Small Bus Econ (2020) 55:123–151

https://doi.org/10.1007/s11187-019-00138-9

Determinants of novice, portfolio, and serial entrepreneurship:

an occupational choice approach
Emanuela Carbonara & Hien Thu Tran &
Enrico Santarelli

Accepted: 10 January 2019 / Published online: 4 February 2019

# Springer Science+Business Media, LLC, part of Springer Nature 2019

Abstract In this paper, we first develop an original associated to a higher likelihood that it is run by any
theory in which, based on their individual skills and type of habitual entrepreneur. Particularly, high entre-
the quality of their business, entrepreneurs can keep preneurial skills together with a high-quality business
their original business (and thus remain novice entre- positively influence the likelihood of an individual to be
preneurs), start and keep a new business in the same or serial or portfolio entrepreneur; (iii) ceteris paribus,
another sector along their current business (therefore firms run by serial or portfolio entrepreneurs tend to
becoming portfolio entrepreneurs), transfer or shut their stay in business longer, although high-quality ones run
original business down to either start a new one (turning by novice entrepreneurs endowed with high entrepre-
themselves into serial entrepreneurs), or enter the labor neurial skills are those with the lowest probability to
market as wage workers. We then use the insights from leave the market.
our theory to develop three main hypotheses that are
finally tested for a 10-year panel dataset (2001 to 2010) Keywords Portfolio entrepreneurship . Serial
of more than 4000 Vietnamese manufacturing firms. We entrepreneurship . Occupationalchoice . Industrialpolicy
estimate an occupational choice model and a survival
model and find that (i) a greater endowment of human JEL codes F02 . L26 . L53
capital is associated with a higher likelihood of a busi-
ness owner to become a serial or a portfolio entrepre-
neur; (ii) a higher quality of the new business is
1 Introduction

We present, and test for the case of Vietnam, an original

E. Carbonara theory of what may lead first-time (novice) entrepre-
Department of Sociology and Business Law, University of
Bologna, Strada Maggiore, 45, 40125 Bologna, Italy neurs to continue to run the original business that they
e-mail: [email protected] have launched, or to become habitual entrepreneurs
who run multiple businesses, either sequentially (serial
H. T. Tran entrepreneurs) or in parallel (portfolio entrepreneurs).1
Telfer School of Management, University of Ottawa, 55 Laurier
Ave E, Ottawa, ON K1N 6N5, Canada It is our opinion that the study of portfolio and serial
e-mail: [email protected] entrepreneurship is relatively underdeveloped in the
extant literature (Carter and Ram 2003), and therefore,
E. Santarelli (*) scope exists for strengthening both the theoretical
Department of Economics, University of Bologna, Piazza
Scaravilli, 2, 40121 Bologna, Italy
e-mail: [email protected] 1
For these three types of entrepreneurship, we follow the definitions
earlier adopted, among others, by Ucbasaran et al. (2006, 2008).
124 E. Carbonara et al.

framework and the empirical evidence upon which it When studying the occupational choices between
relies and by which it is supported.2 portfolio and serial entrepreneurship, not only it is im-
Theoretical research on habitual entrepreneurship and portant to understand what transforms a novice entre-
occupational choices amounts to very few papers (Holmes preneur into a habitual entrepreneur but also what makes
and Schmitz Jr 1990; Plehn-Dujowich 2010; Parker 2014). her/him choose portfolio entrepreneurship over serial
Of these papers, Holmes and Schmitz Jr (1990) and Plehn- and vice versa. Theory-driven findings might prove
Dujowich (2010) focus only on novice and serial entrepre- useful for both individuals facing the decision of choos-
neurs, neglecting portfolio entrepreneurship. Plehn- ing among novice, portfolio, serial entrepreneurship,
Dujowich (2010) pays particular attention to the individ- and policy makers interested in designing and
ual’s occupational choice, rather than to the equilibrium implementing entrepreneurship policies. Placing the oc-
dynamics of firm transfers. However, the linear structure of cupational choice of the entrepreneur within the entre-
the production technology in his model makes it impossi- preneurial opportunity research, the phenomenon of
ble to study portfolio entrepreneurship: in the equilibrium, multibusiness entrepreneurship has been a special inter-
all available capital is invested in a unique enterprise (the est for policy makers and practitioners, not only because
one with the highest expected profitability) and no scenario habitual entrepreneurs possess unique characteristics,
in which an entrepreneur runs multiple businesses simul- but also because, given their experience and expertise,
taneously can arise endogenously. they can identify more entrepreneurial opportunities and
Parker (2014) models portfolio entrepreneurship explic- hold more promise for success and growth as future
itly. His main focus is on the role that recognition and business owners (Westhead et al. 2005, 2009;
exploitation abilities play in the occupational choice of Lafontaine and Shaw 2016; Wang et al. 2017).
individuals. In his model, owning a portfolio of firms is To add to the body of existent literature, we first present
valuable for risk-averse entrepreneurs, as it helps them an original theoretical model that sheds new light on the
diversify their entrepreneurial risk and mitigates income phenomena of interest. It shows that, based on the quality
variance. Thus, risk aversion and diversification are the only of their current business, on the expected quality of their
motivations for portfolio entrepreneurship, and issues relat- prospective new business, on their entrepreneurial skills,
ed to firm quality and production technology are neglected. and the size of their initial investment, first-time entrepre-
All these papers are therefore somehow limited in neurs can alternatively become habitual entrepreneurs (ei-
scope and do not allow us to formulate testable predic- ther portfolio or serial) or remain novice.
tions able to exploit the unique characteristics and rich- The main contribution of this model for advancement
ness of our dataset. of our understanding of entrepreneurship is in consider-
Albeit more abundant, empirical research suffers ing portfolio and serial entrepreneurship as two separate
from the lack of suitable data and from the inadequate occupational choices, rather than putting them together
analysis of some important features of the determinants as habitual entrepreneurship. It also looks at the dynam-
of the entrepreneurial choice (Parker 2012, 2013; ics of the entrepreneur’s investment decisions, while
Sarasvathy et al. 2013; Rocha et al. 2015; Lafontaine other studies normally give a one-shot or static compar-
and Shaw 2016). Most studies, in fact, compare the ative analysis of the various occupational choices.4 Be-
choice and economic outcomes of remaining a novice sides, it is the first model explaining the occupational
with that of becoming a habitual entrepreneur, choice between novice, serial, and portfolio entrepre-
neglecting the choice between portfolio and serial neurship in terms of productive ability, capital invest-
entrepreneurship.3 ment, and the relative quality of both the current and the
new business. Although entrepreneurs are in general
2
Scholars have mostly explored either the choice between entering very heterogeneous, comprising both Bstars^ (high hu-
entrepreneurship for the first time and engaging in paid employment man capital) and Bmisfits^ (low human capital) (Åstebro
(Evans and Leighton 1989; Parker 2009; Santarelli et al. 2009) or
et al. 2011; see also Santarelli and Vivarelli 2007), we
habitual entrepreneurship in general without distinguishing between
serial and portfolio (Lafontaine and Shaw 2016). are able to prove that, while serial entrepreneurs are
3
Among those finding a superior performance of habitual entrepre- highly skilled individuals facing a new opportunity with
neurship, there are Blanchflower and Oswald (1998), Wright et al.
(1998), Åstebro and Bernhardt (2003), Chen (2013), and Rocha et al.
4
(2015). Among those providing opposite evidence, we find Rerup With the exception of Plehn-Dujowich (2010), who, however, does
(2005) and Gottschalk et al. (2017). not consider portfolio entrepreneurship.
Determinants of novice, portfolio, and serial entrepreneurship: an occupational choice approach 125

a very good expected quality, portfolio ones are gener- choice between entrepreneurship and paid employment,
ally medium- to low-skilled subjects that invest part of and several investigate the choice between novice and
their capital in a new venture to mitigate a problem of habitual entrepreneurship. However, within this line of
decreasing productivity in their current business. Novice research, only a few studies focus explicitly on both
entrepreneurs, on the other hand, are successful, skilled serial and portfolio entrepreneurship.
business owners, running a high-quality activity.
We test the predictions obtained from our model using a 2.1 Entrepreneurship or paid employment?
novel dataset from Vietnam. Two main advantages are
associated with the use of this database. First, it enables Two different lines of investigation have studied the
us to construct various measures for entrepreneurial skills choice between entrepreneurship and paid employment:
(personal characteristics of the entrepreneur: education, (i) the first one explains entrepreneurial intentions and
industry experience, managerial experience) and business the tendency to be an entrepreneur from her/his psycho-
quality (features of the business: innovation intensity, share logical, sociological, and cognitive activities (Katz
of technical and managerial employees) which inherently 1992; Kolvereid 1996; Lazear 2005; Carbonara et al.
provide more comprehensive insights. Second, while the 2018) which can be genetically inherited (Zhang et al.
majority of the extant research on the topic deals with 2009; Nicolaou and Shane 2010) and are reflected
developed countries, it enables us to address an important mainly in the willingness to take risk. This approach
research gap by considering the case of a fast-growing can be traced back, among others, to Lucas Jr. (1978),
transition economy. Coherently with our theoretical model, Kihlstrom and Laffont (1979), Cramer et al. (2002); (ii)
we prove that more human capital is associated to a higher the second one, developed by Kirzner (1979), Schultz
likelihood to be a habitual entrepreneur. Moreover, a (1980), Blanchflower and Oswald (1998), and Baumol
higher quality of the new business is associated with a (1990), assumes the entrepreneur to be an individual
higher likelihood of being habitual entrepreneurs, which able to capture opportunities and turn them into new
implies that the higher entrepreneurial skills, the larger the products.
impact of an increase in the quality of the new business on Subsequent literature has tried to apply these two
the likelihood to be serial or portfolio. Finally, habitual approaches to other types of occupational choices be-
entrepreneurs tend to stay in business longer, while novice sides the fundamental choice between entrepreneurship
entrepreneurs, when endowed with high entrepreneurial and paid employment. Contributing to the latter
skills and a high-quality business, are the category with approach, Holmes and Schmitz Jr (1990) build a theo-
the lowest probability to leave the market. retical model based on the idea that business transfers
The paper is organized as follows. Section 2 presents represent very common resource reallocations serving
a concise overview of the literature on the occupational the purpose of facilitating the division of labor among
choices of entrepreneurs. In Section 3, we set up a model entrepreneurs. According to their view, serial entrepre-
of occupational choice and survival that allows to single neurs are individuals who decide to sell (or shut down)
out the main drivers of serial and portfolio entrepreneur- their business if they find that someone else is endowed
ship. Section 4 introduces the hypotheses, based on our with a greater ability to pursue the opportunities that
model, which are tested in the empirical analysis. business is meant for. Their model is supported by
Section 5 describes the distinctive features of the entre- empirical evidence drawn mostly from Pakes and
preneurial eco-system in Vietnam, and Section 6 intro- Ericson (1988). On the same line of research, Plehn-
duces the dataset. The econometric strategy is presented Dujowich (2010) finds that, in equilibrium, a highly
in Section 7, whereas Section 8 discusses the results of skilled entrepreneur shuts down a business of low qual-
the empirical analysis. Finally, Section 9 concludes and ity to become a serial entrepreneur, whereas a low-
illustrates possible directions for future research. skilled entrepreneur shuts down a business of low qual-
ity to enter the labor market.
Contributing to the first approach, Åstebro et al.
2 Literature review (2011) present and test, using data for Korea, a model
in which individuals switching from wage employment
The literature analyzing the occupational choices of to self-employment are more often the most able indi-
entrepreneurs is extensive: many scholars study the viduals, who find self-employment more attractive
126 E. Carbonara et al.

whenever employment with a mismatched firm leads Starting a second venture indicates the existence not
them to earn a poor wage. This occurs since the conse- only of professional abilities or entrepreneurial skills,
quences of such mismatches for wages tend to persist but also Ban appropriate psychological mindset, and a
over time, pushing skilled workers to enter self- belief that entrepreneurship is indeed a preferred career
employment. choice^ (Barnir 2014, p. 3). Therefore, habitual entre-
Parker (2014) offers an important contribution to the preneurs are found to be younger when they started their
relevant literature by presenting one of the very few first business (Westhead and Wright 1998), hold a great-
theoretical studies on the decision between serial and er amount of human and social capital (Wiklund and
portfolio entrepreneurship. In particular, the latter is Shepherd 2008; Li et al. 2009; Sieger et al. 2011), and
justified by the entrepreneur’s risk aversion, which, in have a more diverse experience and more resources
special circumstances, requires portfolio diversification (Westhead et al. 2005; Amaral et al. 2011). Focusing
across multiple opportunities.5 on the role of innate ability in serial entrepreneurship,
Turning to the empirical literature, we can find sev- learning by doing is less important than selection on
eral contributions testing the validity of the two ap- innate ability in explaining both the formation and the
proaches we describe above. Again, very few empirical early performance of serial businesses. This result is
studies focus on portfolio entrepreneurs explicitly. proven by Chen (2013) for a young cohort of U.S. firms
between 1979 and 1994 and by Rocha et al. (2015) for a
2.2 The psychological traits of entrepreneurs sample of Portuguese firms between 1997 and 2003.
Moreover, Chen (2013) finds that the only exception to
Within the psychological and personal characteristics this empirical regularity occurs when an entrepreneur
approach, the existing empirical literature identifies sig- creates a new firm in an industry closely related to her/
nificant differences between novice and habitual entre- his past business experience.
preneurs. Specifically, two main factors influence an Turning to portfolio entrepreneurship specifically, the
entrepreneur’s occupational choice: (i) impetus factors existing literature has analyzed the skills developed by
(related to personal characteristics and relevant life serial entrepreneurs, as compared to portfolio ones. Un-
events) and (ii) situational factors that become relevant der the framework of the resource-based view of the
once impetus factors have been activated (such as envi- firm (Penrose 1959), two broad inputs measure business
ronmental and market changes or organizational perfor- quality: hard inputs (such as financial capital, invest-
mance), affecting her/his perceptions of desirability and ments) and soft inputs (such as technological capabili-
feasibility (Dollinger 2008). Such situational factors can ties or technical skills) (Bridge et al. 1998; Westhead
be narrowed down to the entrepreneur’s personal back- et al. 2003). Such inputs are significantly crucial to
ground and the quality of the business at the organiza- exploit entrepreneurial opportunities. With respect to
tional level. Regarding personal background, habitual soft inputs, while owning multiple businesses simulta-
entrepreneurship is often a by-product of a personal neously enables portfolio entrepreneurs to access wider
learning and development process, in which habitual sources of finance and internalize different kinds of
entrepreneurs exploit experiences and expertise from technological skills (Wright et al. 1998), owning multi-
the first business and apply them to future businesses. ple businesses sequentially allows serial entrepreneurs
to enrich their entrepreneurial experience and sharpen
5
In particular, Parker (2014) analyzes how an entrepreneur’s recogni- their technical expertise. Thus, as a result of prior busi-
tion and exploitation abilities impact on the variance and the mean of ness ownership experience gained by serial entrepre-
the payoffs generated by their entrepreneurial activity. He finds that (i)
neurs, and accumulated diverse skills and expertise
higher opportunity-exploitation ability is associated with portfolio
entrepreneurship rather than with serial and novice entrepreneurship gained by experienced portfolio ones, habitual entrepre-
(see also Parker and van Praag 2010); (ii) a sequence of opportunities neurs are more able to select the best organizational
whose returns covary negatively (diversifying opportunities) promotes
routines and capabilities oriented toward innovation
portfolio entrepreneurship at the expense of both novice and serial
entrepreneurship; (iii) synergies between successive opportunities pro- and business growth (Westhead et al. 2003). Habitual
mote portfolio entrepreneurship over novice entrepreneurship, unless entrepreneurs are then more likely to undertake several
the initial opportunity is sufficiently valuable; and (iv) individuals with innovation activities than to pursue new ventures as a
low (moderate) (high) risk aversion are more likely to be serial
(portfolio) (novice) entrepreneurs, respectively (for experimental lines one-time career change (Robson et al. 2012). With
of evidence, see Koudstaal et al. 2016). respect to hard inputs, portfolio entrepreneurs are able
Determinants of novice, portfolio, and serial entrepreneurship: an occupational choice approach 127

to utilize and leverage the internal financial resources associated subsequent hard and soft resources signifi-
from their existing business(es) to fund their subsequent cantly differentiate the business quality of habitual en-
venture, whereas serial entrepreneurs with a track entre- trepreneurs from that of novice entrepreneurs: busi-
preneurial record and collateral from the previous busi- nesses owned by habitual entrepreneurs are better in
ness can secure external finance to partly fund a subse- overcoming growing barriers and seeking a long-term
quent venture (Westhead et al. 2003). Thus, consistent competitive advantage. As a matter of fact, indication
with Blanchflower and Oswald (1998), it is expected that habitual entrepreneurs are a high-quality fraction of
that successful habitual entrepreneurs have larger novice entrepreneurs arises from a conspicuous body of
amounts of investment capital than novice ones who research suggesting the economic outperformance of
due to the lack of an established track record rely mainly habitual entrepreneurs (Wright et al. 1998; Westhead
on internal sources of capital (personal savings or loans and Wright 1999; Klepper and Simons 2000; Åstebro
from family and friends). and Bernhardt 2003; Agarwal et al. 2004; Hyytinen and
One of the main difficulties with empirical studies of Ilmakunnas 2007; Wiklund and Shepherd 2008; Chen
portfolio entrepreneurship based on the psychological 2013). Comparing the business quality of a portfolio
approach is that the level of analysis needs to be shifted with that of a serial, Westhead et al. (2005) observe,
from the firm to the individual whereas the majority of for a sample of 354 firms in Scotland, that portfolio
the data are at the firm level. Sieger et al. (2011) suggest entrepreneurs have more resources and organizational
that using the firm as the unit of analysis might lead to capabilities (skills, knowledge, etc.) than the serial. Be-
underestimating the prevalence of portfolio entrepre- sides, on average, they seem to offer more attractive
neurship, since owning and managing multiple busi- growth prospects. Tihula and Huovinen (2010), focus-
nesses could be considered a normal diversification ing their analysis on a sample of 245 Finnish firms with
strategy at the firm level. While the goal of strategic 20–49 employees, provide empirical evidence about the
diversification is to maximize managerial efficiency or presence of managerial teams in small firms. They dis-
risk management, reasons for engaging in portfolio en- tinguish solo entrepreneurs (i.e., entrepreneurs who
trepreneurship may include growth aspirations, wealth, themselves run the firm that they own) from entrepre-
value maximization, and providing career opportunities neurs who share responsibility with small groups of
for family members (Mulholland 1997). Using an inter- managers from different functional areas and other key
esting approach based on in-depth exploration of a persons (managerial team). Their findings show that
single-case study, Baert et al. (2016) identify three solo entrepreneurship is more common among novice
groups of a total of eight resource orchestration subpro- entrepreneurs, and there are more management teams in
cesses6 that allow potential portfolio entrepreneurs to firms owned by portfolio entrepreneurs than in firms
explore simultaneously numerous market opportunities owned by serial entrepreneurs.
from which they extract those that are worth exploiting.
2.4 The exit decision

2.3 Entrepreneurship and the exploitation Finally, we turn to the reasons that explain the exit of a
of opportunities firm from the market. The huge literature on firm exit
(cf. for example, Harhoff et al. 1998; Maertz and
We focus now on the second approach that studies the Campion 2004; Mason and Harrison 2006; Cumming
entrepreneurs’ ability to identify and exploit new oppor- 2008; DeTienne 2010; Cefis and Marsili 2012;
tunities and review the empirical literature within that Elfenbein and Knott 2015) has produced a wide range
specific approach. Consistent with Holmes and Schmitz of results some of which are of particular interest for our
Jr (1990), one might wonder whether the process of purposes. Entrepreneurial exit, as defined by DeTienne
division of labor among entrepreneurs produces a dif- (2010, p. 213), is Bthe process by which the founders of
ferent business quality for habitual and novice entrepre- privately held firms leave the firm they helped to create;
neurs. Indeed, these strategic occupational choices with thereby removing themselves, in varying degree, from
6 the primary ownership and decision-making structure of
These are: BAccessing, multiplying, redeploying, incubating,
decoupling, aligning, complementing, and coupling^ (Baert et al. the firm^. As a multilevel phenomenon, entrepreneurial
2016, p. 354). exit can occur in three distinct scenarios: firm exit,
128 E. Carbonara et al.

founder exit, or exit of both firm and founder (Wennberg off their business for financial rewards. The probability
et al. 2010; DeTienne and Cardon 2012). DeTienne et al. of entrepreneurial reengagement after exit is also found
(2015), p. 1) name these three exit categories as entre- to be higher for the latter case (Hessels et al. 2011;
preneurial recycling, entrepreneurial departure, and en- Amaral et al. 2011).
trepreneurial culmination, respectively. In this sense,
entrepreneurial exit can be the result of both poor and
strong performance depending on entrepreneurs’ exit 3 A model of occupational choice and survival
reasons, from poor economic performance to deliberate
exit decisions. In this section, we set up the model that we are going to
DeTienne (2010), p. 209) suggests that three forces test empirically in the remainder of the paper. Extending
explain the termination of the current business: alterna- the tradition initiated, among others, by Schultz (1980)
tive, calculative, or normative. Alternative forces refer and Holmes and Schmitz Jr (1990), we develop signif-
to Balternative opportunities^ (Maertz and Campion icant insights for how idiosyncratic individual attitudes
2004, p. 570) that could be a well-paid job opportunity and business characteristics may lead entrepreneurs to
(become a wage employee) or a new business opportu- make occupational choices in the course of their work-
nity (become a serial entrepreneur). Calculative forces ing life. Such insights will then be used to develop
refer to the ability of achieving their Bgoals and values in testable hypotheses about the qualities and the behavior
the future at their current organization^ (Maertz and of novice, serial, and portfolio entrepreneurs.
Campion 2004, p. 570). Entrepreneurs consider exit Each period, entrepreneurs are endowed with a fixed
evaluating their current venture’s viability and quality. amount K of capital (of at least one unit: K ≥ 1) and
In case of low market demand, weak technological produce using a Cobb–Douglas technology. An entre-
capabilities, or strong competitive environment, exit preneur’s profits are π(q, s) = qKs, where q is the quality
might be preferable. Normative forces, on the other of the firm and s is the entrepreneur’s skill. Thus, πs > 0
hand, refer to the individual’s perception of family or (higher skills increase profitability) and πq > 0 (higher
friends’ expectations regarding their business. Entrepre- quality businesses are more profitable). Also, πqs > 0,
neurs are less likely to close down their business if their i.e., entrepreneurial skills and business qualities are
relatives and friends have confidence or financial complements. Skill affects the marginal productivity of
investments in the business. DeTienne et al. (2015, p. capital and the returns to scale of the entrepreneur’s
6) construct a typology of exit strategies. They suggest business. We assume s ≤ 1.7
that less educated entrepreneurs with a weak psycholog- Quality q is distributed according to the distribution
ical commitment and an insignificant amount of function Q(∙) over the interval [0, ∞), whereas ability s is
invested resources may recognize that a career in entre- distributed according to the function S(∙) over the inter-
preneurship is demanding and adopt a Bvoluntary ces- val [0, 1].
sation strategy^ such as liquidation and discontinuance Time is discrete and infinite. An entrepreneur maximizes
(Harhoff et al. 1998). On the other hand, those the expected value of the future flow of profits V, where
possessing high education and high levels of innovation

∞
would apply a Bfinancial harvest exit strategy^, such as V ¼ E ∑ β πt t
ð1Þ
selling the business if substantial value accrued to the t¼0
entrepreneur (Cumming 2008; Cefis and Marsili 2012).
and 0 < β < 1 is the discount factor.
DeTienne and Cardon (2012) also find that entrepre-
Suppose that, at time t, an entrepreneur owns a busi-
neurial education and experience are positively associ-
ness characterized by quality q. She/he has then four
ated with exit through external acquisition, i.e., selling
possible options: (1) keep her/his current business (re-
off the business for financial surplus. In summary, al-
maining novice); (2) sell or close her/his current
though various reasons could explain firm exit, the
general consensus is that entrepreneurs possessing low- 7
Cases with s > 1 have been left out because they imply increasing
er education and business innovativeness would easily returns to capital. While it may be interesting to study economies in
close down their businesses by liquidation (i.e., bank- which tensions toward market concentration exist, entrepreneurship in
Vietnam is mostly characterized by SMEs. Hence, limiting our theo-
ruptcy) while those with higher education and innova- retical analysis to s ≤ 1 seems more consistent with the data we use to
tiveness would take the same decision if they could sell validate our theory.
Determinants of novice, portfolio, and serial entrepreneurship: an occupational choice approach 129

business and invest in a new one with expected quality ^q In practice, at time t, an entrepreneur choosing to
(becoming a serial entrepreneur); (3) invest in a new become portfolio maximizes the present value V P
business of quality ^q while keeping her/his original one ðq; s; ^q; γ Þ dividing her/his available capital between
(thus becoming a portfolio entrepreneur); and (4) sell or her/his current business and a new one of expected
close her/his current business and leave entrepreneur- quality ^q in order to maximize the present value from
ship, becoming a paid worker.8 this occupational choice, V P ðq; s; ^q; γ Þ.
Define V(q, s) the present value of the future flow of Define γ∗ the value of γ that maximizes (4), with
profits for an entrepreneur with skill s, currently owning 0 ≤ γ∗ ≤ 1. Notice that, with the exception of leaving
a firm of quality q who, at time t, decides to keep it entrepreneurship (option 4 above), all occupational
(novice) choices descend, as special cases, from the maximiza-
∞ ∞ qK s tion of the function in Eq. (4). In particular,
V N ðq; sÞ ¼ ∑ βt πðq; sÞ ¼ ∑ β t qK s ¼ ð2Þ V P ðq; s; ^q; 0Þ ¼ V S ðq; s; ^qÞ: the present expected value
t¼0 t¼0 1−β
of the profits from portfolio entrepreneurship when no
Similarly, an entrepreneur deciding to start a new capital is invested in the current business (γ = 0) is equal
business with expected quality ^q and leave the current to the present expected value from serial entrepreneur-
one (serial) gets a present expected value equal to ship. Similarly, V P ðq; s; ^q; 1Þ ¼ V N ðq; sÞ: the present
  ½ð1−τ ÞK s
expected value of the profits from portfolio entrepre-
V S q; s; ^
q ¼ qK s þ β ^q ð3Þ neurship when all capital is invested in the current
1−β
business (γ = 1) is equal to the present expected value
where τ is the portion of capital the entrepreneur loses from novice entrepreneurship. Since γ∗ maximizes
when setting up the new firm9 and V P ðq; s; ^q; γ Þ, it also selects the alternative that yields
∞
^
q ¼ ∫0 qdQðqÞ ¼ Eq, the expected quality of the new the highest profit. Thus, the entrepreneur chooses to
business.10 become a portfolio entrepreneur if 0 < γ∗ < 1, so that
Assume now that the entrepreneur is willing to start a she/he will operate both firms. She/he will be a serial
new business, but she/he does not want to shut down entrepreneur if γ∗ = 0, so that she/he exchanges her/his
her/his current one (portfolio entrepreneurship). She/he old business for the new one and a novice entrepreneur
therefore invests a fraction γ of her/his resources in the if γ∗ = 1, in which case she/he stays with her/his current
current business and the remaining 1 − γ in the new one, business solely.
running two businesses at the same time. The present The present expected value from the three different
occupational choices (novice, serial, and portfolio) can
expected value from this occupational choice is V P
s ∞ q½ð1−γ Þ½1−τ ÞK s
therefore be summarized by a unique equation
q; γ Þ ¼ qK s þ β qð1−β
ðq; s; ^ γK Þ
þ β ∫0 1− 8  
βdQðqÞ, which can be rewritten as >
> V S
q; s; ^
q if γ ¼ 0
  <  
  qðγK Þs V q; s; ^q; γ ¼ V P q; s; ^q; γ if 0 < γ < 1 ð5Þ
e
>
>
V P q; s; ^
q; γ ¼ qK s þ β :
1−β V N ðq; sÞ if γ ¼ 1
½ð1−γ Þð1−τ ÞK s where the superscript e stands for Bentrepreneurship.^
þβ ^q ð4Þ
1−β Finally, defining w the wage the individual earns in
each period in paid occupation, the present expected
value of leaving entrepreneurship is
8
We assume that, once made, the entrepreneur’s occupational choice is
β
not reversible. This assumption is common in occupational choice V w ðq; s; wÞ ¼ qK s þ w ð6Þ
models (see Plehn-Dujowich 2010) and is also suitable for our dataset, 1−β
which registers whether, at a given date, an entrepreneur is novice (still
owns her/his original business), serial, or portfolio but does not give Consider now an individual that, at time t, is running
account for her/his future decisions.
9
The parameter τ can be interpreted as the Bstart-up cost^ (see Plehn-
a business (this individual could be a novice entrepre-
Dujowich 2010; Holmes and Schmitz 1990). s neur with her/his first business or a serial who has
∞
10
qÞ ¼ qK s þ β ∫0 q½ð1−τ
In fact, V S ðq; s; ^
ÞK 
1−β dQðqÞ, where the founded a new business). The timing is as follows: the
s ∞ s
last integral can be rewritten as ½ð1−τ
1−β
ÞK 
∫0 qdQðqÞ ¼ ½ð1−τ ÞK 
1−β Eq. entrepreneur, who knows her/his skill s and the quality q
130 E. Carbonara et al.

of her/his current business, decides the optimal share of Notice that the entrepreneur chooses to invest in a
capital γ∗ to invest in the business (or businesses) she/he portfolio of enterprises irrespective of the qualities of the
will run in the following period. This determines the old and the new enterprise, q and ^q. When s < 1, capital
maximum amount of profits she/he could obtain from is not very productive: its marginal productivity is small
her/his entrepreneurial activity. Then, she/he makes her/ and decreasing. Transferring some capital from the old
his occupational choice decision, comparing the best to the new business increases productivity, both for the
expected outcome as an entrepreneur with employment. units remaining in the old business and for those in the
Define V ðq; s; ^
qÞ the optimal value of the occupa- new one. Intuitively, the optimal share γ∗ is larger when
tional choice problem for an entrepreneur running a q is large: the higher the quality of the current business,
business of quality q with skill s, who has found a new the larger the investment the entrepreneur is willing to
entrepreneurial opportunity with an expected quality ^q maintain in that activity. Similarly, the larger the expect-
and has to decide whether to remain novice, launch a ed quality of the new business ^q, the bigger the desired
new business, or leave entrepreneurship. investment in it.
The Bellman equation is The results in Proposition 1 imply also that the higher
^
q
and the greater the entrepreneur’s ability s, the bigger
  n   o q

V q; s; ^
q ¼ max V e q; s; ^q; γ ; V w ðq; s; wÞ ð7Þ the share of capital invested in a new business. In fact, as
s increases, the less important is the role of the marginal
productivity of capital in the entrepreneur’s decision and
To determine the value of the function V e ðq; s; ^q; γ Þ, the more she/he tends to invest in the best business
we need to compute the optimal γ∗, which then defines opportunity (in this case, the new enterprise), being the
the type of entrepreneur the individual will be in equi- need to diversify less compelling. Vice versa, when ^qq is
librium, according to expression (5). The maximization
small, γ∗ tends to increase with s, thus the share of
of the value function is then over two occupational capital left in the current business is larger, since, in this
choices: entrepreneurship as defined by the optimal γ∗ case, that is the best opportunity.
and paid work. It is therefore a dynamic programming
problem that considers the choice between discrete Proposition 2. If s = 1, the entrepreneur chooses to
actions.11 dispose of her/his original business
and to invest all her/his capital in a
new enterprise if and only if ^q > 1−τ
q
,
3.1 The entrepreneur’s choice: novice, serial,
becoming a serial entrepreneur.
or portfolio
The intuition is straightforward: being s = 1, this case
Before dealing with the maximization of the value func-
presents a constant marginal productivity of capital on
qÞ, we need to characterize the optimal γ∗.
tion V ðq; s; ^
either activity. The entrepreneur’s occupational choice is
To do so, we maximize the present expected value in
therefore driven only by relative business quality and by
expression (4) with respect to γ. The following propo-
the cost of setting up the new business.13
sitions contain our results.12

Proposition 1. If s < 1, the entrepreneur keeps a share 3.2 Entrepreneurial exit

0 < γ∗ < 1 of the available capital in the
old business but invests the remaining We can now proceed to the maximization of the value
share 1 − γ∗ in the new one, becoming a function in (7) over the two occupational choices: en-
portfolio entrepreneur. The optimal trepreneurship as defined by the optimal γ∗ and paid
share γ∗ is increasing in q and τ, de-
13
creasing in ^q. It is also increasing in s As clarified above, we do not consider cases in which s > 1. Results
would be qualitatively similar to those in Proposition 2. In fact, with
when ^qq is large and τ small. s > 1, the value V P ðq; s; ^
q; γ Þ in expression (4) would be convex in
γ. V P ðq; s; ^q; γ Þ would then be maximized at either γ = 0 or γ =
11
See Ljungqvist and Sargent (2004, chapter 6). 1, according to the relative values of q; ^q, and τ, exactly as in
12
Proofs are in Appendix 1. Proposition 2.
Determinants of novice, portfolio, and serial entrepreneurship: an occupational choice approach 131

work. In order to decide whether to remain entrepre- job.15 Looking at Fig. 1, one should notice that q^P
neurs or to take a paid job, individuals compare their increases with w and decreases with q (the quality of
equilibrium choice as entrepreneurs (novice, serial, and the original firm). In fact, an increase in w shifts V P
portfolio), with the present value of their outside option ðq; s; ^q; γ Þ up, moving the intersection point to the right,
(e.g., earnings as a paid employee). whereas an increase in q moves up both V P ðq; s; ^q; γ Þ
Our goal in this section is to study entrepreneurial and Vw(q, s, w), but V P ðq; s; ^q; γ Þ moves to a greater
exit (or survival). In this paper, we follow the definition ^P
q
provided by DeTienne (2010) and reported in extent. If we define ψP ¼ ∫0 dQð^qÞ the probability that
Section 2.4 above. Thus, entrepreneurial exit, in our the portfolio entrepreneur leaves entrepreneurship, the
model, is determined both by entrepreneurs leaving greater ^qP , the higher ψP. Intuitively, the mean survival
entrepreneurship and taking a paid job and by serial time of a business is negatively related to the probability
ones, who Bleave the firm they helped to create^ to start of exit. Thus, unsurprisingly, a portfolio entrepreneur
a new one.14 tends to Bsurvive^ longer the better the quality of the
Propositions 1 and 2 above have proven that the firms in her/his portfolio and the lower the value of the
entrepreneur chooses to run a portfolio of businesses if outside option, w.
s < 1 and chooses to be either novice or serial if s = 1, Consider then the case with s = 1. In this case, entre-
depending on the relative quality of her/his current and preneurial exit can be determined by both leaving entre-
prospective businesses. preneurship and transferring a firm to start a new one.
Consider first the case in which s < 1. In this case, We concentrate first on the choice to leave entrepreneur-
serial entrepreneurship is ruled out. We focus therefore ship. From Proposition 2, if q ≥ ^qð1−τ Þ, the entrepreneur
on the choice of leaving entrepreneurship. Figure 1 would be novice. Comparing VN(q, s) with Vw(q, s, w),
graphs the value function (7) and shows that the entre- we can see that the entrepreneur exits if q≤ qN ¼ Kw.
preneur Bsurvives^ (i.e., she/he maintains her/his job as qN
Defining ψN ¼ ∫0 dQðqÞ the probability that the novice
an entrepreneur) if ^ q ≥ ^qP and exits (going back to paid
work) if ^q< ^ qP , where ^qP is the value of ^q (the quality
of the new business) at which the individual is indiffer- 15
The function V P ðq; s; ^
q; γ Þ is increasing in ^q and linear and is
ent between being an entrepreneur and getting a paid therefore the upward sloping line in Fig. 1. In fact, applying the
β
envelope theorem, ∂V
P s
q ¼ 1−β ½ð1−γ Þð1−τ ÞK  > 0. The function
∂^
14
This is in line with the data we use in our empirical analysis, V (q, s, w) is invariant with respect to ^q. In Fig. 1, we have
w

according to which entrepreneurial exit encompasses both entrepre- neglected the case in which Vw(q, s, w) (hence w) is so low that
neurs closing their business and those transferring it. Transfer and there is no ^qP > 0 such that the entrepreneur is willing to exit if
closing down can be justified by either leaving entrepreneurship or ^q < ^qP (this case would occur if Vw(q, s, w) lies below the
starting a new firm (serial entrepreneurship). vertical intercept of V P ðq; s; ^q; γ Þ at ^q ¼ 0).

Fig. 1 Individual occupational

choices (s < 1: portfolio vs. paid
work)
132 E. Carbonara et al.

entrepreneur exits, we can conclude that the mean sur-

vival time of a business run by a novice entrepreneur
depends negatively on w and positively on its size K.
Similarly, if ^
q ≥ 1−τ
q
, the entrepreneur would be serial.
Comparing V ðq; s; ^
S
qÞ with Vw(q, s, w), we can see that
the entrepreneur exits if ^q ≤ ^qS ¼ ð1−τwÞs K . Their exit prob-
^
qS
ability is ψS ¼ ∫0 dQðq^Þ. Once again, the mean survival
time of a business run by a novice entrepreneur depends
negatively on the value of their outside option and on τ,
and positively on capital.
Following DeTienne (2010), we need to include in
survival also firm transfers of entrepreneurs that contin-
ue in their entrepreneurial role. In our model, entrepre-
neurs that can transfer or shut down their businesses (not
leaving entrepreneurship) are novice starting a new
business and becoming serials. This happens when
0
V N ðq; sÞ < V S ðq; s; ^qÞ, i.e., when q≤ q ¼ ^qð1−τ Þ. In
order for a novice to survive (in the sense that she/he
keeps her/his original business), it must be q > max {q′,
qN}: the quality q of her/his current enterprise must be
high enough to render keeping it more attractive than
both starting a new firm and leaving entrepreneurship.
If q′ < qN, the probability that a novice Bsurvives^ is
1 − ψN. If q′ > qN, the probability of survival is
q′
μN ¼ 1−∫0 dQðqÞ, which is positively influenced by q
and by τ.
We are now going to address the following questions:
(1) Which type of entrepreneur has the highest proba-
bility to stay in entrepreneurship? (2) Which type of
entrepreneur (novice, serial, portfolio) survives longer?
In dealing with question 1, we keep our assumption
s ≤ 1. In this range, we know from Propositions 1 and 2,
 
that, at γ∗, V P ðq; s; ^
q; γ * Þ≥ max V N ðq; sÞ; V S ðq; s; ^qÞ .
Figure 2a plots the present values of the profits for Fig. 2 a, b Entrepreneurs’ exit decisions and survival
portfolio and serial entrepreneurs and for paid work
against the quality of the future business ^q (the value paid work (this time plotted against the quality of the
for novice entrepreneurs does not depend on ^q and will current business q). All functions are now increasing in
be compared separately). Portfolio entrepreneurs exit if q. Figure 2b shows clearly that the exit threshold for
^
q ≤ ^qP , whereas the serial exit if ^q ≤ ^qS . Figure 2a shows portfolio entrepreneurs is lower, so that their expected
that ^qP < ^qS , which implies, by definition, that ψP < ψS. Bsurvival^ time is again longer.
In general, the probability that portfolio leave entrepre- Consider now serial and novice entrepreneurs. Com-
neurship cannot be higher than the probability that serial paring qN and ^qS (i.e., the quality thresholds that deter-
exit. Therefore, the expected Bsurvival^ time (which, in mine exit and survival of firms led by novice and serial
this case, means time in the entrepreneurial role) is entrepreneurs respectively), it is immediate to see that
higher for portfolio entrepreneurs than for serial ones. qN < ^qS , so that, being q and ^q drawn from the same
Similarly, Fig. 2b compares the expected values of distribution, novice face a smaller probability of leaving
profits of portfolio and novice entrepreneurs and for entrepreneurship than serial. However, this is not
Determinants of novice, portfolio, and serial entrepreneurship: an occupational choice approach 133

enough to conclude that the survival time is longer for 4. A higher quality of the new business (^q) increases
novice, since novices can Bexit^ also by becoming the likelihood that the entrepreneur is habitual (ei-
serial. In particular, if q′ > qN, it may well be then that ther serial or portfolio).
0
q >^ qS , in which case the survival time would be longer 5. The higher the entrepreneurial skills, the larger the
for serial entrepreneurs. This is all the more likely if τ is positive impact of an increase in the quality of the
small, if serial entrepreneurs have a larger endowment of new business on the likelihood to be a serial
capital K or a very good prospective profitability of the entrepreneur.
new enterprise, ^ q. 6. For given skill, a portfolio entrepreneur faces a
As for the impact of skill s on survival time, from longer Bsurvival^ time than her/his serial and novice
our model, it follows that a novice is a highly skilled counterparts.
entrepreneur owning a very good business, while a 7. A highly skilled novice entrepreneur with a high-
serial is a highly skilled entrepreneur facing an extreme- quality business is more likely to keep her/his cur-
ly good new opportunity, so that she/he can face the rent business than a habitual one.
cost τ of setting up a new business. A portfolio entre-
preneur is a low to averagely skilled individual mitigat- Based on the results above, we can formulate the
ing a problem of scarce productivity in her/his original following hypotheses, which we are going to test in
enterprise through diversification. our empirical analysis.
Therefore, as long as good quality and high skill From results 1 and 2, we get our first hypothesis.
induce a high probability of survival, novice entrepre- H1: High entrepreneurial skills and high quality of
neurs have the highest chances to survive, i.e., to con- new business opportunity are more likely to be associ-
tinue as entrepreneurs. ated to habitual entrepreneurship (serial and portfolio).
Results 3, 4, and 5 yield our second hypothesis.
H2: A higher quality of the new business is associat-
ed to a higher likelihood of being habitual entrepreneurs
4 Hypotheses (serial and portfolio).
Particularly, the higher the entrepreneurial skills, the
Given our results on the impact of the characteristics and larger the positive impact of an increase in the quality of
the behavior of novice, portfolio, and serial entrepre- the new business on the likelihood to be a habitual
neurs, we are now able to formulate the hypotheses that entrepreneur.
we are going to test in the empirical part of the paper. Finally, results 6 and 7 yield our third and last
From our theoretical model, we infer that, ceteris hypothesis.
paribus: H3: Firms run by portfolio entrepreneurs face the
longest survival time and those run by serial survive
1. Entrepreneurs tend to remain novice (i.e., continue longer than those run by novice. The combined effect of
to run their current businesses) when they are highly high entrepreneurial skills and good-quality business
skilled (s = 1), and the quality of their current busi- makes firms run by novice entrepreneurs to have the
nesses is high (q ≥ (1 − τ) ^q). Taking into account lowest likelihood to exit.
that divesting capital from a business and
reinvesting it in a new one is costly (τ), it might
happen that q < ^q and still the entrepreneurs prefers 5 An overview of the empirical setting—Vietnam
to retain her/his original firm.
2. Serial entrepreneurs are highly skilled entrepreneurs The transition of Vietnam to a market economy has been
(s = 1) who get a new opportunity of higher quality characterized by the emergence of a strong entrepreneurial
than their current business, so that it is convenient ecosystem (Ronnås and Ramamurthy 2001). Accordingly,
for them to seize the new opportunity notwithstand- this highly entrepreneurial country represents an ideal
ing the setup cost ( ^q ≥ ð1−τ
q
Þ). laboratory for testing the predictions of our model.
3. Portfolio entrepreneurs are low- to medium-skilled Vietnam formally started its successful transition to a
entrepreneurs (s < 1), running businesses of both market economy at the end of 1986 with the adoption of
high and low quality. a Breform and open-door policy.^ During the transition,
134 E. Carbonara et al.

its GDP grew constantly, with a growth rate that was 6 Data description
3.4% in 1986 and reached a peak of 9.5% in 1995. After
a sudden decline in 1999 (5%), the economy regained its Our analysis uses data from five waves of the Danish
growing momentum (back to 8.4% in 2005) and has International Development Agency (DANIDA) surveys
been keeping an average growth rate of around 8% per (carried out in 2002, 2005, 2007, 2009, and 2011),
year since. After joining the Association of Southeast providing detailed information on various aspects of
Asian Nations (ASEAN) in 1995, liberalizing trade with entrepreneurs and their firms. These surveys stemmed
the USA in 2001, and becoming a member of the World from the collaboration of the Central Institute for Eco-
Trade Organization (WTO) in 2007, Vietnam is nowa- nomic Management (CIEM) in Hanoi, Vietnam, and the
days a fully integrated member of the international Ministry of Labor, Invalids and Social Affairs of Viet-
business community. nam, the Department of Economics of the University of
Throughout the entire transition process, the devel- Copenhagen, and the Royal Embassy of Denmark in
opment of entrepreneurship was supported by signifi- Vietnam. The surveys are designed with the objective of
cant institutional changes, including the creation of a collecting and analyzing data representing the entire
pro-entrepreneurship legal framework and the promo- private sector in Vietnam. This means that not only large
tion of market-oriented productive capabilities (Nguyen or formally registered enterprises are interviewed, but
et al. 2008; Nguyen and Mort 2016; World Bank 2009). also a substantial number of nonregistered household/
Initially, Vietnam maintained state entrepreneurship family businesses are studied in order to gain a compre-
while developing private entrepreneurship. Privately hensive understanding of firm dynamics in Vietnam.
owned enterprises (POEs) were first created to perform Despite being carried out at different points in time,
simple economic activities and then spread into the all the surveys use the same questionnaire. Further, the
complex production processes previously dominated analysis of the development of the studied enterprises is
by SOE.16 Immediately after the enactment of the Law possible when they are traced and followed up over
on Foreign Investment in 1988, foreign entrepreneur- time. The 2011 study made use of the sample collected
ship started to emerge and took off during the Bforeign from the 2009 survey which in turn was a follow-up of
investment boom^ period in 1991–1995. These process- the 2007 one, and so on. Each survey round provides
es led to changes in the quantity and qualities of entre- financial information and economic data from the two
preneurship of each type. most recent years. Thus, in aggregate, this yields an
The result of the process described above is that unbalanced 10-year panel dataset, registering also the
Vietnam is characterized by a multifarious and rich entry of new firms and the exit of existing ones. The
entrepreneurial ecosystem, comprising firms of all sizes dataset contains a wide range of variables on the demo-
in many different sectors. Entrepreneurs have the most graphic characteristics of entrepreneurs, their technolog-
diverse background (coming from the management of ical and organizational capabilities, and the economic
previous SOEs or being self-employed start-uppers of performance of their firms (for a comprehensive
microfirms) and possess a wide range of skill, experi- understanding of the surveys, see Rand and Tarp 2007).
ence, and human capital (Nguyen et al. 2008; Nguyen Although the surveys are conducted at the firm level,
and Mort 2016; Santarelli and Tran 2013; Tran and they provide information about current entrepreneurs/
Santarelli 2014). Therefore, Vietnam provides an excel- owners and their past occupational choices. Since we
lent domain to explore the occupational choices of such concentrate precisely on the occupational choices of
a diverse and varied group of entrepreneurs and allows entrepreneurs, we focus on individuals rather than firms.
us to conduct our empirical analysis controlling for a In particular, we study the owners, rather than the man-
wide set of characteristics, both at the individual and at agers of the firms, since it is the owner who is fully
the firm level. responsible for the decisions of maintaining, closing, or
expanding the businesses. Our extracted sample consists
of 18,850 observations covering 4508 entrepreneurs.
16
Private ownership was experimentally permitted to operate in 1987– Table 1 documents the survival rate of entrepreneurs to
1988 in small-scale industries. By the promulgation of the Law on link the five surveys in our sample.
Foreign Investment in Vietnam 1987, the Company Law in 1990, and
the Law on Private Enterprises in 1991, there has been a sharp increase The sample includes 3156 novice entrepreneurs
in the number of private enterprises. (70%), 225 portfolio entrepreneurs (5%), and 1127
Determinants of novice, portfolio, and serial entrepreneurship: an occupational choice approach 135

Table 1 Survival overview

Survey 2002 2005 2007 2009 2011

Categories of surveyed firms Surveyed firms in 2002 1050

Survivors from 2002 644
New entrantsa in 2005 1706
Survivors from 2005 1663
New entrants in 2007 526
Survivors from 2007 1476
New entrants from 2009 534
Survivors from 2009 1097
New entrants in 2011 729
Total analyzed firms 1050 2350 2189 2010 1826
a
New entrants mean new firms entering the survey for the first time regardless of their type as novice, serial, or portfolio

serial entrepreneurs (25%). Appendix 2 (Tables 4, 7 Econometric strategy

5, 6, 7, 8, 9, and 10) presents and interprets some
descriptive statistics and statistical tests for the dif- To test the hypotheses listed in Section 4 empirically, we
ference in age, education, firm age, and legal own- make use of two different equations. In particular, we
ership for the novice, serial, and portfolio entrepre- test hypotheses H1 and H2 (i.e., what leads the entre-
neurs in the sample. No statistically significant preneur to become a serial or a portfolio entrepreneur
differences were found among the three groups given her/his entrepreneurial skills, the quality of her/his
with respect to their main industrial activity. Re- current business and its financial conditions) using a
garding geographical location, habitual entrepre- multinomial logit. We include the interaction of skill
neurs mainly locate in urban cities with an abun- and quality in the model, to analyze their interplay.
dance of business opportunities, whereas novice To test hypothesis H3 (survival), we employ a sur-
entrepreneurs are evenly distributed across vival equation, by means of which we investigate what
provinces. leads a novice entrepreneur to close her/his business
Information on entrepreneurial exit is obtained given her/his entrepreneurial skills, the quality and the
by tracing a firm across different survey waves. financial conditions of her/his current business. Particu-
We can observe when a firm shuts down (entrepre- larly, looking at the three-way interaction term between
neurial culmination) or changes ownership (entre- Bbeing a novice,^ Bentrepreneurial skills,^ and Bcurrent
preneurial departure). But since we are not able to business quality,^ we can determine whether a low- or
tell what happens to the exiting entrepreneur averagely skilled novice entrepreneur facing a low-
(whether she/he moves to paid employment, starts quality business will close down her/his business, while
another business or retires), we cannot observe one facing a high-quality business will keep it alive.
entrepreneurial recycling. Thus, an entrepreneurial
exit is noted when the entrepreneur declares in the 7.1 Methodology
survey that either the business had been shut down
or there is a change in ownership.17 7.1.1 Testing H1 and H2: the occupational choice
equation of novice entrepreneurs

17 Given that we look at the choice of leaving entrepre-

The surveys were designed in the way that all firms are surely traced
over time. Firms exit the surveys for a definite reported reason. This neurship with the survival equation, here the occupa-
keeps the number of enterprises being lost during the sampling to the tional outcome yi for a novice entrepreneur i is one of the
minimal. Indeed, given an average annual exit rate of around 10% three occupational alternatives (continue to be a novice
across the five surveys, only about 20% of these exits are justified with
the sentence Bno specific reason,^ which means that the enterprise entrepreneur, become a serial entrepreneur, or become a
could not be found, or the owner declined to answer the questionnaire. portfolio entrepreneur). We set yi = j if the outcome is
136 E. Carbonara et al.

the jth alternative, j = 1, 2, …,3. The probability that the survival model (logit model).19 We use observations
outcome for entrepreneur i is alternative j, conditional on all the 4508 entrepreneurs in the sample, singling
on her/his entrepreneurial skills (si), on the business out novice entrepreneurs by means of the dummy vari-
quality (qi) and the initial capital investment (Ki) of able Novice (taking value 1 when the entrepreneur is a
her/his current business is novice and 0 otherwise).

pij ¼ Prðyi ¼ jÞ ¼ F j ðsi qi K i ; θÞ; j

7.2 Variables
¼ 1; …; 3; i ¼ 1; …; N ð8Þ
7.2.1 Independent variables
where different functional forms of Fj(.) correspond to
different multinomial models. In line with a great deal of Categories of entrepreneurs A categorical variable pre-
papers addressing similar issues, we apply the multino- senting different, mutually exclusive occupational
mial logit model. choices for an entrepreneur. The variable attains value
0 if the entrepreneur is a novice, 1 if she/he is a serial
7.1.2 Testing H3: the survival of novice entrepreneurs entrepreneur and 2 if she/he is a portfolio. Serial entre-
preneurs are those who answered Byes^ to the question
We define a variable time, measuring the time from the Bbefore establishing the present enterprise, did you own
first year in which the entrepreneur is surveyed until any other enterprise?^. Portfolio entrepreneurs are those
Bdeath^ (i.e., when entrepreneurs exit entrepreneurship, who answered Byes^ to the question Bdoes the owner
closing down the business or transferring the ownership currently have more than one enterprise?^20
to another entrepreneur). Obviously, our 10-year study
is not a time span long enough to observe the death of all Serial entrepreneur/portfolio entrepreneur Dummy
the entrepreneurs in the sample, and thus, our data are variables respectively attaining value 1 if the individual
right censored. The variable exit is an indicator of is either a serial or a portfolio entrepreneur, and 0
whether time refers to business close-down or owner- otherwise.
ship change (value 1) or end of study (value 0). The The following two variables enable us to set the
survival time T may be regarded as a random variable survival time.
with a probability function U(t) and probability density Entrepreneurial exit is an event variable. It is a dum-
function u(t). The survival function or survival curve my variable that attains value 1 if either the entrepreneur
S(t) is given by: shuts down the business or leaves the business (change
in ownership), and 0 otherwise.
S ðt Þ ¼ PðT ≥ t Þ ¼ 1−U ðt Þ ð9Þ
Time is the duration variable. It measures the duration
A further function of interest for survival data is the in years from the starting year of the firm (when the firm
hazard function. It represents the instantaneous failure was established) until she/he closed it down or trans-
rate, i.e., the probability that an entrepreneur experi- ferred ownership (if that is the case during our observa-
ences the event of interest (exit) at a particular point in tion period).
time, given that the event has not yet occurred. The
f ðt Þ
hazard function is given by hðt Þ ¼ S ðt Þ. Combining it 7.2.2 Explanatory variables
with (9), we have − dlogdtðS ðtÞÞ ¼ hðt Þ, so that S(t) =
Based on the model presented above, we introduce three
exp(−H(t)), where H(t) is the integrated hazard function,
groups of explanatory variables:
or cumulative hazard function.
We deploy four different estimation models18: the 19
Given that our duration data are right censored, we cannot analyze
nonparametric Kaplan–Meier estim ator, the them by means of conventional methods such as a linear regression.
semiparametric Cox proportional hazards regression, Survival times tend to have a positively skewed distribution, which
violates the normal distribution assumption of the conventional linear
the parametric Weibull model, and the discrete-time regression.
20
These multiple enterprises could be either in the same or different
18
Appendix 3 contains a formal description of the three models industries. Thus, portfolio entrepreneurs are those running at least two
employed and of their differences. different businesses at the same time, whatever their sector.
Determinants of novice, portfolio, and serial entrepreneurship: an occupational choice approach 137

(1) Entrepreneurial skills sit, proxied by the following size of the firm, measured by the logarithm of the
four variables, measuring human capital character- total assets of the firm. (ii) Debt ratio, measured by
istics: (i) education is the number of schooling the ratio of total debt to total assets and isolates the
years for each entrepreneur. (ii) Industry experi- effect of a firm’s leverage capacity on its perfor-
ence is a dummy taking value 1 if the entrepreneur mance22; and (iii) land ownership is a dummy
has previously worked in the industry and has prior taking value 1 if the entrepreneur owns the land
knowledge of buyers and suppliers, and 0 other- housing the firm’s main production facility, and 0
wise. (iii) Management experience is a dummy otherwise.23
taking value 1 if the entrepreneur has worked in a
managerial position previously, and 0 otherwise. To study the interaction effect of entrepreneurial skills
(iv) Labor force experience is a dummy taking and business quality, we construct two composite indices
value 1 if the entrepreneur has previously worked measuring entrepreneurial skills and business quality, re-
as a paid employee, and 0 otherwise. The rationale spectively. For entrepreneurial skills, the composite Bskill^
for adopting these proxies to reflect entrepreneurial index is the sum of four dummies: an Beducation^ dummy
skills results from extensive studies supporting the attaining 1 if the entrepreneur has high education (technical
importance of human capital factors in boosting college, undergraduate, or graduate degrees), 0 if she/he
entrepreneurial alertness and performance (among has low education (no education, secondary/high school
others, Gimeno et al. 1997; Westhead et al. 2005; education); the Bindustry experience^ dummy, the
Bosma et al. 2004; Poschke 2013; Santarelli and Bmanagerial experience^ dummy, and the Blabor force
Tran 2013; Sorgner et al. 2017). experience^ dummy described above. This Bskill^ index
(2) The quality of the firm qit, proxied by its technolog- ranges from 0 (low education, no experience of any type)
ical and organizational capabilities and measured by to 4 (high education, possessing all industry, managerial,
the following three variables: (i) innovation intensity and labor force experience).
is the ratio of the investment on innovation activities The composite Bquality^ index is the sum of three
to the total annual revenue of the firm. (ii) Share of dummies: Binnovation intensity,^ Btechnical share,^ and
technical employees in the total labor force is the Bmanagerial share^ dummies. These dummies attain
ratio of the number of technical and R&D employees value 1 if the firm’s respective measure is higher than
to the total number of employees of the firm. (iii) the third quartile of the sample and 0 otherwise. The
Share of managers in the total labor force is the ratio Bquality^ index ranges from 0 (the firm’s innovation
of the number of employees in managerial positions intensity, technical share, and managerial share are all
to the total number of employees of the firm.21 The below the sample median values) to 3 (all innovation
rationale for adopting these three proxies could be intensity, technical share, and managerial share are all
originally traced back to the resource-based view of above the third quartile values). Interaction terms are
the firm (Penrose 1959) suggesting that the quality of created using these two composite indices.
a firm results from unique attributes of its competitive
resources. Following Bridge et al. (1998) and 7.2.3 Control variables
Westhead et al. (2003), we value business quality
from two broad inputs: hard inputs, i.e., investment Besides age, gender, and tenure24 of the entrepreneur,
into innovation activities, and soft inputs, i.e., tech- we include firm age, the age of the current firm. We then
nological and managerial knowledge and expertise. isolate the ownership type of the current firm, creating
(3) Total capital Kit in each year, proxied by the loga-
rithm of the total assets of the firm. Three variables 22
While Opler and Titman (1994) find that highly leveraged firms lose
will be adopted to measure the capital investment a substantial market value and impose greater risks to owners and
creditors than their more conservatively financed competitors. Teece
of the entrepreneur: (i) firm size is the economic (1982) finds that debts reduce the chances of bankruptcy through
flexible asset deployment.
21 23
Given the structure of our dataset, we can observe only the quality of Since skill affects productivity and productivity is an aspect of the
the current business which is the new business for serial entrepreneurs, firm (technological) quality, there might be a positive correlation
the main business for portfolio entrepreneurs, and the old business for between skill and firm quality.
24
novice entrepreneurs in case they decide to exit entrepreneurship Duration, in terms of number of years, of the period that an individ-
during the observation years. ual stays in the current business or in entrepreneurship.
138 E. Carbonara et al.

dummies private firms, partnership/cooperatives, limit- entrepreneurs. They are more likely to spend longer
ed liability (all taking value 1 when the firm takes the years in education, possess richer experience from the
corresponding ownership type and 0 otherwise). Finally, industry, and have worked previously as managers.26
a dummy is added to take any divergence or mismatch For instance, if the entrepreneur were to increase her/his
arising from different surveys into account. schooling years by 1 year while holding all other vari-
ables in the model constant, the multinomial log-odds
for serial entrepreneurs relative to novice would increase
by 0.052 units, whereas those for portfolio relative to
8 Estimation results
novice would increase by 0.084 units. This result seems
to be quantitatively similar for serial and portfolio indi-
Appendix 3 (Table 11) presents the descriptive statistics
viduals. Maybe, qualitatively, we can see a better sig-
and pairwise correlation matrix of all the adopted vari-
nificance for serial (all three indicators are significant at
ables. The average exit rate of firms is about 15%, which
the 1% level, whereas only two are significant at the 1%
is relatively low compared to other studies about Viet-
level for portfolio, while managerial experience is sig-
nam (e.g., Vijverberg and Haughton 2002). On average,
nificant at 5% only). Overall, however, our hypothesis
our sampled entrepreneurs are 45 years old and have
H1, according to which high entrepreneurial skills are
spent more than 11 years in education.25 We can see
more likely to be associated to habitual entrepreneur-
from the pairwise correlation matrix, out of 136 pairwise
ship (serial and portfolio), is strongly supported.
correlations, 80 are statistically significant at 1% signif-
Previous experience as a paid employee has a nega-
icance level. However, most of them are very small with
tive impact on the entrepreneur’s motivation to take
correlation coefficients below 0.3. The only two
further entrepreneurial risks rebuilding a failed business
pairwise correlations that are greater than 0.3 are as
or expanding their current one.
follows: (i) the one between debt ratio and investment
With respect to the effect of organizational quality on
capital (0.43), which indicates that external loans are an
the likelihood of engaging in habitual entrepreneurship,
important source of finance for our entrepreneurs’ in-
the impact of innovation intensity is not statistically
vestments, and (ii) the correlation between education
significant. We also allowed for a nonlinear relationship
and firm size (0.318), which suggests that highly edu-
by including its squared value in the regression, but this
cated entrepreneurs are better in attracting financial
did not affect the result. The share of technical em-
resources.
ployees in the total firm labor force exerts a significant
and positive impact for both serial and portfolio entre-
8.1 The occupational choice equation preneurs. However, the share of managerial employees
exerts opposite influences on serial and portfolio entre-
Table 2 presents the results of the occupational choice preneurship: an increase in managerial employees has a
model when Bnovice entrepreneur^ is used as the base positive impact on portfolio, but a negative impact on
category. Regressors are jointly statistically significant serial entrepreneurship. An increase in managerial ex-
at the 1% level across the two treatments we estimate. pertise, while enabling portfolio entrepreneurs to man-
Particularly, the second model (presented in the last two age and allocate resources efficiently among old and
columns) includes the interaction term between the en- new businesses, seems to be an impediment to serial
trepreneur’s skills and her/his current business quality. entrepreneurs’ agility in exploiting new business oppor-
In general, entrepreneurial skills have a significantly tunities. Thus, the first part of hypothesis H2, stating that
positive effect on the propensity to engage in habitual a higher quality of the new business is associated to a
entrepreneurship. Education, industry experience, and higher likelihood of being habitual, is confirmed. The
managerial experience all increase the propensity of an effect seems to be stronger for portfolio. In fact, three
occupational transition to habitual entrepreneurship, and out of three proxies for business quality exert a positive
this effect is consistent for both serial and portfolio
25 26
Ronnas and Ramamurthy (2001: 328) describe a typical urban The two-tailed t-test for the comparison of mean ‘education’ be-
entrepreneur in Vietnam as Ba middle-aged male with at least 10 years tween novice and habitual entrepreneurs significantly rejects the equal-
of education and prior experience in similar fields in a position of ity and supports the superiority of habitual entrepreneurs. Results are
responsibility^. reported in Appendix 2.
Determinants of novice, portfolio, and serial entrepreneurship: an occupational choice approach 139

Table 2 Occupational choice for novice entrepreneurs

Variables Multinomial logit model

Serial Portfolio Serial Portfolio

Schooling years 0.052** (0.007) 0.084** (0.013) 0.056** (0.008) 0.097** (0.015)
Industry experience 0.401** (0.055) 0.805** (0.084) 0.442** (0.061) 0.893** (0.089)
Labor market experience − 6.229** (0.189) − 1.305** (0.086) − 6.182** (0.191) − 1.208** (0.092)
Managerial experience 0.591** (0.118) 0.223* (0.094) 0.627** (0.120) 0.306* (0.156)
Innovation intensity 0.036 (0.091) 0.187 (0.141) 0.102 (0.097) 0.352* (0.151)
Innovation intensity squared − 0.0103 (0.018) − 0.037 (0.029) − 0.021 (0.019) − 0.063* (0.031)
Share of technical employees 5.064** (0.571) 9.239** (0.585) 5.513** (0.614) 10.048** (0.651)
Share of managerial employees − 0.486** (0.157) 1.188** (0.199) − 0.407* (0.162) 1.333** (0.205)
Entre skills * business quality 0.065* (0.028) 0.113** (0.039)
Land ownership − 0.113* (0.048) 0.764** (0.087) − 0.112* (0.048) 0.768** (0.087)
Debt ratio − 0.127 (0.158) 0.451* (0.243) − 0.126 (0.158) 0.463* (0.243)
Investment capital 0.000** (0.000) 0.000** (0.000) 0.000** (0.000) 0.000** (0.000)
Investment capital squared − 0.000 (0.000) − 0.000** (0.000) − 0.000 (0.000) − 0.000** (0.000)
Firm size 0.550** (0.086) − 1.149** (0.117) 0.553** (0.086) − 1.144** (0.117)
Firm size squared − 0.108** (0.017) 0.204** (0.021) − 0.110** (0.017) 0.198** (0.021)
Firm age − 0.000 (0.002) − 0.013* (0.006) − 0.004* (0.002) − 0.013* (0.006)
Age − 0.000 (0.002) − 0.009* (0.004) − 0.001 (0.002) − 0.009* (0.004)
Gender − 0.148** (0.051) 0.160* (0.086) − 0.174** (0.047) 0.158* (0.086)
Tenure − 0.008* (0.004) 0.045** (0.007) − 0.008* (0.004) 0.045** (0.007)
Private firms 0.286** (0.104) − 0.119 (0.178) 0.281** (0.106) − 0.125 (0.178)
Partnership/cooperatives 0.044 (0.159) 0.323 (0.249) 0.032 (0.159) 0.304 (0.247)
Limited liability 0.417** (0.102) − 0.155 (0.173) 0.414** (0.102) − 0.155 (0.173)
Joint stock − 0.252 (0.216) − 0.194 (0.304) − 0.265 (0.217) − 0.191 (0.303)
Intercept − 0.847** (0.186) − 3.424** (0.304) − 0.914** (0.188) − 3.601** (0.311)
Likelihood ratio test χ2(42) = 8658.99** χ2(44) = 8668.81
No. of observations 16,505 16,505

Standard errors are in parentheses

* significant at the 5% level; ** significant at the 1% level

impact on the likelihood to be portfolio, whereas only are positively associated with the likelihood of being a
two proxies (innovation intensity and technical re- habitual entrepreneur.
sources) exert a positive effect on the likelihood to be Regarding the impact of investment capital K, we
serial. consistently find a positive but almost zero effect on
Unsurprisingly, the interaction between entrepreneur- the probability to be habitual. This is consistent with our
ial skills and business quality in columns 3 and 4 is model, where capital does not play a direct role in the
positive and statistically significant for both serial and entrepreneur’s occupational choice. Land ownership has
portfolio entrepreneurship. Holding other factors in the a negative effect on the likelihood of being serial but a
model constant, a high skill combined with a high- positive effect on the likelihood of being portfolio.
quality business increases the multinomial log-odds to Owning the land housing the main production facility
be a serial entrepreneur by 0.065 units, and to be a would deter the motivation to relaunch a new venture to
portfolio by 0.113 units. These findings support the reap short-term profit opportunities, since owners can
second part of hypothesis H2, proposing that high en- rent out the land they own to substitute for their entre-
trepreneurial skills combined with high business quality preneurial income. However, land ownership is
140 E. Carbonara et al.

favorable to business expansion for portfolio entrepre- remaining in their business, entrepreneurs are signifi-
neurs. They can exploit different entrepreneurial oppor- cantly less likely to close down their business. The
tunities on the land they own without any concerns for Nelson–Aalen estimator in Fig. 3b serves to obtain the
the rent or leasing-related risks. The statistically signif- cumulative hazard function by summing up the values
icant and positive parameters of the debt ratio in the of the hazard functions over time. The Kaplan–Meier
portfolio equations indicate that indebted entrepreneurs survival curve (Fig. 3c) presents the survival duration of
are more likely to become portfolio since they can use the sample. It starts from 1 because we have the full
their leverage conditions to undertake other profitable sample of firms at the beginning. Over time, they grad-
but risky investments. ually exit the market, thus the curve steps downwards.
With respect to control variables, while the current There are around 75% of firms which are still in busi-
firm of serial entrepreneurs is larger in terms of the size ness after 2.5 years. After 6 years, this figure reduces to
of firm labor force than the one of novice counterparts, 25%. These findings confirm an empirical regularity
portfolio entrepreneurs have relatively smaller-sized largely supported in the relevant literature (cf. among
firms, possibly because they own and run multiple busi- others, Audretsch et al. 1999).
nesses. The effect of firm size is however nonlinear. Figure 3d presents the survival curves for busi-
There is a consistent finding that younger firms and nesses of novice, serial, and portfolio entrepreneurs,
younger entrepreneurs are more likely to be habitual respectively. A test for the equality of the survival
(but the effect is insignificant for serial entrepreneurs). functions for the three groups of businesses rejects
Male entrepreneurs are more likely to take risks in the null hypothesis of equality at 1% significance
expanding the current business (and thus become port- level.27 The survival estimates for businesses of
folio entrepreneurs), whereas female are more flexible serial entrepreneurs are slightly longer than those
and adaptive in learning from past failures to relaunch a of novice ones. Portfolio entrepreneurs, on the other
new business (and thus become serial entrepreneurs). hand, remain in their businesses much longer than
Serial entrepreneurs tend to have shorter tenure and their counterparts do. Nearly 75% of the portfolio
portfolio ones have longer tenure in running their cur- entrepreneurs are still in business after 6 years.
rent businesses. Finally, the businesses of serial entre- These findings confirm the first part of our hypoth-
preneurs are more likely to be private or limited liability esis H3 (portfolio entrepreneurs face longer survival
companies, whereas portfolio entrepreneurs do not seem time). Moreover, the difference between serial and
to favor any specific ownership type. novice is small and serial tend to survive longer. The
longer survival time of serial entrepreneurs could be
8.2 The survival equation due to the very high quality of the firm run by a
serial entrepreneur or by her/his larger endowment
We acknowledge the limitation that we cannot fully of capital, in line with our Hypothesis H3. Estimates
confirm the exit of individual entrepreneurs, since the in Section 8.1 have shown that habitual entrepre-
code to trace the survival of observations is at the firm neurs tend to be associated to high business quality,
level. Thus, the survival of entrepreneurs is mainly although they do not seem to be systematically
reflected by the survival of their firms. Nevertheless, associated to higher capital investment than novice.
the survey has a question on the likelihood of a change Table 3 presents the estimation results of the
in ownership, so we can also observe owner exit even if semiparametric Cox model, the parametric Weibull
her/his firm is still in operation. In other words, an model, and the discrete-time survival model. We present
entrepreneurial exit in our study means (i) the entrepre- the results of the first 2 models in two specifications:
neur shuts down the business or (ii) she/he transfers the coefficients (log hazard ratios) and exponentiated coef-
business (e.g., selling or retiring). This is in line with the ficients (hazard ratios). Results are generally quite con-
definition of entrepreneurial exit by DeTienne (2010). sistent across models.
Figure 3 presents different graphs for the nonpara- Ceteris paribus, habitual entrepreneurs remain in
metric estimation of firm survival in the sample. In Fig. their business longer than novice ones do. Regard-
3a, the exit rate increases sharply for the first 4 years in ing entrepreneurial skills, while the length of
business, then it is steadily flat around 4 years before
27
falling sharply after 4 years. In other words, after 4 years chi2(2) = 165.14; Pr > chi2 = 0.0000
Determinants of novice, portfolio, and serial entrepreneurship: an occupational choice approach 141

Smoothed hazard estimate Nelson-Aalen cumulative hazard estimate

1.20
.15

1.00
0.80
.14

0.60
0.40
0.20
.13

0 2 4 6 8 0 2 4 6 8 10
analysis time analysis time

a. Graph of hazard ratio b. Nelson-Aalen cumulative hazard curve

Kaplan-Meier survival estimate Kaplan-Meier survival estimates

1.00
1.00

0.75
0.75

0.50
0.50

0.25
0.25

0.00

0 2 4 6 8 10
analysis time
0.00

category = 0 category = 1
0 2 4 6 8 10 category = 2
analysis time

c. Kaplan-Meier survival curve d. Kaplan-Meier survival curve for novice (category=0),

serial (category=1), portfolio (category=2)
Fig. 3 Survival analysis. a Graph of hazard ratio. b Nelson–Aalen cumulative hazard curve. c Kaplan–Meier survival curve. d Kaplan–
Meier survival curve for novice (category = 0), serial (category = 1), portfolio (category = 2)

education is significantly and negatively associated stronger motivation to remain in their business if
with an entrepreneur’s propensity to shut down their they have both high entrepreneurial skills and high
business, industry experience and management ex- business quality. However, considering the two-way
perience are surprisingly found to stimulate entre- interactions among our three interested variables, we
preneurs to close their business sooner than their find some influential results: (i) novice entrepre-
unexperienced peers. In particular, from the Cox neurs being endowed with higher skills can maintain
model, compared to inexperienced entrepreneurs, their business longer; but (ii) novice entrepreneurs
those having industry experience may exit entrepre- owning high-quality business are more likely to exit
neurship at twice the hazard rate per unit time, or the market; and (iii) highly skilled entrepreneurs
those with managerial experience will have higher unsurprisingly have no reasons to exit entrepreneur-
hazard rate about 1.2 times. ship if they own a high-quality business. In order to
To test the last part of Hypothesis H3, we con- find statistical pieces of evidence to either support or
sider the three-way interaction among entrepreneur- reject our Hypothesis H3, we analyze marginal ef-
ial skills, firm-level business quality, and the likeli- fects of a novice entrepreneur exiting entrepreneur-
hood of being novice entrepreneurs. The statistically ship at two specific values of entrepreneurial skills
significant and negative parameters of the interac- and business quality that are one standard deviation
tion across all methodological treatments indicate above the mean and one standard deviation below
that, ceteris paribus, novice entrepreneurs have a the mean. We consistently find the negative and
142 E. Carbonara et al.

Table 3 Survival of entrepreneurs

Variables Cox pro. hazard model Weibull model Discrete-time survival model
(xtlogit)
Log hazard rate Hazard rate Log hazard rate Hazard rate

Novice entrepreneurs 0.221* (0.096) 1.247* (0.120) 0.213* (0.096) 1.235* (0.119) 0.944** (0.189)
Schooling years − 0.024** (0.008) − 0.977** (0.008) − 0.026** (0.008) − 0.973** (0.008) − 0.081** (0.017)
Industry experience 0.547** (0.062) 1.729** (0.107) 0.546** (0.062) 1.726** (0.107) 1.087** (0.131)
Employee experience − 0.027 (0.073) − 1.027 (0.075) − 0.039 (0.073) − 1.039 (0.076) − 0.203 (0.144)
Managerial experience 0.337** (0.078) 1.401** (0.109) 0.368** (0.078) 1.446** (0.113) 0.921** (0.196)
Innovation intensity 0.203** (0.035) 1.225** (0.043) 0.219** (0.035) 1.244** (0.043) 2.562** (0.281)
Share of technical 0.736** (0.083) 2.089** (0.174) 0.768** (0.083) 2.157** (0.181) 4.223** (0.458)
employees
Share of managerial − 1.184** (0.176) − 0.306** (0.053) − 1.222** (0.177) − 0.294** (0.052) − 1.366** (0.283)
employee
Novice * entre. skills * bus − 0.046* (0.016) − 0.898* (0.033) − 0.062* (0.027) − 0.994* (0.035) − 0.184 (0.126)
quality
Novice * entre. skills − 0.085 (0.060) − 0.956 (0.065) − 0.185** (0.061) − 1.204** (0.074) − 0.068 (0.063)
Novice * bus quality 0.061* (0.008) 0.104 (0.009) 0.211** (0.086) 1.234** (0.106) 0.088 (0.092)
Entre. skills * bus quality − 0.019 (0.038) − 0.963 (0.037) − 0.033 (0.039) − 0.966 (0.038) − 0.335** (0.084)
Land ownership 0.164** (0.043) 1.178** (0.052) 0.165** (0.044) 1.180** (0.052) 0.941** (0.084)
Debt ratio 0.739** (0.121) 2.094** (0.252) 0.758** (0.122) 1.710** (0.203) 0.701** (0.126)
Investment capital − 0.000** (0.000) − 0.999** (0.000) − 0.000** (0.000) − 0.999** (0.000) − 0.000** (0.000)
Investment capital squared 0.000** (0.000) 1** (0.000) 0.000** (0.000) 1** (0.000) 0.000** (0.000)
Firm size 0.116** (0.025) 0.908** (0.002) 0.113** (0.026) 1.119** (0.029) 0.505** (0.059)
Firm age − 0.096** (0.003) − 0.908** (0.003) − 0.101** (0.003) − 0.904** (0.003) − 0.017** (0.004)
Age − 0.020** (0.002) − 0.979** (0.002) − 0.021** (0.002) − 0.979** (0.002) − 0.004 (0.004)
Intercept − 3.702** (0.179) − 0.025** (0.004) − 0.199** (0.317)
LR statistic χ2(18) = 2728.12** χ2(18) = 3121.45** χ2(18) = 419.46
Observations 4317 4317 18,382

Standard errors are in parentheses; 4317 observations

* significant at the 5% level; ** significant at the 1% level

statistically significant relationship between being that innovative firms tend to incur high sunk costs
novice entrepreneurs with high skills and high busi- from large R&D investments and operate in a very
ness quality (one standard deviation above the competitive market portions where the innovative-
mean) and exit rate. Therefore, when owning a ness of rivals might be even greater than theirs.
good-quality business, novice entrepreneurs with Another noteworthy finding is the consistent,
high entrepreneurial skills are less likely to exit negative, and nonlinear effect of capital investment
entrepreneurship, which is exactly our Hypothesis on firm exit (although not economically strong).
H3. Other things kept constant, more initial investment
Our survival analysis yields other interesting re- capital leads to higher survival propensity; but once
sults. The technological quality of the firm plays a an optimal level of investment capital is reached, a
negative role in sustaining its survival. The hazard further increase in investment does not lead to a
rate is almost 2.1 times higher for entrepreneurs with proportional increase in entrepreneurial survival.
larger technological resources (Cox model). The re- Finally, heavily indebted entrepreneurs have a
lationship between innovation intensity and the like- lower propensity of survival and larger (in terms of
lihood of survival is not necessarily positive, given labor force) and younger firms are more vulnerable
Determinants of novice, portfolio, and serial entrepreneurship: an occupational choice approach 143

to bankruptcy, while older entrepreneurs seem to Testing the relevance of our theoretical setting for Viet-
survive longer.28 nam, we obtain some noteworthy findings that support our
model propositions. First, there is a significant effect of
entrepreneurial skills (proxied by education, industry ex-
perience, and managerial experience) in increasing the
9 Conclusions propensity of an occupational transition to habitual entre-
preneurship. In addition, both serial and portfolio entrepre-
We studied why some entrepreneurs become serial or neurs are endowed with stronger human capital than their
portfolio while others remain novice. We propose a theo- novice counterparts, and an increase in the quality of the
retical model illustrating the occupational choice of a new business increases the likelihood that it is run by a
novice entrepreneur, given her/his entrepreneurial skills, habitual entrepreneur, with a stronger incidence on portfo-
the quality of the current business and the expected quality lio entrepreneurs. Second, the interaction between entre-
of the prospective future business. In equilibrium, an en- preneurial skills and business quality supports our theory
trepreneur maintains her/his business if she/he is highly that high skills and business quality are generally associ-
skilled and the business is sufficiently profitable. Highly ated to a habitual entrepreneur. Third, novice entrepre-
skilled individuals tend to become serial entrepreneurs neurs’ motivation to remain in their business is stronger
when they encounter a very profitable new opportunity. if they have both high entrepreneurial skills and rich
If skill is either low or medium, they tend to become technical and managerial resources.
portfolio entrepreneurs to mitigate the scarce productivity Our empirical analysis unveils other interesting re-
of their first business. While novice and serial entrepre- sults. While the businesses of both serial and portfolio
neurs generally own good-quality firms, portfolio entre- entrepreneurs are endowed with significant technologi-
preneurs can run both high- and low-quality businesses. cal resources, only portfolio entrepreneurs are more
We also analyzed the interaction between entrepreneur- motivated to invest heavily in managerial resources for
ial skills and the quality of the new business of a habitual greater adaptability and to absorb knowledge spillovers
entrepreneur. We prove that a high quality of the new in new industries/businesses. Investment capital does
business (measured in terms of innovation intensity and not seem to impact on the likelihood of habitual entre-
share of technical and managerial employees in total em- preneurship. Land ownership and leveraging debts are
ployment) increases the likelihood that the entrepreneur favorable to business expansion activities of portfolio
becomes habitual and that such a positive impact tends to entrepreneurs only. Regarding survival, in general, firms
be larger for portfolio entrepreneurs. Moreover, due to the run by habitual entrepreneurs have a longer duration
complementarity between skills and firm quality in the than those run by novice ones. Technological quality is
profit function, the higher the entrepreneurial skills are, negatively associated with firm survival, which is con-
the larger is the positive impact of an increase in the quality sistent with many studies supporting the low survival
of the new business on the likelihood to be a habitual chance of young innovative companies or new
entrepreneur. Finally, a highly skilled novice entrepreneur technology–based firms (e.g., Santarelli and Tran
facing a good-quality business tends to keep it. 2016). Although these capabilities enable entrepreneurs
28
to be responsive to dynamic changes in the market and
We have also performed two robustness checks: one for the occu-
pational choice equation and one for the survival equation. In the first
thus transform their entrepreneurial efforts into observ-
one, we have tried to separate the effects of different skills: education able material outcomes, they are quite costly and chal-
and managerial experience. We have distinguished serial entrepreneurs lenging to develop and manage, especially for inexperi-
between those who launched a new business after closing their previ-
enced novice entrepreneurs. However, managerial ex-
ous one and those who acquired an existing one. Literature (Parker and
Van Praag 2010) claimed that the first ones require education, whereas pertise from the firm management team does help en-
the second require managerial experience. We have then excluded trepreneurs lengthen their survival duration.
those who acquired an existing enterprise and rerun the multinomial
As in other Western advanced countries, habitual entre-
logit regression. In the second robustness check, we sort entrepreneurs
according to their exit modality: some went bankrupt and some others preneurship is common in transition and developing coun-
sold their enterprise. While bankruptcy is typically attributed to skill tries (Smallbone and Welter 2001; Akhter et al. 2016),
and financial resources (or lack thereof), ownership transfer may not. including Vietnam. There are numerous avenues where
We exclude those who sold their firm from the sample and rerun the
survival analysis. The results of the two robustness checks (available further research is required to fully understand the theoret-
upon request) are consistent with our main results. ical and phenomenon-based aspects of both types of
144 E. Carbonara et al.

habitual entrepreneurship. Although we believe that the γ ðs−1Þ ^q

s−1
¼ ð1−τ Þs ð11Þ
key findings of our study (i.e., habitual entrepreneurs ð1−γ Þ q
possess higher entrepreneurial skills, higher business qual-
ity, and longer survival duration) could be generalized Solving Eq. (11) for γ always yields an interior
ðs−1Þ
γ
across other transition countries, their development might solution γ∗ ∈ (0, 1). In fact, when s < 1, R1 ðγ Þ ¼ ð1−γ Þs−1
be limited by unstable macroeconomic conditions and is always decreasing in γ, with lim R1 ðγ Þ ¼ þ∞, and
weak institutional mechanisms. In fact, it is largely ac- γ→0

knowledged in the relevant literature that the institutional R1(1) = 0. Since the right-hand side of (11) is constant
environment and the entrepreneurial ecosystem play im- with respect to γ and positive, there will always exist a
ðs−1Þ
portant roles for shaping entrepreneurial behavior also in γ
value 0 < γ∗ < 1 such that ð1−γ Þs−1
¼ ^qq ð1−τ Þs .
transition economies (Acs et al. 2008; Carbonara et al. The second-order condition for a maximum is
2016). Based on our and previous findings, we therefore
plan to extend our research in three directions. ðs−1ÞqK s γ s−2 þ ðs−1Þ^qð1−τ Þð1−γ Þs−2 < 0 ð12Þ
First, we would like to investigate how both formal and
informal institutions influence the occupational choice of Being (s − 1) < 0, the second-order condition is al-
entrepreneurs. Second, we plan to analyze the linkage ways satisfied and γ∗ is a maximum.
between habitual entrepreneurship and productive entre- Finally, totally differentiating the first-order condi-
preneurship with the aim of shedding light on how habitual tion in (7) with respect to γ, ^q; and τ yields
entrepreneurs produce growth and contribute to the nation- s−1
al economic development. Our model can in fact be further dγ * ð1−τ Þs ð1−γ * Þ
¼h i < 0 ð13Þ
generalized to provide testable predictions about capital d^q qγ *ðs−2Þ þ ^qð1−τ Þs ð1−γ * Þs−2 ðs−1Þ
accumulation and firm growth/size.
Third and most important, we would like to study the since (s − 1) < 0. Similarly
case of entrepreneurs bouncing back from business fail- s−1
dγ * γ*
ure. In the present analysis, we have not distinguished ¼−h i >0 ð14Þ
dq qγ * ð s−2Þ qð1−τ Þs ð1−γ * Þs−2 ðs−1Þ
þ^
the motivations pushing an entrepreneur to become
serial: learning from failure, experimenting with luck,
or truly exploiting a newly recognized opportunity. This and
will be a promising field for future research. s−1
dγ * q^ð1−τ Þs−1 ð1−γ Þ*
¼h i < 0 ð15Þ
dτ qγ *ðs−2Þ þ ^qð1−τ Þs ð1−γ * Þs−2 ðs−1Þ
Acknowledgements Previous versions of this paper have been
presented at the 8th Annual Conference for the Academy of
Thus, γ∗ is increasing in q and decreasing in ^q and τ
Innovation and Entrepreneurship (Toronto, Ryerson University)
and at seminars held at University of Saarland, University of when s < 1.
Luxembourg, University of Trento, and Bucerius Law School Finally,
(Hamburg). We thank seminar participants and, in particular, h h i
*ðs−1Þ s−1
Bettina Müller for helpful suggestions.
dγ * qγ lnðγ Þ−^qð1−τ Þs ð1−γ * Þ lnð1−γ Þ þ ln 1−τ
¼− h i
ds qð1−τ Þs ð1−γ * Þs−2 ðs−1Þ
qγ *ðs−2Þ þ ^

ð16Þ
Appendix 1. Proofs
being γ and τ ∈(0, 1), then all the natural logarithms in
expression (16) are negative numbers. Therefore, the
Proof of proposition 1
numerator is positive if ^q is large relative to q and τ is
small. Given the minus sign in front of the r.h.s., a
Differentiating expression (4) with respect to γ, the first-
positive numerator implies that γ∗ is increasing in s
order condition for the maximization of V P ðq; s; ^q; γ Þ is
(since s < 1 and the denominator is negative). Vice
 s−1 versa, the numerator is negative if ^q is small relative to
qK s γ *ðs−1Þ −^
q½ð1−τ ÞK s 1−γ * ¼0 ð10Þ
q and τ is large. This implies that, in this case, γ∗ is
which can be rewritten as decreasing in s.■
Determinants of novice, portfolio, and serial entrepreneurship: an occupational choice approach 145

Proof of Proposition 2

Given s = 1, the value V P ðq; s; q^; γ Þ in expression (4) is

linear in γ and can be written as
  qK βð1−γ ÞK h i
V P q; 1; ^q; γ ¼ − q−ð1−τ Þ^q ð17Þ
1−β 1−β

which is increasing in γ (and thus maximized at γ =

1) if q ≥^
qð1−τ Þ, whereas it is decreasing in γ (and thus
maximized at γ = 0) if q < ^qð1−τ Þ.■

Appendix 2. t test on the equality of means of age,

firm age, and education among novice, serial,
and portfolio; and tabulation of ownership types
adopted by novice, serial, and portfolio

Analysis of the statistical differences in age, firm age,

and education among novice, serial, and portfolio
entrepreneurs

Table 4 Age—novice and habitual entrepreneurs

Group Mean Std. err. Std. dev. 95% conf. interval

Novice 47.46 0.092 10.457 47.389 47.75

Habitual 46.291 0.131 9.915 46.036 46.547
Combined 47.175 0.075 10.309 47.027 47.322
Diff 1.277 0.163 0.958 1.597
t = 7.838 Degrees of freedom = 18,668
Ha: diff < 0 Ha: diff! = 0 Ha: diff > 0
Pr (T < t) = 1.000 Pr (|T| > |t|) = 0.000 Pr (T > t) = 0.000

We can reject the hypothesis that the mean age of novice and habitual entrepreneurs is equal, and support the alternative hypothesis that mean
age of novice entrepreneurs is higher than that of habitual ones. In other words, habitual entrepreneurs averagely are younger than novice
ones

Table 5 Age—serial and portfolio entrepreneurs

Group Mean Std. err. Std. dev. 95% conf. interval

Serial 46.035 0.143 9.991 45.755 46.316

Portfolio 47.687 0.313 9.377 47.072 48.303
Combined 46.291 0.131 9.915 46.036 46.547
Diff − 1.652 0.360 − 2.358 − 0.946
t = − 4.587 Degrees of freedom = 5766
Ha: diff < 0 Ha: diff! = 0 Ha: diff > 0
Pr (T < t) = 0.000 Pr (|T| > |t|) = 0.000 Pr (T > t) = 1.000

We can reject the hypothesis that the mean age of serial and portfolio entrepreneurs is equal, and support the alternative hypothesis that mean
age of portfolio entrepreneurs is higher than that of serial ones. In other words, habitual entrepreneurs averagely are older than serial ones
146 E. Carbonara et al.

Table 6 Firm age—novice and habitual entrepreneurs

Group Mean Std. err. Std. dev. 95% conf. interval

Novice 13.191 0.088 10.128 13.016 13.364

Habitual 13.709 0.142 10.789 13.431 13.987
Combined 13.349 0.075 10.338 13.201 13.497
Diff − 0.518 0.163 − 0.838 − 0.198
t = − 3.173 Degrees of freedom = 18,828
Ha: diff < 0 Ha: diff! = 0 Ha: diff > 0
Pr (T < t) = 0.001 Pr (|T| > |t|) = 0.002 Pr (T > t) = 0.9992

We can reject the hypothesis that the mean firm age of habitual and novice entrepreneurs is equal, and support the alternative hypothesis that
mean firm age of habitual entrepreneurs is higher than that of novice ones. In other words, firms of habitual entrepreneurs averagely are older
than those of novice ones

Table 7 Firm age—serial and portfolio entrepreneurs

Group Mean Std. err. Std. dev. 95% conf. interval

Serial 13.426 0.157 10.987 13.118 13.735

Portfolio 15.251 0.318 9.498 14.626 15.875
Combined 13.709 0.142 10.789 13.431 13.987
Diff − 1.824 0.392 − 2.593 − 1.055
t = − 4.651 Degrees of freedom = 5764
Ha: diff < 0 Ha: diff! = 0 Ha: diff > 0
Pr (T < t) = 0.000 Pr (|T| > |t|) = 0.000 Pr (T > t) = 1.000

We can reject the hypothesis that the mean firm age of serial and portfolio entrepreneurs is equal, and support the alternative hypothesis that
mean firm age of portfolio entrepreneurs is higher than that of serial ones. In other words, firms of portfolio entrepreneurs averagely are older
than those of serial ones

Table 8 Education—novice and habitual entrepreneurs

Group Mean Std. err. Std. dev. 95% conf. interval

Novice 11.107 0.043 3.316 11.021 11.192

Habitual 11.653 0.031 3.476 11.593 11.713
Combined 11.484 0.025 3.438 11.435 11.533
Diff 0.545 0.054 0.439 0.652
t = 10.049 Degrees of freedom = 18,682
Ha: diff < 0 Ha: diff! = 0 Ha: diff > 0
Pr (T < t) = 1.000 Pr (|T| > |t|) = 0.000 Pr (T > t) = 0.000

We can reject the hypothesis that the number of schooling years of novice and habitual entrepreneurs is equal, and support the alternative
hypothesis that the number of schooling years of habitual entrepreneurs is higher than that of novice ones. In other words, habitual
entrepreneurs averagely have higher education than novice ones do
Determinants of novice, portfolio, and serial entrepreneurship: an occupational choice approach 147

Table 9 Education—serial and portfolio entrepreneurs

Group Mean Std. err. Std. dev. 95% conf. interval

Serial 10.984 0.047 3.312 10.891 11.077

Portfolio 11.776 0.108 3.254 11.562 11.989
Combined 11.107 0.043 3.316 11.021 11.192
Diff − 0.791 0.120 − 1.027 − 0.556
t = − 6.586 Degrees of freedom = 5768
Ha: diff < 0 Ha: diff! = 0 Ha: diff > 0
Pr (T < t) = 0.000 Pr (|T| > |t|) = 0.000 Pr (T > t) = 1.000

We can reject the hypothesis that the number of schooling years of serial and portfolio entrepreneurs is equal, and support the alternative
hypothesis that the number of schooling years of portfolio entrepreneurs is higher than that of serial ones. In other words, portfolio
entrepreneurs averagely have higher education than serial ones do

Table 10 Tabulation of legal ownership

Types of entrepreneurs Household (%) Private (%) Partnership (%) Ltd liability (%) Joint stock (%)

Novice 73.93 7.86 3.47 12.74 1.96

Serial 72.72 9.02 2.67 14.07 1.35
Portfolio 74.94 6.94 2.91 12.53 2.24
Total 73.67 8.12 3.24 13.07 1.81

In general, there is a consistent pattern of preference over legal ownership choices among novice, serial, and portfolio entrepreneurs. More
than 70% of sampled entrepreneurs regardless of their types adopt household ownership for their firms. The next common legal ownership
type is limited liability, accounting for over 12% of sampled firms

Appendix 3. The survival equation. A formal The Cox hazard function for entrepreneur i is hi(t) =
analysis h0(t) exp(si qi Ki, θ), where h0(t) is the baseline hazard
function when all covariates are zero. The parameters θ
In this paper, we use three different estimation models: are estimated by maximizing the partial log likelihood
!
the nonparametric Kaplan–Meier estimator, the expðsi qi K i ; θÞ
semiparametric Cox proportional hazards regression, given by ∑ log , where the first
f ∑i∈rð f Þ expðsi qi K i ; θÞ
and the parametric Weibull model.
summation is over all failures exit f, and the second
The Kaplan–Meier estimator is a nonparametric es-
summation is over all entrepreneurs r(f) who are still at
timator of the survival function S(t). If all the failure
risk at the time of failure.
times are ordered and labeled t(j) such that t(1) ≤ t(2) … ≤
The Weibull model assumes the Weibull distribution
dj
t(n), the estimator is given by S^ðt Þ ¼ ∏ 1− , for T with parameters λ and p, denoted T~W(λ, p), if
jjt ð jÞ ≤ t nj
Tp~E(λ). The cumulative hazard is H(t) = (λt)p, the sur-
where dj is the number of entrepreneurs who exit at time vivor function S(t) = exp(−(λt)p), and the hazard is λ(t) =
t(j), and njis the number of entrepreneurs who are still in λpptp − 1. Both semiparametric and parametric survival
the business at the time and are therefore still Bat risk^ of models are estimated by maximum likelihood estima-
experiencing exit. tion technique.
 148

Appendix 4

Table 11 Summary statistics and matrix of correlation

Mean Std. dev (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16)

(1) 0.145 0.352 1.00

(2) 0.271 0.444 − 0.003 1.00
(3) 0.047 0.212 − 0.03* 0.002 1.00
(4) 11.484 3.438 0.015 0.086* 0.01* 1.00
(5) 0.310 0.462 0.02* 0.06* 0.04* − 0.05* 1.00
(6) 0.472 0.499 − 0.001 − 0.056* − 0.03* 0.26* 0.051* 1.00
(7) 0.074 0.261 0.057* − 0.003 0.001 0.17* 0.018 0.11* 1.00
(8) 0.235 0.742 0.003 0.014 0.011 0.039* 0.124* 0.018 0.042* 1.00
(9) 0.116 0.275 0.055* 0.023* − 0.04* − 0.11* 0.161* − 0.059* 0.028* − 0.065* 1.00
(10) 0.218 0.209 − 0.05* − 0.04* 0.014* − 0.09* − 0.19* − 0.05* − 0.105* − 0.127* − 0.227* 1.00
(11) 10.893 3.903 − 0.015 − 0.006 0.013 0.059* − 0.012 0.015 − 0.001 0.079* 0.041* − 0.04* 1.00
(12) 0.581 0.493 − 0.011 − 0.004 0.061* − 0.16* − 0.036* − 0.058* − 0.057* 0.009 0.059* 0.096* − 0.05* 1.00
(13) 0.075 0.157 0.023* 0.001 0.024* 0.081* − 0.003 0.027* 0.062* 0.025* 0.009 − 0.062* 0.43* − 0.056* 1.00
(14) 1.827 1.064 0.106* 0.021* − 0.07* 0.318* 0.039* 0.099* 0.205* 0.06* 0.075* − 0.273* 0.108* − 0.212* 0.182* 1.00
(15) 45.39 10.43 0.031* − 0.06* 0.011 − 0.06* − 0.126* 0.081* 0.087* − 0.024* − 0.062* 0.051* 0.012 0.052* 0.019* 0.013 1.00
(16) 0.735 0.441 − 0.01* − 0.05* − 0.01 0.095* 0.03* 0.08* 0.016* 0.019* 0.05* − 0.085* 0.000 − 0.027* − 0.01 0.077* 0.016 1.00

Observations 18,850
* significant at the 1% level
(1) Exit; (2) serial entrepreneurs; (3) portfolio entrepreneurs; (4) education (number of schooling years); (5) industry experience; (6) employee experience; (7) management experience; (8)
innovation intensity; (9) the share of technical employees in the total labor force; (10) the share of managers in the total labor force; (11) initial investment capital; (12) land ownership; (13)
debt ratio; (14) firm size; (15) age; (16) gender
E. Carbonara et al.
Determinants of novice, portfolio, and serial entrepreneurship: an occupational choice approach 149

Publisher’s note Springer Nature remains neutral with regard to constitutions influence entrepreneurship. Small Business
jurisdictional claims in published maps and institutional affiliations. Economics, 47(1), 139–162. https://doi.org/10.1007
/s11187-016-9715-z.
Carbonara, E., Santarelli, E., Obschonka, M., Tran, H. T., Potter,
J., & Gosling, S. (2018). Agency culture, constitutional pro-
visions and entrepreneurship: a cross-country analysis.
References Industrial and Corporate Change, 27(1), 507–524.
https://doi.org/10.1093/icc/dtx047.
Carter, S., & Ram, M. (2003). Reassessing portfolio entrepreneur-
Acs, Z. J., Desai, S., & Hessels, J. (2008). Entrepreneurship, ship. Small Business Economics, 21(4), 371–380. https://doi.
economic development and institutions. Small Business org/10.1023/A:1026115121083.
Economics, 31(3), 219–234. https://doi.org/10.1007 Cefis, E., & Marsili, O. (2012). Going, going, gone: exit forms and
/s11187-008-9135-9. the innovative capabilities of firms. Research Policy, 41(5),
Agarwal, R., Echambadi, R., Franco, R., & Sarkar, M. (2004). 795–807. https://doi.org/10.1016/j.respol.2012.01.006.
Knowledge transfer through inheritance: spin–out genera- Chen, J. (2013). Selection and serial entrepreneurs. Journal of
tion, development and survival. Academy of Management Economics and Management Strategy, 22(2), 281–311.
Journal, 47(4), 501–522. https://doi.org/10.2307/20159599. https://doi.org/10.1111/jems.12016.
Akhter, N., Sieger, P., & Chirico, F. (2016). If we can’t have it, then Cramer, J. S., Hartog, J., Jonker, N., & Van Praag, C. M. (2002).
no one should: shutting down versus selling in family busi- Low risk aversion encourages the choice for entrepreneur-
ness portfolios. Strategic Entrepreneurship Journal, 10(4), ship: an empirical test of a truism. Journal of Economic
371–394. https://doi.org/10.1002/sej.1237. Behavior and Organization, 48(1), 29–36. https://doi.
Amaral, A. M., Baptista, R., & Lima, F. (2011). Serial entrepre- org/10.1016/S0167-2681(01)00222-0.
neurship: impact of human capital on time to re-entry. Small Cumming, D. J. (2008). Contracts and exits in venture capital
Business Economics, 37(1), 1–21. https://doi.org/10.1007 finance. Review of Financial Studies, 21(5), 1947–1982.
/s11187-009-9232-4. https://doi.org/10.1093/rfs/hhn072.
Åstebro, T., & Bernhardt, I. (2003). Start-up financing, owner DeTienne, D. R. (2010). Entrepreneurial exit as a critical compo-
characteristics, and survival. Journal of Economics and nent of the entrepreneurial process: theoretical development.
Business, 55(4), 303–319. https://doi.org/10.1016/S0148- Journal of Business Venturing, 25(2), 203–215. https://doi.
6195(03)00029-8. org/10.1016/j.jbusvent.2008.05.004.
Åstebro, T., Cheng, J., & Thompson, P. (2011). Stars and misfits: DeTienne, D. R., & Cardon, M. S. (2012). Impact of founder
self-employment and labor market frictions. Management experience on exit intentions. Small Business Economics,
Science, 57(11), 1999–2017. https://doi.org/10.1287 38(4), 351–374. https://doi.org/10.1007/s11187-010-9284-5.
/mnsc.1110.1400. DeTienne, D. R., McKelvie, A., & Chandler, G. N. (2015).
Audretsch, D., Santarelli, E., & Vivarelli, M. (1999). Start-up size Making sense of entrepreneurial exit strategies: a typology
and industrial dynamics: some evidence from Italian and test. Journal of Business Venturing, 30(2), 255–272.
manufacturing. International Journal of Industrial https://doi.org/10.1016/j.jbusvent.2014.07.007.
Organization, 17(7), 965–983. https://doi.org/10.1016 Dollinger, M. (2008). Entrepreneurship strategies and resources
/S0167-7187(98)00002-2. (4th ed.). Lombard: Marsh Publications.
Baert, C., Meuleman, M., Debruyne, M., & Wright, M. (2016). Elfenbein, D. W., & Knott, A. M. (2015). Time to exit: rational,
Portfolio entrepreneurship and resource orchestration. behavioral, and organizational delays. Strategic Management
Strategic Entrepreneurship Journal, 10(4), 346–370. Journal, 36(7), 957–975. https://doi.org/10.1002/smj.2262.
Barnir, A. (2014). Pre-venture managerial experience and new Evans, D. S., & Leighton, L. S. (1989). Why do smaller firms pay
venture innovation. An opportunity costs perspective. less? Journal of Human Resources, 24(2), 299–318.
Management Decision, 52(10), 1981–2001. https://doi. https://doi.org/10.2307/145858.
org/10.1108/MD-03-2014-0158. Gimeno, J., Folta, T., Cooper, A., & Woo, C. (1997). Survival of
Baumol, W. J. (1990). Entrepreneurship—productive, unproduc- the fittest? Entrepreneurial human capital and the persistence
tive, and destructive. Journal of Political Economy, 98(5), of underperforming firms. Administrative Science Quarterly,
893–921. https://doi.org/10.1086/261712. 42(4), 750–783. https://doi.org/10.2307/2393656.
Blanchflower, D. G., & Oswald, A. J. (1998). What makes an Gottschalk, S., Greene, F. J., & Müller, B. (2017). The impact of
entrepreneur? Journal of Labor Economics, 16(1), 26–60. habitual entrepreneurial experience on new firm closure out-
https://doi.org/10.1086/209881. comes. Small Business Economics, 48(2), 303–321.
Bosma, N., Van Praag, C. M., Thurik, A. R., & de Wit, G. (2004). https://doi.org/10.1007/s11187-016-9780-3.
The value of human and social capital investments for the Harhoff, D., Stahl, K., & Woywode, M. (1998). Legal form,
business performance of startups. Small Business Economics, growth and exit of West-German firms—empirical results
23(2), 227–236. https://doi.org/10.1023 for manufacturing, construction, trade and service industries.
/B:SBEJ.0000032032.21192.72. Journal of Industrial Economics, 46(4), 453–488. https://doi.
Bridge, S., O’Neil, K., & Cromie, S. (1998). Understanding org/10.1111/1467-6451.00083.
enterprise, entrepreneurship and small business. Hessels, J., Grilo, I., Thurik, R., & Van Der Zwan, P. (2011).
Basingstoke: Macmillan. Entrepreneurial exit and entrepreneurial engagement.
Carbonara, E., Santarelli, E., & Tran, H. T. (2016). De jure deter- Journal of Evolutionary Economics, 21(3), 447–471.
minants of new firm formation: how the pillars of https://doi.org/10.1007/s00191-010-0190-4.
150 E. Carbonara et al.

Holmes, T. J., & Schmitz, J. A., Jr. (1990). A theory of entrepre- growth in Vietnam. International Review of Business
neurship and its application to the study of business transfers. Research Papers, 4(2), 203–216.
Journal of Political Economy, 98(2), 265–294. https://doi. Nicolaou, N., & Shane, S. (2010). Entrepreneurship and occupa-
org/10.1086/261678. tional choice: genetic and environmental influences. Journal
Hyytinen, A., & Ilmakunnas, P. (2007). Entrepreneurial aspira- of Economic Behavior and Organization, 76(1), 3–14.
tions: another form of job search? Small Business Economics, https://doi.org/10.1016/j.jebo.2010.02.009.
29(1), 63–80. https://doi.org/10.1007/s11187-005-4783-5. Opler, T. C., & Titman, S. (1994). Financial distress and corporate
Katz, J. A. (1992). A psychosocial cognitive model of employ- performance. Journal of Finance, 49(3), 1015–1040.
ment status choice. Entrepreneurship Theory and Practice, https://doi.org/10.1111/j.1540-6261.1994.tb00086.x.
17(1), 29–37. https://doi.org/10.1177 Pakes, A., & Ericson, R. (1988). Empirical implications of alter-
/104225879201700104. native models of firm dynamics. Manuscript. Madison: Univ.
Kihlstrom, R. E., & Laffont, J. J. (1979). General equilibrium Wisconsin; published in Journal of Economic Theory. 1998,
entrepreneurial theory of firm formation based on risk aver- 79(1): 1–45. doi: https://doi.org/10.1006/jeth.1997.2358.
sion. Journal of Political Economy, 87(4), 719–748. Parker, S. C. (2009). Why do small firms produce the entrepre-
https://doi.org/10.1086/260790. neurs? Journal of Socio-Economics, 38(2), 484–494.
Kirzner, I. M. (1979). Perception, opportunity and profit. Chicago: https://doi.org/10.1016/j.socec.2008.07.013.
University of Chicago Press. Parker, S. C. (2012). Theories of entrepreneurship, innovation and
Klepper, S., & Simons, K. (2000). The making of an oligopoly: the business cycle. Journal of Economic Surveys, 26(3), 377–
firm survival and technological change in the evolution of the 394. https://doi.org/10.1111/j.1467-6419.2012.00728.x.
U.S. tire industry. Journal of Political Economy, 108(4), Parker, S. C. (2013). Do serial entrepreneurs run successively
728–760. https://doi.org/10.1086/316100. better businesses? Evidence from the PSID. Journal of
Kolvereid, L. (1996). Prediction of employment status choice Business Venturing, 28(5), 652–666. https://doi.org/10.1016
intentions. Entrepreneurship Theory and Practice, 21(1), /j.jbusvent.2012.08.001.
47–57. https://doi.org/10.1177/104225879602100104. Parker, S. C. (2014). Who become serial and portfolio entrepre-
Koudstaal, M., Sloof, R., & Van Praag, C. M. (2016). Risk, neurs? Small Business Economics, 43(4), 887–898.
uncertainty, and entrepreneurship: evidence from a lab-in- https://doi.org/10.1007/s11187-014-9576-2.
the-field experiment. Management Science, 62(10), 2897– Parker, S. C., & Van Praag, C. M. (2010). Group status and
2915. https://doi.org/10.1287/mnsc.2015.2249. entrepreneurship. Journal of Economics and Management
Strategy, 19(4), 919–945. https://doi.org/10.1111/j.1530-
Lafontaine, F., & Shaw, K. (2016). Serial entrepreneurship: learn-
9134.2010.00280.x.
ing by doing? Journal of Labor Economics, 34(2), S217–
Penrose, E. T. (1959). The theory of the growth of the firm. New
S254. https://doi.org/10.1086/683820.
York: Oxford University Press.
Lazear, E. (2005). Entrepreneurship. Journal of Labor Economics,
Plehn-Dujowich, J. (2010). A theory of serial entrepreneurship.
23(4), 649–680. https://doi.org/10.1086/491605.
Small Business Economics, 35(4), 377–398. https://doi.
Li, S., Schulze, W., & Li, Z. (2009). Plunging into the sea, again? org/10.1007/s11187-008-9171-5.
A study of serial entrepreneurship in China. Asia Pacific Poschke, M. (2013). Who becomes an entrepreneur? Labor market
Journal of Management, 26(4), 667–680. https://doi. prospects and occupational choice. Journal of Economic
org/10.1007/s10490-008-9102-7. Dynamics and Control, 37(3), 693–710. https://doi.
Ljungqvist, L., & Sargent, T. J. (2004). Recursive macroeconomic org/10.1016/j.jedc.2012.11.003.
theory. 2nd ed. Cambridge: MIT Press. Rand, J., & Tarp, F. (2007). Characteristics of the Vietnamese
Lucas, R. E., Jr. (1978). On the size distribution of business firms. business environment: evidence from a SME survey in
Bell Journal of Economics, 9(2), 508–523. https://doi. 2005. A study prepared under Component 5 – Business
org/10.2307/3003596. Sector Research of the Danida Funded Business Sector
Maertz, C. P., & Campion, M. A. (2004). Profiles in quitting: Program Support.
integrating process and content turnover theory. Academy of Rerup, C. (2005). Learning from past experience: footnotes on
Management Journal, 47(4), 566–582. https://doi. mindfulness and habitual entrepreneurship. Scandinavian
org/10.2307/20159602. Journal of Management, 21(4), 451–472. https://doi.
Mason, C. M., & Harrison, R. T. (2006). After the exit: acquisi- org/10.1016/j.scaman.2005.09.010.
tions, entrepreneurial recycling and regional economic devel- Robson, P. J. A., Akuetteh, C. K., Westhead, P., & Wright, M.
opment. Regional Studies, 40(1), 55–73. https://doi. (2012). Export intensity, human capital and business owner-
org/10.1080/00343400500450059. ship experience. International Small Business Journal, 30(4),
Mulholland, K. (1997). The family enterprise and business strate- 367–387. https://doi.org/10.1177/0266242610364426.
gy. Work Employment and Society, 11(4), 685–711. Rocha, V., Carneiro, A., & Varum, C. A. (2015). Serial entrepre-
https://doi.org/10.1177/0950017097114005. neurship, learning by doing and self-selection. International
Nguyen, Q. A., & Mort, G. S. (2016). Economic reform and Journal of Industrial Organization, 40(1), 91–106.
entrepreneurship in Vietnam: a policy perspective. In J. https://doi.org/10.1016/j.ijindorg.2015.04.001.
Ateljević & J. Trivić (Eds.), Economic development and Ronnås, P., & Ramamurthy, B. (2001). Entrepreneurship in
entrepreneurship in transition economies (pp. 109–127). Vietnam: transformation and dynamics. Copenhagen: NIAS
Cham: Springer. Publishing.
Nguyen, T. H., Alam, Q., & Prajogo, D. (2008). State and market Santarelli, E., & Tran, H. T. (2013). The interplay of human and
relationships: public financial policy support for SMEs social capital in shaping entrepreneurial performance: the
Determinants of novice, portfolio, and serial entrepreneurship: an occupational choice approach 151

case of Vietnam. Small Business Economics, 40(2), 435–458. Vijverberg, W., & Haughton, J. (2002). Household enterprises in
https://doi.org/10.1007/s11187-012-9427-y. Vietnam: survival, growth, and living standards. Discussion
Santarelli, E., & Tran, H. T. (2016). Diversification strategies and paper WPS 2773, The World Bank.
firm performance in Vietnam: evidence from parametric and Wang, T., Thornhill, S., & De Castro, J. O. (2017). Entrepreneurial
semi-parametric approaches. The Economics of Transition, orientation, legitimation, and new venture performance.
24(1), 31–68. https://doi.org/10.1111/ecot.12082. Strategic Entrepreneurship Journal, 11(4), 373–392.
Santarelli, E., & Vivarelli, M. (2007). Entrepreneurship and the https://doi.org/10.1002/sej.1246.
process of firms’ entry, survival and growth. Industrial and Wennberg, K., Wiklund, J., DeTienne, D.R. & Cardon, M.S. (2010).
Corporate Change, 16(3), 455–488. https://doi.org/10.1093 Reconceptualizing entrepreneurial exit: Divergent exit routes
/icc/dtm010. and their drivers. Journal of Business Venturing, 25(4), 361–
Santarelli, E., Carree, M., & Verheul, I. (2009). Unemployment 375. https://doi.org/10.1016/j.jbusven.2009.01.001.
and firm entry and exit: an update on a controversial rela- Westhead, P., & Wright, M. (1998). Novice, portfolio, and serial
tionship. Regional Studies, 43(8), 1061–1073. https://doi. founders: are they different? Journal of Business Venturing,
org/10.1080/00343400801968361. 13(3), 173–204. https://doi.org/10.1016/S0883-9026(97
Sarasvathy, S. D., Menon, A. R., & Kuechle, G. (2013). Failing )90002-1.
firms and successful entrepreneurs: serial entrepreneurship as Westhead, P., & Wright, M. (1999). Contributions of novice,
a temporal portfolio. Small Business Economics, 40(2), 417– portfolio and serial founders in rural and urban areas.
434. https://doi.org/10.1007/s11187-011-9412-x. Regional Studies, 33(2), 157–173.
Schultz, T. W. (1980). Investment in entrepreneurial ability.
Westhead, P., Ucbasaran, D., & Wright, M. (2003). Differences
Scandinavian Journal of Economics, 82(4), 437–448.
between private firms owned by novice, serial and portfolio
https://doi.org/10.2307/3439676.
entrepreneurs: implications for policy-makers and practi-
Sieger, P., Zellweger, T., Nason, R. S., & Clinton, E. (2011).
tioners. Regional Studies, 37(2), 187–200. https://doi.
Portfolio entrepreneurship in family firms: a resource-based
org/10.1080/0034340022000057488.
perspective. Strategic Entrepreneurship Journal, 5(4), 327–
Westhead, P., Ucbasaran, D., Wright, M., & Binks, M. (2005).
351. https://doi.org/10.1002/sej.120.
Novice, serial and portfolio entrepreneur behaviour and con-
Smallbone, D., & Welter, F. (2001). The distinctiveness of entre-
tributions. Small Business Economics, 25(2), 109–132.
preneurship in transition economies. Small Business
https://doi.org/10.1007/s11187-003-6461-9.
Economics, 16(4), 249–262. https://doi.org/10.1023
/A:1011159216578. Westhead, P., Ucbasaran, D., & Wright, M. (2009). Information
Sorgner, A., Fritsch, M., & Kritikos, A. (2017). Do entrepreneurs search and opportunity identification: the importance of prior
really earn less? Small Business Economics, 49(2), 251–272. business ownership experience. International Small Business
https://doi.org/10.1007/s11187-017-9874-6. Journal, 27(6), 659–680. https://doi.org/10.1177
Teece, D. J. (1982). Towards an economic theory of the multi- /0266242609344255.
product firm. Journal of Economic Behavior and Wiklund, J., & Shepherd, D. A. (2008). Portfolio entrepreneur-
Organization, 3(1), 39–63. https://doi.org/10.1016/0167- ship: habitual and novice founders, new entry, and mode of
2681(82)90003-8. organizing. Entrepreneurship Theory and Practice, 32(4),
Tihula, S., & Huovinen, J. (2010). Incidence of teams in the firms 701–725. https://doi.org/10.1111/j.1540-6520.2008.00249.x.
owned by serial, portfolio and first-time entrepreneurs. World Bank (2009). Vietnam development report 2010: modern
International Entrepreneurship and Management Journal, institutions. Joint Donor Report to the Vietnam Consultative
6(3), 249–260. https://doi.org/10.1007/s11365-008-0101-4. Group Meeting, World Bank.
Tran, H. T., & Santarelli, E. (2014). Capital constraints and the Wright, M., Westhead, P., & Sohl, J. (1998). Editors’ introduction:
performance of entrepreneurial firms in Vietnam. Industrial habitual entrepreneurs and angel investors. Entrepreneurship
and Corporate Change, 23(3), 827–864. https://doi. Theory and Practice, 22(4), 5–21. https://doi.org/10.1177
org/10.1093/icc/dtt030. /104225879802200401.
Ucbasaran, D., Westhead, P., & Wright, M. (2006). Habitual Zhang, Z., Zyphur, M., Narayanan, J., Arvey, R., Chaturvedi, S.,
entrepreneurs. Cheltenham: Edward Elgar. Avolio, B., Lichtenstein, B., & Larsson, G. (2009). The
Ucbasaran, D., Alsos, G. A., Westhead, P., & Wright, M. (2008). genetic basis of entrepreneurship: effects of gender and per-
Habitual entrepreneurs. Foundations and Trends in sonality. Organizational Behavior and Human Decision
Entrepreneurship, 4(4), 309–450. https://doi.org/10.1561 Processes, 110(2), 93–107. https://doi.org/10.1016/j.
/0300000014. obhdp.2009.07.002.
Reproduced with permission of copyright owner. Further reproduction
prohibited without permission.