Deportation Threat and Infant Birth Weight: Evidence from California

Prior research investigates the effects of immigration raids and restrictive state-level immigration laws on health, particularly on the birthweight of infants born to immigrant women (Novak, Geronimus, & Martinez-Cardoso, 2017; Torche & Sirios, 2018). These studies indicate the deleterious effects such “shocks” can have on infant health outcomes. Yet the Secure Communities (SC) program – which automatically screens people arrested by local law enforcement for immigration violations – as well as 287(g) – which is a non-automated interior enforcement precursor to SC – have varied in their implementation by time and place. Thus, this paper exploits variation in the implementation of the SC and 287(g) programs in California to examine the relationship between deportation threat and infant birthweight among immigrant women.

We use 2003, 2005, 2007, 2009, and 2011-2012 adult and child linked surveys from the California Health Interview Survey (CHIS). CHIS data was merged with county and birth year level data on immigration enforcement policy and demographics, which come from various sources including the Census Bureau and Transactional Records Access Clearinghouse (which tracks arrests and deportation removals by county). We use three measures of birthweight: birthweight, low birthweight (<2.5kg), and birthweight below 3kg. For our predictors, we aim to capture two primary channels through which deportation threat occurs: 1) the activation of immigration-related policies; and 2) the intensity of deportation threat. For the first channel, we create two measures – an indicator of the activation of SC and a four-level variable that indicates whether none, either SC or 287(g), or both policies were activated. For the second channel, we create a measure indicating (low-medium-high) level of removal rates due to SC. Using difference-in-difference methods, we test the interaction effect between deportation threat and mother’s citizenship status to disentangle the effects on birth weight among undocumented women.

Across all models, we find significant interaction effects for infants who were born after 1 year of the activation of restrictive immigration policies – i.e., the cohorts who have likely been exposed to the immigration policies in the prenatal period. Specifically, the model that uses the activation of SC shows that infants born to undocumented mothers have a higher probability of being low birth weight by 54 percentage points and have a lower birth weight by 1.28 kg compared to infants born to US-born citizen mothers. The model that uses the activation of 287(g) and SC implementation timing finds that SC is mostly driving the significant effects of deportation threat on birthweight outcomes. In addition, the model that uses the removal rate measure shows that the intensity of removal rate matters; compared to infants born to U.S. citizen mothers, infants born to undocumented mothers show poor birth outcomes to a larger degree when the removal rate is at the medium level compared to low levels.

Our analyses on deportation threat by time and place suggest everyday enforcement practices—in addition to “shocks” from raids and restrictive state legislation—are negatively related to birthweight of infants especially among those born to undocumented mothers.