The Effects of Government Quality and Economic Indicators on Self-employment in East Africa: Panel Data Analysis

Self-employment plays a major role in the economic growth of Africa in general and East African countries in particular. It is a major source of new jobs and a way of employing the entrepreneurial abilities of the population. Moreover, self-employment allows people to do what they want and to follow their passion. However, populations in East Africa face various barriers that hinder their ability to start their own business and/or to become self-employed. Thus, the aim of this study is to examine the effects of government quality and economic indicators on self-employment. Data was taken for this study from World Bank’s World Development Indicators, Worldwide Governance Indicators, African development Bank and United Nations Development Program for eight East African countries for 2010-2018. The countries were selected based on the availability of panel data. The fixed effects result indicated that political Stability, control of corruption indices and Voice and accountability, natural logarithm of electricity access, mobile phone subscriptions, people living below income poverty line and primary education enrolment are favorably affecting self-employment in East Africa. The result also shows that the natural logarithm of employment in industry and real GDP growth have negative coefficients. . These findings suggest that in addition to economic and social indicators government quality have greater effect on self-employment in East Africa.

social protection and rights at work (ILO, 2019).
The studies on the political orientations of small-business owners by scholars often hold that this is a group with rather right-wing political preferences. To mention some, self-employment is associated with free market beliefs, and values such as individualism, autonomy, and self-reliance (Evans and De Graaf, 2013;Gayle et al. , 2012;Goss,1991). Youth in East African countries face similar social and economic challenges: high birth rates, poverty, unemployment and underemployment and health problems (ILO, 2019). However, the patterns and severity of these challenges vary by country and sub-country. This study offers a new empirical analysis towards understanding the effect of political stability and corruption on self-employment. A research hypothesis is formulated, by stating a strong positive effect of political stability and corruption on self-employment. In order to prove the claim, empirical data consisting of the control of corruption estimate, political stability and absence of violence estimates were utilized and panel data model was built to assess the impact.
This study is motivated by the following main shortcomings in the literature: the effect of government quality indicator (political stability and control of corruption) and population living below poverty line on selfemployment. Studies on the effects of these covariates on self-employment are very limited and/or scanty in East Africa. Specifically, there is limited understanding of how political stability and absence of violence as well as control of corruption initiates self-employment. Although not well studied, political instability and violence and corruption are rampant and treat to almost all of East African's youth Self-employment. Poor governance, lack of finance, power outages/ limited electricity access, lack of accountability and transparency, low level of institutional control, and extreme poverty and inequality are major obstacles to self-employment in African continent in general and in East Africa in particular.
Generally speaking, self-employment has become a key possible source of new jobs and a way of employing the entrepreneurial abilities of the population in Africa, and it has stimulated many studies and empirical applications. In very broad terms, self-employment can be considered the source of employment in East Africa where poverty and large youth unemployment is prevailing. Despite a flowering literature on the factors underlying the growth of self-employment in Africa, only few studies try to compare and explain self-employment rates across countries.

Literature Review
According to Investopedia, Self-employment is the state of working for oneself rather than an employer. Selfemployed people generally find their own work rather than being provided with work by an employer, earning income from a profession, a trade or a business that they operate. A self-employed individual does not work for a specific employer who pays them a consistent salary or wage. Self-employment may not be subject to tax withholding and, those self-employed are responsible for paying their taxes. Self-employment can provide a great deal of job flexibility and autonomy; however it also comes with a greater degree of employment risk and more volatile income.
Creation of new business is highly affected by a dishonest or illegal behavior of government officials or corruption and in Africa the wide spread political instability is connected with the high rate of corruption as its politicians, leaders, and public servants illegally collect wealth using public office for private gains (Hammed,2018). Almost in all African countries except for few countries, obstacles to self-employment are various and similar. Previous studies attest to this fact (AFD 2017;Baah-Boateng 2016;Townsend et al., 2017). University graduates are increasing from year to year but, The quality of education in East Africa is Low (Kluve et al. 2016;IMF 2017). African political instability and other related problems is basically challenging the growth of self-employment: civil conflicts, bad governance and poor economic conditions (Antony O., 2008). Moreover, a lack of political will to make youth employment a top priority has created coordination and implementation gaps (Osei Tutu, 2004;Annan, 2008).
The study by Justice T. (2018) points out that unreliable electricity supply reduces the probability of being self-employed by 35 percent and indicates that lack of electricity access boosts unemployment and has negative impact on employment rates in Africa. Previous studies found inverse relationship between real GDP per capita and self-employment. Countries' development and mature in free enterprise lead to loss of the importance of selfemployment and there is negative relationship between real GDP per capita growth and self-employment rate (Gindling andNewhouse, 2014, Lofstrom M., 2009;liabadi V., 2017;Hipple, 2010;Bhorat H., 2018).
Obviously, many of the most vital difficulties facing working age people in East Africa are not youth-specific and hinder job creation more generally. The research findings by Gindling and Newhouse (2014) showed that only 7 percent of the self-employed in developing countries are successful. Previous research has predominantly measured self employment using infrastructure, governance and corruption, and access to finance (McKinsey Global Institute 2012;African Development Bank et al., 2012) and only 7 percent of the self-employed in developing countries are successful (Gindling and Newhouse, 2014). Older individuals have more tendencies to enter the self-employment than younger once (Cho Y., Robalino D. and Manuel J., 2015). In the present study, I particularly focus on effects of real GDP per capita growth, population growth, voice and accountability, political stability and absence of violence, control of corruption, people living below income poverty line, education index, mobile cellular subscriptions on self-employment. The connections between self-employment, corruption, political instability and violence are not straightforward. Only limited and contradictory evidence on those connections is available from developing countries (World Bank, 2019).
Furthermore, studies by scholars found that there is strong correlation between mobile cellular subscription and the self-employment rate. For instance, individuals who use mobile phones have 11% points higher probability of being self-employed relative to those who do not use (Moyi E., 2019). The use of mobile money and mobile commerce enhances the growth prospects of self-employment (Aker, 2010;Amegbe et al., Hanu, C., and Nuwasiima, A. 2017). A study conducted by Gary S (2019) concluded that Self-employment and poverty are closely linked, albeit not perfectly. Employment in industrial sector increases the rate of self-employment by 2% (Heintz and Pickbourn, 2012) and study, the predicted probability of self-employment is, on average, around 0.03 higher for educated individuals than for the non-educated. Komunte, M. (2015) found that the incidence of selfemployment was higher among educated individuals.

Methodology 3.1. Data Source and Description
The major source of data for this study is the 2010-2018 World Bank's World Development Indicators in East Africa. In addition, data from African Development Bank (AfDB) and UNDP was used. The sample includes eight East African countries (Djibouti, Eritrea, Ethiopia, Kenya, Madagascar, Rwanda, Tanzania and Uganda). The countries were selected based on the availability of panel data. The dependent variable is self-employment rate which is defined as the percentage of total employment estimated by ILO estimate. Based on previous studies and data availability, the researcher used political stability and absence of violence estimate, control of corruption estimate, population growth rate, real GDP growth rate, unemployment rate, Population living below income poverty line, PPP USD 1.90 a day (%) , mobile phone subscriptions (per 100 people) as independent variables.
Data for the independent variables are collected from various sources. Variables such as Real GDP per capita, population growth, employment in industry rate, primary school enrollment and unemployment rate are from the World Development Indicator (WDI database) and Control of corruption index, political stability and absence of violence/terrorism estimate and voice and accountability estimate from Worldwide Governance Indicators (World Bank). Finally, data on mobile phone subscriptions (per 100 people) was taken from African Development Bank database.

Empirical Model of self-employment
To measure the selected determinants' effects on self-employment rate, the researcher used panel data for 8 East African countries for 2010-2018. An important advantage of panel data compared with time series or crosssectional data sets is that they allow identification of certain parameters or questions, without the need to make restrictive assumptions. For example, panel data make it possible to analyze changes on an individual level. As the error term is likely to include country and time-varying effects, a fixed-effects model, is employed. Fixed effects explore the relationship between predictor and outcome variables within an entity (country, person, company, etc.). Each entity has its own individual characteristics that may or may not influence the predictor variables. When using fixed effect the assumption is that something within the individual may impact or bias the predictor or outcome variables and we need to control for this. This is the rationale behind the assumption of the correlation between entity's error term and predictor variables. Fixed effect removes the effect of those timeinvariant characteristics so we can assess the net effect of the predictors on the outcome variable.
The fixed effects model is simply a linear regression model in which the intercept terms vary over the individual units i, that is, following Verbeek M. (2017) a fixed effects model is: yit = αi + x ′ itβ + uit , uit ~IID(0,δ 2 u) where it is usually assumed that all xit are independent of all uit. We can write this in the usual regression framework by including a dummy variable for each unit i in the model. That is, = + + where dij = 1 if i = j and 0 elsewhere. We thus have a set of N dummy variables in the model. The parameters 1, . . . , N and can be estimated by ordinary least squares in. The implied estimator for is referred to as the Journal of Economics and Sustainable Development www.iiste.org ISSN 2222-1700 (Paper) ISSN 2222-2855 (Online) Vol.11, No.9, 2020 least squares dummy variable (LSDV) estimator. The OLS estimator for , which is called fixed effects estimator, is given by The rationale behind random effects model is that, unlike the fixed effects model, the variation across entities is assumed to be random and uncorrelated with the predictor or independent variables included in the model: "…the crucial distinction between fixed and random effects is whether the unobserved individual effect embodies elements that are correlated with the regressors in the model, not whether these effects are stochastic or not" [Green, 2008, p.183].
If we have reason to believe that differences across entities have some influence on our dependent variable then we should use random effects. An advantage of random effects is that we can include time invariant variables. In the fixed effects model these variables are absorbed by the intercept. The random effects model is: Yit = α + βXit + μi+ uit , uit ~IID(0,δ 2 u) , μi ~IID(0,δ 2 μ) where μi + uit is treated as an error term consisting of two components: an individual specific component, which does not vary over time, and a remainder component, which is assumed to be uncorrelated over time. That is, all correlation of the error terms over time is attributed to the individual effects μi. It is assumed that μi and uit are mutually independent and independent of xjs (for all j and s). This implies that the OLS estimator for α and is unbiased and consistent. In general, random effects assume that the entity's error term is not correlated with the predictors which allows for time-invariant variables to play a role as explanatory variables. In random-effects we need to specify those individual characteristics that may or may not influence the predictor variables. The problem with this is that some variables may not be available therefore leading to omitted variable bias in the model. Random effect allows to generalize the inferences beyond the sample used in the model. Conceptual model of self-employment is given below:

Self-employment and Unemployment in East Africa
In East Africa the growing working-age population coupled with inflation and political instability is posing strong labour market challenges in the coming years. The majority of population in the sub-region is living in extreme poverty and some in moderate poverty and the percentage of population living in extreme poverty is projected to rise, since poverty reduction in the sub-region is going on at a slower pace than elsewhere. In East Africa selfemployment is not evenly distributed. Djibouti has the lowest self-employment rate from the eight East African countries. As Figure 2 below indicates, mean self-employment rate of Djibouti was 48.5557 for 2010-2018, which is the lowest compared to the mean self-employment rates of other East African countries under study. On contrary the largest rate of self-employment was in Madagascar with 89.4292 followed by Ethiopia, Eritrea and Tanzania in that order.
Self-employment creation has kept pace with population growth but joblessness remains high. With too few Journal of Economics and Sustainable Development www.iiste.org ISSN 2222-1700 (Paper) ISSN 2222-2855 (Online) Vol.11, No.9, 2020 formal jobs to absorb East Africa's rapidly increasing labor force and fast urbanization, the share of the active labor force in self-employment is rising faster than in agricultural. Self-employment in the industrial sector has grown fastest, but agriculture is still the largest employer. In many of East African countries the number employed in industry more than doubled, while the number employed in agriculture less than doubled. However, the share of self-employment in agriculture is more than that in industry. Today almost in all East African countries unemployment rate is higher in cities than in rural. Generally, in East Africa the majority of individuals from lowincome households are confined to low quality jobs and rate of unemployment is awfully similar crosswise households. Similarly in the sub-region, individuals from low-income households are inconsistently engaged in self-employment and informal wage work, and in agriculture and low value-added sectors and have nearly no access to better paid wage employments.
This paper focuses on 8 East African countries, the region with the lowest per capita incomes, largest shares of the work force employed in agriculture and lowest agricultural labor productivity. Over this period, the average self-employment rate was declining from 78.10% in 2010 to 74.33% in 2018 across East African countries. According to World Bank data, employment as percentage of total employment is higher in agriculture in East African countries which is near 80% while percentage of total employment in industrial sector is less than 20%. Figure 3 below reports the percentage of self-employed in agriculture, in industry and the overall self-employment rate in East African countries between 2010 and 2018.
The East African countries are plagued by violence, corruption, unemployment and extreme poverty. The figure below shows how African countries compare with one another, taking into account unemployment rate as an indicator. Djibouti is with the highest unemployment rate (worst) while countries with the lowest unemployment rate include Rwanda, Uganda and Madagascar. In Djibouti, the unemployment rate measures the number of people actively looking for a job as a percentage of the labour force. Djibouti is mostly barren, with little development in the agricultural and industrial sectors. The country has a harsh climate, a largely unskilled labour force, and limited natural resources. As such, Djibouti's economy is commanded by the services sector, providing services as both a transit port for the region and as an international transshipment and refueling centre. Countries with almost stable rate of unemployment are Djibouti, Eritrea, Kenya and Ethiopia. Greater fluctuation in the rate of unemployment is found in Madagascar, Uganda and Tanzania.

Fixed Effects or Random Effects Model
To decide between fixed or random effects we can run a Hausman test where the null hypothesis is that the preferred model is random effects vs the alternative the fixed effects (Green, 2008). It basically tests whether the unique errors (ui) are correlated with the regressors; the null hypothesis is they are not.
S  (7) = 46.78 Prob>chi2 = 0.0000 Note: we therefore conclude for fixed effects at any level of significance.
From this result as Prob>chi 2 = 0.0000 is far less than 0.05. Therefore, fixed effect model is appropriate. Under the current specification, the initial hypothesis that the individual-level effects are adequately modeled by a random-effects model is deeply rejected.
The LM test helps decide between a random effects regression and a simple OLS regression. The null hypothesis in the LM test is that variances across entities are zero. This is, no significant difference across units (i.e. no panel effect). chi2(01) = 0.000 ; Prob > chi2 = 0.3400 Source: Computed based on WBES data (2010-2018) Note: We therefore conclude for fixed effects at any level of significance Here the null hypothesis is accepted and conclude that random effects is not appropriate, because Prob > chi2 = 0.3400 exceeds 0.05. This is, no evidence of significant differences across countries, therefore you can run a simple OLS regression.
The fixed effect results show how changes in self-employment status are related to changes in government quality indicators and other covariates by controlling for unobserved, time-invariant individual characteristics. Results of the fixed effects are reported in Table 3 below. The results of the fixed effects panel analysis show that government quality indicators (political Stability and absence violence, control of corruption indices and Voice and accountability) have a positive and significant effect on East African's self-employment rate with coefficients 0.0743, 0.1339 and 0.1811, having p-values 0.000, 0.002 and 0.016, respectively. Similarly, the result of the natural logarithm of electricity access has a positive relationship with natural logarithm of self-employment rate, having coefficient 0.0158 with p-value 0.099. This shows that with more electricity access rate of self-employment will obviously be high.
The estimation result also indicates that natural logarithm of mobile phone subscriptions, % of people living below income poverty line and natural log of population growth have a positive and significant effect on East African's self-employment rate with coefficients 0.0129, 0.1676 and 0.1280, having p-values 0.065, 0.000 and 0.085, respectively. Similarly, the result of the natural logarithm of employment in industry and real GDP growth rate have negative relationship with natural logarithm of self-employment rate, having coefficients -0.0293 and -0.0632 with p-values 0.000 and 0.006, respectively. As the percentage of people living below income poverty line increase, self-employment rate increases. This finding supports the research finding by Gary S., (2019). The positive correlation between mobile phone subscription and self-employment is similar with the findings of Moyi E., (2019), Aker, (2010), Amegbe et al., Hanu, C., and Nuwasiima A., (2017). Journal of Economics and Sustainable Development www.iiste.org ISSN 2222-1700(Paper) ISSN 2222-2855(Online) Vol.11, No.9, 2020 40 The empirical evidence on the association between self-employment and education is very mixed in the literature, and so are the results of my estimation. I test the relationship between primary education enrolment and the self-employment rate: there is positive and statistically significant relationship at 1 percent level with coefficient 0.0195. The positive relation between primary enrolment and self-employment rate supports the hypothesis that some basic skills are required in order to start a business and become self-employed. The result obtained shares similarities with the findings of Carlo P., Roberta R. and Matteo A., (2004) on effect of primary education enrolment on self-employment rate. Carlo P., Roberta R. and Matteo A., (2004) tested the relationship between primary education and the self-employment rate and concluded that there is positive and statistically significant relationship between primary education enrollment and the self-employment rate at 10 percent level. The positive relation between primary enrolment and self-employment rate supports the hypothesis that some basic skills are required in order to start a business.

CONCLUSIONS
The objective of this paper was to establish a number of key facts about the relationship between government quality and self-employment, and economic indicators and self-employment in east Africa. In the 21 st century, self-employment has begun to be stared as a vital potential source of new jobs and a way of employing the entrepreneurial abilities of the population, and it has urged various studies and empirical uses. In awfully extensive terms, self-employment can be regarded as the residual category of useful employment not remunerated by a wage or salary (ILO, various issues). However, few studies are available on determinants of self-employment in East African countries, where the phenomenon is incredibly important, and it is fundamentally related to the informal economy.
The present study analyses the determinants of self-employment in East African countries, in order to identify the existence of cross-national differences and to find effect of government quality indicators, mobile phone subscriptions, population growth, real GDP growth and electricity access on self-employment. This may help to provide some indications on the effect of good governance on self-employment in the East Africa, which has not been attempted, as far as I am aware, in the existing literature.
The main results of the present analysis may be summarized as follows. The study indicated that the government quality indicators (political Stability and absence violence, control of corruption indices and Voice and accountability) are favorably affecting self-employment in East Africa. The governments and political parties should look into this matter and tirelessly work on stabilizing the region that self-employment should not be compromised specially for youth. The study also indicated that natural logarithm of electricity access has a positive relationship with natural logarithm of self-employment rate. Thus, governments and other concerned bodies should support the expansion of grid-connections and electricity access for the population in general and for those in need of job creation. Furthermore, the governments should encourage mobile phone subscriptions and find the way out for those living below income poverty line as the coefficients of these variables are positive and statistically significant. Similarly, the fixed effects result also shows that the natural logarithm of employment in industry and real GDP growth both have negative coefficients. With the expansion of industries and industrial parks the percentage of wage employment increases but self-employment decreases.
Finally, the coefficient on primary education enrolment is positive and statistically significant at 1 percent level because, basic skills are required in order to start a business and become self-employed. The research findings of Carlo P., Roberta R. and Matteo A., (2004) attest to this fact. Therefore, the importance of governments in enhancing basic education is highly recommended and thereby play a more active and effective role in supporting development, self-employment may spark a positive upright circle with higher tax collection, higher government support, and in turn higher self-employment. In Africa in general and East Africa in particular, self-employment remains a major concern and concerted efforts must be made to keep up with growth rates. The flagship African Economic Outlook Report pinpoints industrialization as a key to the region's employment mystery. According to the African Development Bank (2018), East Africa's economy races ahead of its African peers and is leading the continent with GDP growth estimated at 5.7 percent in 2018, followed by North Africa at 4.9 percent, West Africa at 3.3 percent, Central Africa at 2.2 percent, and Southern Africa at 1.2 percent. Economic growth across Eastern Africa will remain at a robust 5.9 percent in 2019, making it a promising investment and manufacturing destination. Within the region, Ethiopia is in the lead as the fastest growing economy with a predicted 8.2% growth in 2019, while Rwanda (7.8%); Tanzania (6.6%); Kenya ( 6%), Djibouti(5.9%), Uganda(5.3%), Madagascar (5.1%) and Eritrea (3.7%) follow behind.