“The effect of entrepreneurship on economic eevelopment in East Africa -- a trend-line analysis of GEM’s necessity driven and opportunity driven TEA rates”

(1)

“The Effect of Entrepreneurship on Economic Development in East

Africa — a trend-line analysis of GEM’s necessity driven and

opportunity driven TEA rates”

Floris Overstegen 10457046

Final version - Friday 22 June 2018

MSc. Business Administration – Entrepreneurship & Innovation University of Amsterdam

(2)

Statement of originality

This document is written by Floris Overstegen who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

Abstract

Most studies regarding the influence of entrepreneurship on economic development are subjected to developed economies. Because of substantial differences in the nature of entrepreneurship and the driving forces behind economic development between developed and developing countries, this paper examines the effect of entrepreneurship on economic development in East Africa. By using GEM’s entrepreneurial data of Botswana, Zambia, and Uganda between 2003 and 2015, this paper conducts a longitudinal study with a trend-line analysis aiming to observe significant differences between the influence of necessity driven and opportunity driven TEA rates on GDP growth rates in East Africa. By the knowledge of the author, this is the first longitudinal study performed in these countries regarding this topic. Expected results were found regarding the development of the trend of variables over the years. Necessity driven entrepreneurship is indicated to change into opportunity driven entrepreneurship in East Africa. However, due to insignificance and low correlations between the variables, entrepreneurship is concluded not to be a main driver for economic development in East Africa. Because the economic state is factor driven, their focus should primarily be on stimulating the fundamentals for a well-functioning business environment. The government is recommended to implement policies for stimulating a movement towards an efficiency driven economy. Once this is established, entrepreneurial activity will have greater effects on East Africa’s economic development.

(3)

The role of entrepreneurship in economic development originates from the question of how entrepreneurs react on transformations in economic stages and what role they play in this transformation. In recent decades entrepreneurship has been analyzed from many different points of view and is shaped into a multidimensional concept which is not easily defined. Although its impact on economic development is widely acknowledged, this relationship still remains a relatively under-researched phenomenon, especially in developing countries. Lingelbach et al. (2005) even stated: “Entrepreneurship in developing countries is arguably the least studied significant economic and social phenomenon in the world today”. This paper aims to contribute by analyzing the influence of entrepreneurial activity on economic development in East Africa.

The main reason for the scope of this research is based on an internship at a Dutch startup called Caspar Coding. Their idea is to connect well-educated East African software developers with European organizations. Europe is currently put up with a huge gap between the demand and supply of software development. The high demand for Western software developers makes them very expensive and for most startups even unaffordable. This is affecting the ease to start up a business, which is the cornerstone to current economic development in Europe (Thurik & Wennekers, 2004). For this reason, Caspar Coding believes the outsourcing of technical development is inevitable. Their goal is to serve this European demand by integrating African tech talent into the European business environment. An important aspect of their business model is the strong focus on a good relationship between the developer and the client. In order for the developer to be most efficient, they have to

(5)

at least 1 year is required. This way they can establish a relationship and also provide the maintenances of the specific software they have built. Additionally, it will also be a rare opportunity for African developers to gain experience and learn from a European perspective of doing business.

A key characteristic of Caspar Coding is their contribution to social responsibility. They are not only interested in serving the European economy with their business model but also feel obligated to contribute to the African economy. In order to do this, they introduced a savings plan for the African developers. Since every developer works under Caspar Coding’s paycheck, 10 percent of their monthly salary will be saved up for them into a personal savings account. Additionally, Caspar Coding will be building business incubators in which the developers can use their saved up money to develop their own business plans and introduce them into the African market. Because of the gained European experience, the high market potential in Africa, and the upfront financial resource, promising ideas and results are expected.

This ‘Incubator Project’ should be a stimulator for entrepreneurship in East Africa and is expected to contribute to economic development (Acs, Desai & Hessels, 2008). An important element in this research regarding entrepreneurship and economic development, is the discovery of very high entrepreneurial activities among developing countries. This discovery resulted in the distinction between necessity driven and opportunity driven entrepreneurship. Necessity driven refers to people who enable in entrepreneurial activity because of financial pressure and no other job options. Opportunity driven refers to contrasting motives like observing opportunities in the market and acting upon it. Only opportunity driven entrepreneurship is assumed to stimulate economic development, though

(6)

these concepts will be discussed more detailed in the next chapters. Most noteworthy is that the ‘Incubator Project’ is expected to stimulate opportunity driven entrepreneurship.

Researchers like Acs, Desai, and Hessels (2008) have concluded that opportunity driven entrepreneurship has a positive effect on economic development. However, these conclusions cannot be assumed to provide the same results in the context of East Africa. Because Caspar Coding is aiming to contribute to their social responsibility by stimulating opportunity driven entrepreneurship, it is important to research whether this will actually have a positive influence on economic development in East Africa. It is also important to understand the influence of necessity driven entrepreneurship on economic development. Therefore, the topic of this research will be the characteristics of entrepreneurship on economic development in East Africa. Central to this study is the answering of the following research question:

What is the effect of entrepreneurship on economic development in East Africa and what is the influence of necessity driven and opportunity driven entrepreneurship?

This research will be a valuable contribution to scientific literature because the topic of entrepreneurial characteristics and economic development has hardly ever been tested in the context of East Africa. Most existing papers about opportunity and necessity driven entrepreneurship and economic growth are measured in developed countries. These countries have a lot of factors that could differ the outcome of the same study in a developing country. Also, a longitudinal study has never been performed in these countries regarding this topic. Hence, interesting outcomes are expected to find.

(7)

To generate a well-analyzed answer to the research question, this research will first discuss some existing theories on entrepreneurship, economic development, their relationship, and the context of East Africa, in the literature review. Then, the methodology section will analyze the research design for this paper by elaborating on the chosen data and sample collection, and the variables used in the model. The main findings will be discussed in the results section, which will be followed by the discussion section where some implications for policies and further research are given. Finally, a conclusion is made and the research question will be answered.

2. Literature Review

Entrepreneurship and its relation to economic development is a relatively new concept in scientific literature. This section will provide a brief overview on the important studies concerning this topic. It contributes to a wider understanding of the general theories and defines which assumptions can be made regarding this topic. This section will also focus on the East African context of the discussed theories.

2.1. Defining entrepreneurship

Although entrepreneurship is a well-known word in our current vocabulary, it is hard to give it a clear definition. Throughout history, entrepreneurship has been studied by many researchers who have given it many different perspectives, descriptions, and definitions. This

(8)

The first academic acknowledgements regarding entrepreneurship were presented by the French-Irish economist Richard Cantillon in 1755 (Iversen, Jørgensen & Malchow-Møller, 2007). He saw the entrepreneur as the opposition to wage workers and landowners who receive a fixed income or rent. The entrepreneur has an uncertain income and for this, he is an arbitrageur who bears risks and equilibrates supply and demand in the economy. Cantillon characterized the entrepreneur as responsible for all exchange and circulation in the economy (Iverson et al., 2007). After Cantillon many highly acknowledged economist has made assumptions about entrepreneurship. Joseph Schumpeter’s thoughts on entrepreneurship, however, resemble a typical definition of entrepreneurship from the managed economy era (Hebert & Link, 1989). In contrast to the existing view of an entrepreneur being a risk bearer, Schumpeter describes entrepreneurs as innovators, people who create new goods or methods of production or even new markets and industries. Given this definition, Schumpeter described entrepreneurship as “the creation of new economic activity, where the development and renewal of a business as a whole depends upon micro-level actors” (Hebert & Link, 1989). Schumpeter’s entrepreneur doesn’t combine the input factors of the production function to achieve the highest efficiency, instead, he disrupts the production function with his innovations. Hence, he moves the economic model into disequilibrium by creating new goods, production methods or markets and industries. So the entrepreneur destroys existing structures by creating new combinations. This process of ‘creative destruction’ is the essence of economic development. The more people engage in entrepreneurship and create a disequilibrium in the circular flow of the economic process, the more it leads into new innovations (Hebert & Link, 1989).

(9)

because knowledge is neither complete nor perfect, and shocks are constantly affecting the economy. In his vision, entrepreneurs are alert for new business opportunities to exploit. This way, they are eliminating the imperfect knowledge in the economy and helping it move towards equilibrium, which is the state where no more opportunities can be discovered (Iverson et al., 2007).

The discussion whether entrepreneurship has an appealing or a repulsive effect on the economy’s equilibrium is not important for this research. Most noteworthy to understand, is that whether the entrepreneur discovers opportunities and reacts upon it, or when he actually creates opportunities, in both ways the process of making these new combinations contributes to the economic development within a country. In the next section, entrepreneurship and its relation to economic development will be discussed more detailed.

2.2. Entrepreneurship and economic development

In the early stage of the 20th century, large firms were seen as the dominant drivers for the economy. The neo-classical theories defined capital and labor as the main sources of economic development (Thurik, 2009). However in the eighties, due to unemployment and stagnating inflation rates, new theories regarding the drivers of economic development raised the attention. In 1957, Robert Solow discovered that a great part of economic development is related to total factor productivity, which is seen as the improvement of existing institutions combined with the creation of new knowledge (Acs & Szerb, 2009). This is seen as the ‘Solow residual’ and rises from the emergence of small businesses. One of the main drivers of this movement was globalisation, which was stimulated by western organizations who sought cheap labor and shifted their manufacturing

(10)

processes to developing countries. Another important element was the rise of technological development, which resulted in a better telecommunication infrastructure available for small firms. Hence, the total western manufacturing industry decentralized and was vertically disintegrated. This led to new innovation possibilities and economic opportunities for entrepreneurs (Wennekers et al., 2010).

According to Schumpeter, entrepreneurs will be disruptive by introducing new products or production processes into the markets (Acs & Szerb, 2009). This competitive behavior will force the established companies within the economy to reshape and improve their productivity. This will drive up the market processes and stimulate the necessary changes in the economy, which leads to higher output and higher market efficiency. At the same time, the emergence of the entrepreneur also stimulates the innovativeness of the output. The outcomes of the entrepreneurial competition, as described above, is assumed to result in economic development (Acs & Szerb, 2009).

The role of the entrepreneur on economic development is also explained by the Global Entrepreneurship Monitor (GEM). An important implication of the GEM is that the entrepreneur’s contribution to an economy varies, depending on the economic stage a country is in (Acs et al., 2004). Originated from the research of Porter et al. (2001), the GEM also sorts their analyzed countries into three different economic stages: the factor driven stage, the efficiency driven stage and the innovation driven stage. The role and reason for entrepreneurial activity vary between the different economic stages.

Factor driven economies typically have a large agricultural sector and are largely dependent on the extraction of natural resources. This is the most common stage for many Least Developed Countries (LDCs). The GEM has recorded high levels of entrepreneurial

(11)

activity in these economies. This is related to motivations of the entrepreneur to start a new business. Hence, in most factor driven economies people are forced into entrepreneurial activity due to financial restraints and the lack of other options. Next parts of this study will provide deeper insights on this specific subject.

Efficiency driven economies are characterized by the greater development of their industrial sector, higher productivity through economies of scale, and more developed financial institutions (Kelley & Herrington, 2012). In this stage there is a high emergence of large national businesses. For this reason, entrepreneurial activity is observed to reduce and the motivation for entrepreneurship shifts towards higher levels of opportunity driven entrepreneurship. Examples of efficiency driven economies are countries like China, East European countries and most of the Latin Americas.

Finally, innovation driven economies are marked by an expanding service sector with increasing research and development and a rise in knowledge-intensive businesses (Kelley & Herrington, 2012). In this stage there is a larger share of opportunity driven entrepreneurship with a high focus on smaller firms and a decreasing share of manufacturing. This is mainly stimulated by the improvements in information technologies and the productive infrastructure. GEM’s global reports indicate that the innovative entrepreneurial firms are significant drivers for economic development in these countries.

2.3. Measuring entrepreneurship

For many years, it was hard to determine the actual effect of entrepreneurship on economic development. Despite the wide agreement of a positive relationship, it could not be

(12)

analyzed, there was a lack of empirical evidence (Wennekers et al., 2010). This had an influence on the theory built on this subject. Although many researchers claimed a direct relationship between the two variables, it could never be academically supported due to a lack of empirical measurement of the variable: ‘entrepreneurship’. One of the reasons was that it was hard to find a clear definition of the concept but also a lack of cross-national measurements to support any claims.

Hart (2003) claims, that in a simple form, entrepreneurship can be measured by the creation of new firms or self-employment. By simply collecting the data of newly enrolled startups in an economy, a rough indication can be made of a country’s entrepreneurial activity. However, as Naudé (2014) describes, this simplistic perception of measuring entrepreneurship has been argued by many academics. For instance, in Schultz’ (1975) definition of an entrepreneur, entrepreneurship doesn’t need to result in the creation of new firms, it can also be seen as a part of a managing function within existing firms. Hence, entrepreneurship should have a wide set of measuring point to be empirically accurate.

A primary measure of entrepreneurship, developed by the GEM, is the Total Early-Stage Entrepreneurial Activity (TEA) rate. The TEA rate illustrates the relative amount of individuals engaged in blossoming entrepreneurship and new firm ownership in the adult population. Everyone in this population who is engaged in any behavior related to the business creation contributes to the national level of entrepreneurial activity. However, every entrepreneur is different in their profile and impact on the economy. For this reason, the GEM separates their data in a unique range of indicators to ensure fairly analyzed results on every observed country. So instead of just counting the number of entrepreneurs in a country, they also consider aspects like the level of employment creation, growth ambitions, age, and

(13)

gender. The indicators to measure the TEA rate are collected through the GEM’s Adult Population Survey (APS). This survey is conducted through a random representative sample of at least 2000 adults between 18 and 64 years old, at the same time of every year. It is strictly administered by academic teams of the participating countries and translated into the native language to ensure validity. Area stratified probability sampling is used to ensure representativeness of the sample. The survey is conducted by either accredited research companies or by a national team of GEM’s qualified survey vendors. Upon completion, the raw data is sent directly to the Global Data Team for quality check and uniform statistical calculations. After approval of the participating countries, the statistics are compiled into

(14)

annual global reports. The raw datasets are also released and are seen as high-quality data used often by researchers and policymakers (Eijdenberg & Masurel, 2013). A model of the indicators GEM uses to observe a country’s entrepreneurial activity is shown in Figure 1.

2.4. Necessity driven and opportunity driven entrepreneurship

After reviewing the data of the GEM, Reynolds et al. (2001) state how they were surprised by the high rate of entrepreneurial activity amongst developing countries. In some countries, the observed TEA rates were significantly higher than those of developed countries. As shown in Figure 2 and proved by Wennekers et al. (2010), the relationship between TEA (on the y-axis of Figure 2) and GDP per capita (x-axis) is U-shaped. Hence, high and low levels of GDP are related to high levels of entrepreneurial activity, medium levels of GDP are related to low levels of entrepreneurial activity. The discovery of this unexpected high entrepreneurial activity in developing countries led to the separation of entrepreneurship into necessity driven and opportunity driven. This separation explains the motivations, or as Eijdenberg and Masurel (2013) call it, push and pull factors of an individual to become a self-employed business owner. Necessity driven entrepreneurs are forced into entrepreneurial activity because they are unsatisfied with their current position and have few other options. On the other hand, opportunity driven entrepreneurs are motivated to exploit an identified business opportunity but otherwise had other options of employment (Langevang, Namatovu & Dawa, 2012). The data of GEM shows that a greater level of poverty results in higher levels of necessity driven entrepreneurship.

(15)

Acs, Desai, and Hessels (2008) conclude in their study, that opportunity driven entrepreneurship has a positive effect on economic development and that the income of an opportunity driven entrepreneur relates positively with the growth in GDP. They also suspect that economic prosperity would decrease the necessity for entrepreneurship. Because economic development usually has a positive effect on the supply of jobs, people will be less forced into self-employment. By analyzing GEM data of 37 (mostly developed) countries, Wong et al. (2005) showed insignificant results of necessity driven entrepreneurship and its relation with economic development. They stated that the presence of necessity driven entrepreneurship has no effect on a country’s economic development. Because their research Figure 2: Level of TEA related to GDP per capita U-Shaped. Source: Acs et al. (2004) Global Entrepreneurship Monitor

(16)

is of an exploratory nature though, the impact of necessity driven entrepreneurship cannot be neglected completely. This impact can also differ between countries and different economic states.

Hence, the GEM has been a vital element in analyzing the aspects of entrepreneurship. The discovery of high levels of entrepreneurial activity in countries with low levels of GDP per capita led to the perception of necessity driven and opportunity driven entrepreneurship. High levels of GDP lead to pulling factors for entrepreneurship, where low levels lead to pushing factors. The revelation of this U-shaped trend-line led to increasing academic attention for a more accurate analysis of the impact of entrepreneurship on economic development.

2.5. Entrepreneurship and economic development in East Africa

As Herrington and Kelley (2012) describe, the research for entrepreneurship has never been more relevant than in sub-Saharan Africa since there are huge changes taking place and the growth in GDP per capita is one of the highest in the world. Although high poverty and unacceptable unemployment rates are still prevalent aspects in most African countries, these positive movements push them into the right direction. The emergence of the entrepreneurial activity in these counties could be a part of the solution to the problems they are still facing. It could also be one of the drivers of the fast changes and growth in these countries.

In countries with a low GDP per capita, TEA rates are typically high and have a high proportion of necessity driven entrepreneurship. This is certainly the case in East Africa. For example, Zambia and Uganda have exceptionally high TEA rates of 40% and 35%

(17)

respectively, whereas the global average is 13,5%. They are in fact some of the highest TEA rates in the world. Despite these extraordinary high rates, these countries are still factor driven economies with a lot of problematic restrains for doing business. For instance, tax regulations disturb the ease of setting up an own business and they experience a lot of corruption. Also, the access to finance, which is vital for entrepreneurship to arise, is very limited. For this reason, it is assumed that in East Africa the basic requirements like macroeconomic stability, health, primary education, institutions, and infrastructure, will provide the fundamental conditions required for a well-functioning business environment. These requirements are and should be the focus of development efforts in factor driven economies like in East Africa. Once these factors are sufficiently established, the country’s economy moves towards the efficiency driven stage where funding and development efforts should focus toward the efficiency enhancers. These factors include financial market sophistication, technology, development, higher education and training and labour market efficiency (Herrington & Kelley, 2012). These assumptions reflect the very high rates of necessity driven entrepreneurship in East Africa. The entrepreneurial environment is not optimal as long as they are in the middle of establishing their basic fundamentals. However, this does not indicate that entrepreneurship in East Africa lacks in any impulse to stimulate economic developments. On the contrary, Kiggundu (2002) shows in his research how the emergence of technology and other external opportunities are being introduced to Africa. This resulted in an enhancement of the entrepreneurial environment which stimulated the opportunistic motives for entrepreneurship.

In summary, exceptionally high rates of entrepreneurial activity are observed in East Africa. Because they have a factor driven economy, this activity is mainly necessity driven.

(18)

They still have to focus on establishing their basic fundamentals in order to transform into an efficiency driven economy. However, they are in the middle of the process and the role of entrepreneurship will rise along these developments.

2.6. The influence of business incubators

In the research of Ogutu and Kihonge (2016), they define a business incubator as a “conducive environment where startups and new ventures are nurtured, ideas are developed to commercialization thereby building profitable, sustainable and scalable enterprises without exposing them to the harsh realities of the business environment and high startup costs. In the incubator tenants are offered training, mentorship and business support services”. Business incubators play a substantial role for entrepreneurs and the survival of their startups. In many parts of the world, incubators are seen as an important tool for promoting the development of startups, especially technology-oriented growth firms seem to experience a lot of these benefits (Bergek & Norrman, 2008). Since incubators have the characteristic to attract large investments, it is important to analyze the elements of successful incubator models. For this reason, research about incubators substantially contains the measurement of its outcome. Most literature doesn’t focus on the influence of incubators on macroeconomic aspects, like economic development or the degree of entrepreneurial activity, although there are some academic papers concerning this topic. Bergek and Norrman (2008) describe that despite the conducted researches about the influence of incubators on economic development, it is hard to generate assumptions due to a limitation of data. Because incubation outcomes may take years to materialize, they are difficult to analyze.

(19)

Nevertheless, Ogutu and Kihonge (2016) have focussed their research on the impact of incubators on economic and entrepreneurial development. They have investigated the possible connection between the number of startups, incubators, and the GDP growth within a country. By analyzing the regional population distribution of incubators against the selected countries’ GDP and TEA rate, they have explained the global role of business incubators in economic and entrepreneurial development. As a main result, they found a significant relationship between the number of incubators in a country and the level of GDP growth. In their model, results concerning the number of incubators and TEA rates were not significant. This could indicate a validity or reliability error in their measurements or an actual lack of relation between incubators and TEA rates. Another important finding resulted from the startup’s lifecycle analyze in their research. They have found the success and survival rate to be significantly high among incubated businesses, especially in developing countries. In the context of East Africa, they have concluded a success rate for regular Kenyan startups of 40%. For startups evolved in a business incubator, this success rate is about 70%. Since startups are the cornerstone of employment creation and poverty reduction, their rate of succession are assumed to have an impact on economic development (Ogutu & Kihonge, 2016). Hence in this study, an increase in the number of business incubators is assumed to have a positive effect on the entrepreneurial activity within a country. Moreover, it is assumed they will stimulate opportunity driven entrepreneurship, as they tend to attract sustainable opportunistic ideas and nurture them into startups.

(20)

2.7. Hypotheses

By taking the literature review into account, several assumptions can be made on the concepts around this topic. However, these assumptions have to be tested to validate them into the context of East Africa. Hence, this section will describe the hypotheses that will be tested throughout the rest of this research. The analysis of these tests will form a solid base in the process of answering the central research question. After the hypotheses are presented, they will be visualized by the conceptual model.

The first hypothesis reflects to the well-acknowledged assumption about the positive relationship between entrepreneurship and economic development (Acs et al., 2008; Wong et al., 2005). Although previous studies performed different research on the concept, this study will also assume a positive relationship between entrepreneurship and economic development. Nevertheless, this should be tested in order to draw conclusions regarding the scope of this research. Hence, the first hypothesis of this research is stated as follows:

H1: Entrepreneurship is positively related to economic development in East Africa.

The second hypothesis results from the given fact that entrepreneurship is separated into opportunity driven and necessity driven motives. Taking the studies of Reynolds et al. (2001) and Wennekers et al. (2010) into account, a simple confirmation of hypothesis 1 is not accurate enough to conclude on entrepreneurship in East Africa. This fact is therefore already included in the research question. Similar to hypothesis 1, conclusions on entrepreneurial motives and economic development of previous studies are assumed. Acs et al. (2008) and Langevan et al. (2012) both showed a positive relationship between opportunity driven entrepreneurship and economic development in their research. Yet these assumptions have to

(21)

be tested into the context of East Africa to draw conclusions for this research. Hence, the second hypothesis of this study is composed as follows:

H2: Opportunity driven entrepreneurship is positively related to economic development in East Africa.

consequently to hypothesis 2, an assumption about necessity driven entrepreneurship has to be made. Following up from the literature review, it shows that very little research on necessity driven entrepreneurship is performed on developing countries who experience very high rates. Wong et al. (2005) showed in a research on 37 countries a negative to zero relationship between necessity driven entrepreneurship and economic development. Since this has not been tested on a sample time frame in the context of East Africa, it is expected to find unique results. However, this study assumes necessity driven entrepreneurship will also have a negative, or no influence, on economic development in East Africa. Hence, the third and final hypothesis of this research is stated as follows:

H3: Necessity driven entrepreneurship is negatively / or not related to economic development in East Africa.

The hypotheses are visualized in the conceptual model in Figure 3, which in addition to my literature review functions as a framework for this research.

(22)

3. Methodology

This section will analyze the research design for this paper by elaborating on the chosen data and sample collection. After these aspects are analyzed, the variables used in the model will be discussed.

(23)

3.1. Introduction

Before defining the research design, first, it is interesting to understand the characteristics of researching entrepreneurship in East Africa. As Lingelbach et al. (2005) described, entrepreneurship in LDCs is probably the least studied economic and social element in the world. Most studies are performed in the context of developed countries because there is more data provided and since empirical studies on entrepreneurial effects are relatively new in today’s academics, there is a higher need for the research outcomes in developed economies (Lingelbach et. al, 2005). However, data collection and interest in developed countries have grown recently. Herrington and Kelley (2012) showed how GDP growth rates in sub-Saharan countries are amongst the highest in the world. This interest is supported by the GEM, who started collecting data on some developing countries since the beginning of this century. The literature review discussed some vital studies on the relationship between entrepreneurship and economic development. Most of these studies have made use of the GEM’s data on entrepreneurship. In the studies concerning developing countries though, the small sample sizes and data constraints have caused most of them into an exploratory nature. Despite their interesting results, the insignificance made it difficult to make accurate generalizations (Yao, Müller & Wang, 2005).

The research design chosen in this study is based on the low availability of entrepreneurial data in developing countries, and the research question and desired results taken into consideration. To analyze the effect of entrepreneurship on economic development, this study uses a time frame comparison between the variables in order to find patterns over time. Growth in GDP will represent economic development in this study. Analyzing growth translates in observing changes during a certain time frame. Given this method of

(24)

measurement of economic development, entrepreneurship also has to be measured over a time frame in order to identify correlation. Hence, to find the best results on the influence of entrepreneurship on economic development, a longitudinal study will be used. This is the first longitudinal study performed in these countries regarding this topic. For the data analysis of this study, the low availability of entrepreneurial data on developing countries has to be considered. Since there are not many years in which this data is collected, a linear regression analysis will not be suitable (Yao, Müller & Wang, 2005). Also, a statistical analysis alone will not be sufficient because of possible validity errors. Hence, to support the results of the small sample of this research, a trend-line analysis will be used. This way the correlations will be statistically and visually observed, aiming to find the most representative results as possible (Hojem & Ottenbacher, 1988).

3.2. Sample and data collection

To resemble the context of East Africa, this study will focus on the countries Botswana, Uganda, and Zambia. The reason for this choice is primarily driven by the availability of the GEM’s data. Although Botswana and Zambia are not officially part of the East African Community (EAC), they do share the same kind of economic characteristics that can influence the variables tested in this model (Herrington & Kelley, 2012). The selected countries also experience high TEA rates of both necessity and opportunity driven entrepreneurship. Advantageous for this research in analyzing these countries, is that they have participated in the GEM more often. For this reason, a time frame of 12 years can be researched, which will make the validity of this longitudinal study most efficient (Yao,

(25)

To measure entrepreneurship this study will mainly use secondary entrepreneurial data yearly collected by the GEM. As mentioned in the literature review, the GEM separates its data in opportunity driven and necessity driven entrepreneurship which is very useful for this research. The primary measure of entrepreneurship used by GEM is the TEA rate. TEA indicates the prevalence of individuals engaged in nascent entrepreneurship and new firm ownership in the adult population. The key indicators of the TEA are measured through an Adult Population Survey (APS). The sample to represent a countries’ population consists of at least 2000 randomly picked representative adults between the ages of 18 and 64 years old. An advantage of using GEM’s data on entrepreneurship is that it documents how entrepreneurship is affected by national conditions and that they take a more socioeconomic approach in collecting data. Where other data sets simply count new business registrations, the GEM monitors the different attitudes, activities, and aspirations of the entrepreneur. Hence, an identification of different types and phases of entrepreneurship within a country is observed (Bosma, 2013). This is explained and visualized by Figure 1 in the literature review of this study. The TEA rate, TEA Necessity rate, and TEA Opportunity rate data used in this research, are all obtained from the GEM 2001-2015 APS Global Key Indicators Dataset.

Economic development will be measured as the percentage growth in GDP of the concerning country. Data of these growth rates will be collected from the open database of the World Bank. They are a reliable source and are amongst the world’s greatest collectors of economic data. This data is widely used in many types of research, so for this study, it will be a reliable measure for the variable. The level of income is measured through the country’s Gross National Income per capita, also collected by the World Bank. As well as the data on the level of technology. This is measured as the share of high technology export from a

(26)

Data on patents will be provided by the World Intellectual Property Organization (WIPO). The WIPO is an organization responsible for promoting the protection of intellectual property throughout the world and for the administration of various multilateral treaties (Kwakwa, 2002). Patent applications are worldwide patent applications filed through the Patent Cooperation Treaty procedure or with a national patent office for exclusive rights for an invention, a product or process that provides a new way of doing something or offers a new technical solution to a problem. A patent provides protection for the invention to the owner of the patent for a limited period, generally 20 years. Patent-based statistics reflect the inventive performance of countries, regions, and firms, as well as other aspects of the dynamics of the innovation process such as co-operation in innovation or technology paths (Herce, 2001). For this study, the total amount of patent filings of domestic residents within a country are taken.

Data on innovation is supplied by the Global Innovation Index (GII), who provides detailed metrics about the innovation performance of 127 countries and economies around the world. Its 81 indicators explore a broad vision of innovation, including political environment, education, infrastructure, and business sophistication (Dutta, 2012).

Globalisation data is provided by the KOF which is a Swiss organization who collects data on globalisation for 122 countries. Their data is measured on economic, political, and social criteria and is used for globalisation measurements in many types of research.

Unemployment rates are observed from TheGlobalEconomy.com, which is a reliable collector of external data. They offer interactive data from many different databases and bundle it into their own database. They do not empirically collect data themselves, though

(27)

they have been acknowledged and recommended by the American Economic Association (AEA).

3.2. Variables and Measurements

This section will elaborate on the variables used in this model and how they should measure and relate to the economic theories. First, an explanation of the dependent and independent variables is given, then the moderating and control variables will be analyzed on their expected outcomes.

3.2.1. Dependent Variable

Growth in GDP is the dependent variable representing economic development in this model. It is measured as the annual percentage growth rate of GDP at market prices, based on constant local currency. Aggregates are based on constant 2010 U.S. dollars. GDP is the sum of gross value added by all resident producers in the economy plus any product taxes and minus any subsidies not included in the value of the products. It is calculated without making deductions for depreciation of fabricated assets or for depletion and degradation of natural resources. This makes it a proper measurement of the economic state of a country (World Bank, 2013).

3.2.2. Independent Variables

TEA rate is the measurement used in this study to indicate the level of entrepreneurship. This number can be interpreted as the amount of individuals in the working

(28)

age population who are actively involved in business startups, either in the phase of starting a new firm (nascent entrepreneurs), or in the phase spanning 42 months after the birth of the firm (owner-manager of new firms) (van Stel, Carree & Thurik, 2004). For example, when 32 people of 100 in a country are active in entrepreneurial activity, the country’s TEA rate will be 32.

This rate is divided into necessity driven and opportunity driven entrepreneurship including the corresponding TEA rates, which will be the other independent variables in this study. So TEA Necessity is the percentage of those involved in TEA who are involved in entrepreneurship because they had no other option for work. And TEA Opportunity is the percentage of those involved in TEA who (i) claim to be driven by opportunity as opposed to finding no other option for work; and (ii) who indicate the main driver for being involved in this opportunity is being independent or increasing their income, rather than just maintaining their income (GEM 2001-2015 APS Global Key Indicators Dataset).

3.2.3. Moderating Variables

To have a better understanding of the direct relationship between the dependent and independent variables, this study will make use of some moderating variables. A moderator has been defined as one which systematically modifies either the form and/or strength of the relationship between a dependent and independent variable. As such, the moderator variable concept holds important implications for understanding and predicting this relationship (Sharma et al., 1981).

(29)

Purchasing Power Parity rates. An international dollar has the same purchasing power over GNI as the U.S. dollar has in the United States. Data are in current international dollars (Worldbank, 2017). Income is expected to have a positive correlation with TEA Opportunity and a negative correlation with TEA Necessity since a rise in a necessity driven entrepreneur’s income is assumed to reduce the pressure of forced employment.

Another moderating variable in this model is a country’s level of Technology. This level is measured as the percentage of high technology exports from the total exports of a country. This indicates the degree of technological development of a country and is expected to stimulate the TEA rate. Also, because high levels of technology are assumed to enhance the productive infrastructure, opportunities are easier to discover and more realistically pursuable. Hence, a positive correlation with TEA Opportunity is also expected.

The last moderating variable in this model is the level of Unemployment. This is seen as the share of the labor force without work but available for and seeking employment. The unemployment rate is argued to be the main driver for necessity driven entrepreneurship and is also an important factor in deferring economic growth (Deli, 2011). For this reason, unemployment is expected to be positively correlated with TEA Necessity.

3.2.4. Control Variables

Another element in this model to create a better understanding of the relationship between the dependent and independent variables are the control variables. In this model, the control variables are specifically used to control the dependent variable GDP Growth for possible external influences.

(30)

The first control variable is the yearly number of Patents in a country. This is measured as the total number of patents filed only by residents of the country. Because many countries, in particular least developed countries, have only just begun to address the challenges of setting up an appropriate patent system in place to reap economic and social benefits, limited data is expected to collect. However, the development of these countries' resources and infrastructure and their capacity to benefit from the rapid growth of intellectual property as a valuable economic asset in the world economy remain an urgent concern (Herce, 2001).

The level of Innovation is the next control variable in this model. Innovation describes the development and application of ideas and technologies that improve goods and services or make their production more efficient. Similar to entrepreneurship, innovation is also a multidimensional concept which is not easily measured. However, the Global Innovation Index gives a well-analyzed indication of a country’s innovative characteristics. Because they are closely related, innovation will influence the entrepreneurial activity within a country and is also argued to be a positive driver for economic development (Verspagen, 2005). For this reason, innovation is expected to show interesting results as a control variable, which fits well within this model.

The last control variable in this model is Globalisation, measured as an index by the KOF Index of Globalisation. As mentioned in the literature review, globalisation has been an important driver for economic development worldwide. The globalisation index is expected to positively correlate with economic growth.

(31)

4. Results

This section will present the main results of the analyzed data. The results are separated into the three examined countries. For every country, a correlation matrix and charts of the model with their corresponding linear trend-line equation are given. The correlations shown in the correlation matrix will function as an indicator of the strength and direction of the relationship between the variables in the model. Given the low availability of data, statistical significance is unlikely to achieve. Additionally, the charts will support the analysis of the hypotheses. All the graphs are constructed with the timeline on the x-axis and the units of the variables on the y-axes. Because the charts will analyze several variables with different units of measurement on one figure, they are provided with a primary and secondary y-axis. Note that the dependent variable GDP Growth is always on the primary y-axis. All hypotheses are being tested, and conclusions will be made through the elements described above. Since this is the first longitudinal study performed in these countries regarding this topic, interesting results are expected to find.

4.1. Botswana

As stated in the hypothesis, entrepreneurship is expected to have a positive relationship with economic development. However, the TEA rate of Botswana is negatively correlated with GDP Growth in this model (Table 1). This correlation has a coefficient of -0.900 and is significant at the 0,05 level (two-tailed). Although this significant relationship is unexpected, an explanation can be given with the support of the graph in Figure 4. The analyzed time frame is showing an increase in the TEA rate of Botswana. A beta of 2.865

(32)

indicates a consistent growth of entrepreneurship and the corresponding ! of 0.41 represents a good statistical fit. Simultaneously, TEA necessity is showing a strong growth in this time frame, while TEA Opportunity seems to drop. This could imply that the observed entrepreneurial growth in Botswana is driven by necessity driven entrepreneurs. The drop in GDP Growth corresponds with the assumed negative relationship between necessity driven

(33)

entrepreneurship and economic development. Despite its insignificance, the correlation coefficient of -0.418 can verify the above statement. Hence, these findings support hypotheses H2 and H3.

Furthermore, the relationship between the dependent and independent variables is moderated by an analysis between the TEA Rates and the moderating variables, unemployment, technology, and income. The graphical analysis is presented in Figure 5. The chart shows a positive relationship between necessity driven entrepreneurship and income. Since a low income is argued to be a main driver for necessity driven entrepreneurship, this result is unexpected. Nevertheless, the GNI per Capita is a multidimensional variable which can be influenced by many different aspects of an economy. The insignificant correlation coefficient of 0.527 also indicates the absence of a strong relationship between the two

(34)

variables in this model. In addition, unemployment and income both move in a constant trend. For this reason, it is hard to make assumptions regarding their influence on the independent variables.

Finally, the dependent variable GDP growth is controlled on external influences by the control variables Patents, Globalisation and Innovation in this model. For this analysis, a greater time-frame is monitored to ensure the best statistical fit. Although Table 1 doesn’t show any significant correlation between any of these variables, the trend-lines in Figure 6 are all moving as expected. Low availability of data on Patents and Innovation could influence the credibility of these results, thus, globalisation is the most reliable variable to control in this model. With trend-line coefficients of -0.2568 and -0.3719, they indicate to behave very similarly. Hence, the control variables in this model have shown that the

(35)

dependent variable is not irregularly influenced by other factors. This consolidates the results of the independent variables in this study.

From the analysis of the results, we can conclude that Botswana has had a growth of their overall entrepreneurial activity. However, this was particularly stimulated by necessity driven entrepreneurship. This led to the findings of some unexpected results regarding the independent, dependent, and moderating variables. The low availability of data resulted in a lot of insignificance within this model. Given this fact, it is hard to draw highly decisive conclusions about these results. Hence, they will function as a reinforcement and aiming point for the total analysis of entrepreneurship in East Africa.

4.2. Zambia

Equivalent to Botswana, Zambia has a negative correlation between TEA and GDP growth (Table 2). Despite its insignificance, this indicates that entrepreneurship is stimulated

(36)

entrepreneurial activity and necessity driven entrepreneurship. On the contrary, the TEA rates of opportunity driven entrepreneurship has dropped. As expected, this results in a drop of the GDP growth rates. Ignoring the insignificance of the correlations, a trend-line analysis confirms this study’s assumptions about the effect of necessity and opportunity driven entrepreneurship on economic development. Thus, hypotheses H2 and H3 are being supported in this example.

Considering the moderating variables, the TEA rate seems to have a strong correlation with all three moderators. This validates the statistical fit of the moderating variables in this model. Like in Botswana, Figure 8 shows an unexpected positive relationship between income and necessity driven entrepreneurship. This could be explained by an overall drop in the unemployment rate. Furthermore, the trend-line of is technology downward sloping., Figure 7: Line chart with trend-lines of GDP growth and TEA rates of Zambia

(37)

equivalent growth in opportunity driven entrepreneurship during this time-frame. The correlation coefficient of 0.695 suggests a strong relationship between the two variables, despite its insignificance. Although these results elaborate on the effect of the independent variables in this model, the lack of data withholds their credibility. Hence, they are only accounted for as valuable insights for this research.

Figure 9 shows the trend of the control variables and the dependent variable. Equivalent to Botswana, Globalisation and GDP growth experience a similar movement. Also, innovation moves in an expected trend with economic development. In this case, however, Patents are highly fluctuating over the years. This results in a trend-line with a low statistical fit of ! =0.09. A corresponding correlation of -0.03 with GDP growth indicates, that for Zambia the number of patents does not have a significant influence on their economic development. For innovation, it is also hard to assume its influence due to its lack of data.

R2

(38)

Nevertheless, the results don’t show any odd irregularities regarding the dependent variable, which validates the relative relationship with the independent variables in this model.

The results of Zambia have given a good overview of the influence of entrepreneurship on economic development. The overall analysis shows there are many similarities with the analysis of Botswana. Also in this case, the lack of data affects the credibility to make decisive statements. Hence, the findings will also function as a reinforcement and aiming point for the total discussion of entrepreneurship in East Africa.

4.3. Uganda

First of all, the results of Uganda are expected to be more sufficient and reliable due Figure 9: Line chart with trend-lines of control variables on dependent variable of Zambia

(39)

because the GEM started collecting entrepreneurial data of Uganda since 2003. On other variables, there was no significant difference in data availability. Observing the TEA rates trend-lines in Figure 10, it is clear that necessity driven entrepreneurship has dropped and opportunity driven entrepreneurship has been growing over the past years in Uganda. The overall entrepreneurial activity fluctuates but moves in a nearly constant trend. Considering the growth of opportunity driven and the drop in necessity driven entrepreneurship, a rise in economic development is being expected. However, the overall trend-line of GDP growth is downward sloping. It is important to note this does not resemble an economic recession since the measurement unit is a percentage growth rate. This means there will be economic development as long as the GDP growth rate is above zero. The trend-line has a beta of -0.2751 and an ! of 0.2062, which indicates a weak statistical fit with high fluctuations and a slow decline of the growth. Nevertheless, according to the assumptions and hypotheses of this research, an upward sloping GDP growth trend-line was being expected.

The relationship of the moderators with the independent variables in Figure 11 is showing a plausible indication of the movements in the TEA rates. Unemployment is moving

(40)

in a decreasing but nearly constant pace, resulting in a trend-line beta of -0.2751. No concrete assumptions about the influence of this variable can be made because of its steadiness. Technology has a decreasing trend-line beta, though it is experiencing some great outliers resulting in a very low ! of 0.0338. Clearly, the peaks of technology are accompanied by movements of growth in opportunity driven entrepreneurship. This could indicate that high levels of technology have enhanced the productive infrastructure and increased the ease of opportunity discoveries. Though despite its insignificance, this is not supported by the correlation coefficient of -0.122. The most vital indicator for the movement of the TEA rates, however, is the growth in income. Since the first measurement in 2003, income per capita has almost doubled in Uganda. Table 3 also shows the highest correlation with TEA necessity and TEA opportunity of -0.488 and 0.513 respectively. Despite the insignificance, this does confirm the explanatory strength of this variable. Hence, it is fair to say that the growth in

R2

(41)

income is related to the growth in opportunity driven and the decrease in necessity driven entrepreneurship.

Furthermore, the control variables displayed in Figure 12 should give better insights into the decreasing trend of the dependent variable. Clearly, the trend-lines of Patents and Globalisation are upward sloping and Innovation is downward sloping with betas of 0.3103, 0.4934 and -1.24 respectively. Because Patents and Globalisation have more observation points, they are more accurate in controlling on the dependent variable. In contrast to the expectation, their upward slope should not result in a downward sloping GDP growth. Nevertheless, betas of 0.31013, 0.4934 and a GDP growth beta of -0.2744 do not indicate a very disturbing deviation from the assumption. Also, a positive GDP growth rate still implies the existence of economic development. Hence, the slow decline in economic development in Uganda could be explained by the declining innovation rate but is most likely influenced by Figure 11: Line chart with trend-lines of independent and moderating variables of Uganda

(42)

many other elements. The increasing rate of Globalisation and Patents refer to the fact that GDP Growth is above zero, which implies economic development.

As a result from the analyses above, it is clear that Uganda experienced a rapid growth of opportunity driven entrepreneurship in the past years. At the same time, necessity driven entrepreneurship has been decreasing, whereas the overall entrepreneurial activity roughly stagnated. This indicates a process where necessity driven entrepreneurship changes into opportunity driven entrepreneurship. Another remarkable occurrence is the high growth of the income per capita. As assumed, these events should have resulted in an upward trend of the GDP growth rate. However, the observed trend-line is downward sloping. This indicates there are other external elements influencing the economic development in Uganda, especially since two of the three control variables also showed an upward trend. Nevertheless, is the positiveness of the GDP growth rate implicating the existence of Figure 12: Line chart with trend-lines of control variables on dependent variable of Uganda.

(43)

economic development and, therefore, the assumptions and hypotheses of this research will not be rejected.

5. Discussion

This chapter will integrate the academic literature review with the analyzed results of this research into a relevant discussion about the influence of entrepreneurship on economic development in East Africa.

5.1. Main findings into the context of East Africa

The focus of this study is the influence of entrepreneurship on economic development in East Africa. In order to generate conclusions, this study has analyzed opportunity driven and necessity driven entrepreneurship and economic growth of the countries Botswana, Zambia, and Uganda. Herrington and Kelley (2012) stated that Botswana and Zambia are not officially part of the East African Community (EAC), but they do share the same kind of economic characteristics that can influence the variables tested in this model. In this part, the main findings of this research will be generalized into the context of East Africa.

Observing the three countries, it is clear they all have experienced high rates of entrepreneurial activity and high growth rates of GDP over the last years. Only Botswana has had two years of economic recession in the analyzed time-frame (see Appendix). These findings correspond with the assumptions made by Reynolds et al. (2001) and Wennekers et

(44)

countries in resembling East Africa. The results of Uganda, which are assumed to be the most reliable, show a rising opportunity driven line, and a declining necessity driven trend-line. At the same time, the overall TEA rate slightly rises but moves fairly constant. This indicates a process in East Africa where necessity driven entrepreneurship is changing into opportunity driven entrepreneurship. By acknowledging the literature and hypotheses, this movement is has stimulated economic development in East Africa. Yet, the GDP growth rate of East Africa is slightly downward sloping. According to Wong et al. (2005), this refers to the observed high rates of necessity driven entrepreneurship. Hence, this leads to an acceptance of hypothesis H3 of this study. Despite its overall decline, the GDP growth rate is exclusively positive, which indicates a long trend of economic development because a positive GDP growth rate indicates economic development. Besides the high rates of necessity driven entrepreneurship, high rates of opportunity driven entrepreneurship were also observed. These rates have shown to partly explain the overall economic development in East Africa. This corresponds with the assumptions of Acs et al. (2008) and Langevan et al. (2012) and supports hypothesis H2 of this study. Total entrepreneurial activity is slightly growing over the years in East Africa, which refers to its high growth in GDP to a certain extent (Acs & Szerb, 2009). Although this study has shown a vital relationship between entrepreneurship and economic development in East Africa, it does not provide sufficient evidence to claim a significant importance. Hence, entrepreneurship is not assumed to be a main driver for economic development in East Africa, however, a positive relationship is observed, so hypothesis H1 of this study is supported.

Regarding the moderating and control variables, income and unemployment have proven to be highly related with the independent variables. They both show the highest

(45)

insignificance, they indicate a moderating effect in this model. As Wennekers et al. (2010) stated though, entrepreneurship is a very broad and multidimensional concept. This indicates there will be many variables with a moderating effect in such a model. As for the control variables, mainly expected trend-line movements were found which indicates an expected relation between the control variables and the dependent variable. However, Table 1, 2 and 3 show low and unexpected correlations of the variables. This is a result of the high fluctuations in the control variables and refers to the unstable economic elements in East Africa, but also to the low availability of data. Moreover, the low correlations between the control variables and GDP growth could also be explained by the possibility that patents, globalisation, and innovation are not prominent drivers for economic development in factor driven economies. Additionally, these variables are all interlinked with the concept of entrepreneurship, which refers to the assumptions that entrepreneurship is not a main driver for economic development in such economies. Kelley and Herrington (2012) describe how factor driven economies should focus on stimulating the fundamentals of a well-functioning business environment, in order to gain entrepreneurial development and economic growth. Basic fundamentals, in this case, are for example a country’s macroeconomic stability, primary education, health, and infrastructure. Once these fundamentals are sufficiently established, the country can move from a factor driven to an efficiency driven economy, where development should be more focused on efficiency enhancers like financial market sophistication, technology, and labour market efficiency. Hence, the selected variables of this model are not the most important drivers of economic development in the observed countries, though the results show a vital relationship between entrepreneurship and growth in GDP.

(46)

5.2. Limitations

As mentioned before, this study had to deal with a low availability of data which has led to some limitations in the results. Besides these limitations, this section will also illustrate other factors that could have weakened the statistical accuracy of this research.

First, this study draws conclusions about East Africa by generalizing results from the countries Botswana, Zambia, and Uganda. These countries were chosen because they share a fairly similar economic state and because of their availability of data. For instance, the GEM has collected data of Uganda from 2003-2015 and of Botswana and Zambia from 2012-2015 and 2010-2013 respectively. This is still a limited dataset for a proper longitudinal study to generalize conclusions. However, the data used in this study appeared to be the best fit for resembling East Africa. Nevertheless, for a truly accurate analysis of the East African economy, all the countries within East Africa need to be observed. Unfortunately, the GEM and other databases, did not provide data for all those countries during this research.

Secondly, considering the short time-frames of data, a trend-line analysis was used to observe the variables in the model. The high fluctuations in necessity driven and opportunity driven entrepreneurship resulted in some low statistical fits of the trend-lines and low ! . This weakens the validity of the trend-line analysis and resulted in less accurate results. For this reason, in some cases, the trend-lines were neglected and the movement of observation points were analyzed. The high fluctuations of the observed TEA rates question the accuracy of the GEM data, since it is very unlikely that entrepreneurship changes that dramatically every year.

A final aspect which may have limited the accuracy of the results, is the choice of R2