This lecture focus on the effects of pregnancy loss. There is a burgeoning economics literature on shocks and investments during pregnancy and early childhood¹. A common, yet under-examined shock is a pregnancy loss;² an often traumatic ending to one in four recognized pregnancies (Everett, 1997; Meaney et al., 2017; Farren et al., 2020; Quenby et al., 2021). The majority of early pregnancy losses occur during the first trimester and arise from random chromosomal abnormalities that affect the viability of the fetus (Larsen et al., 2013).³ Given the high prevalence rate of miscarriages, the welfare consequences are likely to be large, particularly when compared to less frequent and short-term shocks that are typically examined in the literature. 

Although employed in the economics literature as an exogenous variation to birth timing and maternal outcomes (Hotz, McElroy, and Sanders, 2005; Miller, 2011; Buckles and Munnich, 2012), there is scarce evidence of the effects of a pregnancy loss on parents and on subsequent children. This is the focus of work in progress with Aline Butikofer, Deirdre Coy and Orla Doyle, that uses administrative records on all pregnancies that lasted at least 12 weeks in Norway between 1999 and 2018 to investigate the impact of pregnancy loss on maternal investment and subsequent children’s birth and health outcomes. We link these data to tax, health care services and Social Security registers to study parental health and labor market outcomes.

It is estimated that up to between 10 to 15 percent of pregnancy losses before week 12 are driven by individual risk factors such as previous pregnancy loss, assisted conception, high parental age, low BMI, substance use, and some disorders and chronic diseases (Garcı́a-Enguı́danos et al., 2002; Maconochie et al., 2007), while the remainder are random. This randomness allows us to overcome the identification challenges due to common determinants that influence the likelihood of early miscarriage and health and health related behaviours. Thus, we rely on two strategies. First, we compare the outcomes of families who experience a pregnancy loss to those who did not. Second, to address potential omitted-variable bias, we restrict our sample to families with two children who experience at most one pregnancy loss between the two births and account for fixed maternal characteristics that could be associated with parent and child outcomes and the risk of pregnancy loss (see Currie and Schwandt, 2013). 

Pregnancy loss may impact parent and child outcomes through several channels. It may change maternal investment in the subsequent pregnancy due to fear of recurrent pregnancy loss which may induce mothers to change their level of investment in later pregnancies (see Lee, McKenzie-McHarg, and Horsch, 2017). However, about 20 percent of women who experience a pregnancy loss develop some form of depression and/or anxiety (Nynas et al., 2015), which may reduce maternal investments during subsequent pregnancies. Given the largely random nature of miscarriages, especially early and first losses, increasing prenatal investment is unlikely to prevent a subsequent miscarriage from occurring. 

Increased investment may have a positive impact on birth and health outcomes, which in turn facilitates early skill development with long term impacts across the lifecycle (Cunha and Heckman, 2007; Currie and Almond, 2011). Higher levels of prenatal investment in the form of supplementation use, diet and nutrition, and reduced stress have been shown to impact the developing child.⁴ Our paper contributes to this literature by considering whether pregnancy loss serves as a health shock which changes prenatal investment with consequences for subsequent child outcomes. 

The paper also contributes to studies examining the short and long term impact of shocks induced by natural disasters, terrorist attacks, conflict, and parental death on child outcomes (see Currie and Rossin-Slater, 2013; Quintana-Domeque and Ródenas-Serrano, 2017; Mansour and Rees, 2012). One study which measures stress directly, using cortisol from blood samples, shows that children exposed to higher levels of stress in utero have worse cognitive, health and educational outcomes, without affecting birth outcomes (Aizer, Stroud, and Buka, 2016). Other studies, use the death of a parent during pregnancy as a proxy for stress and grief, find small effects on birthweight and no effects on later on educational attainment, earnings, or health in adulthood (Black, Devereux, and Salvanes, 2016; Persson and Rossin-Slater, 2018). 

Pregnancy loss itself has been used as an exogenous shock to study the effects of changes in the age at first birth (see Hotz, Mullin, and Sanders, 1997; Hotz, McElroy, and Sanders, 2005; Ermisch and Pevalin, 2005) and as instrument for additional spacing between consecutive births (Buckles and Munnich, 2012; Karimi, 2014). One of the few studies to examine the consequences of pregnancy loss itself is Rellstab, Bakx, and Garcia-Gomez (2022) that studies the mental health consequences of miscarriage in the Netherlands. They find that early pregnancy losses increase the use of mental health services in the year the miscarriage took place, but there are no impacts on parent’s labor market outcomes. 

This is work in progress, but our findings so far show that pregnancy loss increases maternal investment in the subsequent pregnancy through increased supplementation, decreased smoking, and increased GP visits. These visits are not driven by worse physical or mental health. We also find that maternal labor market engagement during pregnancy and up to two years after the birth declines. These effects are driven by higher educated women.

Fotnoter:

1. Negative prenatal shocks, such as malnutrition, natural disasters, radiation, and parental death, are often associated with poorer cognitive, behavioral, and educational outcomes, with lasting effects into adulthood (see Almond, Currie, and Duque, 2018research on the fetal origins hypothesis is flourishing and has expanded to include the early childhood (postnatal). ↩︎
2. Pregnancy loss is alternatively labelled miscarriage or spontaneous abortion. Only one percent of miscarriages occur after the first 12 weeks of gestation. ↩︎
3. A chromosome abnormality is a missing, extra, or irregular portion of chromosomal DNA. ↩︎
4. See for example Almond and Mazumder (2011) and Almond, Mazumder, and van Ewijk (2015) for evidence of prenatal shocks to nutritional investment; and Bharadwaj, Johnsen, and Løken, 2014; Hajdu and Hajdu, 2018, for the impacts of the introduction of the workplace smoking ban. Heightened prenatal stress may also directly impact the developing child as it can influence fetal programming (Seckl and Meaney, 2004; Nakamura, Sheps, and Clara Arck, 2008). Stress on the child’s brain negatively affects neurodevelopment, cognitive development, temperament, and psychiatric disorders (see Van den Bergh et al., 2020). ↩︎