Mental well-being is shaped by various social, environmental, genetic, neurodevelopmental, and psychological elements, which must be considered over an individual's lifetime [1].
In this study, we ask, how does malnutrition in utero affect depression symptoms in adulthood? We specifically focus on the effects of varying early life nutritional conditions caused by the China Great Famine (1959–1961) and use a nationally representative household survey dataset from China to examine the causal relationship.
1.1. The fetal origins hypothesis
The fetal origins hypothesis proposed by Barker suggests that a focus on the prenatal period is driven by evidence that adult health can be shaped during fetal development [2]. Research has shown that malnutrition during pregnancy can affect the epigenome, physiology, metabolism, and physical structure of the developing fetus, potentially leading to differences in chronic illness risk later in life [3]. Moreover, in the fields of psychiatry and psychology, the fetal origins hypothesis is well recognized as emphasizing the importance of early developmental influences as determinants of mental health and human capabilities. To examine the relationship between early-life circumstances and later health outcomes, researchers often utilize natural or quasi-experiments, such as famines, diseases, wars, or catastrophic weather events [4]. Among these events, undernutrition is considered as a natural result of famine [5]. Therefore, famine can be regarded as natural experiments that can provide valuable insight into the fetal origins hypothesis, more specific, how early-life malnutrition affects the adult health in the long run [6, 7].
1.2. Mental Health and Famine
The relationship between adult mental health and exposure to famine in early life is believed to be negative. Studies have shown that in-utero exposure to the Dutch Famine can increase the risk of schizophrenia by twofold [8]. Similar findings were reported in the Great China Famine cohort [9]. Exposure to famine has also been found to be significantly correlated with depression later in life [10–12]. A meta-analysis conducted in 2019 by Dana et al. concluded that those exposed to famine are significantly more likely to suffer from psychological problems in adulthood [13]. Overall, famine exposure in utero has been associated with mental health problems in adulthood. Nevertheless, there’s limited causal evidence of the relationship between in utero malnutrition and later depression symptoms.
1.3. The China Great Famine
The China Great Famine is generally considered to the worst famine in world history due to its high mortality rate and long duration [14]. Specifically, China Great Famine caused between 16.5 and 30 million deaths [14]. Due to the famine, the population cohort between 1959 and 1961 is significantly smaller compared to both preceding and succeeding years (as shown in Fig. 1). This makes the China Great Famine a powerful opportunity to study the effects of malnutrition in utero.
[Figure 1]
1.4. Empirical Challenges of Identifying the Causal Link
Despite numerous studies being conducted on famine, there remain many empirical challenges when considering its causality of long-term effects. The first issue is selective fertility and mortality, where the traits of famine survivors often differ from the general population and are linked to the characteristics of their families during fetal life. This may lead to an underestimation or overestimation of famine's effects if survivors are compared to the control group. To address this bias, Meng and Qian estimated the impact on different quantiles of the outcome’s distribution [15]. To alleviate this bias, Almond et al. added respondent’s mother’s education level as controls [16]. Another challenge is omitted variables. Research suggests that regions with lower agricultural productivity are more susceptible to famine [15], or that regions that experience severe famine are more likely to implement successful recovery policies. In either scenario, the causal relationship is complicated by the influence of omitted variables.
1.5. Gender Specific Effects and Son-preference Culture
Famine-health association is found be sex-specific. Some research suggests that prenatal exposure to famine has particularly negative effects on men's health, such as cardiometabolic conditions and mental health [17, 18]. However, some other studies have shown that women's mental health may be affected even more [11, 19, 20, 12]. And, in female survivors, early undernutrition has a greater impact on abdominal obesity [5]. A systematic review also indicates that famine exposure during pregnancy may increase the risk of certain cardiometabolic conditions in offspring, particularly among women [7].
The gender-specific effects observed could be attributed to mortality selection during famine and son-preference culture [21]. It has been well-documented that during famine, women tend to experience lower rates of excess mortality, referred to as the "female mortality advantage" [21]. According to this hypothesis, the surviving men would have better health outcomes than the famine-unaffected. The reasons behind the vulnerability of males have been discussed in various ways [13]. On the other hand, it could be result of the culture of son preference [22]. Whenever household resources are significantly restricted by the famine, gender bias tends to become prominent and households tend to treat members in an unequal manner. Traditionally, women who are deemed less productive would receive less nutrition and poorer medical care. As a result, surviving male may have better health outcomes compared to their female counterparts [4].
Given the long-standing tradition of son preference in China, this study holds significant relevance to this issue. To date, only a few studies have attempted to address this issue, primarily by examining the sex ratio of second births [23, 24]. Inevitably, this approach is hampered by significant measurement errors, and none of these studies address the issue directly. The impact of son preference in shaping gender heterogeneity remains an area in need of further investigation.
In this paper, firstly, we apply a cohort difference-in-differences (DID) as our baseline estimation strategy to investigate the relationship between malnutrition in utero and adult mental health. Our findings support the hypothesis that exposure to severe malnutrition during fetal development has persistent and long-lasting impacts on adult mental health. To counteract selection bias resulting from selective fertility, we control for the individual's family background and community quality. Our estimated effects are not influenced by this adjustment. To address this issue caused by omitted variables, we utilize two instruments to measure regional famine intensity. After establishing the causal relationship, we examine other potential causal pathways, including spiritual characteristics of the mother, adverse events, and the spousal effect. Our findings indicate that the estimated causal effects remain robust even after controlling for these competing channels. Moreover, we provide a clear illustration of heterogeneous effects across family background and gender, and directly confirm the presence of son-preference culture in explaining gender heterogeneity.