Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, ECONOMICS AND FINANCE (oxfordre.com/economics). (c) Oxford University Press USA, 2018. All Rights Reserved. Personal use only; commercial use is strictly prohibited (for details see Privacy Policy and Legal Notice).

date: 16 December 2018

Famines, Hunger, and Later-Life Health

Summary and Keywords

Modern-day famines are caused by unusual impediments or interventions in society, effectively imposing severe market restrictions and preventing the free movement of people and goods. Long-run health effects of exposure to famine are commonly studied to obtain insights into the long-run effects of malnutrition at early ages. This line of research has faced major methodological and data challenges. Recent research in various disciplines, such as economics, epidemiology, and demography, has made great progress in dealing with these issues. Malnutrition around birth affects a range of later-life individual outcomes, including health, educational, and economic outcomes.

Keywords: famine, malnutrition, hunger winter, instruments, long-run health effects, developmental origins of health and diseases

Introduction

This article is about exposure to hunger or undernutrition early in life and its effects on health and mortality later in life. There is quite a large body of literature on the long-run effects of exposure to adverse shocks early in life (e.g., the survey by Almond & Currie, 2011). The idea is that adverse shocks during the intrauterine and postneonatal period may affect the development of vital organs and the immune system in such a way that it may lead to increased vulnerability to chronic conditions later in life (Barker, 1992). Besides this direct effect, a shock early in life may trigger adverse outcomes during the life course and therefore may result in worse health at advanced ages. For instance, a poor start may lead to worse childhood health, and this may in turn lead to worse educational outcomes in childhood and subsequently worse labor-market outcomes and health later in life (Ben Shlomo & Kuh, 2002).

Regardless of the underlying mechanisms, assessing causal effects in this area is not straightforward, as the results may be easily confounded by unobserved factors that influence the individual socioeconomic and medical conditions early in life, as well as later-life health outcomes. For instance, genetic factors may simultaneously influence the parents’ income and the health of the infant later in life. To be able to detect causal effects, one needs independent (exogenous) variation in early-life conditions and relate this to outcomes later in life. In the absence of randomized trials, researchers have relied on indicators (instruments) Z at the more aggregate level that affect the individual context early in life (X), but have no direct effect on health and mortality later in life (Y). If one finds an association between the indicators Z and later-life health Y, then this is evidence of a causal effect of X on Y. Indicators Z that satisfy this “exclusion restriction” property do not give rise to endogeneity and simultaneity biases because they are exogenous from the individual’s point of view.

The economics and epidemiological literature has used different types of instrumental variables (IVs). Season of birth was first used by Doblhammer (2004) to study the effect of disease exposure on mortality later in life. Those born in the spring are supposedly exposed in utero to the winter period, a period with elevated disease exposure risks. She finds that the lifespan of those born in the spring is a few months shorter than those born in the fall. Somewhat further from the malnutritional literature, Almond (2006) uses the influenza epidemic of 1918 and finds that the offspring of pregnant mothers who were exposed to influenza have worse health and socioeconomic outcomes later in life. Later studies that exploited exposure to epidemics found similar results. In one example, van den Berg, Lindeboom, and Portrait (2006) use historical data from the 19th-century Netherlands to examine whether cyclical variation in macroeconomic conditions affect individual mortality later in life. Recessions, particularly in the 19th century, may have strong negative effects on household income, the provision of food, and living conditions for pregnant women and children. Indeed, they find negative effects on life expectancy and show that this effect is bigger in more severe recessions.

Important for the present article is that in the context of undernutrition, the literature has often relied on famines to proxy or instrument individual undernutrition. Of relevance are the specific features of the famines and the context in which they took place. We therefore first describe some most studied famines in the next section “Famines Studied in the Literature on Nutrition and Health.” Further, the use of famines imposes other challenges on empirical analyses. In the section entitled “Methodological Issues,” we discuss these challenges and specify conditions for valid inferences in this context. Next, in the section entitled “Findings from the Famine Literature,” we evaluate some of the most well known famines and discuss studies that have used these famines in the literature and discuss avenues for future research.

Famines Studied in the Literature on Nutrition and Health

Normally functioning societies and economies do not generate famines. In the presence of responsibly acting governments and accessible trade opportunities with the outside world, even massive crop failures can be dealt with without famines. As explained by O’Grada (2009) and Sen (1981), modern famines are caused by unusual impediments or interventions in society. These effectively amount to the imposition of severe market restrictions and the prevention of free movement of people and goods. In particular, trade opportunities and food transport are cut off abruptly, and external aid is prohibited or absent. A wide array of exacerbating factors may come into play. For example, nutritional conditions tend to be more vulnerable if food provision in the population is skewed toward some staple food, if a substantial part of the food is imported, or both. Another exacerbating factor is if there is no confidence in the currency, as this leads to hoarding.

Clearly, conditions leading to famines are particularly likely during or in the aftermath of war, especially if the governing institutions are not primarily concerned with the well-being of the population. The textbook example of this is the Dutch “hunger winter” famine lasting from November 1944 to April 1945. Prior to World War II, food standards had been high in the Netherlands, both in terms of caloric value and composition of the diet. There were no notable disruptions in food availability during the first years of the German occupation of the Netherlands, which started in 1940. In September 1944, though, in response to political and military developments, the occupying forces initiated an embargo prohibiting any food transport to the densely populated western part of the country. In combination with the early onset of the harsh winter of 1944–1945, the freezing of waterways, and the generally bad state of the transportation infrastructure, the embargo effectively closed off the western part of the country from any imports of food, fuel, medication, and other vital supplies. During the winter, individuals had to live on rations as low as 500 kcal per day. This situation lasted until the end of the occupation, which coincided with the end of World War II in early May 1945. Immediately, rations rose to 2,400 kcal per day.

The increased death rate in the first half of 1945 over the rate in 1944 amounts to 269% for men and 173% for women (Dols & Van Arcken, 1946). Banning (1946) reports a higher incidence of tuberculosis and hunger edema and an increased infant mortality rate. Inhabitants of large cities were struck hardest by the famine. However, Banning (1946) notes that in small towns, the mortality rate rose to almost as high a level as those in large cities. Banning (1946) mentions that of the potential candidate children examined, 29% had been severely undernourished, while 31% suffered moderate undernourishment. About 27% of the children lost about 10% of their weight.

Similar famines occurred during World War II in Greece (in 1941–1942) and Leningrad (in 1942), as well as right after World War II in Germany. Like the Dutch 1944–1945 famine, the Leningrad famine and the Greek famine were a direct result of food blockades by one of the major warring parties (Hionidou, 2006). The German famine in 1945–1948 primarily happened because (a) the agricultural land in the east, which had been a major provider of food, had been occupied by Poland and the Soviet Union; and (b) Germany was divided into four occupation zones that were administered separately, with highly limited interzone trade because of political and bureaucratic barriers and lack of transportation.

This famine provides a good illustration of the fact that a multitude of factors may influence the severity of a famine. This included the influx of millions of migrants from areas that were occupied by the victorious countries and the extent to which the victors imprisoned prime-aged men and shipped off industrial plants as reparation payments. Also, the destruction of infrastructure and the extent to which victors provided emergency food relief affected the severity of the famine. All this creates substantial variation across regions and over time with food shortages (see, e.g., Klatt, 1950; Farquharson, 1985; Trittel, 1990; Häusser & Maugg, 2009; and Reichardt & Zierenberg, 2009, for overviews).

The German famine toward the end of World War I, in 1916–1919, also fits this mold (see, e.g., Vincent, 1985). At the time, Germany was not occupied. However, by mid-1916, the Allied powers had successfully enacted a complete naval blockade of Germany that restricted the maritime supply of raw materials and foodstuffs. This was exacerbated by a crop failure in 1916. The blockage continued after the armistice to induce Germany to sign an unfavorable peace treaty.

In 1943, the Bengal province in India was close to the front line of World War II. The governmental regime was colonial and gave a top priority to the provision of food to its army. This, in combination with various natural disasters, led to a famine in 1943, which in turn went along with widespread epidemics (Sen, 1981).

The Irish famine of 1845–1849 did not occur in the context of a war. Instead, it was triggered by a potato blight that destroyed successive potato harvests. However, a crucial cause of the famine was the unwillingness of the British government to intervene on behalf of the population of Ireland which at the time was part of the United Kingdom. Around the time of the Irish famine, there was a similar famine in regions of the Netherlands that were particularly dependent on potato as staple food. The reason that the potato blight led to that famine was the absence of immediate government intervention. However, later interventions proved helpful, and as a result, the famine was less devastating than the famine in Ireland (Lindeboom, Portrait, & van den Berg, 2010).

More recently, the health effects of the Chinese famine of 1959–1961 and the Ukraine famine of 1932–1933 have been studied in detail as well. These two famines did not occur in the context of war either; rather, they originated by ruthless governmental attempts to societal engineering. In both famines, governments tried to enforce a major restructuring of the economy at the expense of food provision (see Vaiserman & Lumey, 2013, for the Ukraine famine, and references in the next section, “Methodological Issues,” for the Chinese famine).

Finland experienced a number of crop failures leading to a famine in 1866–1868. Here, the backwardness of the agricultural sector, together with a lack of adequate policy response from the authorities, were primarily responsible for turning the crop failures into a famine (O’Grada, 2001). As argued by Doblhammer, van den Berg and Lumey (2013), it is interesting to examine why the neighboring country Sweden did not experience a famine in these years, apart from the fact that the Finnish harvest failures were more severe. As it turns out, the Finnish relief funds were inadequate, the severity of the situation was underestimated at the governmental level, and policy actions were delayed until 1868.

This listing of famines leads to a number of insights that are important from a methodological point of view. First, it is clear that since famines occur in periods of severe societal disruption, opportunities for data collection during famines are rare. One of the reasons why the Dutch famine has been widely used in the literature is that it was so short that the quality of the population registers for the exposed cohorts was not heavily affected. In many other cases, research has to rely on retrospective surveys among individuals interviewed many decades after the famine. Another insight is that the extent of hunger experienced during famine years is heterogeneous across time and space. A third one is that famines never come alone. The market failures that cause a famine often also affect life in different ways, such as by restricting education or disrupting family life. The famines themselves make a population more susceptible to epidemics. A fourth insight is that famines may affect the composition of newborns (see also Dyson, 1991). As already mentioned in the Introduction, this is important when comparing the fate of exposed and unexposed cohorts.

Methodological Issues

One rarely observes whether an individual has actually been exposed to undernutrition in utero or early in life, and therefore the literature often uses exposure to a famine around the time of birth (see Lumey, Stein, & Susser, 2011, for a review of the literature). The idea is that during such periods, the availability and quality of food for households are substandard, so a comparison of the later-life outcomes of a famine-exposed individual with those of a nonexposed individual provides evidence of a causal effect of undernutrition on later-life health. In the context of a simple linear model, this often boils down to a regression of later-life health (Y) on famine indicators at or around birth (Z) and, if necessary, a set of additional covariates (A):

Y=α+βZ+γA+ε
(1)

Note that in this context, the vector A needs to be restricted to variables measured at the time of the famine exposure, as the treatment may also influence other contemporaneous variables and thus bias estimates of β. The set of indicators is usually defined by geographical indicators, time dummies, or both. Equation (1) can also be seen as the reduced form of the “second-stage” relation between Y and an indicator for undernutrition (X) and the “first-stage” relation X and the instrument Z. The parameter estimate of β in Equation (1) can thus be interpreted as an Intention To Treat (ITT) effect. The compliance rate [i.e., the coefficient of the regression of X on Z (say, δ)] is needed to obtain the IV estimate, the Local Average Treatment Effect (LATE). More specifically, βTV=β/δ. This means that the lower the compliance rate, the lower the “power” of the first-stage regression, the further the ITT is from the LATE effect.

In the previous section, we observed that famines often display heterogeneity in exposure and that data collected in times of famine may be of limited quality. Because of this, it is not always possible to unambiguously assign individuals to a group that is exposed to famine and a group that is not. In such situations, this could at best lead to low compliance rates, and therefore small ITT effects. In particular, classification problems may arise when the famine is merely the most extreme part of a longer period of food insecurity. In such cases, the start and end of the famine are not precisely determined. If the fact that some of the controls actually experienced bad nutritional circumstances is ignored, then this will tend to cause an underestimate of the effects. In such cases, food price information can be used to construct a refined measure of exposure to undernutrition (see e.g., Bengtsson & Lindström, 2000; van den Berg et al., 2017).

In the previous section, we also observed that famines often accompany other types of societal disruption. If this point is ignored, this affects the interpretation of the famine as an effect through undernutrition. For instance, in times of war, healthcare systems may function suboptimally because the demands on them are higher than usual or because the supply of resources (staff, medicine) may be lower than usual. The disease environment and the susceptibility to infectious diseases may also change during a period of starvation. It is therefore important to take these other types of hardship into account and, if possible, to find other instruments for them.

Another observation from the previous section concerns the fact that the composition of newborns may be different during a famine. This creates econometric problems of selectivity. In a wider sense, this includes stillbirths and abortions, which also affect the composition of newborns. Usually, it is argued that ignoring these types of selection leads to an underestimation of the effect of the famine on later-life outcomes. Occasionally, studies have access to external data that enable an assessment of selection on observable characteristics or compositional markers. For instance, the composition in terms of parental socioeconomic status or the sex ratio among newborns may be observed. It is suggested that if there is no selectivity in terms of such factors, it also less likely that there is compositional selectivity in terms of unobserved confounders.

A different way to deal with selective fertility is by identifying groups of individuals who share the same value of the unobserved component that drives the selection mechanism. The idea is that outcomes depend on this unobserved component, which also influences the probability of childbirth in a famine. Under certain assumptions, the differences in outcomes for two individuals known to share the same unobserved component is informative on the long-run famine effect if one of the individuals is exposed and the other is not. This effectively involves a fixed-effect panel data analysis. In practice, one may assume that two individuals born in the same region or city share the same unobserved component, or, indeed, that classmates or siblings share the same unobserved component. Whether such assumptions are credible depends on the context.

We now return to the key issue that has not been addressed in most of the literature, which we discussed at the start of this section—namely, that exposure to a famine is not equivalent to exposure to a nutritional shortage. During a famine, a fraction of all households does not face food shortages, for example because a household belongs to the ruling or wealthy class, or because it is self-sufficient in terms of food, or, in a war context, because it is allied with those responsible for the cause of the famine. Similarly, in nonfamine eras, a fraction of households faces food shortages because of poverty. This means that the comparison of famine-born individuals to nonfamine-born individuals does not provide a quantitative estimate of the average causal effect of nutritional shortages around birth. Most likely, the latter effect is underestimated in absolute value by such a comparison.

To advance on this, it is necessary to observe the nutritional status in the households at the time during which the child is in utero or at the childhood age of interest. However, as we have seen, during famines, data are typically not collected, as societies are in a state of disruption.

Van den Berg, Pinger, and Schoch (2015) developed a method to deal with this, which was based on self-reported, retrospective data on the individual occurrence of a period of severe hunger at certain childhood ages. These can be provided by exposed and nonexposed individuals. When, as is usually the case, the age at exposure is too low to have a recollection of hunger, the retrospective data can also be provided by older individuals who lived through the same period. Two-sample IV estimation techniques can then be used to estimate average causal effects of nutritional shortages during certain childhood ages on health outcomes later in life. In terms of the IV treatment evaluation literature, the IV is the exposure to a famine early in life, and the treatment is the experience of nutritional shortage early in life. With heterogeneous effects, the IV estimation provides LATE estimates. This procedure has been applied by van den Berg et al. (2015) to data on the Dutch, Greek, and German famines related to World War II, and by Deng & Lindeboom (2017) to the Chinese famine. Often, the estimates of the effect of famine on later-life outcomes have to be multiplied by a factor of 3 to 5 to obtain the causal effect of hunger or malnutrition on those outcomes.

An additional complication in empirical famine studies concerns the way that researchers observe their data. Usually, the observation scheme is such that individuals are sampled decades after the occurrence of the famine, and that it is inferred whether these individuals have been exposed to a famine. This conditioning on the sample of survivors decades after exposure to the famine imposes another selectivity problem. Usually, it is assumed that this leads to a downward bias of the estimated famine effect, as it is expected that such mortality will primarily affect frailer individuals. However, such mortality selectivity effects may be very strong, even masking long-run effects.

For example, in one study, van Ewijk and Lindeboom (2017) find that mortality (and fertility) selection can be substantial under moderate warlike situations. Observation schemes that allow for continuous follow-up are desirable in such cases because they allow one to explicitly model mortality throughout life and to estimate age-specific effects of famine exposure on later-life health. In recent work that examines the 1946–1947 Dutch potato famine, van den Berg et al. (2017) use birth and death certificates to construct continuous lifetimes and find substantially elevated neonatal (0–1 years) and infant (1–5 years) mortality rates. This implies that most of the selection effects take place at early ages, and that this impact can be substantial.

Findings From the Famine Literature

Health and Mortality

The literature on the long-run effects of famine exposure generally agrees that famine exposure affects weight, diabetes, hypertension, and mental health, while some studies also find effects on height, cognition, cardiovascular disease, and mortality. Lumey, Stein, and Susser (2011) provide an excellent review of observational studies based on the Dutch, Chinese, and Leningrad famines. Their review shows a persistent association across the three famines between famine exposure early in life and weight, diabetes, and schizophrenia.

The Dutch famine is the most studied famine in the epidemiological literature, and various effects have been found on adult health measures, including obesity (Lumey et al., 2011), epigenetic changes (Heijmans et al., 2008), glucose problems and diabetes (Ravelli, Meulen, Michels, Osmonds, & Barker, 1998; Roseboom, de Rooij, & Painter, 2006), and abnormal blood pressure and hypertension (Stein et al., 2007). The evidence on cognition is mixed. Ampaabeng and Tan (2013) use the 1983 Ghanaian famine to look at the long-run effects of famine exposure in early childhood (ages 0–2 years) on cognitive development. They find a direct and negative effect on cognitive performance, as measured by IQ score. These effects in turn affect performance on reading and math tests in adolescence and adulthood. St. Clair et al. (2005) find a substantially elevated risk of schizophrenia among survivors of the Chinese famine. Huang et al. (2010), also looking at the Chinese famine, find that exposed women have an increased risk of mental illness, but for men, they do not find such effects. They argue that this may be due to a stronger natural selection in utero of male versus female fetuses during a severe famine period. Li and Lumey (2017) provide a systematic review of studies of the Chinese famine. They confirm the results for schizophrenia (as well as effects on weight, type 2 diabetes, hyperglycemia, and metabolic syndrome).

Some famine exposure studies specifically focus on adult height (as a proxy for health) as an outcome measure. Oppers (1963) finds a negative effect of exposure to the Dutch hunger winter famine at ages 7–14 on adult height among men. Koupil et al. (2007) do not find an effect of the Leningrad famine on height, but this may be because the famine was so prolonged and severe that it generated a strong dynamic selection in height outcomes. This most likely also explains the absence of significant effects of the Chinese famine on height by Gorgens, Meng, and Vaithianathan (2012).

On the other hand, Chen and Zhou (2007) find effects on height for infants (younger than 6 years) exposed to the Chinese famine. A recent study by van den Berg et al. (2015) uses two sample IV methods on data from the Dutch, German, and Greek famine to estimate LATE. The literature generally uses reduced-form regressions and thus estimates ITT effects (see also the discussion in the section entitled “Methodological Issues”). Their LATE effects are about four times larger than ITT effects usually found in the literature. More specifically, undernutrition in the age interval from in utero to age 4 results in about a 3-cm reduction in adult height in males. Deng and Lindeboom (2017) also examine hunger recall information of the Chinese famine and find large LATE effects on hypertension and a score for metabolic problems (based on obesity, high blood pressure, diabetes, and heart disease).

The findings on mortality later in life are mixed. Some studies did not find differences in mortality and survival at older age for cohorts born during famines. The pioneering national cohort study of the Great Finnish Famine (1866–1868) by Kannisto, Christensen, and Vaupel (1997) did not find any effects. Recall from the previous section, however, that a study allowing for unobserved heterogeneity saw the opposite result. Painter et al. (2005) examine those born during the Dutch Hunger Winter over the subsequent 57 years and did not find effects on mortality. A regional study of the Chinese famine (Song, 2009) did not find mortality effects either. However, Lindeboom, Portrait & van den Berg (2010) use historical data about the Dutch potato famine of 1846–1847 and find strong evidence for long-run effects of exposure to the famine. These results are stronger for boys than for girls. Boys and girls lose on average 4 and 2.5 years of life, respectively, after age 50 after exposure at birth to the potato famine. Lower social classes appear to be more affected by early-life exposure to the potato famine than higher social classes. Unlike the previous studies, they were able to follow individuals from the moment of birth until mortality, which enabled them to assess the effects of the famine at various stages of the life cycle and correct for selective mortality.

Economic Outcomes

Work on the effect of famine exposure on economic outcomes is relatively understudied. Most of the work on the long-run effects of famine exposure early in life on later-life socioeconomic outcomes comes from the economics literature. Chen and Zhou (2007), Meng and Qian (2009), and Almond, Edlund, Li, and Zhang (2010) use the Chinese famine to look at the effects of famine exposure on a range of educational and labor-market outcomes at prime ages (up to age 45). Chen and Zhou (2007) find that exposure at early ages leads to reduced labor supply and lower earnings. In addition, large effects on educational attainment are found by Meng and Qian (2009). Almond et al. (2010) look at the 2000 U.S. census and find that higher famine intensity, as measured by time and region-specific mortality rates, is associated with greater risk of illiteracy and of being out of the labor force. In addition, they find that men who were exposed marry at later ages, and women that were exposed marry less-educated men. Interestingly, they also find that famine exposure lowers the boy-to-girl sex ratio.

Neelsen and Stratmann (2011) look at the effect of exposure to the Greek famine (1941–1942) in utero, during infancy, and at age 1 and find reductions in the number of years of education and the probability of completing upper-secondary schooling in the exposed. The effects are largest for cohorts exposed in utero and infancy. Finally, Scholte et al. (2015) use the Dutch hunger winter famine to examine the impact of exposure to the famine at different gestational periods. They find a significant effect of exposure during the first trimester of gestation on employment outcomes about 55 years after birth. This is in line with the negative effects on cognition and mental health of exposure revealed in other studies of the Dutch hunger winter (as discussed earlier). Exposure in the second and third trimester of gestation leads to higher hospitalization rates. Scholte et al. (2015) also look at marital outcomes and find that males exposed in the second trimester have a disadvantage in the marriage market, as revealed by their lower age at marriage and noncustomary age pairing.

Conclusions

Throughout human evolution, famines and periods with severe hunger were fairly common. As such, the adaptive response to famine exposure in early childhood has contributed to the formation of human biological identity. However, from a modern societal perspective, famines are rather rare. They occur only in the presence of severe impediments or interventions in society, notably in war-related contexts.

Long-run effects of exposure to famines are interesting in and by themselves. However, such effects are more commonly studied to obtain insights into the long-run effects of malnutrition at early ages. This research agenda has been facing major methodological challenges. To single out effects of malnutrition at specific stages early in life, the exposure period should be short, well-defined, and unanticipated. In reality, the extent of malnutrition is heterogeneous across time and space, and this often goes along with other disruptions in daily life. In addition, famines may affect the composition of newborns, complicating a comparison of long-run outcomes between exposed and unexposed cohorts. We have shown that the literature in various disciplines, such as economics, epidemiology, and demography, has made great progress along these lines. By now, it is well established that malnutrition around birth affects a large range of late-life health outcomes. There is mounting concurrent evidence for effects on educational outcomes and on economic performance before retirement ages.

We expect this literature to expand and to further broaden its scope. It is interesting to study whether there are interventions between the first year of life and the ages at which outcomes are revealed that can mitigate the adverse long-run effects. This may include educational measures and health measures. Recent advances in epigenetics may open up possibilities for interventions in the epigenome, effectively reversing some of the biological expressions of early-life exposures. At the same time, the evaluation of policy interventions may benefit from the study of adverse nutritional shocks that are less extreme than famines. After all, it is not always clear whether effects of extreme events, even if suitably extrapolated, are helpful inputs into cost-benefit analyses of policies.

Other topics for further research concern effects on offspring further down the family tree. Some studies have found effects on subsequent generations. From the health and economic points of view, the existence of such effects may have dramatic implications for the costs and benefits of the prevention of malnutrition. Recently, studies have emerged that document adverse effects on the offspring of mothers that were exposed in utero to the Dutch or Chinese famine. However, it is not clear a priori whether generations subsequent to the exposed cohort are always worse off as a result of the exposure. Some pioneering work based on famines due to crop failures in isolated villages in North Sweden in the early 19th century shows that exposure in the years just before puberty actually may have beneficial effects on the health of the grandchildren (see, e.g., Kaati, Bygren, Pembrey, & Sjöström, 2007). The authors postulate epigenetic mechanisms, as well as an underlying evolutionary explanation, for this finding. Clearly, this line of work opens up many opportunities for further research.

References

Almond, D., & Currie, J. (2011). Killing me softly: The fetal origins hypothesis. Journal of Economic Perspectives, 25(3), 153–172.Find this resource:

Almond, D., Edlund, L., Li, H., & Zhang, J. (2010). Long-term effects of the 1959–1961 China famine: Mainland China and Hong Kong. In T. Ito & A. K. Rose (Eds.), The economic consequences of demographic change in East Asia (pp. 321–348). Chicago: University of Chicago Press.Find this resource:

Almond, D. V. (2006). Is the 1918 influenza pandemic over? Long-term effects of in-utero influenza exposure in the post-1940 US population. Journal of Political Economy, 114, 672–712.Find this resource:

Ampaabeng, S. K., & Tan, C. M. (2013). The long-term cognitive consequences of early childhood malnutrition: The case of famine in Ghana. Journal of Health Economics, 32, 1013–1027.Find this resource:

Banning, C. (1946). Food shortage and public health, first half of 1945. Annals of the American Academy of Political and Social Science, 245, 93–110.Find this resource:

Barker, D. J. P. (1992). Fetal and infant origins of adult disease. London: BMJ Publishing Group.Find this resource:

Ben-Shlomo, Y., & Kuh, D. (2002). A life course approach to chronic disease epidemiology: Conceptual models, empirical challenges, and interdisciplinary perspectives. International Journal of Epidemiology, 31(2), 285–293.Find this resource:

Bengtsson, T., & Lindström, M. (2000). Childhood misery and disease in later life: The effects on mortality in old age of hazards experienced in early life, southern Sweden, 1760–1894. Population Studies, 54, 263–277.Find this resource:

van den Berg, G. J., Lindeboom, M., Popławska, M., & Portrait, F. R. M. (2017). Bad potatoes: A sibling analysis of famine exposure and effects on longevity. Working paper, University of Bristol, Bristol, UK.Find this resource:

van den Berg, G. J., Lindeboom, M., & Portrait, F. (2006). Economic conditions early in life and individual mortality. American Economic Review, 96, 290–302.Find this resource:

van den Berg, G. J., Pinger, P. R., & Schoch, J. (2015). Instrumental variable estimation of the causal effect of hunger early in life on health later in life. The Economic Journal, 126, 465–506.Find this resource:

Chen, Y., & Zhou, L. (2007). The long-term health and economic consequences of the 1959–1961 famine in China. Journal of Health Economics, 26, 659–681.Find this resource:

Deng, Z., & Lindeboom, M. (2017). Long run effects of undernutrition in early childhood: More evidence from the Chinese famine. Working paper, Vrije Universiteit Amsterdam.Find this resource:

Doblhammer, G. (2004). The late life legacy of very early life. Berlin: Springer.Find this resource:

Doblhammer, G., van den Berg, G. J., & Lumey, L. H. (2013). A re-analysis of the long-term effects on life expectancy of the Great Finnish Famine of 1866–68. Population Studies, 67, 309–322.Find this resource:

Dols, M. J. L., & van Arcken, D. J. A. M. (1946). Food supply and nutrition in the Netherlands during and immediately after World War II. Milbank Memorial Fund Quarterly, 24(4), 319–358.Find this resource:

Dyson, T. (1991). On the demography of South Asian famines: Part I. Population Studies, 45(1), 5–25.Find this resource:

van Ewijk, R., & Lindeboom, M. (2017). Why people born in World War II are healthier. Working paper, Mainz University, Mainz, Germany.Find this resource:

Farquharson, J. (1985). The Western Allies and the politics of food—agrarian management in postwar Germany. Berg Publishers Ltd.Find this resource:

Gørgens, T., Meng, X., & Vaithianathan, R. (2012). Stunting and selection effects of famine: A case study of the great Chinese famine. Journal of Development Economics, 97(1), 99–111.Find this resource:

Häusser, A., & Maugg, G. (2009). Hungerwinter. Deutschlands humanitäre Katastrophe 1946/47. Berlin: Prophyläen.Find this resource:

Heijmans, B. T., Tobi, E. W., Stein, A. D., Putter, H., Blauw, G. J., Susser, E. S., . . ., Lumey, L. H. (2008). Persistent epigenetic differences associated with prenatal exposure to famine in humans. Proceedings of the National Academy of Sciences of the United States of America, 105(44), 17046–17049.Find this resource:

Hionidou, V. (2006). Famine and death in occupied Greece, 1941–1944. Cambridge, UK: Cambridge University Press.Find this resource:

Huang, C., Zhu, L., Meng, W., & Reynaldo, M. (2010). Early life exposure to the 1959-1961 Chinese famine has long-term health consequences. The Journal of Nutrition, 140, 1874–1878.Find this resource:

Kaati, G., Bygren, L. O., Pembrey, M., & Sjöström, M. (2007). Transgenerational response to nutrition, early life circumstances, and longevity. European Journal of Human Genetics, 15(7), 784–790.Find this resource:

Kannisto, V., Christensen, K., & Vaupel, J. W. (1997). No increased mortality in later life for cohorts born during famine. American Journal of Epidemiology, 145, 987–994.Find this resource:

Klatt, W. (1950). Food and farming in Germany: I. Food and nutrition. International Affairs, 26(1), 45–58.Find this resource:

Koupil, I., Shestov, D. B., Sparén, P., Plavinskaja, S., Parfenova, N., & Vågerö, D. (2007). Blood Pressure, Hypertension and Mortality from Circulatory Disease in Men and Women who Survived the Siege of Leningrad. European Journal of Epidemiology, 22, 223–234.Find this resource:

Li, C., & Lumey, L. H. (2017). Exposure to the Chinese famine of 1959–1961 in early life and long-term health conditions: A systematic review and meta-analysis. International Journal of Epidemiology, 46(4), 1157–1170.Find this resource:

Lindeboom, M., Portrait, F., & van den Berg, G. J. (2010). Long-run effects on longevity of a nutritional shock early in life: The Dutch potato famine of 1846–1847. Journal of Health Economics, 29(5), 617–629.Find this resource:

Lumey, L. H., Stein, A. D., & Susser, E. (2011). Prenatal famine and adult health. Annual Review of Public Health, 32, 24.1–24.26.Find this resource:

Meng, X., & Qian, N. (2009). The long-run consequences of famine on survivors: Evidence from a unique natural experiment using China’s Great Famine. NBER working paper 14917.Find this resource:

Neelsen, S., & Stratmann, T. (2011). Effects of prenatal and early life malnutrition: Evidence from the Greek famine. Journal of Health Economics, 30(3), 479–488.Find this resource:

O’Gráda, C. (2001). Markets and famines: Evidence from nineteenth-century Finland. Economic Development and Cultural Change, 49(3), 575–590.Find this resource:

O’Gráda, C. (2009). Famine: A short history. Princeton, NJ: Princeton University Press.Find this resource:

Oppers, V. M. (1963). Analyse van de acceleratie van de menselijke lengtegroei door bepaling van het tijdstip van de groeifasen. Doctoral dissertation, University of Amsterdam.Find this resource:

Painter, R. C., Roseboom, T. J., & Bleker, O. P. (2005). Prenatal exposure to the Dutch famine and disease in later life: An overview. Reproductive Toxicology, 20(3), 345–352.Find this resource:

Painter, R. C., Roseboom, T. J., Bossuyt, P. M. M., Osmond, C., Barker, D. J. P., & Bleker, O. P. (2005). Adult mortality at age 57 after prenatal exposure to the Dutch famine. European Journal of Epidemiology, 20, 673–676.Find this resource:

Ravelli, A. C., van der Meulen, J. H., Michels, R. P., Osmonds, C., & Barker, D. J. (1998). Glucose tolerance in adults after prenatal exposure to famine. Lancet, 351, 173–177.Find this resource:

Reichardt, S., & Zierenberg, M. (2009). Damals nach dem Krieg. Eine Geschichte Deutschlands 1945 bis 1949. Munich: Goldmann.Find this resource:

Roseboom, T., de Rooij, S. & Painter, R. (2006). The Dutch famine and its long-term consequences for adult health. Early Human Development, 82(8), 485–491.Find this resource:

Scholte, R., van den Berg, G. J. & Lindeboom, M. (2015). Long-run effects of gestation during the Dutch Hunger Winter famine on labor market and hospitalization outcomes. Journal of Health Economics, 39, 17–30.Find this resource:

Sen, A. (1981). Poverty and famines: An essay on entitlement and deprivation. Oxford: Oxford University Press.Find this resource:

Stein, A. D., Kahn, H. S., Rundle, A., Zybert, P. A., van der Pal-de Bruin, K., & Lumey L. H. (2007). Anthropometric measures in middle age after exposure to famine during gestation: evidence from the Dutch famine. American Journal of Clinical Nutrition, 85, 869–876.Find this resource:

Song, S. (2009). Does famine have a long-term effect on cohort mortality? Evidence from the 1959–1961 Great Leap Forward Famine in China. Journal of Biosocial Science, 41, 469–491.Find this resource:

St. Clair, D., Xu, M., Wang, P., Yu, Y., Fang, Y., Zhang, X., . . . He, L. (2005). Rates of adult schizophrenia following prenatal exposure to the Chinese famine of 1959–1961. JAMA, 294, 557–562.Find this resource:

Trittel, G. J. (1990). Hunger und Politik—Die Ernährungskrise in der Bizone (1945–1949). Campus Verlag.Find this resource:

Vaiserman, A., & Lumey, L. H. (2013). Early life nutrition and adult health and development: Lessons from changing dietary patterns, famines, and experimental studies. Hauppauge, NY: Nova Science.Find this resource:

Vincent, C. (1985). The politics of hunger: The Allied blockade of Germany, 1915–1919. Athens: Ohio University Press.Find this resource: