Appendix Table 4.

Robustness - IV-Mother Fixed Effects, Full-term Gestation Instrument

	Low Birth Weight (<2500g)	Gestation (weeks)	Any Abnormal Conditions of Newborn (1996–2004)	Any Complications of Labor or Delivery (1996–2004)
A. Excluding Large Cities (Houston and Dallas)
Hurricane within 30km During 1st Trimester	−0.0003 (0.0100)	−0.0427 (0.0692)	0.0360** (0.0136)	0.0200 (0.0193)
Hurricane within 30km During 2nd Trimester	−0.0008 (0.0092)	−0.0355 (0.0685)	0.0180 (0.0143)	0.0250 (0.0194)
Hurricane within 30km During 3rd Trimester	0.0089 (0.0100)	−0.0335 (0.0645)	0.0416** (0.0147)	0.0391** (0.0197)
N	303,896	303,946	187,319	187,319
B. Mothers with No Risk Factors During Any Pregnancy
Hurricane within 30km During 1st Trimester	−0.0087 (0.0100)	−0.0378 (0.0723)	0.0392** (0.0147)	0.0113 (0.0213)
Hurricane within 30km During 2nd Trimester	0.0061 (0.0098)	−0.0863 (0.0721)	0.0104 (0.0157)	0.0205 (0.0207)
Hurricane within 30km During 3rd Trimester	0.0190+ (0.0106)	−0.0432 (0.0680)	0.0377** (0.0155)	0.0411+ (0.0210)
C. Models Interacting Exposure with Presence of a WIC Clinic
Hurricane within 30km During Pregnancy	0.0010 (0.0079)	−0.0572 (0.0555)	0.0252** (0.0117)	0.0310** (0.0154)
WIC Clinic in Zip Code * Hurricane within 30km During Pregnancy	0.0020 (0.0113)	0.0310 (0.0779)	0.0078 (0.0164)	−0.0044 (0.0224)
N	485,048	485,111	301,683	301,683

Notes: Each column is from a separate regression. In panel A, mothers residing in Austin, Dallas, Houston, and San Antonio are omitted. In panel B, mothers with risk factors are omitted. In Panel C the WIC interaction indicates that there was a clinic operating in the zip code at some point during the mother’s pregnancy. See notes under Tables 1 and 4 for information about the sample, the storms and hurricanes, and the estimation methods and controls. For each outcome, births by mothers who have at most one child with non-missing data for that outcome are omitted.

Significance levels:

⁺p<0.10

^**p<0.05

^***p<0.001