Political Instability and Economic Growth in Ethiopia: An Empirical Analysis

This paper analyzes the effect of political instability on economic growth in Ethiopia since 2004 using ARDL regression approach to identify the existence of this correlation and the channels through which this impact happened. The paper found that political instability has a significant effect on economic growth and this impact is more significant through the direct channels more than the indirect effects through its effect on economic environment or investment and financial flows.


Introduction
Historically, Political instability was seen as one of the main obstacles facing African pathway to achieve Economic growth and Development. Many studies try to investigate this relationship and identify the channels through which political instability could affect economic growth.
In Ethiopia, political stability has been considered as one of the main obstacles that hampered economic growth in this country. Although the country manages to achieve a high economic growth recently political instability situation continuously threat the sustainability of this growth through its direct and indirect economic effects. In this paper we try to investigate empirically the linkage between political instability and growth in Ethiopia since the fall of the authoritarian regime of Mengistu using the ARDL regression model.
A lot of studies tried to analyze the nature relationship between political instability and economic growth. These literatures could be divided into two groups: the first focuses on the impact of economic growth on political instability. These studies argue that poverty, lack of resources, bad economic performance, income inequality, and other economic problems usually cause instability and public and social distortions as found in study of (Arriola, 2009) that focused on the main economic and political sources of instability in Africa using the data of 40 African countries during . The study found that economic growth could be considered as one of the main economic causes of political instability.
On the other hand, some studies found that accelerated rates of development in some cases leave some groups behind in what is called in economic literature: growth with no trickle-down effect. This kind of growth is usually associated after a while with political instability and increasing rates of violence. This case usually happens in rentier states when high economic growth generated form natural resources don't distribute equally between citizens and rents concentrated with a certain group in what is called "Dutch Disease" as found in (Lujala, 2010) and (Dwumfour, 2018).
The second group of studies argues that political instability could be considered as one of the main constraints that hinder economic growth and development through its direct and indirect impacts on economic variables. In other words, political instability affects economic growth through its effects on its dynamics as affecting foreign direct investment, financial markets, economic policies efficiency..etc. it also affects development and inclusion of economic growth through its effects on human capital development and income inequality and distribution in the way that could cause mor social and political distortions and lead to a voracious cycle of political instability and bad economic performance as found in (Elhawary, 2010) (Samih, 2019) (Abdiweli, 2001) (Hira, 2017) (Kirmanoglu, 2003) (Gunhild, 2008).
To analyze the correlation between political instability and economic growth in Ethiopia this paper is divided into three parts; the first highlights the conceptual framework of political instability, indicators and sources. And then empirically analyzing the linkage between political instability and economic growth in Ethiopia and the last is for results and discussion.

Political Instability: Conceptual framework
One of the oldest definitions of political stability was offered by Lipstel (1960) who argued that a country is considered as stable if 'it has been a liberal and consistent democracy or dictatorship for twenty-five years. Ake (1975) focused on the type of political exchanges, thus he defined political stability as 'the regularity of the flow of political exchanges. According to his definition, Regular behaviors are those who do not 'violate' the system or the pattern of political exchanges. In other words, political stability is maintained if members of the society restrict themselves to behavioral patterns that fall within the limits imposed by political role expectations. Contrarily, any behavior that diverges from these limits is a manifestation of political instability (Bacha Kebede et al., 2020).
Political instability is also defined as the probability that the government will be destabilized or overthrown by unconstitutional or violent means, including domestic violence and terrorism (Alesina et al., 1996). Gyrimah-Brempony & Traynor (1999) defined political instability as 'the situations, activities or patterns that threaten to change or actually change the political system in a non-constitutional way (Aftab Hussain Tabassam et al.) The definition was taken one step further by Margolis (2010) as he expands the contours of instability. For him, political stability is far more than just state stability as all political objects other than the state struggle with the question of stability of power relationships. Margolis' understanding of political instability involves two tiers namely the formal and the informal. Where the formal and the informal coincides, political stability is achieved. In his words, 'when the formal roles and structures set by authority match those constructed by informal social interaction, an object is stable.' Alternatively, when either set of roles or structures change so they conflict, an object is unstable.
In the same vein, Morrison and Stevenson (1972) define political instability as 'a condition in national political systems in which the institutionalized patterns of authority break down, and the expected compliance to political authorities is replaced by violence intended to change the personnel, policies, or sovereignty of the political authorities. However, non-violent behaviors that violate institutional and legal means of political change are also instances of political instability.
Literature has also attempted to differentiate between types of 'instability' for example, Morgolis (2010) draws a distinction between 'political stability' and 'state stability' and deems the former as broader than the later. Furthermore, Morrison and Stevenson (1972) distinguish between three types of political instability namely elite instability, communal instability and mass instability. Regardless the theoretical definitions of the term, measuring political instability has proven to be a challenging undertake. Since political instability in a given country cannot be measured directly, empirical studies aiming to measure political stability rely on an array of indicators also known as 'proxies' such as coups d'état whether they are successful or not (Fosu,1992;2002), political assassinations, revolutions, mass protests, and many other indicators that aim to capture the phenomenon (Richard Jong A-Pin, 2006) Furthermore, indices have been developed to measure political (in)stability. One of the well-known measures is the Political Stability Index (PS), which measure the expected orderliness of political transitions according to established rules. In this index (PS) is closely related to other variable such as control of corruption (CC), Rule of Law (RL) and Government effectiveness (GE) (Bacha Kebede et al., 2020)

Political Instability in Ethiopia: Trends, Dimensions, and Drivers
In our analysis of political instability in Ethiopia, we use the world bank political stability and absence of violence index. Based on this index we will highlight the trends concerning political stability in Ethiopia between 2005 and 2019. Given the fact that these general trends mask the main dimensions and drivers of political stability, this section briefly discusses different aspects of political instability as well as the main factors that drive political instability during the period under investigation.
Political stability and absence of violence index rates countries based on their score that could take any value between -2.5 (weakest in terms of political stability) and 2.5 (strongest). Throughout the period under investigation in this article 2005-2019, it is obvious that that Ethiopia suffers from a high degree of political instability, or as described by Bacha Kebede and his colleagues (2020) as a state that 'has been unstable ever since its unification.' The average negative value -even when political stability seems to improve-shows that there is a tendency for political actors to deviate from the patterns defining their political roles. In other words, there was a 'mismatch between the formal roles and structure set by the authority and those constructed by the informal social interactions' causing almost 'perfect instability' (Bacha Kebede et al., 2020) The year 2005 signaled the beginning of deteriorating political stability that reached its lowest point in 2007 at -1.8. this was followed by slow and little improvement between 2008 and 2014 when PSI reached the value of -1.3. From this point, a sharp decline started again and continued throughout till 2019 as the country was engulfed again toward political instability.
However, this general picture drawn from the PS index, tells us very little about the drivers of political instability nor the factors that contributed to relative improvement. It also tells us very little about the different dimensions and types of political stability throughout the period under investigation. It is against this backdrop, that we will turn our focus to the different proxies for political instability in Ethiopia namely: violence associated with electoral processes, civilian protests and unrest, and ethnic communal conflicts. We will then briefly discuss the main factors that could explain this state of instability in Ethiopia.

Electoral process and instability in Ethiopia
Throughout the 15 years that followed the fall of the authoritarian regime of Mengistu, there were five multi-party elections in Ethiopia, none of them was detrimental to Ethiopia's political stability as the 2005 elections, which explain the declining trend in political stability that started in 2005 till 2007.
The general elections of 2005 were exceptional in more than one aspect (Lahra Smith, 2007). First, it was the first election since the 1991 transition that saw a significant participation of opposition political parties, after they boycotted elections in 1992 and 1995. In 2000 elections, some major opposition parties decided to take part, but they were unable to challenge the ruling party effectively and won only twelve seats in the House of People's Representatives, which remain largely dominated by the ruling coalition. Second, it was not the mere participation of opposition parties that made a difference, but the quality of this participation that seems to challenge the EPRDF control over political life. Leading political opposition namely the United Ethiopia Democratic Forces (UEDF), and the Coalition for Unity and Democracy (CUD), as well as smaller parties such as the Oromo Federalist Democratic Movement (OFDM) all Campaigned on specific promises and policy positions. In the debates that preceded the elections, they tackled difficult issues such as land ownership, economic development, and even the underpinning principles of ethnic federalism itself (Lahra Smith, 2007) Third, the political environment was more open compared to previous elections. The ruling coalition agreed to introduce some reforms such as amending electoral law, allowing international observers, and opening up the political space so that opposition parties could enjoy balanced media coverage and fair and equal access to state-controlled media (Christian Opitz et al., 2013) However, what started as an unprecedented experience in the history of Ethiopia's politics changed drastically to be one of the most violent electoral processes. Preliminary results indicated that the opposition was likely to obtain a surprisingly large numbers of seats in the parliament. When confronted with an unexpected challenge to its dominance, the EPRDF used its power over the NEBE to manipulate the results (Christian Opitz et al., 2013). Final results showed an increase in the opposition share of seats from just 12 to 173 seats with the EPRDF still holding onto its parliamentary majority. These results were challenged by the opposition and its supporters and the opposition decided to resort to civil disobedience, a step that the government considered unconstitutional and a pretext to armed rebellion. Consequently, the government started its clamp-down on opposition with oppositions parties' leaders arrested, and hundreds of people massacred during anti-EPRDF demonstrations. (Melakou Tegegn, 2008& Lahra Smith, 2007 Following 2005 elections, the ruling EPRDF managed to eliminate any challenge from opposition parties. In 2010 opposition parties won only one seat in parliament and in the subsequent 2015 elections, the EPRDF won 100% of the seats in parliament (ACLED, 2017)

Inter-ethnic and communal conflicts
A second line of conflict has been the persisting violent inter-ethnic and communal conflicts. Being one of the most multi-ethnic societies, a consistent feature was conflict between communities over a wide range of issues, ranging from disagreement over killil (region of federal unit) borders and grazing lands, as well as power struggles among communal groups to assert their control over regions (Semir Yusuf, 2019).
Unlike the violence associated with the elections, which appears only during periods of electoral processes, community conflicts are a constant feature of Ethiopia's political life. Even though the adopted ethnic federalism system created largely ethnic-based territorial units, these units are multi-ethnic in their composition, rendering competition between these ethnicities overpower, dominance and resources inevitable.
In the Amhara region for example, the Qemant community is demanding autonomy within the Regional State. Although the Amhara government granted self-rule to the Qemant in 69 districts in 2017, there are fears amongst the community that Amhara government officials would gradually seek to control surrounding Qemant kebels. The Qemant issue should not also be isolated from a long-standing Amhara-Tigray deteriorating relations and territorial disputes the leaders of the National Movement of Amhara (NAMA) perceive the issue as a 'form of proxy war waged by the TPLF' (William Davison et al., 2018& Semir Yusuf, 2019 Another case to illustrate this aspect of political instability, is the conflict between contending Oromo and Amhara nationalisms over the Oromo Special Zone in the Amhara region. This zone enjoyed self-autonomy for over two decades, however, plans have emerged to scrap that zonal status, creating intercommunal conflict in the region (Semir Yusuf, 2019).

Civilian Display of Protest
Protest movements are functions in socioeconomic deprivation and/or political marginalization as argued by Leonardo R. Arriola (2013). According to him measures reflecting great control of resources, whether organizational or material are associated with higher likelihood of protest. Similarly, greater repression of dissent also increases the likelihood of protest related violence.
The most recent wave of political instability that engulfed Ethiopia from 2015 up until 2018 was characterized by mass mobilization and protest movements. From November 2015, anti-government protests-commonly referred to as the 'Oromo protests'-swamped the country. These protests initially erupted in Oromia against the controversial 'Addis Ababa Integrated Regional Development Plan' Also known as the Master Plan. Mass mobilization for these movements started earlier around April-May 2014, when students across several locations in Oromia Regional State protested the Master plan.
Data from Armed Conflict Location and Event Dataset (ACLED) showed fluctuations in the intensity of protests, numbers of protests and riots dropped sharply from 56 in October 2016 to only 2 in January 2017. (ACLED, 2017). This significant reduction in riots and protests, however, was accompanied by an increase of political and ethnic militia activity, as well as battles involving security forces and foreign-based rebel groups, especially in Oromia, Amhara and Tigray. (ACLED, 2017). This could be explained in the light of government's response to these protests that range from imposing the state of emergency, the use of legal tools to criminalize protesting as an act of terrorism, and the excessive use for force by law enforcement agencies. Confrontations between protestors and security forces resulted in resulted thousands of casualties and tens of thousands of arrests for terrorism offences. Few remarks worth mentioning regards the mass protest movements. First: that what started as 'Oromo protests' have soon spread to several parts of the country. In the SNNPR for example university students held various protests and sit ins. Those who joined the mobilization and protest movements across the country resented the government response to the Oromo movement. This act of solidarity is particularly striking between the Amhara and the Oromo given the long-standing historical tensions as well as differing political agendas. Second, A combination of factors has contributed to the 2015 protests amongst which economic exclusion, lack of economic opportunities, soaring inequality, high rate of youth unemployment, corruption and failure on the part of the state to adequately respond to these problems. third, protest movements were in large urban-based. Rural populations are perceived to have achieved small but significant economic and sociocultural gains as material improvements in social services and rural infrastructure have exceeded those of previous regimes in modern Ethiopian History. On the other hand, the Urban population is extremely hostile to regimes in power, and much easier to mobilize (Lahra Smith, 2007). This remark is consistent with various theoretical traditions explaining political instability. For example, proponents of frustration-aggression theory, claim that modernization produces instability because of 'the frustrating ways in which people are pulled out of their different traditional cultures into modern economic and national political systems' (Morrison and Stevenson, 1972). Similarly, proponents of the theory of cultural pluralism, see modernization as a cause of political instability by superimposing lines of economic stratification on lines of cultural cleavage (Morrison and Stevenson, 1972)

Methodology
To investigate the relationship between political instability and economic growth in Ethiopia, Auto Regression Distributed lag (ARDL) Cointegration Approach is adopted. The ARDL Bounds cointegration test was developed by Pesaran and Shin (1999) and extended by Pesaran et al. (2001).
This paper uses annual data on Ethiopian Economy between 2005 and 2019, the variables include both political and economic variables represent the direct relationship between economic growth measured by annual GDP per capita growth rate and political stability measured by world bank political stability and absence of violence index; and the indirect impacts through other variables as corruption measured by corruption perception index, government effectiveness index of the world bank, and foreign direct investment. The following table shows the variable used and its sources.

Data Description
For a better understanding of Ethiopian performance in the areas related to our discussion during (2005-2019) the following graphs shows the annual data of Ethiopia.
From the previous graph, we can conclude that although Ethiopia manage to achieve a relatively high economic growth rates during the study period but in real terms the real GDP per capita growth rates that reflects more the economic conditions of Ethiopian citizens tells another story. The Ethiopian government could not maintain the high growth rates, and the general trend of the per capita growth was decreasing over the study period. This could be understood given that the growth rate driver in Ethiopian economy was not a reflection of a real transformation in its economic structure rather an external push especially in the investment related to infrastructure leads to a non-permanent or temporary growth in service sector value added.
On the other hand, the political stability status in Ethiopia measured by the political stability index shows a decreasing trend during the study period reflecting an enhancement in the political stability status in Ethiopia. Although a sharp deterioration in the index in the period (2014-1017) because of the protests of the Oromo against the government development plans that include the acquisition of Oromo land to Addis Ababa territory. The direct relationship between Economic Growth and Political Stability in Ethiopia during (2005-2019) as shown in figure (2) was negative and strongly significant with (-0.57) correlation coefficient.

Model Estimation and Results
although ARDL model doesn't assume the same level integration of the variables, the unit root test is still important as pre-test to make sure that the variables are either stationary at level or at the first difference I(0) or I(1). By applying augmented dickey fuller unit root test we find that both GDP per-capita and FDI are stationary at level I(0), while ps, ge, cpi are stationary at the first difference I(1). So ARDL bounds test approach pf cointegration could be applied since all variables are I(0) and I(1). To estimate the long-term relationship or cointegration ARDL model is estimated. From the estimation result the PS is the only significant factor that affecting GDP. This result suggests that the direct effect of political stability on economic growth is more significant than any indirect effects in Ethiopian case.
For estimating the direct impact of PS adopting the bounds test of cointegration between gpd per capita and political stability, the test assures the existence of a cointegration or a long-term relationship between political stability and growth as F-statistics is higher than the upper limit of the bounds test.
To estimate the short-term effect of political stability on economic growth, an error correction model is estimated. The result suggests the existence of the effect of political stability on economic growth in the short term with a correction level of -1.71.
The previous model tried to estimate both direct and indirect impacts of political stability on economic growth in Ethiopia. The result suggests the existence of a direct impact of political stability on Ethiopian economic growth and to analyze this effect in more detail, the paper estimate both long term and short-term impacts of political stability using ARDL bounds test approach of cointegration and error correction model. The results suggest that political stability have a significant impact on economic growth both in short and long term.

Conclusion
This paper has analyzed the effect of political instability on economic growth in Ethiopia since 2004 using ARDL regression approach to identify the existence of this correlation and the channels through which this impact happened.