations,
while the regressions reported in row 3 have 289 observations. All regressions control for the age, race, gender, and
educational attainment of the household head; the total number of persons, children, elderly persons, and disabled
persons in the household; a vector of dummy variables indicating the household’s year of arrival in the United
States (if immigrant); a dummy variable indicating if the household head is a refugee; the state’s unemployment
rate in the particular survey year; and all interactions between the unemployment rate, the period fixed effect, and
the vector of dummy variables indicating the person’s immigration status (i.e., native, citizen, or non-citizen).
The regressions also include a vector of state fixed effects interacted with the dummy variable indicating if the
observation was drawn from the post-1996 period. The state fixed effects are also interacted with the dummy
variables that indicate the person’s immigration status classification. All regressions are weighted by the sample
size of the cell