1
BUSINESS SURVIVAL and the ASSOCIATED FACTORS:
EMPIRICAL EVIDENCE from YOUTH-OWNED MICRO and SMALL ENTERPRISES in ETHIOPIA
Tassew Woldehanna Wolday Amha Manex Bule
Associate Professor Executive Director Research Fellow
Addis Ababa University Association of Ethiopian Ethiopian Inclusive Finance
[email protected] Microfinance Institutions and Research Institute
[email protected] [email protected]
Helen Berga
Research Fellow
Ethiopian Inclusive Finance
and Research Institute
ABSTTRACT
Micro and small enterprises are expected to play a crucial role in the development process of a country through
employment creation, increasing income and poverty alleviation. This study investigates the effects of person, firm
and industry related characteristics that are typically identifiable at the time of establishing the enterprise on the
MSEs' risk of drop-out. Moreover, it examines factors that can explain the difference in the firm survival among
micro and small enterprises, and opportunity and necessity entrepreneurs. The paper employ a retrospective cross
sectional survey data from 941 enterprises, which are founded in the year 2008 and later. The non-parametric
estimation results reveal that there is an increasing risk of drop-out during the start up phase of the enterprises.
Moreover, the study confirms that there exist significant differences in survival rate and drop-out risk between micro
and small enterprises. A Cox proportional hazard estimate shows that gender, age, previous labor market experience,
motivation, entrepreneurial education and training, initial size, legal form, size of start-up capital, industry type and
formality are important predictors of the MSEs' survival in Ethiopia. A separate Cox proportional hazard estimation
provides that gender, entrepreneurial training, size of start-up capital are the major factors accounted for the better
success of small enterprises compared to micro enterprises. The findings from this study have important
implications for prospective entrepreneurs, business owners, support providers, and policy makers on how to
improve self-employment and create conducive business climate.
Key words: Ethiopia, Micro and small scale enterprises survival, non-parametric hazard model, Cox proportional
hazard model.
Introduction
The importance of enterprises as the principal source of growth and employment cannot be overstated. Enterprises,
particularly small and medium enterprises (SMEs), are decisive as a major source of income and employment and
2
are at the heart of an economic activity and development for developing countries (Mead and Liedholm, 1999; ILO,
2007). However, if the growth and survival of the newly established firms are not ensured, the expected positive
results will rather be replaced with negative outcomes of unemployment, wastage of resources and time in the part
of the owner and economic loss in general.
Extensive theoretical and empirical literatures aimed at identifying micro-level factors, determining the emergence
and success of enterprises. Though they are on the industrialized countries context, most of the studies have shown
that owner and firm related characteristics are the basic factors that determine the success of a firm (see, e.g. Brtiderl
et al, 1992; Storey, 1994; Coleman et al. 2010). Mata and Machado (1996) analyzed the impact of industry
characteristics on start ups success. Moreover, Pfeiffer (1994) concluded that a newly established firm survival is
more likely depend on initial financial endowment, their human capital, risk aversion, the wish for independence,
and the support of their social and family networks. Studies focused on developing economies consistently
highlighted, imperfection in the credit and financial markets, a non-transparent regulatory environment, lack of
infrastructure, bureaucracy burden, and the high incidence of bribing as the pervasive challenges to enhance the
productivity improvement and survival of SMEs in most developing countries (Hoskisson et al.,2000; Mousley
2003; and Aterido, Hallward-Driemeier and Pagés 2009).
The Ethiopian government has made significant effort to make a paradigm shift on youth employment - from job
seekers to job creators. The five-year Micro and Small Enterprises (MSEs) development strategy was designed and
implemented to particularly create jobs to solve the high unemployment rate observed in the country (where urban
unemployment was reported to be 17.5% in 2012 by CSA) and ensure economic growth and transformation1.
Currently, MSEs are becoming a significant job supplier in the urban labor market, as they have created 2.5 million
new jobs in the last four years, particularly for unemployed youth (FeMSIDA, 2014).
Despite the support and focus given to the sector, large number of MSEs are expected to eventually close or
stagnated at starting phase. As reports and studies documented, there are internal and external factors which affect
the success of the MSEs such as shortage of working capital, lack of marketing skills, poor location of business
cites, failure to take risk, and over emphasis to short term profit are internal factors affecting the survival rate of
MSEs. In addition, lack of access to efficient infrastructure, access to factor of production, bureaucracy burden, and
lack appropriate skill and training are also documented as external factors hindering the success of the MSEs (see
e.g. WB, 2014; Dayanandan, 2012; Soderbom, 2010).
The co-existence of MSEs can be interesting, concomitantly the performance of MSEs in terms of survival and
growth needs to be assess in detail. Moreover, identifying the growth and survival factors of MSEs is important as it
establishes the base for preparing a policy framework and strategy that safeguards the success of MSE operators.
However, as far as we know, there is none empirical efforts have been done to study the MSEs growth and survival
using a nation-wide data. The existed knowledge-gap and the focus for the sector development motivated us to
empirically analyze the factors associated with the survival of MSEs in Ethiopia. More specifically, the paper
intended to identify the socio-demographic and economic characteristics of the owners and firms at startup period
that can determine the survival of youth-owned MSEs. We also motivated to investigate owners, firms and industry
characteristics that can explain the difference among micro and small enterprises survival rate.
The study produces and provides policy variables that can be input to the MSEs development strategy and be helpful
to MSE promoters on revisiting their support services. Generally, examining the associating factors of firm survival
will provide basic information to prospective entrepreneurs, business owners, support providers, and policy makers
on how to promote self-employment and create conducive business climate. On the other hand, since there are
hardly surveys and research reports intended at identifying factors influencing the survival of MSEs in Ethiopia
using survival analysis, the study is expected to contribute to the empirical literature by analyzing the survival rate
and its relationship to time invariant covariates using Cox proportional hazard model apart from the descriptive
statistics.
1
The youth population in Ethiopia is rapidly growing, as it has more than doubled between 1990 and 2007, from 6
million to 13 million (CSA 2007).
3
The data used in this paper is a nationwide cross-sectional survey data which is rich with retrospective information.
The survey was conducted in fifteen major cities of the country on 2014. Statistical and econometric models are
used to answer the research questions and achieve the objective. Mainly, non-parametric and semi-parametric
survival models are employed to estimate the survival rate of micro and small enterprises and the explanatory
variables for the survival rates.
This paper is organized as follows: Chapter 2 presents the conceptual framework and the empirical literature review.
Chapter 3 presents the data and descriptive statistics. Chapter 4 presents the econometric model specification
employed in the paper. Chapter 5 reports empirical results and interprets the estimated parameters. Chapter 6
presents the conclusion part. Chapter 7 is about the policy implications.
Conceptual framework and review of empirical studies
Understanding and studying the survival rate of youth-owned enterprises require developing the conceptual
framework, which identifies the relationship between the different factors (core reasons behind drop-out) and the
probability of sustaining a business. Reviewing the relevant literature on firm survival in different countries provides
useful input in refining the method of analysis and the hypotheses. Attempts are also made to develop appropriate
methodology to analyze survival rate of youth-owned MSEs in Ethiopian.
The conceptual framework developed for this study indicates that the survival of youth-owned enterprises are
mainly affected by the inadequacies in the macro-level factors, which include the performance of the economy
(economic growth, inflation and relative price variability); functionality of the capital, labor and product markets;
and policy and institutional development (policy and regulatory environment, support program, governance and
bureaucracy, infrastructure development, etc) (see Figure 1). On the other hand, the survival rate of youth-owned
enterprises is also influenced by micro-level factors such as owner's characteristics (gender, age, education, skill and
experience, entrepreneurial motive); firm characteristics (firm size and age, location, ownership form); business
strategic (human capital quality, planning, marketing and financial management, and ICT); and industry
characteristics (sector, competition, market linkages, industry performance, technological change).
At the micro-level, there is a need to assess the sustainability of an enterprise in terms of the structure of the sector
or market in which it operates and, in particular, its relationship with suppliers and customers along the value chain.
The micro-level determinants of firm survival (sustainability) include what goes within the enterprise or its
immediate environment (the management of human and financial resources and use of physical resources like
energy, transport and communications systems) and to the direct interface between enterprises and their customers
and suppliers (ILO 2007).
The weak performance of the economy (business cycle, and inflation and relative price volatility), market failure
(imperfections in the financial, labor and goods markets), and institutional factors (non-transparent regulatory
environment, lack of infrastructure, and incidence of bribing) are key factors that lower the survival rate of
enterprises. In other words, inefficient regulatory environment with regard to labor market rules, taxation, red tape
producers, property rights and bankruptcy laws, lack of support and incompetence in implementing existing support
strategies, bribing and lack of adequate infrastructural endowment are major factors which influence the survival of
an enterprise (Quatraro and Varelli, 2013).
Empirical Literature Review
The main factors which influence the survival of firms at micro level can be broadly grouped as owner, firm,
strategy and industry characteristics (Fadahunsi, 2012).
Owner characteristics
Many empirical researches have tended to explain the relationship between the entrepreneurs characteristics and
firm survival and growth. As earlier studies have shown that there are several owner characteristics that can affect
the success of a business. They are typically identifiable prior to the start up and include a range of personal and
4
psychological characteristics and these include gender, education and training, entrepreneurs mind setup, previous
work experience, and ethnicity (Fadahunsi, 2012).
Gender: The study by Aspray and Cohoon (2007) shows, the impact of owner‟s gender on the performance of the
firm is inconclusive. While some studies indicated that women-owned firms perform less in terms of different
performance measures such as: number of employees (Rosa et al. 1996), revenue (Rosa et al., 1996; Johnson and
Storey, 1993; Brush et al., 2003) profitability and survival (Robb, 2002; Carter et al., 1997; Fairlie and Robb, 2008).
Other studies found that there is hardly any difference between men and women owned firms (Menzies et al. 2004;
Kalleberg and Leicht, 1991; Johnson and Storey, 1993). Studies which are conducted in developed countries
documented different evidence towards the impact of owner‟s gender on the survival of business. Kalleberg and
Leicht (1991) found that businesses headed by women were not more likely to go out of business, nor less
successful, than those owned by men.
On the other hand, Fairlie and Robb (2008) shows that business which are owned by females have lower survival
rate that those owned by males. They found that female-owned businesses are less successful than male-owned
businesses because they have less startup capital, and human capital acquired through prior work experience in a
similar line of business and prior work experience in family business. Nevertheless, gender of the owner appears to
be a significant determinant of firm performance in many studies. Despite the importance of analyzing the gender
difference in enterprise performance in developing countries, very few researches have been conducted in these
countries. A study of Coad and Tamvada (2008), conducted in developing country setting, shows that enterprises
managed by women are less likely to grow and more likely to decline.
Age: Empirical studies conducted by various researchers reveal that younger entrepreneurs are likely to be more
willing to assume risks and grow their business compared to older individuals. According to Davidsson (1991),
younger individual may have a higher need for additional income. An older individual who continues to be the
owner- manager of a small firm is more likely to have reached his/her initial aspirations. However, while younger
individuals have more motivation to expand their business they also may have fewer financial resources and fewer
networks. The limited empirical evidence suggests that the owner-manager's age tends to be negatively related to
growth of enterprises (Boswell, 1973; Davidsson, 1991). A study by Amran (2001) also shows that a negative
relationship between owner‟s age and business performance suggesting that matured owners underperform, while
the young owners are more aggressive in enhancing the firm value.
Education: Most previous studies show that formal education has a positive impact on the business performance in
general and survival in particular. On survival rate of enterprises, studies found that firms having more highly
educated owners were more likely to survive (see, e.g. Bates, 1995; Cressy, 1996; and Lussier and Pfeifer,2001). As
Taylor (1999) found, formal higher education qualification is not an important factor to determine self-employmnet
duration in Britain. However, there are also converse arguments that owner or manager of SMEs, who had degree,
generally achieved lower rate of success than those less well educated (Barkham et al.,1996).
Previous experience: Many studies found that owner-managers‟ employment experience prior to the start up tend to
link with SMEs success. Most studies reveal that firm performance is positively influenced by the owner‟s level of
human capital measured by education and, prior work and managerial experience (see e.g. Schiller and Crewson,
1997; Honig, 1998; Chandler and Hanks, 1998; Kangasharju and Pekkala, 2002; Pena, 2002). Taylor (1999) also
noted that prior work and self-employment experience increases the likelihood of firm survival. Whereas, Kalleberg
and Leicht (1991) found that no direct relationship between previous experiences as self-employed and short-term
business success.
Motivation: The reason why people decide to be self-employed has been investigated for the last two decades.
Birley and Westhead (1994) have identified seven different motives, namely, need for independence, need for
approval, need for personal development, welfare considerations, perceived instrumentality of wealth, tax reduction,
and following role models. In more comprehensible way, Carter et.al.(2003) have extracted six entrepreneurship
reasons: innovation, independence, recognition, roles, financial success, and self-realization. There has been a recent
trend to distinguish an entrepreneur‟s motivation to „opportunity‟ and „necessity‟ entrepreneur (see Reynolds et al.,
2001). Necessity entrepreneurs exhibit higher probability of exit as compared to opportunity entrepreneurs
(Caliendo and Kritikos, 2009). Block and Sandner (2009) found that being an entrepreneur out of necessity or
opportunity driven motives does not have significant impact on duration in self-employment.
5
Entrepreneurship education and training: Entrepreneurship education and training has been found to be
important factor to stimulate entrepreneurial activity and performance and in turn firm growth and survival. Many
argue that the individual psychological factors and the cultural context are of the basic importance in persuading the
entrepreneurial behavior of an individual. Entrepreneurship education imparts entrepreneurial qualities and skills
and mental awareness that the entrepreneur can use during each phase of business development (Alberti, Sciasca,
and Poli, 2004). As Dana, (2001) noted that countries like Brazil, India, Malaysia, Singapore and the UK have
introduced entrepreneurial education program since 1990. Crowling (2009) has found a positive effect of school
based entrepreneurial training on job creation by small firms in UK. He also found a positive relation between
entrepreneurs who received government backed entrepreneurial training and exporting intensity of small firms.
Ethnicity: Ethnicity of the owner is other important determinants of firm survival. Concerning ethnicity, controlling
for industry, sector and various types of human capital, Coleman (2005) found that firms owned by white women,
black women, and black men were still significantly smaller than firms owned by white men in terms of total sales
and total assets. Firms owned by Hispanic men were significantly smaller than firms owned by white men in terms
of total assets but not in terms of sales. Cooper et al. (1994) also documented that being part of a racial minority is
linked to lower probabilities of both survival and growth. Establishing business outside birth place might have
impact on the survival of an enterprise. On the one hand, it might have positive impact on the survival of an
enterprise as the feeling of being new and the expectation of hardship emanating from it might motivate the owner to
use his/her maximum potential to sustain the business. On the other hand, lack of knowledge about the local market,
preference of peoples and possible discrimination that can be faced both from the society and local officials might
also reduce the survival of an enterprise.
Firm characteristics
Firm's characteristics refer to factors identifiable at the start-up phase which have an important influence on the way
the business is managed. These characteristics, such as size, age, ownership form and location, might also determine
the survival rate of an enterprise.
Size: The studies of Mata et al. (1995); and Agarval and Audretsch (2001) showed a positive relationship between
start-up size and survival, as larger size entry is a signal of commitment and self-confidence, which reduces the
occurrence of an entry mistake and the risk of a failure due to diseconomies of scale less likely. Similarly, firm‟s
age is found to be positively correlated with survival (Dunne and Hughes, 1994; Yasuda, 2005; Calvo, 2006), as
experienced and mature firms are more able to deal with market dynamics and so more likely to survive. Thus,
failure rates tend to be higher for firms that are new or small (Cooper et al., 1994; Disney et al., 2003; Holmes et al.,
2008).
Location: Location of an enterprise may also influence the survival of an enterprise. Locating business in an area
where there is high sectoral linkage (both from input and output side) and near to markets might increase the chance
of survival as it gives entrepreneurs access to key networks and sources of information, market, raw material supply
and support. Previous studies have found that entrepreneurs who are tied in with key networks gain better access to
resources which, in turn, increase their chances of survival (Stearns et al., 1995).
Ownership form: Legal structure of a business has an influence on growth and success with limited companies
enjoying faster growth than either sole-proprietorship or partnership. Kalleberg and Leicht (1991)in their study of
small firms in the United States found that firms with limited liability have higher growth rate. Woldie et.al. (2008)
have found an association between growth and the legal status of SMEs in Nigeria.
Industry characteristics
The survival of enterprises is also influenced by industry characteristics, including the type of the sector the firm
operates, level of competition within the industry, linkage, market and, technological change and innovation (see
e.g. Bruderl et al., 1992; Audretsch & Mahmoud, 1995; Headd, 2000; Bates, 2005; Carter and Van Auken, 2006).
Coleman et al. (2010) indicated that firms involved in retail and service sectors are more prone to failure due to low
barriers to entry and high levels of competition. On the other hand, innovative or technology-based firms have better
6
chances to survive, while others stress that the risks associated with technology actually increase the likelihood of
failure.
Cluster: In Africa, the most common type of micro and small enterprises cluster is a spontaneously grown natural
clusters which can an outcome of government zoning regulations and policies (World Bank, 2011). The result from
this study revealed that, on average, in Africa cluster-based small and micro enterprises have higher labor
productivity and sales per worker.
Empirical model specification: survival model
Survival rate is the most often practical index to measure the success of a firm and the analysis examines the elapsed
time since entry to the state at time t = 0. There are two major problems in survival analysis: (i) the distribution of
survival times are typically skewed because of the non-negative value of times; and (ii) the data observed is usually
right-censored, i.e. at the time of observation the relevant event had not yet occurred for all individuals in the sample
(Stephen Jenkins, 2004). Though the former could easily be dealt with using standard econometric techniques, it is
unable to handle censoring using other econometric models such as OLS, logit and probit. Hence, the solution is to
specify survival model accounted for the sequential nature of the data, and are able to handle censoring.
Let T represents survival time and as a random variable with a cumulative distribution function )Pr()( tTtF
and a probability density function ./)()( dttdFtf The reverse cumulative distribution function of T provides a
new function called survival function S(t) and it is defined as
).(1)Pr()( tFtTtS (1)
The hazard rate, the instantaneous risk of demise at time t conditional on the firm having survived up to that point in
time, is the motive to be estimated in our survival analysis. The hazard function is an absolute slope of log survival
function and it defines as follows
dt
tSd
tF
tfth
)(ln
)(1
)()(
(2)
and the cumulative hazard function H(t) is
).(ln)( tStH (3)
The survival function describes the probability of surviving past time t. The hazard function describes the relative
likelihood of the event occurring at time t, conditional on the subject‟s survival up to time t. The hazard rate shows
the instantaneous rate of failure at time t and ignores the accumulation of hazard up to time t (see Stephen Jenkins,
2004).
Non-parametric approach
The above modeled survival and hazard functions can be estimated without making assumptions about the shape of
the relevant functions. The Kaplan-Meier estimator is a non-parametric estimator that does not make any
assumptions about the distribution of exit times and how covariates shift the hazard functions (Stephen Jenkins,
2004). Hence, the proportion of surviving to survival time tj and the corresponding integrated hazard function is
estimated by
7
ttj j
j
j
jn
dtS
|
1)(
(4)
)(ln)( jj tStH
(5)
where dj denotes the number of failed enterprises at tj and nj is the number of enterprises at risk of failure
immediately prior to tj and defines as the sum of those who have a censored or completed spell of length tj or
longer. The Kernel-smoothed estimator is a weighted average of )(tH
over event times that are within a
bandwidth distance b of t. The Kernel-smoothed estimator considers the jumps of )(tH
and variance of )(tH
at
the event times as follows:
))(())(())((
)()()(
1
1
jjj
jjj
tHVtHVtHV
tHtHtH (6)
where t0=0.
As the analysis on the effect of enterprise‟s initial size and the entrepreneur‟s motive on its survival probability is
one major objective of our paper, we grouped the data set into Micro and Small enterprises and opportunity and
necessity entrepreneurs. Accordingly, for ease of interpretation, we grouped enterprises into six months intervals and
execute the so called Lifetable estimator. Basically, the Lifetable estimator processes the same idea as the Kaplan-
Meier method, but it was explicitly designed to handle situations where there is grouped survival time data.
Semi-parametric hazards model
The focal point in our survival analysis is the hazard function allowing for time invariant covariate effects as an
econometric framework. The Cox proportional hazard model has considerable flexibility and is the most widely
used hazard regression model in empirical studies. This model, proposed by Cox(1972), enables us to estimate the
relationship between the hazard rate and covariates without having to make any assumptions about the shape of the
baseline hazard function, referring the model as a semi-parametric model.
A survival model examines the relationship between the survival distribution and the covariates and mostly entails
the specification of a linear-like model for the log hazard
)...exp()(
...)(log
2211
2211
jkkjjj
jkkjjj
xxxth
xxxth
(7)
where x‟s are the covariates and represents a log-baseline hazard because if all x‟s are zero then .)( eth j
The Cox model leaves the baseline hazard function )(log)( 0 tht unspecified. Thus, the hazard function is
parameterized as follows
)...(
02211)()( jkkjj xxx
j ethth
(8)
8
where h0(t) is the so-called baseline hazard which depends only on t. The change in the explanatory variables,
therefore, induces multiplicatively proportional shift in the hazard rate relative to the baseline hazard. This implies,
the model assumes that the hazard functions of any individuals with different values on one or more covariates differ
only by a factor of proportionality.
Based on equation (8), the hazard ratio for two individuals, j and i, defines as follows
)}(...)()(exp{)(
)(222111 ikjkkijij
i
jxxxxxx
th
th
(9)
Where, the hazard ratio is independent of time t. In the way to compute the hazard ratio, the baseline hazard h0
cancels out of the computation, implying that the relationship between the hazard rate and the explanatory variables
is estimated without taking h0 in to account. Consequently, the Cox model is become a proportional hazard model,
but the proportional hazard assumption should be tested.
Instead of the maximum-likelihood method, the Cox model can be estimated by the method of Partial Likelihood
(PL), because it requires less assumption and so is more robust (Fox, 2002). The PL only considers the contribution
from failure times, not from the right-censored times. This implies that the Cox model is conspicuously ignoring all
information available at times when no failure occurs. Cox (1972) has found little efficiency loss due to ignoring
spells in which no failure occurs.
Data and descriptive statistics
Data
The study uses the data collected from a nationwide survey on MSEs residing in major urban areas of the country.
This cross sectional survey provides retrospective information about the initial startup data including the owner and
firm related characteristics. Moreover, information regarding the ongoing and past business strategy of the
enterprise, since their startup, is incorporated. The survey was conducted in fifteen major cities of the country. Data
is collected from 1109 randomly selected youth (age between 18 and 34) owned MSEs, 909 of them being existing
enterprises and 200 closed enterprises. Stratification was made by gender of the owner, the enterprise size (Micro
and Small) and the enterprise type (manufacturing, construction, urban-agriculture, service, and trade).
We used population sampling process, implying both the existing enterprises, the start-ups observed at the time of
the survey, and the closed enterprises, the start-ups that have been already dropped out at the time of the survey,
were interviewed. Closed firms, in this particular study, refer to those that have discontinued their operations as they
are intended during establishment. This means, if the owner of the enterprise, which have been closed down, starts a
new business - different type of business and industry, then we considered the first enterprise as a closed enterprises
in our sample. The enterprise owners were asked about the month and year of the business start-up and of the
termination date, in case the start-up has already closed. Such information allows us to establish the business
operation spell for each firm, and the spell might be either completed or right censored at the time of survey. The
survey has detailed information on socio-demographic and economic background of the owner/manager as well as
firm and business strategy characteristics. In addition, information regarding the motivation for setting up the
business, the quality of the support programs and the respective gap, challenges experienced by enterprises and the
labor market status of the youth before and after self-employment are available in the survey dataset.
The retrospective information determines the validity of the research, since it has been subscribed to self-
justification bias To account for the retrospective bias, the data used in this paper consists of 941 enterprises which
are founded in the year 2008 and later, expecting the startups' owners can remind and provide information with less
inaccuracy during this period.. Thus, the sample includes 941 enterprises, 569 enterprises are grouped as micro and
9
372 are as small enterprises based on registration certificate2, their initial number of workers
3. Moreover, out of the
total sample, we observe 18 percent completed business operation spell and the remaining 82 percent is right-
censored at time of interview.
Descriptive analysis for the pooled data
Based on the theoretical and conceptual framework discussed in section 2 and relying on the context of the country's
social and economic settings, a range of variables are hypothesized as the associated factors of MSEs survival or
failure. Table 1 presents the hypothesized variables that are identifiable at the time of start up and their description
along with main summery statistics. The first part of the table describes the socio-demographic and economic
characteristics of the owner.
The table shows that 26 percent of the enterprises are owned by female (i.e. as a solo and group) entrepreneurs,
implying the dominance of male entrepreneurs in MSEs in Ethiopia. The mean age of the owners at the time of the
startup was 24 years old. Out of the total sample, 25 percent of the entrepreneurs have only elementary school
attainment. During the startup period, 17 percent of the entrepreneurs had at least a college diploma and only 10
percent of the entrepreneurs have a technical vocational education and training (TVET). Almost more than half of
the owners were active in the labor market before self-employment. There are few owners who had self-employment
experience prior to the startup. Regarding previous experience, 33 percent and 50 percent of the owners had
apprentice and business experience prior to the startup, respectively. Looking at the training type that the owners
took, we observe that 35 percent of them are obtained short-term training before establishing the business. It turns
out that only few, 14 percent, of the founders have taken entrepreneurship training before setting up the business.
Considering the motivation behind being self-employed, 69 percent of the owners were opportunity driven
entrepreneurs.
Concerning firm and industry characteristics, the average initial number of workers which is a proxy for initial size
of the firm was eight, implying the dominance of micro level enterprises in our sample. The majority of the
enterprises, 83 percent, were raised a start-up capital of less than or equal to Birr 25,000. However, only 6 percent
can raise a start-up capital of more than Birr 100,000 during establishment. At the start up period, 38 percent, 31
percent and 26 percent of the enterprises were established as a sole-proprietorship, partnership and cooperative legal
form, respectively. Regarding industry type, it turns out that most of the enterprises are product providers (almost 71
percent) of which 37 percent and 28 percent of the enterprises are engaged in metal and wood work and construction
sector, respectively.
Moreover, 66 percent of the owner had received at least one type of support program (it can be physical, material or
financial support) at the start up period either from government or non-government support service providers. With
respect to their formality, 81 percent of the enterprises were registered.
Descriptive analysis for the group samplings
In order to reveal differences between owner related and business related characteristics among micro and small
enterprises, t-test mean equality is conducted. The reported p-values show whether the means are not equal at a
significant level of 5%.
Table 2 compares micro with small sized firms under business operation survivors and drop-outs samples. We
observe that female entrepreneurs are apt to establish a micro sized enterprise. The share of female entrepreneurs is
significantly higher among the micro enterprises (almost 47%) in the group of business operation drop-outs. In the
2
Regional government offices provide a certificate for new start ups upon registration and it includes information about
the initial size of the business.
3
Micro enterprise is define as an enterprise employing less than five workers, while small is defined as an enterprise
which started by employing more than five but less than thirty workers.
10
survivor and drop-out sample, micro and small sized enterprises do not differ significantly regarding educational
attainment.
Regarding previous work experience, we observe that opportunity entrepreneurs, in the survivor sample, have on
average better experience than necessity entrepreneurs. survivor sample is significantly higher among opportunity
(almost 43%) than necessity entrepreneurs (almost 8%). Similarly, the share of owners with business experience is
significantly higher among opportunity (almost 56%) than necessity entrepreneurs (almost 37%). The share of
owners with apprentice experience in the drop-out sample, is significantly higher among micro enterprises (almost
37%) than small enterprises (almost 20%).
Comparing the type of training the founders attained before the startup, we observe that small enterprise founders
have a significantly higher share of attaining important trainings than micro enterprise founders. For example, in the
survivor sample, a significantly higher share of owners with entrepreneurship training at the startup period found
among small (almost 18%) than micro enterprises (almost 13%). This indicates that the small enterprise founders
organize themselves better than micro enterprise founders to get in to the business environment by taking important
business related trainings prior to starting the business. With respect to the motivation of business operation, we find
that 69 percent of entrepreneurs are opportunity driven entrepreneurs. This share is significantly lower among small
(almost 38%) than micro enterprises (almost 61%) in the drop-out sample.
The table also compares firm and industry related characteristics in the business operation survivor and drop-out
groups. Concerning legal form, the share of solo businesses is significantly higher among micro than small
enterprises in both survivor and drop-out samples as it accounts 54 percent and 47 percent, respectively. In contrast,
the share of cooperative businesses is significantly higher among small than micro enterprises in both survivor and
drop-out samples as it accounts 41 percent and 68 percent, respectively. The result implies two basic things; first,
since the required capital and the level of management, in terms of resource, controlling system and marketing
strategy, for small size start ups are higher compare to micro start ups, the business founders desire to start a
business by sharing financial, human and physical resources; second, in order to minimize the risk sharing amount in
time of failure, the necessity entrepreneurs probably prefer a cooperative business.
With respect to start-up capital, we observe that, in the survivor sample, small business owners are likely invest on
average higher amount compared to micro enterprises. In contrast, the share of owners who invest amount of less
than Br 25,001 is significantly higher among micro enterprises than small enterprises in drop-out sample. Regarding
industry type, 70 percent of startups were engaged in the manufacturing sector (product providers). This share is
significantly higher among small than micro enterprises in the business operation survivor group and lower in the
business operation drop-out group. We observe a significantly higher share of service sector among micro than small
enterprises in both survivor and drop-out samples.
As to support programs, it turns out that the share of owners who obtained a support service is significantly higher
among small enterprises compared to micro enterprises in the survivor sample. This result might have different
rationale, a) either they are always available and well informed in the process of getting the support programs, b)
either the system is designed, consciously or unconsciously, in favor of small enterprise founders. With respect to
formality, the share of registered startups among small enterprises is significantly higher than micro enterprises in
the drop-out sample.
The survival model estimation
The non-parametric estimation results
Both non-parametric and parametric microeconometrics techniques are used to analyze the survival rate of MSEs in
Ethiopia. Attempts are made to study the effect of owner, firm and industry related characteristics on the enterprise
survival, in general, and what factors explain the deferential of survival among micro and small enterprise founders,
in particular.
11
Under non-parametric survival function analysis, the survival and hazard rate (smoothed and cumulative) estimates
are worth to consider. The Kaplan-Meier estimation method is used for the non-parametric survival and cumulative
hazard analysis - we estimate the cumulative hazard by taking the negative log of the Kaplan-Meier estimated
survivor function (see equation 5). But, we employed the Kernel-smoothed hazard estimation method for smoothed
hazard analysis as it efficiently detect the times where the most rapid changes of the hazard function occur. Figure 2
shows the estimated survival and hazard rates.
Figure 2a shows that survival rate declines as duration goes by. On average almost 88 percent of MSEs survive their
first two years in business, however the survival function is sloped strongly afterword and the probability of survival
reaches to almost 68 percent after six years of business operation. From Figure 2b, there is a sharp rise in the
cumulative hazard at the beginning of the analysis time and then decelerates, reflecting higher hazard rate during the
startup period. As Figure 2c shows, the risk of drop-out is increasing during the startup phase of the enterprise till
reaches its peak at 24 months of business operation duration.
The survival functions for the groups, micro and small enterprises, are estimated using Lifetable estimator. Figure 3a
shows that the survival probability among micro enterprises decreases rapidly after the first six months of business
operation. The estimate for micro/small group shows that about 92 percent of the small enterprises and about 90
percent of the micro enterprises survive the first two years of business operation, which is almost equal. However,
we observe that after four years of business duration the survival rates among the groups are notably growing apart;
77 percent for the small enterprises and 66 percent for the micro enterprises (see Appendix Table A).
Thus, the results reveal strong evidence that startup initial firm size has effect on survival probability. Entrepreneurs
who establish new small sized firms survive more compared to those who set up a micro size enterprises in Ethiopia.
To test for equality of the survival functions, we run the log rank test. The null hypothesis of no survival difference
between the groups rejected at a 10 percent level of significance (Chisq. p=0.0707). Figure 3b depicts that drop-out
risk is higher among micro than small enterprises at a given business operation duration. Moreover, it has an
inverted-U shape for the micro enterprises. Initially, there is increasing risk of drop-out during the first 2 years of
business operation and then the level of the risk remain high for the next 2 consecutive years before it declines over
time. In contrast, we observe increasing risk only during the first business operation year and decreasing afterward
among small sized enterprises.
Overall, our non-parametric survival analysis suggests two basic facts on MSEs operation in Ethiopia: one, there is
survival differential among micro and small enterprises; two, the risk of drop-out is high during the first 2 business
operation years. This calls efficient and inclusive support interventions for micro business founders and mostly
during the startup phase of the MSEs in Ethiopia. However, what factors derive the survival of MSEs and what
factors are account for the deferential of survival rates among micro/small enterprise needs a regrious analysis. To
answer these questions we conduct a semi-parametric analysis.
The Cox proportional hazard estimation
In this section, the results from the semi-parametric Cox- proportional model are discussed for the whole sample
followed by the results from the groups analysis classified based on the size of the enterprises.
We execute tests and graphical diagnostics whether the fitted Cox regression model adequately explains the data
using the scaled Schoenfeld residuals. The tests for the assumption of proportional hazard reveal a strong evidence
of proportional hazard for each covariates and for the global test (see Appendix Table B). The chi-square statistics
soundly fail to reject accommodating proportional hazard for the model as a whole (p-value=0.988). Similarly,
graphical diagnostic tests of the Schoenfeld residuals also support the assumption of proportional hazard for all
covariates as there is non-systematic departure of covariates from their smoothing-spline fitted line over time - we
observe non-trend pattern with time in the plots for all covariates (see Appendix Figure A).
The log form of the owner‟s age and the initial number of workers are employed in the fitted model. Table 3
illustrates the estimated coefficients from the Cox-proportional models. The first column presents the estimates from
the pooled sample, and the next two columns display the estimation results from the grouped samples; from
micro/small enterprises sampling.
12
Starting with socio-demographic and economic characteristics of the owner, we find that gender has a significant
positive association with risk of drop-out. Enterprises owned by women have higher risk of drop-out than men
entrepreneurs, which is consistence with the result from Coad and Tamvada (2008). As Davidsson (1991) and
Amran (2001) found, there is a significant negative relationship between the business founders' age and business
survival. This suggests that entering a self-employment at matured age is associated with a higher drop-out rate than
younger individuals.
With respect to education qualification, we find that it does not appear to be significant factor for MSEs survival. As
Taylor (1999) argued, this suggests that to enter self-employment and survive, entrepreneurs do not need to have a
high level of formal education. Moreover, there is no.significant positive effect of technical and vocational
education and training (TVET) on the probability of a firm's survival Our fitted model proves that previous labor
market experience appears to be an important factor for survival. Consistence with the result of Taylor (1999), active
labor market participation prior to the startup reduces the risk of drop-out. But, as Kalleberg and Leicht (1991)
found, we do not find a significant relationship between self-employment experience prior to the startup and
business survival. Previous general business and apprentice experience do not have significant impact on a firm's
survival.
In line with our non-parametric survival analysis and the result from Caliendo and Kritikos (2009), it turns out that
opportunity motivation significantly reduces the risk of drop-out, implying that the opportunity driven
entrepreneur‟s probability of survival is significantly higher than necessity entrepreneurs. Regarding entrepreneurial
training, we find that it has a significant negative effect on drop-out rate, suggesting that entrepreneurial education
and training prior to startup appear to be important factor to stimulate entrepreneurial activity and business survival.
Concerning firm and industry characteristics, startup size had the expected positive effect on MSEs survival,
implying the larger the size of the set up enterprise the lower is the probability of drop-out. The impact of financial
resource at the time of startup captures by the size of start-up capital, and it turns out that larger amount of start-up
capital reduces the risk of drop-out. Entrepreneurs who invested more than Birr 100,000 has a significantly higher
probability of survival than the others, suggesting that adequate amount of start-up capital is a key factor for a firm's
survival. Unlike Kalleberg and Leicht (1991), we find that a solo type of business set up has a significant positive
effect on MSEs survival than cooperatives. This might be associated with the complete control and decision making
power that the owner enjoys and being free from corporate tax.
Concerning industry type, MSEs engaged in the manufacturing sector has a lower risk of drop-out. This is
consistent with Taylor (1999) result, implying that the service sector is exposed to higher risk of drop-out. With
respect to support, supports obtained by the owner at time of startup do not appear to have significant effect on the
survival of MSEs. Last but not least, our analysis proves that formal system is a significant determinant of MSEs
survival. It turns out that registration at the startup time reduces the probability of drop-out.
Overall, gender, age, previous labor experience, motivation, entrepreneurial education and training, initial size, legal
form, size of start-up capital, industry, and formal system therefore are important factors in determining the
probability of MSEs survival in Ethiopia.
The groups survival rate analysis
The result from Table 3 reveals that gender has a significant effect on the probability of risk of drop-out for micro
enterprises. As we find out from the descriptive analyses, women have a significantly higher share of founding
micro enterprises, but the result from the survival model shows that they experience a significant higher exit rate. It
suggests that the gender issue is one of the reasons why micro enterprises survive less than small enterprises do. Age
has a significant positive effect on the probability of drop-out in all the samplings, implying the higher the age of the
entrepreneur is the higher the risk of drop-out.
Like the result from the pooled sample, previous education qualification of the owners has no significant effect on
the survival rate among all the samplings. We find an unexpected result with respect to the effect of previous
experience. Apprentice experience prior to startup has a significant positive effect on the hazard rate among micro
enterprises. The justification for this result demand a further study on the subject matter, but the probable
13
explanation can be the gap between the level and quality of the apprentice experience with respect to the nature of
and demand for micro enterprises.
The study confirms that entrepreneurs out of opportunity motivation do experience a significantly lower level of
hazard rate either they established as a micro or small enterprises. This finding is consistent with the result from the
non-parametric Lifetable estimator and the Cox model for the pooled data. Entrepreneurship training drives the
small enterprise owners to experience low level of drop-out risk compare to the micro enterprise founders. The
result suggests that, prospective entrepreneurs aiming to form micro enterprises can benefited from low level of
hazard rate by taking entrepreneurial trainings prior to the startup. Moreover, we confirm that entrepreneurial
training is one of the explanations for the survival deferential between micro and small enterprises.
Our analysis depicts that large value of start-up capital has a significant and pronounced negative effect on the risk
of drop-out from the small enterprises sample. This implies that it is an important factor accounted for the survival
rate deferential between micro and small enterprises.
Interestingly, there is a significant negative association between solo business type and the probability of risk of
drop-out from all samplings, though it is not quite statistically significant among small enterprises. The relationship
between engaging in a manufacturing sector and the risk of drop-out appears to be negative among the micro
sampling. We also find that formality has a significant negative association with the risk of drop-out among micro
enterprises.
Our Cox proportional model estimation identifies what owner, firm and industry characteristics at startup period
explain the deferential of survival rate among micro and small enterprises. Accordingly, we find that gender,
entrepreneurial training and size of start-up capital are the major factors accounted for the better success of small
enterprises compared to micro enterprises.
Conclusions
The Ethiopian government has made significant effort to make a paradigm shift on youth employment - from job
seekers to job creators. The five-year Micro and Small Enterprises (MSEs) development strategy was designed and
implemented to particularly create jobs to solve the high unemployment rate observed in the country and ensure
economic growth and transformation. Currently, MSEs are becoming a significant job supplier in the urban labor
market, as they have created 2.5 million new jobs in the last four years, particularly for unemployed youth.
This paper examines the survival of the MSEs and the associated factors based on a data obtained from 941 MSEs in
Ethiopia. Both descriptive and survival econometric models are used for analyzing the data. The analysis is done for
the whole sample of enterprises and separately for two sub-groups. The sub-group includes 569 micro and 372 small
enterprises and is grouped on the basis of their registration certificate and initial number of workers. Factors
associated with the survival of MSEs are analyzed using the Cox-proportional hazard model. The papers estimates
the effect of the characteristics that are typically identifiable at the time of establishing the enterprise on the MSE's
survival.
The results from the pooled data show that gender, age, previous work experience, motivation, entrepreneurship
training, initial size (proxied by initial number of workers and size of start-up capital), ownership type, type of
industry the firm operating in and legal status (registration) are important factors in explaining the probability of
MSEs survival in Ethiopia. In particular, enterprises which are owned by women and older individuals have lower
survival rate while those enterprises which are established by youths with previous work experience, driven by
opportunity motivation, have entrepreneurship training, with sole-proprietor ownership, engaged in production, and
registered have higher survival rate than otherwise enterprises.
The analysis also confirms that there is survival differential among micro and small enterprises and the risk of drop-
out is high during the first two business operation years of MSEs in Ethiopia. Moreover, the analysis based on the
sub-groups, reveals that gender, entrepreneurial training and size of start-up capital are the major factors accounted
for the better success of small enterprises compared to micro enterprises.
14
Policy implications
Having identified the main factors which affect the survival rate of urban MSEs in Ethiopia, it is possible to forward
some policy implications that the government and MSE promoters are responsible for further development of MSEs
as the GTP asserted. First, the evidence that women entrepreneurs are highly linked with high risk of drop-out calls
for greater need of tailored support programs for women entrepreneurs to address the specific challenges they are
facing. This could be intense training sessions for women entrepreneurs, design new women biased financial
products, and arranging government subsidies in the form of tax exemption during a start-up period or start-up
allowance and guarantee for women business founders. Second, entrepreneurial education is found to be a key factor
for the success of MSEs, thus there is a need to provide the training as a package in any TVET programs or in the
form of short-term training when ever new business founders knock the door for the support programs. MSE
promoters also should give emphasis for entrepreneurial education beyond their support in terms of physical and
financial capital.
Third, it has been noted in the result that MSEs engaged in manufacturing sector have better chance of survival and
thus, policy makers should keep their eyes on product providing MSEs, as they are the breeding-ground to create
industry led economy.
Last but not least, micro enterprise founders are suffering from less survival rate compared to small scale enterprise
owners. This could be due to the inclination of most support programs towards small enterprises or overlooking of
the importance of micro enterprises in the process of creating viable and sustainable medium and large industries
from both sides - the MSE promoters and micro scale enterprise owners. Unless and otherwise the responsible body
should give much emphasis on creating conducive environment for micro enterprise, the process of leading MSEs to
medium and large industries will be on the wrong track. We, therefore, propose to the MSE promoters, in general
and FeMSEDA, in particular, to adhere a separate support procedure for micro business founders during screening,
training and follow-up stages, so that it will be possible to identify the specific gaps and solutions related to micro
enterprises.
15
Micro-level characteristics
Owner
Firm
Industry
Firm Survival
Performance of the economy Functionality of market Institutional development
Macroeconomic and Institutional Factor
Inclusive support
- Gender
- Age
- Education
- Managerial skill
and expertise
- Entrepreneurial
learning
- Training
- Motivation
- Ethnicity
- Risk aversion
- Size
- Age
- Location
- Ownership
form
- Formal
system
- Type of sector
- Competition
- Market Linkage
(cluster)
- Industry
performance
- Technological
change and
innovation
- Business cycle
- Inflation
- Relative price volatility
- Capital market
- Labor market
- Product market
- Policy and regulatory environment
- Quality of support program and
implementation capacity
- Governance
- Bureaucracy
- Custom regulation
- Infrastructure development
Facilitators: state and non-state actors
Business strategy
- Human capital
- Planning
- Financial
management
- Marketing
management
- Internet
(ecommerce)
Figure 1 Conceptual framework to study the determinants of firm survival
16
Figure 2: The non-parametric survival and hazard rate estimation
Figure 3: The survival and hazard rate deferential between micro and small enterprises
17
Table 1: Definition of variables and their corresponding mean and standard deviation value
Covariates Description Mean Std.dev
Part A : Owner related
characteristics Socio-
demographic and economic
Female 1 if the owner is female( solo and group) 0.255 0.436
Age Age at the time of establishment 24 4.42
Elementary school 1 if the owner has elementary qualification 0.249 0.432
College diploma and above 1 if the owner had college diploma and above 0.166 0.372
TVET 1 if the owner has attended in any technical
vocational education and training programs 0.100 0.300
Active in a labor market
1 if the owner has any work employment
experience in the labor market prior to start the new
business 0.514 0.500
Self-employment experience 1 if the owner was self-employed 0.122 0.328
Experience and training
Apprentice Experience 1 if the owner or the manger had an apprentice
experience prior to this business 0.329 0.154
Business experience 1 if the owner or the manger had a business
experience prior to start the business 0.501 0.500
Entrepreneurship training 1 if the owner took entrepreneurship training before
establishing the business 0.143 0.351
Short term training 1 if the owner took short-term training before
establishing the business 0.350 0.477
Motivation
Opportunity
1 if the owner was opportunity driven entrepreneur 0.690 0.463
Part B: Firm and industry related
characteristics
Initial firm size
Initial number of workers Initial number of workers 8 10.6
Start-up capital
Large capital 1 if the start-up capital was more than 100,000 0.055 0.229
Micro capital 1 if the start-up capital was less than 25,000 0.833 0.373
Legal form
Sole-proprietorship 1 if the business started as a solo business 0.383 0.486
Partnership 1 if the business was established by partners 0.312 0.464
Cooperative 1 if the business was established as a cooperative 0.262 0.440
Industry type
Metal and wood work 1 if the enterprise engaged in metal and wood
works 0.258 0.438
Leather products 1 if the enterprise engaged in leather production
industry 0.015 0.121
18
Textile and cloth 1 if the enterprise engaged in textile and cloth
production industry 0.048 0.214
Construction 1 if the enterprise engaged in construction 0.197 0.398
Service 1 if the enterprise engaged in service sector 0.114 0.318
Product provider 1 if the enterprise was established as a product
provider 0.705 0.456
Service provider 1 if the enterprise was established as a service
provider (e.g. service and trade) 0.295 0.456
Support 1 if the owner received any kind of support from
any recognized support provider 0.661 0.474
Formal system
Registered 1 if the enterprise registered officially during
establishment 0.810 0.393
19
Table 2: t-test mean equality of owner, firm and industry characteristics of survivors and dropouts by the initial
size of the enterprises and by the type of entrepreneurs1
Survival Drop-out
Variable All MSEs Micro Small Micro Small P12 P2
2
Socio-demographic and economic
Female 0.255 0.279 0.157 0.466 0.212 0.0000 0.0004
Age (In number of years old) 24 24 24 26 26 0.4566 0.2673
Elementary 0.249 0.247 0.231 0.291 0.273 0.3198 0.3979
College diploma and above 0.166 0.172 0.170 .0126 0.167 0.4750 0.2327
TVET 0.100 0.109 0.118 0.029 0.030 0.3624 0.4826
Active in the labor market 0.514 0.543 0.502 0.447 0.470 0.1405 0.3852
Self-employment experience 0.122 0.105 0.134 0.155 0.136 0.1113 0.3682
Experience and Training
Apprentice Experience 0.024 0.350 0.314 0.370 0.200 0.1500 0.0087
Business experience 0.501 0.521 0.471 0.485 0.515 0.0836 0.3541
Entrepreneurship training 0.143 0.131 0.183 0.126 0.076 0.0242 0.1512
Short term training 0.350 0.264 0.418 0.398 0.561 0.0000 0.0194
Motivation
opportunity 0.690 0.781 0.644 0.612 0.379 0.0000 0.0015
Initial firm size
Small 0.396
Legal form
Sole-proprietorship 0.383 0.541 0.183 0.466 0.061 0.0000 0.0000
Partnership 0.312 0.290 0.343 0.359 0.258 0.0585 0.0844
Cooperative 0.262 0.122 0.418 0.175 0.667 0.0000 0.0000
Start-up capital
Large capital 0.055 0.024 0.111 0.039 0.045 0.0000 0.4172
Micro capital 0.833 0.906 0.729 0.874 0.742 0.0000 0.0146
Industry type
Metal and wood work 0.258 0.294 0.271 0.155 0.106 0.2470 0.1825
Leather products 0.015 0.017 0.013 0.000 0.030 0.3266 0.0382
Textile and cloth 0.048 0.051 0.059 0.000 0.045 0.3307 0.0145
Construction 0.197 0.146 0.271 0.136 0.303 0.0000 0.0040
Service 0.114 0.127 0.069 0.204 0.091 0.0048 0.0255
Product provider 0.700 0.670 0.807 0.544 0.727 0.0000 0.0083
Service provider 0.300 0.330 0.193 0.456 0.273 0.0000 0.0083
Others
Support 0.661 0.556 0.794 0.592 0.894 0.0000 0.0000
Registered 0.810 0.755 0.912 0.718 0.864 0.0000 0.0137
Observation 466 306 103 66 1 Backgrounds are found from the retrospective information. Numbers are shares unless stated otherwise.
2 p-values refer to t-test of mean equality in the variables between micro and small firms in the status of survival
and failure, respectively.
20
Table 3: Partial likelihood estimates for the Cox proportional hazard models
Covariates MSEs Micro Small
Female 0.683*** 0.907*** 0.370
(0.0001) (0.0000) (0.2536)
Age (at startup) 3.244*** 3.635*** 2.714**
(0.0000) (0.0000) (0.0011)
Elementary school -0.111 -0.065 -0.034
(0.5434) (0.7941) (0.9129)
College diploma and above 0.037 -0.023 0.252
(0.8773) (0.9379) (0.4907)
TVET -0.523 -0.462 -0.612
(0.2622) (0.4448) (0.4084)
Active in labor market -0.388* -0.400 -0.332
(0.0440) (0.1194) (0.2970)
Self-employment experience 0.309 0.494 0.047
(0.2426) (0.1377) (0.9213)
Business experience 0.191 -0.042 0.625†
(0.2922) (0.8572) (0.0510)
Apprentice experience 0.306 0.574* -0.281
(0.1136) (0.0157) (0.4538)
Opportunity -0.837*** -0.712** -0.974**
(0.0000) (0.0027) (0.0014)
Entrepreneurship training -0.543* -0.178 -1.210*
(0.0347) (0.5640) (0.0147)
Initial number of workers -0.198*
(0.0382)
Large start-up capital -0.824† 0.358 -1.433*
(0.0620) (0.5780) (0.0295)
Micro start-up capital -0.272 -0.134 -0.377
(0.2454) (0.7191) (0.2422)
Sole-proprietorship -0.898*** -1.016*** -1.026†
(0.0002) (0.0001) (0.0915)
Cooperative 0.312 0.055 0.478
(0.1381) (0.8571) (0.1370)
Product provider -0.598*** -0.712** -0.470
(0.0006) (0.0010) (0.1518)
Support -0.092 -0.219 0.238
(0.6413) (0.3370) (0.5865)
Registered -0.503* -0.485* -0.411
(0.0126) (0.0449) (0.2948)
LHR test 0.0000 0.0000 0.0000
Wald test 0.0000 0.0000 0.0000
Log rank test 0.0000 0.0000 0.0000
Note: Figures in parenthesis are p-values and †, *, ** and *** indicate statistically significant
at the 10%, 5%, 1% and 0.1 % level, respectively.
21
Appendices
I. Tables
Table A: Survival and hazard rate from the Lifetable estimator for Small/Micro enterprises
Small Micro
Survival rate difference (%) Duration At risk Survival hazard At risk Survival hazard
0-6 372 1 0.0027 569 1 0.0024 0
6-12 358 0.9837 0.0034 533 0.9856 0.0055 -0.19
12-18 320 0.9636 0.0072 430 0.9534 0.0103 1.02
18-24 280 0.9227 0.0044 348 0.8963 0.0037 2.64
24-30 251 0.8987 0.0094 286 0.8767 0.0103 2.20
30-36 210 0.8494 0.0043 233 0.8243 0.0064 2.51
36-42 181 0.8280 0.0040 183 0.7932 0.008 3.48
42-48 152 0.8083 0.0024 151 0.7561 0.0063 5.22
48-54 124 0.7967 0.0059 113 0.728 0.0153 6.87
54-60 103 0.7691 0.0018 83 0.6641 0.0046 10.5
60-66 80 0.7607 0.0024 63 0.6461 0.0000 11.46
66-72 60 0.7499 0.007 47 0.6461 0.0000 10.38
72-78 35 0.7190 0.0056 30 0.6461 0.0000 7.29
78-NA 25 0.6954 0.0000 19 0.6461 0.0000 4.93
Note: The survival rate difference is the difference of cumulative proportion of small and micro entrepreneurs‟
survival rate up to specific time interval.
22
Table B: Proportional hazard test using Chi-square statistics: Cox
model analysis-Part one
Covariates P-value
Female 0.112
Age (at startup) 0.559
Elementary school 0.523
College diploma and above 0.347
TVET 0.967
Active in labor market 0.959
Self-employment experience 0.333
Business experience 0.487
Apprentice experience 0.552
Opportunity 0.824
Entrepreneurship training 0.988
Initial number of workers 0.471
Sole-proprietorship 0.329
Cooperative 0.731
Large start-up capital 0.954
Micro start-up capital 0.286
Product provider 0.611
Support 0.985
Registered 0.591
Global 0.951
23
II. Figures
Figure A: Graphical diagnostic Schoenfeld residuals test: Cox model analysis
24
Figure A: Continued
25
Figure A: Continued
26
Figure A: Continued
27
References
Agarval, R. and Audretsch, D.B. (2001), Does Entry Size Matter? The Impact of the Life Cycle and Technology on
Firm Survival, Journal of Industrial Economics, 49, 21-43.
Alberti, F., Sciasca, S., & Poli, A. (2004). Entrepreneurship education: Notes on an ongoing debate. 14th Annual
IntEnt Conference. Italy: University of Napoil Federico II.
Amran N. A. (2011). The effect of Owner‟s Gender and Age to Firm Performance: A Review on Malaysian Public
Listed Family Businesses. Journal of Global Business and Economics 2(1):104-116.
Aspray, W. and Cohoon, J. (2007). Gender Differences in Firm Size, Growth, and Persistence: A Review of
Research Literature on Women‟s Entrepreneurship in the Information Technology Field. National Center for
Women and Information Technology. Entrepreneurial Report Series.
Aterido, R., Hallward-Driemeier, M. and Pagés, C. (2009), Big Constraints to Small Firms‟ Growth? Business
Environment and Employment Growth Across Firms, World Bank Policy Research Working Paper 5032,
Washington DC, World Bank.
Audretsch, D.B. and Mahmood, T. (1995). New firm survival: New results using a hazard function. The Review of
Economics and Statistics 77 (1): 97-103.
Bates, T. (1995). A Comparison of Franchise and Independent Small Business Survival Rates. Small Business
Economics 7, 377-388.
Bates, T. (2005). Analysis of young, small firms that have closed: Delineating successful from unsuccessful
closures. Journal of Business Venturing 20, 343-358.
Birley, S. and Westhead, P. (1994). A taxonomy of business start-up reasons and their impact on firm growth and
size. Journal of business venturing 9(1):7-31.
Blackbur, R. A., Hart, M. and Wainwright, T. (2013). Small business performance: business, strategy and owner-
manager characteristics. Journal of Small Business and Enterprise Development 20 (1): 8-27.
Block, J. and Sandner, P. (2009). Necessity and Opportunity Entrepreneurs and their Duration in Self-employment:
Evidence from German Micro Data. Berlin: SOEP papers on Multidisciplinary Panel Data Research.
Boswell, J. (1973). The Rise and Decline of Small Firms, London: George Allen & Unwin.
Bruderl, J., P. Preisendorfer, and R. Ziegler, (1992). Survival Changes of Newly Founded Business Organizations.
American Sociological Review 57(2):227-242.
Caliendo, M., & Kritikos, A. (2009). “I Want to, But I Also Need to”: Start-Ups Resulting from Opportunity and
Necessity. IZA Discussion Paper No. 4661 .
Calvo, J.L. (2006), Testing Gibrat‟s Law for Small, Young and Innovating Firms, Small Business Economics, 26,
117-23.
Carpenter, R. E. and Petersen, B. C. (2002). Is the Growth of Small Firms Constrained by Internal Finance. The
review of economics and statistics 84(2): 298-309.
28
Carter, N., Gartner, W., Shaver, K. and Gatewood, E. (2003). The career reasons of nascent entrepreneurs. Journal
of business venturing 18, 13-39.
Carter, R. andVan Auken, H. (2006). Small firm bankruptcy. Journal of Small Business Management 44 (4), 493-
512.
Central Statistical Agency (CSA). Population and Housing Census in Ethiopia. Addis Ababa, 2007.
Chandler, G. N. and Hanks, S. H. (1998). An Examination of the Substitutability of Founders Human and Financial
Capital in Emerging Business Ventures. Journal of Business Venturing 13, 353-369.
Coad, A. and Tamvada, J. P. (2008). The Growth and Decline of Small firms In Developing Countries. Papers on
Economics and Evolution No. 0808.
Coleman, S. (2005). The Impact of Human Capital Measures on Firm Performance: A Comparison by Gender, Race
and Ethnicity. The Journal of Entrepreneurial Finance 10(2):38-56.
Coleman, S., Cotei, C. and Farhat, J. (2010). Factors Affecting Survival, Closure and M&A Exit for Small
Businesses. Available at SSRN: http://ssrn.com/abstract=1768728
Cooper, A. C., Gimeno-Gacson, F. J. and Woo, C. Y. (1994). Initial Human and Financial Capital as Predictors of
New Venture Performance. Journal of Business Venturing, 9: 371-395.
Cowling, M. (2009, January). The impact of entrepreneurship training and small business expeience on future
entrepreneurial activity in the UK. IES working paper. Bringhton, UK: Institute for Employmnet Studies.
Cressy, R. (1996). Are Business Startups Debt-Rationed? The Economic Journal 106, 1253-1270.
Dana, P. (2001). The education and training of entrepreneurs in Asia. Education & Training 43( 8/9 ):405-415.
Davidsson, P. (1991). Continued entrepreneurship: ability, need and opportunity as determinants of small firm
growth. Journal of Business Venturing, 6: 405-429.
Dayanandan, R. (2012). Sustainability of Micro and Small Enterprises: Imperatives, Myths and Realities.
International Journal of Entrepreneurship & Business Environment Perspectives. 1(2): 129-142.
Disney, R., Haskel, J., Heden, Y., (2003). Entry, exit, and establishment survival in U.K. manufacturing. Journal of
Industrial Economics 51 (1):91-112.
Dobbs, M. and Hamilton, R. (2007). Small business growth: recent evidence and new directions. International
Journal of Entrepreneurial behaviour and research, 13 , 296-322.
Dunne, P. and Hughes, A. (1994), Age, Size, Growth and Survival: UK Companies in the 1980s, Journal of
Industrial Economics, 42, 115-40.
Fadahunsi, A. (2012). The growth of small business: Towards a research agenda. American journal of economics
and business administration 4(1):105-115.
Fairlie, R. W. and Robb, A. M. (2008). Gender Differences in Business Performance: Evidence from the
Characteristics of Business Owners Survey. IZA Discussion Paper Series No. 3718.
29
FeMSEDA. (2013/14). Micro and Small Enterprises Sector Development Yearly Statistical Bulletin. Addis Ababa,
Ethiopia.
Garcia-Teruel, P. J., & martinez-Solano, P. (2007). Effects of working capital management on SME profitability.
International journal of mangerial finance, 3(2 ):164-177.
Goi, C. L. (2009). A Review of Marketing Mix: 4Ps or More? International Journal of marketing studies, 1(1) .
Gopal, C. C. (2008). Finanacial Management. New Delhi: New Age International Publishers.
Hassen, Y. A., & Svensson, A. (2014). The role of e-commerce for the growth of small enerprises in Ethiopia. The
electronic journal of information system in developing countries, 65 , 1-20.
Headd, B. (2000). Business success: Factors leading to surviving and closing successfully. U.S. Small Business
Administration, Office of Advocacy.
Holmes, P., Stone, I., Braidford, P., (2008). An analysis of new firm survival using a hazard function. Working
Paper
Honig, B. (1998). What Determines Success? Examining the Human, Financial, and Social Capital of Jamaican
Microentrepreneurs. Journal of Business Venturing 13, 371- 394.
Hoskisson, R.E., Eden, L., Lau, C.M. and Wright, M. (2000), Strategy in Emerging Economies, Academy of
Management Journal, 43, 249-67.
Ibidunni,O..,, Oluwole, I., & Ayodotun S, I. (2014), Product Innovation, a survival strategy for small and medium
enterprises in nigeria, European Scientific journal, 10, 1 , 194-209.
ILO (2007), The promotion of sustainable enterprises. International Labor Conference, 96th
Session, Report VI.
ISBN 978-92-2-118143-9 ISSN 0074-6681.
Jenkins, Stephen P. (2004). Survival Analysis. Unpublished manuscript, Institute of Social and Economic Research,
University of Essex, Colchester, Uk. Downloadable from
http://www.iser.essex.ac.uk/teaching/degree/stephenj/ec968/pdfs/ec968lnotesv6.pdf
Kalleberg, A.L. and Leicht, K. T. (1991). Gender and Organizational Performance: Determinants of Small Business
Survival and Success. Academy of Management Journal. 34(1): 136-161.
Kangasharju, A. and Pekkala, S. (2002). The Role of Education and Self- Employment Success in Finland. Growth
and Change 33, 216-237.
KiatGan, C. and Almsafir, M. K. (2013). The determinanats of SME succession in Malaysia, from enterpreneurship
perspective. Journal of advanced social research, 3(12 ):350-361.
Kotler, P. and Armstrong, G. (1999). Principles of Marketing, 8th edition. New Delhi: Prentice-Hall of India Private
Limited.
Kraus, S., Harms, R. and Schwarz, E. J. (2006). Strategic planning in smallerenterprise-new emperical findings.
Mangement research news, 29(6 ):334-344.
30
Lussier, R.N. and Pfeifer, S. (2001). A Crossnational Prediction Model for Business Success. Journal of Small
Business Management 39 (3), 228-239.
Marimuthu, M., Arkiasamy, L. and Ismail, M. (2009). Human Capital Development and Its Impact on firm
Peformance: Evidence from Developmental Economics.The Journal of International Social Research, 2(8 ):265-
272.
Mata, J. and Machado, J. (1996). Firm Start-up Size: a Conditional Quantile Approach. European Economic Review.
Elsevier 40(6): 1305-1323.
Mata, J., Portugal, P. and Guimaraes, P. (1995), The Survival of New Plants: Start-up Conditions and Post-entry
Evolution, International Journal of Industrial Organization, 13, 459-82.
McMahon, R. G. (2001). Growth and Performance of Manufacturing SMEs: The influence of fianancial
management characterstics. International small business journal, 19(3 ):10-28.
McMahon, R. G. and Stanger, A. M. (1995). Understanding the small enterprise fianancial objective function.
Entrepreneurship: Theory and Practice, 19 (4):21-39.
Mongare, M. E. and Nasidai, S. E. (2014). The impact of information communication technology on inventory
control systems in transport organization: A case study of kenya ferry services. European Journal of Logestic and
Supply Chain Management, 2(1 ):17-41.
O'Dwyer, M., Gilmore, A. and Carson, D. (2009). Innovative marketing in SMEs. European Journal of Marketing,
43(1/2 ): 46-61.
Paramasivan, C., & manian, T. S. (2009). Financial management. New Dehli: New age International Limited
Publishers.
Parker, S.C. and Belghitar, Y. (2006). What happens to nascent entrepreneurs? An econometric analysis of the
PSED. Small Business Economics 27, 81-101.
Pena, I. (2002). Intellectual Capital and Business Start-up Success. Journal of Intellectual Capital 3 (2), 180-198.
Pfeiffer, F.; Reize, F. (2000) “Business Start-ups by the Unemployed – An Econometric Analysis Based on Firm
Data” Labour Economics; 7, 629-63.
Quatraro. F. and Varellie M. (2013): Entrepreneurship in Developing Country Context. Department of Economic
Statistics, Working Paper Series 14/13/.
Read, L. (1998). The fianancing of small business: A comparative study of male and female business owners.
London : Routledge.
Reynolds, P., Camp, M., Bygrave, W., Autio, E. and Hay, M. (2001). Global entrepreneurship monitor 2001
excutive report. Babson college, London business school.
Savery, L. K., & Luks, J. A. (2004). Does training influence outcomes of organizations? Management developmnet ,
23(2 ):119-123.
Schiller, B.R. and Crewson, P. E. (1997). Entrepreneurial Origins: A Longitudinal Inquiry. Economic Inquiry 35(3),
523-531.
31
Shafique, M., Rizwan, M., Jahangir, M., Mansoor, A., Akram, S. and Hussain, A. (n.d.). Determinants of
entrepreneurial sucess/failure from SMEs perspective. Journal of Business and Management, 2278-487X, 2319-
7668 , 83-92.
Skrt, B., & Antoncic, B. (2004). Strategic planning and small firm growth: an emperical examination. Managing
global transitions 2 (2) , 107-122.
Stearns, T.M., Carter, N.M., Reynolds, P.D. and Williams, M.L. (1995). New Firm Survival: Industry, Strategy and
Location. Journal of Business Venturing 10, 23-42 © 1995 Elsevier.
Storey, D. (1994), Understanding the small business sector. London: Routledge
Solomon, T. G., Weaver, K. and Jr, L. W. (1994). A historical examination of small business management and
entrepreneurship pedagogy. Simulation and gamming 25(3 ): 338-352.
Taylor, M. P. (1999). Survival of the Fittest? An analysis of self-employmnet duration in Britain. The Economic
Journal, 109 ,C140-C155.
Warren, M. (2004). Farmers online: drivers and impediments in adoption of internate in UK agricultural businesses.
Journal of small business and enterprise development, 11 (3 ): 371-381.
Woldie, A., leighton, P. and Adesua, A. (2008). Factors influencing small and medium enterprises (SMEs): An
exploratory study of owner/manager and firm characterstics. Banks and Bank Systems, 3 (3 ).
Yasuda, T. (2005), Firm Growth, Size, Age, and Behavior in Japanese Manufacturing, Small Business Economics,
24, 1-15.
World Bank. (2011). Industrial clusters and micro and small enterprises in Africa. Washington, D.C. The World
Bank.