Table 2

Effect of Food Stamps on Benefit Receipt and Medium-Run Health

ASEC
Panel A: Food Stamp Participation
TI Elig.0.080***
(0.018)
Confidence interval = [0.044, 0.115]
Mean Y0.25
N 5,949
Panel B: Food Stamp Benefit Amount
TI Elig.184.781**
(82.300)
Confidence interval = [23.473, 348.089]
Mean Y731.54
N  5,949
NHIS
Panel C: Overall Health (1 = Excellent,…, 5 = Poor)
Years TI Elig. Ages IU– >4−0.036***
0.012)
Confidence interval = [−0.060, −0.013]
p = 0.00Adj-p = 0.06
Mean Y1.9
N 9,072
Panel D: “Poor,” “Fair,” or “Good” Health
Years TI Elig. Ages IU– >4−0.017***
(0.005)
Confidence interval = [−0.027, −0.007]
p = 0.00Adj-p = 0.05
Mean Y0.29
N 9,072
Panel E: Physical Health Index
Years TI Elig. Ages IU– >4−0.003
(0.016)
Confidence interval = [−0.034, 0.029]
p = 0.83Adj-p = 0.83
Mean Y−0.03
N 3,271
Panel F: Developmental Health Index
Years TI Elig. Ages IU– >4−0.076***
(0.024)
Confidence interval = [−0. 123, −0.029]
p = 0.00Adj-p = 0.06
Mean Y0.04
N 3,635
NHIS
Panel G: Mental Health Index
Years TI Elig. Ages IU– >4−0.013
(0.019)
Confidence interval = [−0.050, 0.024]
p = 0.52Adj-p = 0.72
Mean Y−0.04
N2,808
  • Notes: Data from the 1998–2015 NHIS and the 1995–2007 ASEC. The NHIS sample includes children born in the United States between 1989–2005, observed at ages 6–16, whose parents are treated immigrants and whose mothers have a high school education or less. The ASEC sample includes children born in the United States between 1989–2005, observed at ages zero to four, whose parents are treated immigrants and whose mothers have a high school education or less. Treated immigrants defined as those born outside of the United States and who report arriving in the United States 1985–1996. Regressions using the NHIS include state of birth and year of birth fixed effects, and regressions using the ASEC include state of residence and year of observation fixed effects. The models include controls for demographic characteristics (child’s age, gender, race/ethnicity, and number of siblings, as well as mother’s education, years lived in the United States, and age at childbirth) and controls for the state unemployment rate and Medicaid/SCHIP generosity at the time of birth and the time of the survey (or only time of survey for the ASEC). The results are weighted using the NHIS and CPS-provided weights. Standard errors are clustered by state of birth in the NHIS, and state of residence in the ASEC and shown in parentheses. p-values for each NHIS estimate are shown in the second column, and the third column displays the p-values after adjustment for multiple hypothesis testing, described in more detail in the text. *p < 0.10, **p < 0.05, ***p < 0.01.