Journal of Business Venturing 36 (2021) 106063

Contents lists available at ScienceDirect

Journal of Business Venturing

journal homepage: [Link]/locate/jbusvent

Childhood adversity and the propensity for entrepreneurship: A

T
quasi-experimental study of the Great Chinese Famine
Zhiming Chenga,
⁎,1
, Wei Guob, Mathew Haywardc, Russell Smythd, Haining Wange,
⁎

a
Social Policy Research Centre & Centre for Social Research in Health, The University of New South Wales, Sydney, New South Wales, Australia
b
Department of Social Work and Social Policy, School of Social and Behavioral Sciences, Nanjing University, Nanjing, Jiangsu, China
c
Department of Management, Monash Business School, Monash University, Melbourne, Victoria, Australia
d
Department of Economics, Monash Business School, Monash University, Melbourne, Victoria, Australia
e
Center for Chinese Public Administration Research, School of Government, Sun Yat-sen University, Guangzhou, Guangdong, China

ARTICLE INFO ABSTRACT

Keywords: Studies on the determinants of entrepreneurship emphasize that challenged adults tend to become
Entrepreneurship entrepreneurs. However, research has not addressed the childhood origins surrounding the pro­
Childhood pensity for entrepreneurship. This article links childhood adversity to the propensity of individuals
Internal migration to become migrant entrepreneurs later in life. We test hypotheses derived from this theory in the
China
context of whether, and when, children who survived the Great Chinese Famine of 1959–1961
became migrant entrepreneurs. Results strongly indicate that those who survived greater hardship
during the Famine are more likely to become entrepreneurs, especially when they were younger
during the famine years. We also find that being younger at the time of migration increased the
likelihood of becoming entrepreneurs in their new locale. Overall, this study casts light on why,
how and when childhood adversity shapes the propensity for entrepreneurship.

Executive summary

What is the gap in the literature? The underdog theory of entrepreneurship posits that adversity contributes to the propensity
for entrepreneurship, but it emphasizes challenges faced in adulthood (Miller & Le Breton-Miller, 2017). Some entrepreneurs note
that poverty and other adversity-experienced during childhood contributed to their decision to become entrepreneurs later.
However, the question of whether, and when, childhood adversity contributes to the likelihood that one becomes an entrepreneur
later in life represents a material gap in the entrepreneurship literature. Moreover, the Miller and Le Breton-Miller (2017) fra­
mework assumes that challenging conditions exist and then observe that people with these conditions and experiences feature as
entrepreneurs. We build on this insight by suggesting that experiencing adversity in childhood is informative of how individuals
developing coping and adapting skills that are predictive of entrepreneurship.
Why should the gap be filled? The gap should be addressed because the literature may be missing an essential perspective for
explaining and predicting the presence of entrepreneurship. Vast numbers of individuals experience trauma, ranging from poverty
to parental abuse in childhood. There are policy implications insofar as entrepreneurship (defined as self-employment) is a viable
career path for challenged individuals that can potentially be nurtured by policymakers. Filling the gap is a step towards more

Corresponding authors.
⁎

E-mail addresses: [Link]@[Link] (Z. Cheng), weiguo@[Link] (W. Guo), [Link]@[Link] (M. Hayward),
[Link]@[Link] (R. Smyth), wanghn36@[Link] (H. Wang).
1
Each author contributed equally to the study.

[Link]
Received 19 August 2019; Received in revised form 17 August 2020; Accepted 30 September 2020
Available online 28 October 2020
0883-9026/ © 2020 Elsevier Inc. All rights reserved.
Z. Cheng, et al. Journal of Business Venturing 36 (2021) 106063

robust testing and extension of the underdog theory of entrepreneurship insofar as it applies to youth and adults.
How does the paper fill the gap? The paper uses a quasi-natural experimental design for causally studying entrepreneurship
tendencies among millions of children who survived the trauma of the Great Chinese Famine. Heterogeneity arose in these
childhood experiences because some provinces were more severely affected by the Famine than others, and some children were
younger than others during the Famine. With this dataset, we document that childhood adversity contributes to the propensity for
entrepreneurship. We can also identify childhood conditions that are most conducive to such propensity, guided by the notion that
adversity promotes entrepreneurship in the underdog theory of entrepreneurship.
What are the contributions to theory? We show that individuals' challenges have their origins in early life and identify a) that
hardship and migration link to entrepreneurship and b) the range in which exposure to adversity has the most substantial effect on
subsequent propensity to become an entrepreneur. In this way, we contribute to, and inform, Miller and Le Breton-Miller's (2017)
proposal that ‘underdogs’ or disadvantaged adults, whether they are already in the workplace or not, are more likely to become
entrepreneurs and are more likely to be successful at this pursuit. They suggest that these individuals must become more resourceful
to survive and succeed, including working harder, garnering social support, and undertaking work outside of conventional em­
ployment. They focus on contemporaneous (economic, sociocultural, cognitive, emotional and physical) disadvantages and em­
phasize that one or more of these exacerbates life challenges. However, such challenges are not linked to childhood.
What are the contributions to practice? If children who face adversity and migrants are particularly likely to become en­
trepreneurs, they may benefit from training on the nature, merits, and demands of self-employment. Second, training could
respond to the hardships that these individuals face in childhood. Third, policymakers who seek to encourage entrepreneurship
and innovation in their countries could be mindful of the role of childhood hardship in the propensity for entrepreneurship.

‘I never had proper shoes before the age of 20 and I was almost always hungry. Because I was very poor, I was very diligent and
worked very hard. That creates values … I can look at my career and my life with a calm attitude.’
Liu Yonghao, survivor of the Great Chinese Famine and founder of New Hope Group, one of China's largest agribusiness com­
panies (Waldmeir, 2015)

1. Introduction

In the ‘underdog’ theory of entrepreneurship, Miller and Le Breton-Miller (2017: 3) proposed that ‘negative personal circum­
stances of an economic, sociocultural, cognitive, and physical/emotional nature may have a … powerful role to play in getting people
to become effective entrepreneurs’. An underdog theory complements perspectives that individuals' backgrounds and personalities
explain and predict their propensity to become entrepreneurs (Alvarez and Barney, 2007; Gartner, 1985). Many adults are ethnically,
emotionally, and socially disadvantaged and so cannot readily find employment (De Clercq and Honig, 2011). Instead, they seek
more control over their work through self-employment. In particular, Miller and Le Breton-Miller (2017, 9–10) note that migrants are
especially prone to entrepreneurship because they are disadvantaged in the labor markets to which they migrate. For instance, in the
United States, immigrants are almost twice as likely to become entrepreneurs relative to native-born Americans (Acs and Armington,
2006; Portes and Zhou, 1996; Vandor and Franke, 2016). All of these studies concern challenges facing adults.
However, the role of childhood in explaining and predicting the propensity for entrepreneurship is lacking in this framework. The
paucity of such research is a significant omission in the literature on entrepreneurship because childhood adversity shapes the
challenges faced by adults. Throughout history, scores of young people experienced privation and became displaced from their
childhood homes. In the twentieth century, persecuted Jewish children moved from their homelands; Romanian children became
overseas orphans before the end of the Cold War; and millions of Bangladeshi, Chinese, Congolese and Ethiopian children fled their
childhood homes as a result of famine (Kaler and Freeman, 1994; Yang, 2013b). Moreover, mounting evidence shows that childhood
environments shape subsequent cognitive and noncognitive abilities that influence their suitability for employment and en­
trepreneurship (Heckman et al., 2013; Heckman, 2006). Scholars of adolescent development show that such hardship forces in­
dividuals to become more self-reliant, resilient, and resourceful, each of which is a quality that explains and predicts the propensity
for entrepreneurship (Baker and Nelson, 2005; Bonanno, 2004; Caspi et al., 2003).
By extending the underdog theory of entrepreneurship, we seek to examine whether and when childhood adversity leads people
to become migrant entrepreneurs. We study this issue by using a quasi-experimental design in which we treat the Chinese Great
Famine (1959–1961) as an exogenous adverse shock, responding to recent calls for entrepreneurship research to use such designs to
establish causal effects (Wiklund et al., 2019; Williams et al., 2019). Specifically, this approach allows us to study whether surviving
the Famine in utero and childhood fosters migrant entrepreneurship in adulthood. Traumatic early childhood experiences have an
‘imprinting effect’ that can stay with survivors for their lives, as shaped by the nature of privations and their age at the time. Overall,
we seek to contribute to the underdog theory of entrepreneurship by explaining why and when childhood adversity increases the
propensity for entrepreneurship. In particular, we shall link such propensity to (a) excess death rates arising from the Famine in local
communities affecting children, (b) the age at which children endured the Famine, and (c) the age when they migrated.
In particular, we extend the Miller and Le Breton-Miller (2017) framework in three important ways. The first is that Miller and Le
Breton-Miller (2017) assume that challenging conditions exist and then observe that people with these conditions and experiences are over-
represented among entrepreneurs. This study informs how these individuals came to experience these challenging conditions and show
that their origins lie in early life adversity. The second is that Miller and Le Breton-Miller (2017) give several examples of underdog
entrepreneurs, of which migrants are one. By examining the migration of young people, we show how childhood creates challenging

2
Z. Cheng, et al. Journal of Business Venturing 36 (2021) 106063

backgrounds that lead to certain conditions and experiences that necessitate the acquisition of coping or adaptive skills that support en­
trepreneurship. Third, we provide evidence on the relationship between age at migration and propensity to become an entrepreneur.
We begin by providing context about the Famine before showing how the core elements of the underdog theory of en­
trepreneurship motivate hypotheses.

2. The Great Chinese Famine of 1959–1961

China is an important setting to examine whether young survivors of trauma become entrepreneurs because it experienced one of
the most devastating famines of the twentieth century. The Famine resulted from natural disasters and the Great Leap Forward of 1957,
an attempt by the Chinese government to accelerate collectivization and industrialization (Meng et al., 2015; Yang, 2013b). This plan
forced rural residents to move from farm-work into industrial work so that they could help produce industrial goods, including steel.
Meanwhile, local administrators fabricated reports of record-high grain harvests to ‘sell’ the success of agricultural communes and curry
favor with Mao Zedong and the central government. As a result, Chairman Mao's regime overestimated the availability of grain and
over-procured grain output for urban areas. By the spring of 1959, Famine was widespread in a third of Chinese provinces.
From 1959, the joint effects of starvation and related disease led to dramatically increased death rates in the provinces. An estimated
45 million people (of a population of 665 million) died of starvation and related illness, although there are reports that these statistics
were grossly under-reported by the central government (Yang, 2013b). Survivors of the Famine had higher incidences of type 2 diabetes,
hyper-glycemia, metabolic syndrome, and schizophrenia relative to those born afterward (e.g. Li and Lumey, 2016). Maternal mal­
nutrition exacerbated disease and congenital disabilities among those born during the Famine (Almond et al., 2010).
Encouraged by the Chinese government, survivors of the Famine fled their rural abodes for Chinese cities in droves after the
opening-up reforms in 1978. By 2017, 245 million Chinese migrants lived and worked in a locale other than their place of household
registration or ‘hukou’ (National Health Commission of China, 2018). Moreover, as many as one-quarter of rural migrants who moved
to urban areas became entrepreneurs (Giulietti et al., 2012). Over a sustained period, migrant entrepreneurs fuelled China's high
economic growth rate, supporting economies in traditional urban areas and increasing populations in cities such as Guangzhou,
Shanghai, Shenzhen, Wuhan, and Yangzhou (Cheng, 2014; Liu et al., 2018).
The following proposes that the challenged upbringing of famine survivors who migrated to the cities increased their propensity
to become entrepreneurs.

3. The underdog theory of entrepreneurship, youth privation and the propensity for entrepreneurship

According to the underdog theory of entrepreneurship, challenging backgrounds lead to certain conditions and experiences that
necessitate the acquisition of coping or adaptive skills conducive to entrepreneurship.2 The two sources of challenged backgrounds
featured in our study are privations arising from the Famine and migration. The framework offered by Miller and Le Breton-Miller
(2017) and the different framework that we offer appear in Fig. 1A and B.
We first examine how hardship experienced by children during the Famine, in terms of privation and age during the Famine,
resulted in entrepreneurship. Then we consider how age at migration shaped these hardships.

3.1. Core challenges facing youth survivors of the famine who subsequently migrate

Subjects of interest for this study are children survivors of the Famine, many of whom migrated to Chinese cities for work in
adulthood. As discussed, these survivors suffered extreme poverty during the famine years. Greater poverty during childhood is a
leading cause of lifelong social, emotional, physical, and cognitive challenges. Poverty increases the likelihood of becoming divorced
and imprisoned; it contributes to emotional disorders, including depression, anxiety, and schizophrenia; it leads to medical problems
such as heart disease and hepatitis, and it is associated with attention deficit disorder and poor self-regulation (Bonanno, 2004).
Individuals in our sample also migrated to Chinese cities at a young age, leaving their homes, friends, and family to make a better life
and they confronted foreign customs, people and work conditions (Cheng et al., 2014a; Cheng and Wang, 2013; Wang et al., 2015).

3.2. Conditions and experiences associated with the famine and subsequent migration

It is beyond the present scope to comprehensively elaborate upon the conditions and experiences facing famine survivors who
migrated to Chinese cities. However, studies of child development suggest that three kinds of conditions are ubiquitous and con­
sequential for survivors of childhood poverty.
First, studies of the Famine show that children suffered from disrupted relationships with their parents from an early age. During the
Famine, parents focused on their own survival and that of their dependents (i.e. infants, the elderly and the infirmed), factors that con­
tributed to childhood neglect (Coale and Banister, 1994; Yang, 2013b). In this vein, studies of childhood survivors of privation, including in
Pomerania and Romania, show that greater poverty leads to greater parent-child detachment, including the abandonment of children

2
Few studies have applied this theory to date. One application of this theory showed that underprivileged Chinese adults were more likely to bribe
officials to obtain resources (Baron et al., 2018). Another showed that Indian adults with severe disabilities were more likely to become en­
trepreneurs or self-employed (Saxena and Pandya, 2018).

3
Z. Cheng, et al. Journal of Business Venturing 36 (2021) 106063

Fig. 1. A. Underdog theory of entrepreneurship for challenged adults (Miller & Le Breton-Miller, 2017).
B. Underdog theory of entrepreneurship for survivors of youth privation.

(Repetti et al., 2002; Steinberg, 2001; Steinberg and Morris, 2001). Children face the insecurity that their physical, emotional, and social
well-being is precarious (Chugani et al., 2001; Evans and Kim, 2013; Völzke et al., 2011). For instance, attachment anxiety is a disorder from
childhood that refers to the anxiety felt in relationships, hindering meaningful relationships (Solomon and George, 1999).

4
Z. Cheng, et al. Journal of Business Venturing 36 (2021) 106063

Second, poverty leads to stunted child development associated with delays in establishing a sense of self and identity (see
Steinberg and Morris, 2001 for a review). Prendergast and Humphrey (2014: 250) refer to a ‘stunting syndrome’ in developing
countries involving ‘multiple pathological changes marked by linear growth retardation in early life are associated with increased
morbidity and mortality, reduced physical, neurodevelopmental and economic capacity and an elevated risk of metabolic disease into
adulthood’. Stunting is a cyclical process because females who were themselves stunted in childhood tend to have stunted offspring.
What results is an intergenerational cycle of poverty and reduced human capital that is difficult to break. In this vein, stunting from
poverty is linked to brain damage and psychiatric disorder, including depression, anxiety, and personality disorder (e.g. Brown and
Pollitt, 1996).
Third, survivors of childhood hardship become displaced from their hometowns and locations. When they migrate, they confront
the ‘culture shock’ of unfamiliar customs and people (Oberg, 1960). Migration deprives them of the family, friends, and social
networks of their home environments. This can be particularly traumatic for younger people who are yet to develop strong social and
emotional skills. In their study of entrepreneurship patterns among migrants, Vandor and Franke (2016: 390) note that ‘the first
encounter with a new culture generates such a flood of information perceived as overwhelming and stressful (Furnham and Bochner,
1986)’. In China, internal migration control was institutionalized in 1958 through the hukou system that enforces internal passports
that classify individuals according to their place of birth and denies rural people the same rights as city dwellers. In effect, rural
citizens who migrated to cities without permits are treated as illegal immigrants. The system compounds challenges for migrants
because it institutionalized widespread discrimination by urban residents and employers towards rural migrants (Cheng et al., 2015a;
Cheng et al., 2015b; Kuang and Liu, 2012).
Overall, individuals facing poverty and privation from childhood famine and migration face very considerable life challenges. As a
means of mitigating suffering and seeking a better life, they develop coping and adaptive skills. These skills include self-reliance,
resilience, resourcefulness, and risk-taking that are conducive to entrepreneurship, as we shall now elaborate upon (see Bonanno,
2004; Rutter, 1989 for reviews).

3.3. Adaptive and coping behaviors associated with the famine and subsequent migration and entrepreneurship

As individuals fend for themselves, they become more self-reliant, developing skills that include meeting strangers and learning
how to be detached and establish personal values (Ryan and Lynch, 1989; Samuels and Pryce, 2008). Samuels and Pryce (2008)
describe this as a ‘what does not kill you makes you stronger’ mindset in which children from impoverished backgrounds forge a
sense of self and identity, one that is forged without support systems and role models. Self-reliance also involves the capacity to
find essential resources independently. Children of the Famine were left to scavenge for food, water, and clothing relative to
counterparts in the cities where those resources are taken for granted (Bonanno, 2004; Scherer et al., 1989). Of course, self-reliance
is strongly related to the propensity for entrepreneurship because most businesses are self-funded and sole-operated
(Bauernschuster et al., 2012).
As individuals survive hardship, they also become more resilient. Development psychologists describe resilience in terms of
‘protective factors’ deployed by children exposed to ‘unfavorable or aversive life circumstances’, including loss and setbacks
(Bonanno, 2004: 20; Garmezy, 1993; Luthar et al., 2000). This includes coping with grief from the loss of loved ones (Mancini and
Bonanno, 2009). Rutter (1989) shows that survivors of traumatic conditions develop psychological resilience or skills to reduce the
risk of impact on self; strategies for mitigating negative chain effects of losses on loved ones; early establishment and maintenance of
self-esteem and self-efficacy; and the search for opportunities to recover. In China, Cheung (2014: 152) suggests that ‘greater stress
and more challenges due to migration may hone the migrants' personal growth, allowing them to develop a resilience that manifests
as eudaimonic well-being (Ryff et al., 2003)’. Considering the myriad setbacks and failure associated with self-employment, resilience
or the ability to recover from setbacks is integral to being an entrepreneur (Ayala and Manzano, 2014).
In the case of external and internal migration, these coping behaviors further include learning new ways of living in foreign
environments and developing new relationships (Deshingkar and Start, 2003; Furnham and Bochner, 1986). Cross-cultural skills
enable migrants to interact in their newfound environments and learn about new business opportunities (Vandor and Franke, 2016).
For instance, Chinese migrants tended to develop social relations in their host cities as a coping resource to reduce psychological
stress (Cheung, 2013). In China, discrimination associated with the hukou system limited formal employment opportunities, forcing
migrants to acquire and apply such skills in self-employment.
Resourcefulness is also a property of privation because the fewer resources individuals possess, the more resourceful they become
in finding them (Lévi-Strauss, 1962). Often, children of the Famine could not find food, clothes, or water nearby. As they grew older,
their families increasingly expected them to provide for dependents. Lacking resources, they depended almost entirely upon their
resourcefulness to obtain food, water, clothing, and shelter to survive. Resourcefulness included making contacts who could provide
resources, finding locations with food and resources, getting there, and using leftovers from others. During the Chinese Famine, rural
residents sought help from urban relatives and purchased food from the black market. They also show that survivors of the Famine
scavenged for unripened crops and that they stole grain to survive (Manning and Wemheuer, 2011; Yang, 2013b).
Moreover, Chinese migrants faced widespread discrimination resulting from the hukou system, forcing them to resourcefully find
work in the cities (Loyalka, 2012), including dirty, dangerous, and demeaning jobs that were shunned by pre-existing city workers
(Roberts and Wei, 1999; Wang et al., 2016). In urban China, disadvantaged rural migrants lived by building resourceful relationships
with urban residents and institutions for their use, despite regular state crackdowns (Loyalka, 2012; Xiang, 2004). Resourcefulness is
strongly associated with entrepreneurship because it involves the ability to make something from nothing, thus reducing the barriers
to entrepreneurship (Baker and Nelson, 2005; Sarasvathy, 2009). The personal attribute of ‘eating bitterness’ (chiku in Chinese) – the

5
Z. Cheng, et al. Journal of Business Venturing 36 (2021) 106063

ability to endure and overcome difficulties and forge ahead – is a noted feature of rural survivors of the Great Chinese Famine
(Manning and Wemheuer, 2011) and rural-urban migrants (Loyalka, 2012).3
Poverty is also related to risk-taking, including that required to start a new business. Haushofer and Fehr's (2014: 862) review of
the economic literature on risk-taking in developing countries concludes that poverty leads to ‘stress and negative affective states’
resulting in ‘habitual behaviors’ versus ‘goal-directed behaviors’ like entrepreneurship. However, migrants in this sample sought self-
employment as a matter of necessity, consistent with evidence that those with less to lose from taking risks tend to be risk-seeking
(Harris et al., 2002).
Overall, the combination of privation during the Famine, together with subsequent migration, created a class of prospective
entrepreneurs who entered Chinese cities. Relative to those who were more sheltered from the Famine fallout, those who experienced
greater adversity were likely to have developed stronger self-reliance, resilience, resourcefulness, and risk-taking from a younger age.
Collectively, these factors would tend to increase their propensity for entrepreneurship. In this study, childhood adversity is strongly
indicated by excess death rates in home locations (i.e., excess death rates during the famine relative to the years preceding it) as
indicated by hypothesis 1.
Hypothesis 1. The more severe the adversity experienced by children survivors of the Famine, the more likely they will become
migrant entrepreneurs.
Next, we examine the age at which survivors of the Famine are most vulnerable to the effects of trauma experienced at young
ages. In terms of the underdog framework, this speaks to the question: at what age(s) does the early life shock have its most
significant effect of posing social, emotional, physical, and cognitive challenges for survivors later in life? There is evidence that
younger children, even those in utero, are most vulnerable to trauma from poverty. Malnutrition of mothers during childbirth can
result in irreversible brain damage to unborn children (McLaughlin et al., 2014; Teicher et al., 2016). Ages from birth to two are
critical to the development of basic senses that can be disrupted by trauma. In a comprehensive review, Black et al. (2017: 77)
conclude that ‘proxy measures of stunting and poverty indicate that 250 million children (43%) younger than five years in low-
income and middle-income countries are at risk of not reaching their developmental potential’. In those years, neural processes can be
disrupted in a manner that compromises early learning systems, thus adversely affecting subsequent health and development.
Consistent with the underdog theory of entrepreneurship, we submit that younger survivors of the Famine would face more acute
cognitive and emotional challenges (Ampaabeng and Tan, 2013; Kim et al., 2017; Neugebauer et al., 1999; St Clair et al., 2005), and
would be less well suited to employment (Chen and Zhou, 2007; Scholte et al., 2015). Moreover, they are more strongly motivated to
mitigate suffering by acquiring skills required for self-reliance, resilience, and resourcefulness (Dikötter, 2015; Rahmato, 1991; Yang,
2013b; Zhou, 2012). With these skills, they have a stronger propensity to become entrepreneurs. For instance, Shiri et al. (2010)
suggest that those individuals who were more traumatized, and experienced a higher degree of adversity, may also be those who were
more open to more considerable changes in their belief profiles through a motivated process that represents the transition from victim
to survivor. Indeed, Lothe and Heggen (2003) and Lothe (2010) find that young survivors of the 1984–1985 Ethiopian famine have
developed strong resilience and self-reliance from the co-experience of hope and despair (‘no growth without catastrophe’) and that
some of them have managed to become an employer entrepreneur. These factors lead us to suggest that:
Hypothesis 2. Children who were younger during the Famine are more likely to become migrant entrepreneurs.
We now turn to the effect of age at migration in light of evidence that individuals who are younger upon migration lack the
maturity to deal with their newfound circumstances (Cheung, 2013, 2014), exacerbating the adverse conditions and experiences
described above. Moreover, those who are younger at migration will have a greater need for self-reliance to the extent that they lack
social and emotional skills to develop meaningful relationships (Shen et al., 1998). For instance, Yang (2013a: 56) suggests that ‘older
migrants have had more time to gain local reputations, find decent employment, and establish social networks with local residents
and, thus, become more economically successful’ versus younger migrants.
Stronger resilience skills emerge for those who lack life experiences to tackle adversity arising from unfamiliar circumstances.
They will also be more resourceful in that they arrive in cities with fewer resources than older counterparts, on average (Bonanno,
2004). For example, Mu (2018: xviii) suggests that, with the ‘loss of social capital in schools, neighborhoods, and community of
origin’, young rural migrants ‘have to learn to cope with challenges during their rural-to-urban transition’ and that migration at a
young age provides them with ‘a unique and critical learning moment for growing stronger despite adversities’. Law and Peng (2008)
find that young rural migrants strategically employ communication technology to form mobile cyber-kinship networks as a resource
of market information and business opportunities amid desolation and insecurity upon arrival.
Lacking the emotional and social resources of older migrants, we hypothesize that those who are younger when migrating are
disadvantaged in finding formal employment (Cheng et al., 2014b; Gagnon et al., 2014). Instead, we expect that they would tend to
draw on self-reliance, resilience, and resourcefulness skills to become entrepreneurs at a younger age — younger age upon migration
increases the likelihood of entrepreneurship:
Hypothesis 3. The younger the age of migration of famine survivors, the higher the propensity for entrepreneurship.
We now turn to the data and sample and outline the empirical strategy used to test these hypotheses.

3
Manning and Wemheuer (2011) and Loyalka (2012) are book length studies of this issue for survivors of the Great Famine and rural-urban
migrants respectively and both are titled Eating Bitterness.

6
Z. Cheng, et al. Journal of Business Venturing 36 (2021) 106063

4. Methods

4.1. Data and sample

We use data from the 2016 China Migrants Dynamic Survey (CMDS), a nationally representative survey of migrants administrated
by the Migrant Population Service Centre of the National Health Commission of China.4 The sampling frame of 2016 CMDS was the
whole population of migrants aged 15 and older, did not hold a local hukou, and lived in their current locality for more than a month
as of May 2016. The CMDS employs a stratified multi-stage probability proportional to size sampling method.
The 2016 CMDS surveyed 169,000 individuals across 31 mainland provinces/autonomous regions/municipalities, plus the
Xinjiang Production and Construction Corps.5 The CMDS collects detailed information on individuals, households, and communities
related to demography, employment, consumption, income, and other characteristics, making it the most suitable dataset for our
purposes. Most other available survey datasets focus on residents with a local hukou or who are in the labor force or cover only
individual provinces; thus, they do not provide a representative sample of the population of migrants.
In robustness checks and extensions on our main results, we supplement the use of the 2016 CMDS with the 2017 CMDS as well as
several other large datasets; namely, the 2013 China Household Income (CHIP) dataset, 2015 China Social Survey (CSS), 2015 China
General Social Survey (CGSS), and 2009 Rural-Urban Migration in China (RUMIC) rural dataset, where data are not available in the
2016 CMDS.
We focus on rural-urban and rural-rural migrants and exclude urban-rural and urban-urban migrants who hold an urban hukou
for several reasons. The first reason is that the Famine hit rural areas much harder than urban areas due to the food procurement and
allocation system, which favored urban residents during the Famine. Thus, it would be difficult to account for such preferential
treatment for urban residents in estimating the effects of the Famine. Rural residents did not usually receive additional relief during
the Famine, making those cohorts exposed to Famine relatively homogenous, compared to their urban counterparts. Moreover,
because of land reform and agricultural collectivization in the 1950s, which deprived rural households of private ownership of land
and other productive assets, household socioeconomic status was more equalized in rural areas. Focusing on rural migrants also
avoids the confounding effects of the Cultural Revolution (1966–1976), which was mainly an urban phenomenon (Giles et al., 2019).
The second relevant factor is that, between 1949 and the early-1990s, most urban locals were assigned a job by the state in the
formal sector (e.g., government agencies and state- and collective-owned enterprises). As a result, only a relatively small proportion
of urban residents became entrepreneurs. The work-life histories of urban residents are also complicated because many experienced
layoffs in the state sector and were affected by the reform of housing and social insurance systems (Cheng and Beresford, 2012; Wang
et al., 2014). These changes did not affect rural migrants.
A third practical consideration is that the coverage of social insurance is high among urban locals, and a significant proportion had
retired by the time that the CMDS was administered. In contrast, rural residents receive little, or no, pension, and many remain in the labor
force after age 60. For instance, a recent report finds that 37% of rural females and 22% of rural males remained working after age 60,
compared to only 11% of urban males and 5% of urban females who remained working after 60 (National Survey Research Centre, 2014).

4.2. Empirical strategy

To examine our core hypothesis (hypothesis 1), we first obtain baseline estimates by applying a difference-in-difference (DID)
estimator to examine the long-term causal effect of Famine on entrepreneurship by exploiting variation in the exposure to Famine both
across provinces and across birth cohorts. Our empirical specification is of the form:
6 6
yipc = + c cohortic + p EDRip + c (cohortic × EDRip) + X + ipc
c=2 c=2 (1)

where y is a dichotomous outcome variable if migrant i, who is originally from province p and was born in cohort c, is an entrepreneur
(yes = 1; no = 0). We define an individual as an entrepreneur if he/she is self-employed (solo entrepreneur) or an employer (business
owner). In robustness checks, we also further examine solo entrepreneurs and business owners separately (Nikolaev and Maldonado-
Bautista, 2019). Cohort is a categorical variable for cohorts 1–6, with cohort 1 (born 1966–1968) being the reference group. We define
six birth cohorts by age-related child development stage, relative to the famine period (see Table 1). We divide the sample into those
born 5–7 years after the Famine (cohort 1; born 1966–1968); born 2–3 years after the Famine (cohort 2; born 1963–1965); born, or in
utero, during the Famine (cohort 3; born 1959–1962); born 1–3 years before the Famine (cohort 4; born 1956–1958)6 born 4–8 years
before the Famine (cohort 5; born 1951–1955); and born 9–14 years before the Famine (cohort 6; born 1945–1950).

4
Formerly the National Health and Family Planning Commission. For more information about the CMDS, see [Link].
5
The Xinjiang Production and Construction Corps (XPCC) formed in 1954 as a paramilitary and economic organization in Xinjiang autonomous
region. The administrative status of XPCC is on par with a provincial capital. The XPCC administrates seven county-level cities and five towns. In
2016 its GDP was 213 billion RMB, and the population was 2.8 million.
6
Following the existing literature (see e.g., Chen and Zhou, 2007), we include samples who were conceived during the Famine in cohort 3. The
literature suggests that in utero exposure to Famine can harm mental and physical health and cognitive ability (see e.g., Chen and Zhou, 2007;
Schulz, 2010; Song, 2013). Reclassifying these individuals into cohort 2 (unborn during the Famine) does not qualitatively change our results
(results from this exercise are available from the authors upon request).

7
Z. Cheng, et al. Journal of Business Venturing 36 (2021) 106063

Table 1
Descriptive statistics of cohorts.
Cohort 1 Cohort 2 Cohort 3 Cohort 4 Cohort 5 Cohort 6 Total

Year of birth 1966–1968 1963–1965 1959–1962 1956–1958 1951–1955 1945–1950

Age during the famine Unborn Unborn 1–3 4–6 7–11 12–17
Life stage during the famine Unborn Unborn Fetus/infancy/ Early childhood Middle Adolescence
(1959–1961) toddlerhood childhood
Age during the survey (2016) 48–51 51–53 54–57 56–60 61–65 66–71
Entrepreneur (reference: employee) 55.64% 53.43% 50.02% 51.49% 40.92% 41.94% 52.92%
Observations 5852 3921 1871 872 738 217 13,471

The severity of Famine is measured by EDR, the excess death rate in 1960 in the rural migrant's home province p. Following Chen
and Zhou (2007), EDR is the difference between the death rate in 1960, the most severe year of the Famine, and the average death
rate in the three years preceding the Famine (1956–1958). EDR is measured at 1‰, such that a one-unit increase in the EDR means an
increase of one excess death per thousand people. We do not have the EDR statistics for Tibet autonomous region and Hainan
province. Since Hainan province was part of Guangdong province between 1950 and 1988, we follow Lin and Yang (2000) and use
the EDR in Guangdong province as a proxy for Hainan. Because we do not have the EDR for Tibet, we exclude migrants from Tibet.
There is a significant variation in the EDR across provinces (see Fig. 2). In robustness checks, we also use alternative measures of the
severity of Famine, and our conclusions remain unaffected. X is a vector of control variables for individual, household, and desti­
nation characteristics. The definition and summary statistics, for all variables, are provided in appendix Table A1.
In the DID estimator, region-invariant differences across cohorts are controlled for by comparison across regions, while cohort-
invariant differences across regions are controlled for by comparison across cohorts. If the Famine affected the probability of be­
coming an entrepreneur in later life, the interaction term cohort × EDR (hereafter ‘the interaction term’) would be statistically
insignificant for cohort 2 (Note this cohort did not experience the Famine), and statistically significant for (some of) cohorts 3–6 who
experienced the Famine in their childhood with cohort 1 as the reference group.
Interaction terms in nonlinear models (e.g. logit and probit models) are interpreted differently to those in linear regression models.
In nonlinear models the marginal effect of a change in both interacted variables is not equal to the marginal effect of changing just the
interaction term. Different observations may also have different signs for the interaction term. Therefore, the interaction term's sta­
tistical significance cannot be determined solely from the z-statistic reported in the probit regression results. Following the approach in
Ai and Norton (2003) and Norton et al. (2004), we calculate the interaction effects of cohort × EDR in a probit model and also
graphically check the signs of the interaction terms across the entire range of observations to ensure the validity of the DID estimator.
One concern may be that the results estimated from Eq. (1) may be biased due to systematic differences in control variables
between the Famine and post-famine cohorts. As a robustness check, we first group cohorts 1 and 2 as the post-Famine cohort and
cohorts 3–6 as the Famine cohorts. We then use coarsened exact matching (CEM) to match the covariates between the post-Famine
and Famine cohorts to create better balance between treatment and control groups (Iacus et al., 2012). We drop unmatched ob­
servations from the data and then apply a DID estimator in the following form, where Famine equals one if a migrant was in the
Famine cohort, or zero if a migrant was in the post-Famine cohort:
yip = + faminei + EDRp + (faminei × EDRp) + X + ip (2)
As an alternative to our DID estimates, we instrument for the EDR by applying an instrumental variable probit (IV-probit)
estimator using those in cohorts 3–6 who experienced the Famine. An advantage of the IV approach is that it does not need to rely on
cohort variations as in the DID approach. The empirical specifications are:
yip = + EDRp + X + ip (3)

EDRp = + Zp + X + p (4)
In Eq. (4) Zp is one or more IVs related to EDRp, the severity of the Famine in the migrant's home province p, depending on the
exact specification. We use four IVs in three separate IV-probit models.
The first IV is when county officials exaggerated grain harvests in 1958 in the migrant's home province p. The second and third IVs
measure differences in the institutional and political environment across provinces leading up to the Famine. The second IV is the
proportion of counties that had reported exaggerations of grain harvests in 1958 in the migrant's home province p. The third IV is the
takeover sequence of the migrant's home province p by the People's Liberation Army of CCP between 1946 and 1950. The fourth IV is
the participation rate in communal dining halls in the migrant's home province p. For these IVs to be valid, they should significantly
correlate with the EDR in the migrant's home province, but be orthogonal to the error term in Eq. (3). These potential relationships
between the IVs and Famine severity evolved 55–70 years ago before the 2016 CMDS survey. Hence, they are unlikely to be directly
related to migrants' career choice, except through their relationship with famine severity in the home province.
The first two IVs are constructed based on newspaper reports of exaggerations of grain harvests by county officials in each
prefecture, as published in the 1958 archives of the People's Daily (renmin ribao), the mouthpiece of the CCP (Xu et al., 2016).7 Xu

7
The first two IVs are constructed based on the county-level exaggerations data used in Xu et al. (2016), who generously shared their data with us.

8
Z. Cheng, et al. Journal of Business Venturing 36 (2021) 106063

Fig. 2. Excess death rate (‰) by province in the Chinese Great Famine, 1959–1961.

et al. (2016) geocoded to the county level a total of 558 instances of grain harvest exaggerations, defined as grain harvests of 1000
catties per mu (or 3000 kg per acre) or more (Kung and Lin, 2003) as reported in the People's Daily from June to September in 1958,
which was the most intensive period of falsification of grain harvests. Given that we measure the EDR at the provincial level due to
the lack of prefecture and county level geocodes in the CMDS data, we aggregate and calculate the two IVs at the provincial level. The
number of times that grain harvests were exaggerated, and the proportion of counties that reported exaggerations, in a given province
should be positively related to famine severity in that province. Local officials fabricated grain harvests to be seen to be endorsing the
Great Leap Forward, showing their loyalty to the party-state and seeking career progression and other benefits. The inflated pro­
duction figures led to excessive compulsory grain procurement, failed agricultural experimentation, and reduced sown acreage
(Ashton et al., 1984). Local officials gradually stopped falsifying grain harvest reports towards the end of the season of harvest in
1958, and by 1961 this practice had entirely ceased.
Our third IV is similar to that used in Gooch (2017), who employs the sequential takeover of provinces from the Kuomintang (the
Nationalist Party) by the CCP during the Chinese Civil War (1946–51) based on the work of Westad (2003). Gooch (2017: 140) states
that ‘the liberation sequence was considered during the process of collectivization throughout the 1950s by Mao and others. Con­
sequently, later liberated areas had a higher density of non-local leaders positioned to implement the Great Leap Forward, starting in
1958. During the collectivization process, non-local leaders allowed rural citizens' well-being to disintegrate to a relatively lower
level than had local leaders.’
The fourth IV is the participation rate in communal dining halls at the provincial level in 1959. Some studies suggest that the
communal dining system, which engendered excessive consumption and waste of food, was one of the Famine's major causes (Chang
and Wen, 1997; Liu et al., 2014). As a scheme operated during the Great Leap Forward, authorities discontinued the communal
dining system after the Famine.
As a further robustness check on the DID and IV-probit results, we apply the special regressor method in Dong and Lewbel (2015),
which accommodates endogenous, or mismeasured, regressors and heteroskedastic errors. The special regressor model includes a
particular regressor, in addition to an IV, that is exogenous and appears as an additive term in the model, is continuously distributed
and has the same or larger support than Xφ + εipc, and preferably (but not strictly necessarily) has a thick-tailed distribution (e.g.
kurtosis ≥ 3). Any normally distributed regressor would satisfy this condition. Dong and Lewbel (2015) use age as the particular
regressor to predict migration probability in the U.S. We also use age (kurtosis = 3.2353) as the special regressor to estimate such
model.

9
Z. Cheng, et al. Journal of Business Venturing 36 (2021) 106063

To test hypotheses 2 and 3, we use extensions or variations of the above framework. To examine hypothesis 2, we interact EDR
with the respondent's age during the Famine in the IV-probit regression. To examine hypothesis 3, we interact EDR with respondent's
age at first migration in the IV-probit regression.

5. Results

Table 2 provides the results for our first hypothesis.8 In model 1, the interaction terms for cohorts 3 and 5 are statistically
significant. In model 2, the interaction terms for cohorts 3, 4, 5 and 6 are statistically significant, while the interaction term for cohort
2 is statistically insignificant. Model 2 is our preferred model because it controls for several correlates with entrepreneurship and has
relatively better model fit statistics. The results in Table 2 are consistent with the first hypothesis. A one-unit increase in the EDR (i.e.,
1‰; an increase of one excess death toll per thousand people) in the province in which the cohort was born and grew up as children
increases the probability of becoming an entrepreneur by 0.18, 0.22, 0.23 and 0.35 percentage points for cohorts 3, 4, 5 and 6,
respectively. A one standard deviation increase in the EDR increases the probability of becoming an entrepreneur by 1.5 percentage
points for cohort 3 and 1.3 percentage points for cohorts 4–6, respectively.9
Table 3 presents the CEM-DID, IV-probit, and special regressor results. These findings are consistent with the DID estimates in
model 2 in Table 2 that exposure to more severe Famine has a positive effect on the probability of becoming an entrepreneur. In each
of the IV-probit estimates presented in models 2–4, a one-unit increase in EDR leads to an average increase of 0.2–1.2 percentage
points in the probability that a rural migrant exposed to the Famine becomes an entrepreneur.10 For each of the IV-probit models, the
first stage F statistic well exceeds 10, satisfying the requirement for a valid IV when there are one or two endogenous regressors (Stock
et al., 2002).11 Overall, the results in both Tables 2 and 3 are consistent with our first hypothesis and the underdog theory's main
prediction.12
Table 4 presents IV-probit estimates for testing hypothesis 2. We interact respondent's age during the Famine with EDR. The
coefficient on the interaction term between age during the Famine and EDR is negative and statistically significant. Thus, children
who were younger during the Famine are more likely to become entrepreneurs, which is consistent with hypothesis 2.
Table 5 presents IV-probit estimates for testing hypothesis 3. The coefficient for the interaction term between EDR and age of first
migration is negative and statistically significant. This result suggests that the older the age at which individuals first migrated post-
Famine, the less likely that they are to become entrepreneurs. Hence, we are able to confirm hypothesis 3.
It is likely that birth adversity is a necessary, but not sufficient condition for a significant increase in subsequent entrepreneurial
behavior by migrants. To put it differently, specific demographic or human capital characteristics are likely to act in combination
with being born in the Famine to make individuals more likely to become entrepreneurs as migrants. At one level, identifying these
characteristics is an important proviso on the generalizability of our claims about birth adversity; however, at another level, it helps
us to understand more precisely where the birth adversity effects are strongest and also suggests the mechanisms through which birth
adversity works to increase likelihood of later entrepreneurial behavior by migrants. We examine this issue in Table 6, in which we
consider the moderating effects of education, gender, number of children, and once being a left-behind child on the relationship

8
Note that we control for the EDR and cohort dummies in this and all other DID specifications.
9
We graphically examine the pattern of interaction effects across cohorts 2–6 in appendix Fig. A1. Fig. A1 shows that the interaction terms have a
positive sign for all observations in cohorts 3–6. The interaction effects among most observations are statistically significant at the 90 percent level
or higher in cohorts 3–6. The interaction effects for cohorts 3–6 exhibit similar general patterns, indicating the effects of Famine are relatively
consistent across these cohorts.
10
The marginal effects from the IV-probit is higher than those from the DID in part because IV-probit and DID estimate the effects among cohorts
3–6 and cohorts 1–6, respectively.
11
We estimate the first-stage F statistics from a two-stage least square model. The potential weak instrument problem is only relevant for the
linear reduced form for the endogenous explanatory variable. Therefore, diagnostics for weak instruments are estimated as for the linear model case.
12
We undertake several robustness checks to examine the sensitivity of the results for our first hypothesis. Results feature in the online appendix.
We do the following: (1) estimate the interaction term in Table 2 in a non-linear way (see e.g. Vaaler, 2011) (Fig. A2); (2) conduct a placebo test, in
which we use a subsample of rural migrants who were born between 1963 and 1977 and construct five birth cohorts, each at three-year intervals
(Table A2); (3) rule out self-selection into both migration and entrepreneurship where the cohorts exposed to the famine are systematically more
likely to both migrate and choose to be an entrepreneur (Table A3); (4) rule out potentially systematically low labor force participation rates among
older cohorts being the result of compulsory retirement or generous pensions (Table A4); (5) rule out the possibility that the selection of respondents
who answered the health module in CMDS is related to whether they are an entrepreneur (Table A5); (6) control for health status and chronic
disease among a randomly selected sub-sample (Table A6); (7) replace cohorts with individual birth years from 1954 to 1964 (Table A7); (8) replace
EDR with the cohort size shrinkage index (CSSI) and fertility reduction index (FRI) (Table A8); (9) control for neighbourhood characteristics (Table
A9); (10) rule out that being born in winter is associated with the likelihood of becoming an entrepreneur or that being an adolescent during the
Cultural Revolution is associated with the likelihood of becoming an entrepreneur for those who experienced the famine (Table A10); (11) control
for factor scores derived from responses to survey modules on acculturation and discrimination, social activity, and social participation (Table A11);
(12) separate migrants into sub-samples of solo entrepreneur and employer entrepreneur (Table A12); (13) rule out that migrants exposed to more
severe famine earn higher incomes or work longer hours (Table A13); (14) rule out that the likelihood of becoming an entrepreneur later in life
reflects that the famine reduced the level of old-age support and aged care available from children born in cohorts exposed to the famine (Table
A14); (15) assess the effects of severity of the famine on the probability of migrating alone and searching for jobs through labor market inter­
mediaries (Table A15); (16) assess the effects of severity of the famine on risk-aversion and optimism (Table A16); and (17) control for the local
economic and employment environment (Table A17).

10
Z. Cheng, et al. Journal of Business Venturing 36 (2021) 106063

Table 2
Probit estimates of entrepreneurship.
This table presents full results of the difference-in-difference probit estimates of the effects of famine experience on the probability of being an
entrepreneur. Model 1 does not include any control variables except the cohorts, EDR and the interaction between them. Model 2 controls for a full
set of control variables as in appendix Table A1.
(1) (2)
Without controls With controls

Coefficients Coefficients Interaction effects

Cohort 2 × EDR −0.000200 (−0.143) 0.00112 (0.661) 0.000395 (0.655)

Cohort 3 × EDR 0.00436⁎ (1.838) 0.00514⁎⁎ (2.036) 0.00184⁎⁎ (1.986)
Cohort 4 × EDR 0.00414 (1.417) 0.00612⁎⁎ (2.031) 0.00219⁎⁎ (2.002)
Cohort 5 × EDR 0.00567⁎ (1.735) 0.00647⁎⁎ (1.980) 0.00232⁎⁎ (1.957)
Cohort 6 × EDR 0.00698 (1.503) 0.00972⁎ (1.689) 0.00346⁎ (1.665)
EDR −0.00424⁎ (−1.818)
Cohort 2 0.00112 (0.661)
Cohort 3 0.00514⁎⁎ (2.036)
Cohort 4 0.00612⁎⁎ (2.031)
Cohort 5 0.00647⁎⁎ (1.980)
Cohort 6 0.00972⁎ (1.689)
Male −0.0284 (−1.050)
Age −0.0755 (−0.641)
Age squared 0.000662 (0.615)
Married −0.105⁎⁎ (−2.040)
Education 0.0185⁎⁎ (2.153)
Communist Party member −0.0744 (−1.233)
Total household size 0.0631⁎⁎ (2.000)
Household size in destination 0.0209 (0.890)
Number of biological children 0.105⁎⁎⁎ (3.338)
Number of minor children −0.0211 (−0.576)
Lagged household annual income per capita 0.189⁎⁎⁎ (3.872)
Homeownership in hometown 0.0223 (0.590)
Homeownership in destination 0.0264 (0.387)
Health record 0.0462 (0.601)
Health training 0.00313 (0.0660)
Social insurance 0.159⁎⁎⁎ (2.707)
Professional/managerial job 1.141⁎⁎⁎ (9.473)
Family migration 0.576⁎⁎⁎ (8.463)
Parent(s) has migration experience 0.176⁎⁎ (2.120)
First-time migration 0.167⁎⁎⁎ (3.315)
Years in current destination 0.0170⁎⁎⁎ (3.843)
Total years of migration
1–2 0.0589 (0.834)
3–4 0.140⁎⁎ (2.140)
5–9 0.195⁎⁎⁎ (2.679)
10–14 0.170⁎⁎ (2.338)
15–19 0.173⁎ (1.957)
20–29 0.271⁎⁎ (2.340)
≥30 0.421⁎⁎⁎ (3.081)
Type of migration
Intra-province 0.00952 (0.0798)
Intracity 0.0856 (0.746)
Migration intention in five years
Return to hometown −0.246⁎⁎⁎ (−4.354)
Migrate to another locality −0.0465 (−0.413)
To be decided 0.0576 (1.642)
Urban neighbourhood 0.333⁎⁎⁎ (4.922)
Constant 0.250⁎⁎⁎ (3.354) −2.092 (−0.659)
Observations 13,409 12,494
Akaike's information criterion (AIC) 18,438.73 15,698.76
Bayesian information criterion (BIC) 18,527.92 15,929.18
Pseudo R2 0.0070 0.0948

Notes: z-statistics in parentheses; standard errors are clustered at the county/district level; N = 12,494 is used in calculating the BIC.
⁎⁎⁎
p < 0.01.
⁎⁎
p < 0.05.
⁎
p < 0.1.

11
 Z. Cheng, et al.

Table 3
Post-matching probit, instrumental variable probit and special regressor method estimates of entrepreneurship among the famine cohorts.
This table presents post-matching probit, instrumental variable probit and special regressor estimates of the effects of famine intensity on the probability of being an entrepreneur among the famine
cohorts. Model 1 provides difference-in-difference probit estimates of the effects of famine on the probability of being an entrepreneur, after match the covariates between the post-famine and famine
cohorts to create better balance between treatment and control groups. Models 2–4 provide results of the effects of famine on the probability of being an entrepreneur from instrumental variable probit
models, which use different sets of instrumental variables for the first stage estimate on EDR. Model 5 provides the results of the effects of famine on the probability of being an entrepreneur from the
special regressor method. Specifications for models 1–4 include a full set of control variables as in appendix Table A1. Specification for models 5 includes a full set of control variables as in appendix Table
A1, except for age squared because age as the special regressor must be continuously distributed after conditioning on the other regressors in the special regressor method. Full results are available from
the authors.
(1) (2) (3) (4) (5)
CEM-DID probit IV-probit IV-probit IV-probit Special regressor method

Coefficients Interaction Coefficients Marginal Coefficients Marginal Coefficients Marginal Coefficient Marginal
effect effect effect effect effect

12
Famine cohort × EDR 0.00328⁎⁎ (2.09) 0.00116
EDR Yes 0.0377⁎⁎⁎ (6.224) 0.0122 0.0123⁎ (1.885) 0.00437 0.0198⁎⁎⁎ (6.600) 0.00706 0.159⁎⁎⁎ (3.170) 0.00247
Control variables Yes Yes Yes Yes Yes
First stage
Number of exaggerations 0.106⁎⁎⁎ (5.556)
Proportion of exaggerations 36.14⁎⁎⁎ (10.18)
PLA takeover sequence 0.688⁎⁎⁎ (11.58)
Communal dining hall 0.372⁎⁎⁎ (27.66)
participation rate
Observations 10,701 3118 2502 2791 3755
First stage F-statistic 135.559 136.460 755.535

z-Statistics in parentheses; standard errors are clustered at the county/district level.

⁎⁎⁎
p < 0.01.
⁎⁎
p < 0.05.
⁎
p < 0.1.
Journal of Business Venturing 36 (2021) 106063
Z. Cheng, et al. Journal of Business Venturing 36 (2021) 106063

Table 4
Instrumental variable probit estimates of entrepreneurship among the famine cohorts.
This table presents instrumental variable probit estimates of the effect of age during the famine as of
1961, conditional on the famine intensity experienced, on the probability of being an entrepreneur
among the famine cohorts. Instrumental variables are proportion of exaggerations and PLA takeover
sequence. Results based on other instrumental variables as used in Table 3 are qualitatively similar
(results are available upon request). All specifications include a full set of control variables as in ap­
pendix Table A1, except age, which is already controlled by the age during the famine. Full results are
available from the authors.
Coefficients

EDR × age during the famine −0.00667⁎⁎⁎ (−8.430)

EDR Yes
Age during the famine Yes
Control variables Yes
Observations 3118

z-Statistics in parentheses; standard errors are clustered at the county/district level.

⁎
p < 0.1.
⁎⁎
p < 0.05.
⁎⁎⁎
p < 0.01.

between childhood adversity and becoming an entrepreneur. Being better educated, female, having more children or having left-
behind experience strengthens the relationship between childhood adversity and becoming a migrant entrepreneur.

6. Discussion

We have sought to make three principal contributions to the understanding of individuals' propensity to become entrepreneurs.
First, we contribute to the underdog theory of entrepreneurship by extending it to childhood adversity. Miller and Le Breton-
Miller (2017) proposed that adult underdogs become entrepreneurs out of necessity or choice. They focused on contemporaneous
(economic, sociocultural, cognitive, emotional, and physical) disadvantages as predictors of the propensity for entrepreneurship. The
specific conditions and experiences which Miller and Le Breton-Miller (2017) give as increasing the likelihood that one becomes an
entrepreneur are disparate. They include mental and physical health problems, cognitive disabilities, and a lack of human capital. We
sought to provide a theoretical basis that underpins the root source of such conditions and experiences. A literature in economics and
related disciplines shows that disadvantages experienced later in life stem from adversity in early life (see e.g. Currie and Vogl, 2013,
for a review of much of this literature). By exploring early life adversity in the context of the Great Famine, the propensity for
entrepreneurship relates to self-reliance, resilience, resourcefulness, and risk-taking that children adopt to mitigate suffering and
survive.
Second, we consider how younger, migrants are more likely to become entrepreneurs. We extend Miller and Le Breton-Miller's
(2017) observations to show that the source of the challenges migrant entrepreneurs faced originated in childhood adversity and their
adaptive and coping skills. Relatedly, we provide evidence on when childhood adversity is likely to increase the propensity for
migrant entrepreneurship. In our framework, younger migrants were more able to develop self-reliance, resilience, resourcefulness
and risk-taking at a younger age in a manner that increased the likelihood that they would become entrepreneurs.

Table 5
Instrumental variable probit estimates of entrepreneurship.
This table presents instrumental variable probit estimates of the effect of age of first migration,
conditional on the famine intensity experienced, on the probability of being an entrepreneur among
the famine cohorts. Instrumental variables are proportion of exaggerations and PLA takeover se­
quence. Results based on other instrumental variables as used in Table 3 are qualitatively similar
(results are available upon request). All specifications include a full set of control variables as in
appendix Table A1. Full results are available from the authors.
Coefficients

EDR × age of first migration −0.00593⁎⁎ (−2.08)

EDR Yes
Age of first migration Yes
Control variables Yes
Observations 3118

z-Statistics in parentheses; standard errors are clustered at the county/district level.

⁎
p < 0.1.
⁎⁎
p < 0.05.
⁎⁎⁎
p < 0.01.

13
Z. Cheng, et al. Journal of Business Venturing 36 (2021) 106063

Table 6
Instrumental variable probit estimates of moderating effects on entrepreneurship.
This table presents instrumental variable probit estimates of the effects of being female (model 1), years of education (model 2), number of
children (model 3) or experience of once being a left-behind child (model 4), conditional on the famine intensity experienced, on the probability of
being an entrepreneur among the famine cohorts. Instrumental variables are proportion of exaggerations and PLA takeover sequence. Results based
on other instrumental variables as used in Table 3 are qualitatively similar (results are available upon request). All specifications include a full set of
control variables as in appendix Table A1. Full results are available from the authors.
(1) (2) (3) (4)

Coefficients Coefficients Coefficients Coefficients

EDR × female −0.0469⁎⁎⁎ (5.11)

EDR × education −0.0157⁎⁎⁎ (−3.94)
EDR × no. of children −0.0447⁎⁎⁎ (5.53)
EDR × left-behind child −0.0410⁎⁎⁎ (−3.81)
EDR Yes Yes Yes Yes
Control variables Yes Yes Yes Yes
Observations 3118 3118 3118 3118

z-Statistics in parentheses; standard errors are clustered at the county/district level.

⁎
p < 0.1.
⁎⁎
p < 0.05.
⁎⁎⁎
p < 0.01.

Third, we offer a perspective on why entrepreneurship is a vital avenue for individuals from the ‘school of hard knocks’ to make a
better life (Chu et al., 2019). We do not wish to diminish or glorify the Famine tragedy and its direct role in causing death and
lowering life expectancy. However, entrepreneurship enabled them to find work and, sometimes, wealth when they were dis­
criminated against and denied jobs in the Chinese cities. Many entrepreneurs relate their success to challenging early life experiences,
and we provide systematic evidence on how this increases the propensity for entrepreneurship.
Liu Yonghao, the founder of the New Hope Group, whom we quoted to lead this article, offers a case in point of someone who
became an entrepreneur after surviving the Famine and moving to a large Chinese city. Born in 1951, Liu was the youngest son of
low-income rural parents in Sichuan province who send Liu's brother to adoption because they could not provide for him. At 17, Liu
migrated to Chengdu, the capital of Sichuan province, where he could not find employment. Instead, he earned 2 cents a day
operating a street stand that repaired basic radios. Liu eventually pooled his life savings, together with those of his brothers (roughly
$500), to raise and sell quail and then saw an opportunity for supplying animal feed, as a precursor to founding the New Hope Group.
The full details of this account (i.e., survival of Famine, migration to Chinese cities at a young age and setbacks from self-employ­
ment) are further elaborated upon by Yang's (2013a) first-hand narratives of the plight of children in Xinyang prefecture, Henan
province.
A further contribution is to use a quasi-experimental design to unravel causal effects on the relationship between en­
trepreneurship and adversity. This design is useful in addressing endogeneity concerns (e.g., reverse causality, omitted variable, and
measurement error) (Williams et al., 2019). A similar design examines how law changes in one jurisdiction result in the departure of
individuals and intellectual property from that jurisdiction (e.g. Marx et al., 2009). Other settings suitable to quasi-experimental
studies include extreme weather events, wars, and other societal disruptions such as pandemics.

6.1. Limitations and suggestions for future research

While we can rule out other possible explanations for our main results, our study has limitations. Of course, while we can observe
whether subjects become entrepreneurs, we cannot comment on the nature and success of their self-employment. Another limitation
is that given our methodological approach, it is difficult to identify the precise channels through which early life adversity leads
individuals to become entrepreneurs. Table 6 addresses the moderating effects of some key demographic and human capital char­
acteristics on whether childhood adversity leads individuals to become entrepreneurs. In the appendix (in Table A14) we examine the
potential role of the level of old-age support and aged care available from children born in cohorts exposed to the Famine. However,
we cannot determine whether individuals became entrepreneurs because they could not find other forms of employment versus out of
preference.
Regarding the latter, we cannot pinpoint the role played by three of the four mechanisms that we described earlier: self-reliance,
resourcefulness and resilience. However, we find some evidence that rural cohorts who experienced more severe Famine are less risk-
averse. Beyond this, we rely on evidence from other studies that link adversity during childhood to the development of these coping
skills. One suggestion for future research is to design, administer, surveys that contain self-reliance, resourcefulness and resilience to
examine their role as mediators through which childhood adversity affects the propensity for entrepreneurship. Such surveys could
also collect information on setbacks experienced during adulthood to disentangle these effects from childhood adversity. This would
extend our analysis, and enrich our understanding of the challenge-based model, by providing further empirical evidence on the
channels through which early life adversity affects whether there is a propensity for entrepreneurship.

14
Z. Cheng, et al. Journal of Business Venturing 36 (2021) 106063

6.2. Practical implications

Results from this study have practical implications in light of evidence that children who experienced adversity at a younger age
are more likely to become entrepreneurs. Existing studies on early life disadvantage have tended to emphasize the negative long-term
implications of such a disadvantage for income and health. The resulting policy prescription has been early-life intervention to
counteract these challenges. For example, Heckman (2006) argues that deficits resulting from early-life adversity are mitigated by
investing in quality early childhood education. However, through the lens of the underdog model of entrepreneurship, our result
provides a different perspective: Early life adversity fosters positive outcomes in creating successful entrepreneurs in adulthood. If
children who face adversity are particularly likely to become entrepreneurs, they may benefit from training that more closely
educates them about the merits and demands of self-employment. Such awareness and training could be instrumental in developing
countries with more limited opportunities for formal employment.

CRediT authorship contribution statement

Zhiming Cheng: Conceptualization, Methodology, Software, Data curation, Resources, Writing - original draft, Writing - review &
editing
Wei Guo: Data curation, Resources, Writing - original draft
Mathew Hayward: Conceptualization, Methodology, Writing - original draft, Writing - review & editing
Russell Smyth: Conceptualization, Methodology, Writing - original draft, Writing - review & editing
Haining Wang: Conceptualization, Methodology, Writing - original draft, Writing - review & editing.

Declaration of competing interest

None.

Acknowledgments

We thank the three anonymous referees and Professor Nikolaus Franke (field editor) for their insightful comments and useful
suggestions on earlier versions of the paper. Earlier versions of the paper were presented at the George Institute for Global Health and
the 15th International Conference of the Western Economic Association International, at which participants provided helpful feed­
back. Zhiming Cheng acknowledges support from the University of New South Wales Scientia Fellowship (grant number: PS45957).
Wei Guo acknowledges support from the National Natural Science Foundation of China (grant number: 71921003). Haining Wang
acknowledges support from the National Social Science Fund of China (grant number: 20BGL236). This study uses datasets from the
2016 and 2017 China Migrants Dynamic Survey, 2013 China Household Income Project, 2015 Chinese Social Survey, 2015 Chinese
General Social Survey and 2009 Rural-Urban Migration in China. The authors also thank Hongwei Xu (Queens College - CUNY), who
generously shared with us his compiled data on newspaper reports of exaggerations of grain harvests. Any remaining errors are our
own.

Appendix A. Supplementary materials

Supplementary materials to this article can be found online at [Link]

References

Acs, Z.J., Armington, C., 2006. Entrepreneurship, Geography, and American Economic Growth. Cambridge University Press, Cambridge.
Ai, C., Norton, E.C., 2003. Interaction terms in logit and probit models. Econ. Lett. 80 (1), 123–129.
Almond, D., Edlund, L., Li, H., Zhang, J., 2010. Long-term effects of early-life development: evidence from the 1959 to 1961 China Famine. In: Ito, T., Rose, A.K. (Eds.),
The Economic Consequences of Demographic Change in East Asia. The University of Chicago Press, Chicago.
Alvarez, S.A., Barney, J.B., 2007. Discovery and creation: alternative theories of entrepreneurial action. Strateg. Entrep. J. 1 (1–2), 11–26.
Ampaabeng, S.K., Tan, C.M., 2013. The long-term cognitive consequences of early childhood malnutrition: the case of famine in Ghana. J. Health Econ. 32 (6),
1013–1027.
Ashton, B., Hill, K., Piazza, A., Zeitz, R., 1984. Famine in China, 1958–61. Popul. Dev. Rev. 10 (4), 613–645.
Ayala, J.-C., Manzano, G., 2014. The resilience of the entrepreneur. Influence on the success of the business. A longitudinal analysis. Journal of Economic Psychology
42, 126–135.
Baker, T., Nelson, R.E., 2005. Creating something from nothing: resource construction through entrepreneurial bricolage. Adm. Sci. Q. 50 (3), 329–366.
Baron, R.A., Tang, J., Tang, Z., Zhang, Y., 2018. Bribes as entrepreneurial actions: why underdog entrepreneurs feel compelled to use them. J. Bus. Ventur. 33 (6),
679–690.
Bauernschuster, S., Falck, O., Gold, R., Heblich, S., 2012. The shadows of the socialist past: lack of self-reliance hinders entrepreneurship. Eur. J. Polit. Econ. 28 (4),
485–497.
Black, M.M., Walker, S.P., Fernald, L.C., Andersen, C.T., DiGirolamo, A.M., Lu, C., McCoy, D.C., Fink, G., Shawar, Y.R., Shiffman, J., 2017. Early childhood devel­
opment coming of age: science through the life course. Lancet 389 (10064), 77–90.
Bonanno, G.A., 2004. Loss, trauma, and human resilience: have we underestimated the human capacity to thrive after extremely aversive events? Am. Psychol. 59
(1), 20.
Brown, J.L., Pollitt, E., 1996. Malnutrition, poverty and intellectual development. Sci. Am. 274 (2), 38–43.
Caspi, A., Harrington, H., Milne, B., Amell, J.W., Theodore, R.F., Moffitt, T.E., 2003. Children’s behavioral styles at age 3 are linked to their adult personality traits at

15
Z. Cheng, et al. Journal of Business Venturing 36 (2021) 106063

age 26. J. Pers. 71 (4), 495–514.

Chang, G.H., Wen, G.J., 1997. Communal dining and the Chinese famine of 1958–1961. Econ. Dev. Cult. Chang. 46 (1), 1–34.
Chen, Y., Zhou, L.-A., 2007. The long-term health and economic consequences of the 1959–1961 famine in China. J. Health Econ. 26 (4), 659–681.
Cheng, Z., 2014. The new generation of migrant workers in urban China. In: Cheng, Z., Wang, M.Y., Chen, J. (Eds.), Urban China in the New Era: Market Reforms,
Current State, and the Road Forward. Springer, New York, pp. 125–153.
Cheng, Z., Beresford, M., 2012. Layoffs in China's city of textiles: adaptation to change. Journal of Contemporary Asia 42 (2), 155–181.
Cheng, Z., Wang, H., 2013. Do neighbourhoods have effects on wages? A study of migrant workers in urban China. Habitat International 38, 222–231.
Cheng, Z., Smyth, R., Sun, G., 2014a. Participation and expenditure of migrants in the illegal lottery in China’s Pearl River Delta. Eurasian Geogr. Econ. 55 (4),
333–361.
Cheng, Z., Wang, H., Chen, Y., 2014b. Labor contract, trade union membership, and workplace relations: a study of migrant workers in Guangdong Province, China. In:
Hao, Z., Chen, S. (Eds.), Social Issues in China: Gender, Ethnicity, Labor, and the Environment. Springer, New York, pp. 183–206.
Cheng, Z., Nielsen, I., Smyth, R., 2015a. Determinants of wage arrears and implications for the socio-economic wellbeing of China’s migrant workers: evidence from
Guangdong province. In: Iredale, R.R., Guo, F. (Eds.), Handbook of Chinese Migration. Edward Elgar Publishing, Cheltenham, pp. 105–125.
Cheng, Z., Smyth, R., Guo, F., 2015b. The impact of China’s new Labour Contract Law on socioeconomic outcomes for migrant and urban workers. Hum. Relat. 68 (3),
329–352.
Cheung, N.W., 2013. Rural-to-urban migrant adolescents in Guangzhou, China: psychological health, victimization, and local and trans-local ties. Soc. Sci. Med. 93,
121–129.
Cheung, N.W., 2014. Social stress, locality of social ties and mental well-being: the case of rural migrant adolescents in urban China. Health Place 27, 142–154.
Chu, J., Png, I.P.L., Yi, J., 2019. Entrepreneurship and the School of Hard Knocks: Evidence From China’s Great Famine.
Chugani, H.T., Behen, M.E., Muzik, O., Juhász, C., Nagy, F., Chugani, D., 2001. Local brain functional activity following early deprivation: a study of post­
institutionalized Romanian orphans. Neuroimage 14 (6), 1290–1301.
Coale, A.J., Banister, J., 1994. Five decades of missing females in China. Demography 31 (3), 459–479.
Currie, J., Vogl, T., 2013. Early-life health and adult circumstance in developing countries. Annual Review of Economics 5 (1), 1–36.
De Clercq, D., Honig, B., 2011. Entrepreneurship as an integrating mechanism for disadvantaged persons. Entrepreneurship and Regional Development 23 (5–6),
353–372.
Deshingkar, P., Start, D., 2003. Seasonal Migration for Livelihoods in India: Coping, Accumulation and Exclusion. Overseas Development Institute, London.
Dikötter, F., 2015. Coping with famine in Communist China (1949–62). European Review of History: Revue européenne d'histoire 22 (6), 917–928.
Dong, Y., Lewbel, A., 2015. A simple estimator for binary choice models with endogenous regressors. Econ. Rev. 34 (1–2), 82–105.
Evans, G.W., Kim, P., 2013. Childhood poverty, chronic stress, self-regulation, and coping. Child Dev. Perspect. 7 (1), 43–48.
Furnham, A., Bochner, S., 1986. Culture Shock: Psychological Reactions to Unfamiliar Environments. Methuen, London.
Gagnon, J., Xenogiani, T., Xing, C., 2014. Are migrants discriminated against in Chinese urban labour markets? IZA Journal of Labor Development 3, 1–23.
Garmezy, N., 1993. Children in poverty: resilience despite risk. Psychiatry 56 (1), 127–136.
Gartner, W.B., 1985. A conceptual framework for describing the phenomenon of new venture creation. Acad. Manag. Rev. 10 (4), 696–706.
Giles, J., Park, A., Wang, M., 2019. The Great Proletarian Cultural Revolution, Disruptions to Education, and the Returns to Schooling in Urban China. Econ. Dev. Cult.
Chang. 68, 131–164.
Giulietti, C., Ning, G., Zimmermann, K.F., 2012. Self-employment of rural-to-urban migrants in China. Int. J. Manpow. 33 (1), 96–117.
Gooch, E., 2017. Estimating the long-term impact of the Great Chinese Famine (1959–61) on Modern China. World Dev. 89, 140–151.
Harris, K.M., Duncan, G.J., Boisjoly, J., 2002. Evaluating the role of “nothing to lose” attitudes on risky behavior in adolescence. Social Forces 80 (3), 1005–1039.
Haushofer, J., Fehr, E., 2014. On the psychology of poverty. Science 344 (6186), 862–867.
Heckman, J.J., 2006. Skill formation and the economics of investing in disadvantaged children. Science 312 (5782), 1900–1902.
Heckman, J., Pinto, R., Savelyev, P., 2013. Understanding the mechanisms through which an influential early childhood program boosted adult outcomes. Am. Econ.
Rev. 103 (6), 2052–2086.
Iacus, S.M., King, G., Porro, G., 2012. Causal inference without balance checking: coarsened exact matching. Polit. Anal. 20 (1), 1–24.
Kaler, S.R., Freeman, B., 1994. Analysis of environmental deprivation: cognitive and social development in Romanian orphans. Journal of Child Psychology Psychiatry
35 (4), 769–781.
Kim, S., Fleisher, B., Sun, J.Y., 2017. The long-term health effects of fetal malnutrition: evidence from the 1959–1961 China great leap forward famine. Health Econ. 26
(10), 1264–1277.
Kuang, L., Liu, L., 2012. Discrimination against rural-to-urban migrants: the role of the hukou system in China. PLoS One 7 (11).
Kung, J.K.-s., Lin, J.Y., 2003. The causes of China’s great leap famine, 1959–1961. Econ. Dev. Cult. Chang. 52 (1), 51–73.
Law, P.-l., Peng, Y., 2008. Mobile networks: migrant workers in Southern China. In: Katz, J.E. (Ed.), Handbook of Mobile Communication Studies. The MIT Press,
Cambridge, pp. 55–64.
Lévi-Strauss, C., 1962. The Savage Mind. University of Chicago Press, Chicago.
Li, C., Lumey, L., 2016. Exposure to the Chinese famine of 1959–61 in early life and current health conditions: a systematic review and meta-analysis. Lancet 388, S63.
Lin, J.Y., Yang, D.T., 2000. Food availability, entitlements and the Chinese famine of 1959–61. Econ. J. 110 (460), 136–158.
Liu, Y., Wen, J.G., Wei, X., 2014. Communal dining system and the puzzle of the Great Leap Famine: re-examine the causality between communal dining and the
famine. China Agricultural Economic Review 6 (4), 698–716.
Liu, C.Y., Ye, L., Feng, B., 2018. Migrant entrepreneurship in China: entrepreneurial transition and firm performance. Small Bus. Econ. 1–16.
Lothe, E.A., 2010. Hope or horror? A follow-up study of resilience in survivors from the 1985 Famine in Ethiopia. In: McDonald, J., Stephenson, A.M. (Eds.), The
Resilience of Hope. Rodopi, New York, pp. 165–188.
Lothe, E.A., Heggen, K., 2003. A study of resilience in young Ethiopian famine survivors. J. Transcult. Nurs. 14 (4), 313–320.
Loyalka, M.D., 2012. Eating Bitterness: Stories From the Front Lines of China’s Great Urban Migration. University of California Press, Berkeley.
Luthar, S.S., Cicchetti, D., Becker, B., 2000. The construct of resilience: a critical evaluation and guidelines for future work. Child Dev. 71 (3), 543–562.
Mancini, A.D., Bonanno, G.A., 2009. Predictors and parameters of resilience to loss: toward an individual differences model. J. Pers. 77 (6), 1805–1832.
Manning, K.E., Wemheuer, F. (Eds.), 2011. Eating Bitterness: New Perspectives on China’s Great Leap Forward and Famine. UBC Press, Vancouver.
Marx, M., Strumsky, D., Fleming, L., 2009. Mobility, skills, and the Michigan non-compete experiment. Manag. Sci. 55 (6), 875–889.
McLaughlin, K.A., Sheridan, M.A., Lambert, H.K., 2014. Childhood adversity and neural development: deprivation and threat as distinct dimensions of early ex­
perience. Neurosci. Biobehav. Rev. 47, 578–591.
Meng, X., Qian, N., Yared, P., 2015. The institutional causes of China’s great famine, 1959–1961. Rev. Econ. Stud. 82 (4), 1568–1611.
Miller, D., Le Breton-Miller, I., 2017. Underdog entrepreneurs: a model of challenge–based entrepreneurship. Entrepreneurship Theory and Practice 41 (1), 7–17.
Mu, G.M., 2018. Building Resilience of Floating Children and Left-behind Children in China: Power, Politics, Participation, and Education. Routledge, New York.
National Health Commission of China, 2018. 2018 Report of Migrant Population Development. National Health Commission of China, Beijing.
National Survey Research Centre, 2014. Report of the 2014 China Longitudinal Aging Social Survey Beijing: Renmin University of China.
Neugebauer, R., Hoek, H.W., Susser, E., 1999. Prenatal exposure to wartime famine and development of antisocial personality disorder in early adulthood. J. Am. Med.
Assoc. 282 (5), 455–462.
Nikolaev, B.N., Maldonado-Bautista, I., 2019. Personality traits and income attainment of self-employed people—accounting for model uncertainty. J. Bus. Ventur.
Insights 11, e00111.
Norton, E.C., Wang, H., Ai, C., 2004. Computing interaction effects and standard errors in logit and probit models. Stata J. 4, 154–167.
Oberg, K., 1960. Cultural shock: adjustment to new cultural environments. Practical Anthropology 4, 177–182.
Portes, A., Zhou, M., 1996. Self-employment and the earnings of immigrants. Am. Sociol. Rev. 61 (2), 219–230.

16
Z. Cheng, et al. Journal of Business Venturing 36 (2021) 106063

Prendergast, A.J., Humphrey, J.H., 2014. The stunting syndrome in developing countries. Paediatrics and International Child Health 34 (4), 250–265.
Rahmato, D., 1991. Famine and Survival Strategies: A Case Study From Northeast Ethiopia. Nordic Africa Institute, Uppsala.
Repetti, R.L., Taylor, S.E., Seeman, T.E., 2002. Risky families: family social environments and the mental and physical health of offspring. Psychol. Bull. 128 (2), 330.
Roberts, K.D., Wei, J., 1999. The floating population of Shanghai in the mid-1990s. Asian Pac. Migr. J. 8 (4), 473–510.
Rutter, M., 1989. Pathways from childhood to adult life. J. Child Psychol. Psychiatry 30 (1), 23–51.
Ryan, R.M., Lynch, J.H., 1989. Emotional autonomy versus detachment: revisiting the vicissitudes of adolescence and young adulthood. Child Dev. 340–356.
Ryff, C.D., Keyes, C.L., Hughes, D.L., 2003. Status inequalities, perceived discrimination, and eudaimonic well-being: do the challenges of minority life hone purpose
and growth? Journal of Health and Social Behavior 275–291.
Samuels, G.M., Pryce, J.M., 2008. “What doesn’t kill you makes you stronger”: survivalist self-reliance as resilience and risk among young adults aging out of foster
care. Child Youth Serv. Rev. 30 (10), 1198–1210.
Sarasvathy, S.D., 2009. Effectuation: Elements of Entrepreneurial Expertise. Edward Elgar Publishing, Cheltenham.
Saxena, S.S., Pandya, R.S.K., 2018. Gauging underdog entrepreneurship for disabled entrepreneurs. Journal of Enterprising Communities 12 (1), 3–18.
Scherer, R.F., Adams, J.S., Carley, S.S., Wiebe, F.A., 1989. Role model performance effects on development of entrepreneurial career preference. Entrepreneurship
Theory and Practice 13 (3), 53–72.
Scholte, R.S., 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. J. Health Econ. 39, 17–30.
Shen, Q., Lu, Y.-w., Hu, C.-y., Deng, X.-m., Gao, H., Huang, X.-q., Niu, E.-h., 1998. A preliminary study of the mental health of young migrant workers in Shenzhen.
Psychiatry Clin. Neurosci. 52 (S6), S370–S373.
Shiri, S., Wexler, I.D., Simonsson, P.R., Kreitler, S., 2010. Fear and hope: the relationships between traumatic stress symptoms and positive psychological impact
following exposure to politically motivated violence. In: McDonald, J., Stephenson, A.M. (Eds.), The Resilience of Hope. Rodopi, New York, pp. 153–164.
Schulz, L.C, 2010. The Dutch Hunger Winter and the developmental origins of health and disease. Proc. Natl. Acad. Sci. 107, 16757–16758.
Solomon, J., George, C., 1999. The measurement of attachment security in infancy and childhood. In: Cassidy, J., Shaver, P.R. (Eds.), Handbook of Attachment:
Theory, Research, and Clinical Applications. Vol, vol. 2. Guilford Publications, New York, pp. 383–416.
Song, S., 2013. Assessing the impact of in utero exposure to famine on fecundity: evidence from the 1959–61 famine in China. Popul. Stud. 67 (3), 293–308.
St Clair, D., Xu, M., Wang, P., Yu, Y., Fang, Y., Zhang, F., Zheng, X., Gu, N., Feng, G., Sham, P., 2005. Rates of adult schizophrenia following prenatal exposure to the
Chinese famine of 1959–1961. J. Am. Med. Assoc. 294 (5), 557–562.
Steinberg, L., 2001. We know some things: parent–adolescent relationships in retrospect and prospect. J. Res. Adolesc. 11 (1), 1–19.
Steinberg, L., Morris, A.S., 2001. Adolescent development. Annu. Rev. Psychol. 52 (1), 83–110.
Stock, J.H., Wright, J.H., Yogo, M., 2002. A survey of weak instruments and weak identification in generalized method of moments. J. Bus. Econ. Stat. 20 (4), 518–529.
Teicher, M.H., Samson, J.A., Anderson, C.M., Ohashi, K., 2016. The effects of childhood maltreatment on brain structure, function and connectivity. Nat. Rev.
Neurosci. 17 (10), 652.
Vaaler, P.M., 2011. Immigrant remittances and the venture investment environment of developing countries. J. Int. Bus. Stud. 42 (9), 1121–1149.
Vandor, P., Franke, N., 2016. Why are immigrants more entrepreneurial? Vol. 2020.
Völzke, H., Alte, D., Schmidt, C.O., Radke, D., Lorbeer, R., Friedrich, N., Aumann, N., Lau, K., Piontek, M., Born, G., 2011. Cohort profile: the study of health in
Pomerania. Int. J. Epidemiol. 40 (2), 294–307.
Waldmeir, P., 2015. Living the Chinese dream. Vol. 2019.
Wang, M.Y., Cheng, Z., Zhang, P., Tong, L., Ma, Y., 2014. Old Industrial Cities Seeking New Road of Industrialization: Models of Revitalizing Northeast China. World
Scientific, Singapore.
Wang, H., Guo, F., Cheng, Z., 2015. Discrimination in migrant workers’ welfare entitlements and benefits in urban labour market: findings from a four-city study in
China. Popul. Space Place 21 (2), 124–139.
Wang, H., Cheng, Z., Smyth, R., 2016. Are Chinese workers compensated for occupational risk? J. Ind. Relat. 58 (1), 111–130.
Westad, O.A., 2003. Decisive Encounters: The Chinese Civil War, 1946–1950. Stanford University Press, Stanford.
Wiklund, J., Nikolaev, B., Shir, N., Foo, M.-D., Bradley, S., 2019. Entrepreneurship and well-being: past, present, and future. J. Bus. Ventur. 34 (4), 579–588.
Williams, D.W., Wood, M.S., Mitchell, J.R., Urbig, D., 2019. Applying experimental methods to advance entrepreneurship research: On the need for and publication of
experiments. J. Bus. Ventur. 34, 215–223.
Xiang, B., 2004. Transcending Boundaries: Zhejiangcun: The Story of a Migrant Village in Beijing. Brill, Leiden.
Xu, H., Li, L., Zhang, Z., Liu, J., 2016. Is natural experiment a cure? Re-examining the long-term health effects of China’s 1959–1961 famine. Soc. Sci. Med. 148,
110–122.
Yang, J., 2013a. Social exclusion and young rural-urban migrants’ integration into a host society in China. The Annals of the American Academy of Political and Social
Science 648 (1), 52–69.
Yang, J., 2013b. Tombstone: The Great Chinese Famine, 1958–1962. Farrar, Straus and Giroux, New York.
Zhou, X. (Ed.), 2012. The Great Famine in China, 1958–1962: A Documentary History. Yale University Press, New Haven.

17