Corruption, Political Instability and Economic Development in the Economic Community of West African States (ECOWAS): Is There a Causal Relationship?

Despite the abundant research on economic development, corruption and political instability, little research has attempted to examine whether there is a causal relationship among them. This paper examines the causal relationship among corruption, political instability and economic development in the ECOWAS using the Granger causality test within a multivariate cointegration and error-correction framework for the 1996-2012 period. The findings indicate that political instability Granger-causes economic development in the short term, while political instability and economic development Granger-cause corruption in the long term. In addition, we employed the forecast error variance decomposition and impulse response function analyses to investigate the dynamic interaction between the variables. The results demonstrate positive unidirectional Granger causality from political instability to economic development in the short term and positive unidirectional Granger causality from political instability and economic development to corruption in the long term in ECOWAS countries. Thus, ECOWAS governments should employ policies to promote political stability in the region.


Introduction
There is broad consensus that the Economic Community of West African States (ECOWAS) is one of the least developed regions in the world. Furthermore, the region has continued to face rising corruption and political instability, which in turn contribute to the region's underdevelopment through adverse effects on government revenue, production, savings, investment, growth, income distribution and poverty (see Aisen & Veiga, 2013;Asiedu & Freeman, 2009;Alesina & Perotti, 1996;Edwards, 1996;Fosu, 1992;Ghura, 2002;Gyimah-Brempong, 2002;Gyimah-Brempong & Dapaah, 1996;Gyimah-Brempong & Traynor, 1999; Oto-Peralías, Romero-Ávila, & Usabiaga, 2013). Although researchers have established that corruption is harmful to an economy (see Gyimah-Brempong, 2002;Mauro, 1995), we have observed the coexistence of high levels of corruption and economic development in some Asian countries, such as China. This observation is not surprising, as Leff (1964) and Huntington (1968) suggested that corruption can have positive impact on the economy via increased efficiency in countries where the bureaucracy is inefficient and constitutes a barrier to investment. This statement reflects the "grease the wheels" hypothesis. For instance, entrepreneurs pay bribes to government officials (grease the wheels) to reduce the time that they spend in queues to obtain business permits, licenses or contract approval. This system in turn increases efficiency and investment and, as a result, economic growth (Méon & Sekkat, 2005). In addition, Goldsmith (1999) cited cases in which entrepreneurs greased the wheels and won contracts even at inflated prices, thus facilitating construction of the intercontinental railway system in the United States.
However, the level of economic development (or income level) plays an important role in promoting or reducing corruption and political instability. Mauro (1995) noted that low-income (less developed) countries tend to be corrupt and politically unstable. Hence, higher-income countries (with improvements in economic conditions) tend to have lesser corruption (Montinola & Jackman, 2002;Van Rijckeghem & Weder, 2001;Schumacher, 2013) and higher political stability (Adelman & Morris, 1968;Helliwell, 1994). Corruption and political instability, including economic development, seem to reinforce one another, at least in the ECOWAS region.
Whereas the military allude to corruption and poor economic conditions, among other things, for seizing power or staging a coup, it has also been accused of engaging in massive corruption and failing to bring the majority of citizens out of poverty (Edi, 2006). Moreover, some scholars hold the view that corruption is an important determinant of political instability (Mauro, 2004;Mbaku & Paul, 1989). Other authors claim that political instability accounts for the high level corruption found in many countries (Billger & Goel, 2009;Campbell & Saha, 2013;Park, 2003;Serra, 2006;Shabbir & Anwar, 2007;Zhang, Cao, & Vaughn, 2009) and that political stability tends to moderate the adverse effects of corruption in an economy (Habib & Zurawicki, 2001).
The various reports of Transparency International (TI), which publishes the corruption perception index (CPI), suggest that most ECOWAS countries are highly corrupt. Of the 15 countries in the ECOWAS region, only Cape Verde is among the top 50 in the TI ranking in the past few years. Similarly, the Political Risk Service International Country Risk Guide (ICRG) political risk rating, which reflects the extent of political uncertainty, demonstrates that ECOWAS countries are politically unstable. Many ECOWAS countries have had an average rating of less than 60% for several years, indicating that they have been experiencing serious political problems.
Despite the abundant research on economic development, corruption and political instability, few attempts have been made to examine whether there is a causal relationship among them. Most studies have focused on the relationship between two of the variables, while studies examining the association among the three variables are almost non-existent. For instance, researchers have investigated the causal effects of the relationships between corruption and development/growth (Bentzen, 2012;Blackburn & Forgues-Puccio, 2007;Gyimah-Brempong, 2002;Mauro, 1995;Ugur & Dasgupta, 2011), political instability and development/growth (Aisen & Veiga, 2013;Comeau, 2003;Fosu, 2002aFosu, , 2002bMbaku, 1988), and corruption and political conditions (Mbaku & Paul, 1989;Montinola & Jackman, 2002 Hence, the main objective of this paper is to examine the causal relationship among corruption, political instability and economic development in ECOWAS countries. The remainder of the paper is organized as follows. Section two reviews the relevant literatures.
Section three describes the theoretical framework and model, and section four presents and discusses the results. Finally, section five concludes the paper.

Review of relevant literature
Most studies on corruption, political instability and economic development have focused on their causes and consequences. In fact, some authors have found political factors and/or the level of development (income level) to be important determinants of corruption. For instance, Park (2003) employed multiple regression analysis to examine the determinants of corruption across countries. The author discovered that economic freedom, socio-political stability, a tradition of law abidance and national cultures are the major factors explaining corruption in the countries considered in his study. (2007) Billger & Goel, 2009;Campbell & Saha, 2013;Emerson, 2006;Goldsmith, 1999;Iwasaki & Suzuki, 2012;Lederman, Loayza, & Soares, 2005;Serra, 2006;Shabbir & Anwar, 2007).

Del Monte and Papagni
Meanwhile, some studies have established that corruption has a damaging impact on political conditions in a country, whereas others have concluded that higher economic development tends to reduce the level of political crisis in a society. For example, Mbaku and Paul (1989) found a positive relationship between rent seeking (corruption) and destabilization of political activities. Habib and Zurawicki (2001) concluded that corruption has an adverse effect on both domestic and foreign investments and found that the degree of openness and political stability of the host country moderates the influence of corruption. Montinola and Jackman (2002) discovered that corruption is lesser in dictatorships than in partially democratized countries. In addition, they found that higher levels of democracy reduce corruption. Other important determinants of corruption include membership in the Oil Producing and Exporting Countries (OPEC) and low wages for public sector employees in low-income countries. Claderon and Chong (2006) confirmed that democratic regimes have a negative association with rent seeking behavior in Uruguay. Moreover, a number of studies have shown that an improvement in economic development (or a higher growth rate) tends to reduce the political instability in a country (see Bollen & Jackman, 1985;Gasiorowski, 1998;Gupta, Madhavan & Blee, 1998;Gyimah-Brempong & Traynor, 1999).
The impact of corruption and political instability on economic development or growth has also been empirically investigated. Some studies have confirmed a negative impact of corruption on growth (see Gyimah-Brempong, 2002;Mauro, 1995;Ugur & Dasgupta, 2011). For instance, Mauro (1995)

Author(s)
Objective(s) Empirical evidence Park (2003) Examined the determinants of corruption across countries Socio-political stability leads to lower corruption Serra (2006) Investigated the determinants of corruption in 62 countries Higher levels of political instability are associated with higher corruption, while the length of preservation of democratic institutions has a negative relationship with corruption Shabbir and Anwar (2007) Evaluated economic and non-economic determinants of corruption in 41 developing countries The level of economic development has a negative effect on corruption Billger and Goel (2009) Determined whether greater democracy and economic freedom lead to less corruption in nearly 100 countries Greater democracy reduces corruption Lederman et al. (2005) Examined the determinants of corruption with a primary focus on political institutions that increase accountability across countries from 1975 to 1999 Democracies along with parliamentary systems, political stability, and freedom of press are negatively related to corruption Mbaku and Paul (1989) Tested the hypothesis that corruption destabilizes political activities in African countries Corruption leads to the destabilization of political activities Bollen and Jackman (1985) Examined economic and non-economic determinants of political democracy in a sample of almost 100 countries Economic development was significant in all regression analyses Gupta et al. (1998) Analyzed the relationship among democracy, political instability and economic growth in a sample of 120 countries Growth in income per capita has a positive impact on democracy, but the effect on political violence is negative Gyimah-Brempong and Traynor (1999) Explored the relationship between political instability and economic growth in SSA Higher economic growth leads to lesser political instability Gyimah-Brempong (2002) Evaluated the impact of corruption on economic growth and income distribution in African countries Corruption decreases economic growth directly and indirectly via reduced investment in physical capital Ugur and Dasgupta (2011) Examined the effect of corruption across countries Corruption has a negative effect on per capita GDP growth. Anoruo and Braha (2005) Investigated the impact of corruption on economic growth in 18 African countries Corruption reduces economic growth directly by slowing productivity and indirectly by lowering investment Aisen and Veiga (2013) Assessed the effect of political instability on economic growth in a sample of 169 countries from 1960 to 2004 Higher political instability leads to lower GDP per capita growth rates via its effect on productivity growth as well as physical and human capital accumulation Level of economic development has a significant impact on corruption Butkiewicz and Yanikkaya (2006) Estimated the relationship between economic growth and five measures of democracy in 100 countries from 1970 to 1999 Democratic countries have higher growth rates

Author(s) Objective(s) Empirical evidence
Fosu (2002) Studied the different effects of various elite political instability situations (which include coups d' état, abortive coups or coup plots) on economic growth in 31 SSA countries from 1960 to 1986 Abortive coups and coup plots rather than successful coups have a negative effect on economic growth Mbaku (1988) Examined the impact of political instability on economic development in SSA countries Lack of political stability has negatively impacted economic performance Ades and Chua (1997) Evaluated the effect of regional instability on economic growth in 98 countries from 1960 to 1985 Existence of negative spillovers in politically unstable neighboring countries  Investigated the relationship between political instability and GDP per capita growth in a sample of 113 countries from 1950 to 1982 Growth tends to be lower in countries and periods with a strong tendency for government collapse have adverse effects on economic performance (that is, sluggish economic growth), leading to the collapse of the government. For example, corruption among government officials or bureaucrats reduces the amount of social services such as healthcare and education that alleviate poverty and inequality (Gupta, Davoodi, & Alonso-Terme, 2002) and therefore leads to social discontent, protests, strikes, and political violence. Sustained dissatisfaction among citizens will lead to a collapse of (or a change in) government. In developed countries, a change in government occurs through the electioneering process (Gyimah-Brempong & Dapaah, 1996) and in line with constitutional provisions. By contrast, in developing countries such as those in the ECOWAS, an unconstitutional change in government through military intervention has always been the case. Military takeover in the ECOWAS region appears to be the norm rather than exception (Edi, 2006), and it tends to have a destabilizing impact on political stability in these countries because of their fragile political structures (Adelman & Morris, 1968).
Moreover, frequent changes in government induce public or elected officials to practice rent-seeking behavior because of the high uncertainty surrounding their tenure of office. In a previous study, Shleifer and Vishny (1993) cited in Park (2003) argued that if public officials realize that their term in office will be shortlived because of political instability, they will become irresponsible and become involved in rent-seeking behavior. Using the same line of argument, Park (2003) opined that high uncertainty and anxiety among public officials (arising from political instability) would lead them to seek gain through corrupt means to protect their social status even after they no longer have their positions. Moreover, in explaining Lipset and Raab's (1970) concept of "status strain", Park (2003) emphasized that the fear of a decline in status will compel people to do anything (including engage in corrupt behaviors) to maintain their status and property. Economic development (or income level) largely promotes or reduces corruption and political instability in a country. Mauro (1995) proposed that low-income countries are likely to be corrupt and politically unstable. In fact, low income levels or civil servant wages encourage rent-seeking behavior because people see corruption as an opportunity to improve their socio-economic well-being. By contrast, an improvement in economic conditions (such as higher wages or income) tends to lower corruption (Montinola & Jackman, 2002;Van Rijckeghem & Weder, 2001;Schumacher, 2013). Similarly, declining economic fortunes encourages discontent and socio-political instability. However, higher economic development, such as higher incomes, improves people's well-being and promotes political stability (Adelman & Morris, 1968;Helliwell, 1994).
Meanwhile, economic development can be influenced by the level of corruption and political instability through their negative impact on savings, investment and production, among other effects. For instance, political instability disrupts production activity, reduces investment and negatively influences economic performance (Alesina & Perotti, 1996;Aisen & Veiga, 2013;. According to Butkiewicz and Yanikkaya (2005), one of the best measures to improve the economic well-being of people in the poorest nations is to prevent political instability. In the same vein, corruption discourages investment and production, leading to sluggish growth (Asiedu & Freeman, 2009;Gyimah-Brempong, 2002;Mauro, 1995

Corruption, Political Instability and Economic Development
Where i refers to a given country and t a given year; i α , i β and i δ are coefficients; and U is the error term.
The data used in this paper were obtained from three main sources: TI, ICRG and the World Bank's World Development Indicators (WDI). Specifically, data on political instability were collected from the ICRG. Political instability has been measured by the number of successful coups, the number of people killed in domestic mass violence incidents as a fraction of the total population, the number of attempted but unsuccessful coups, or the number of politically motivated assassinations (Alesina & Perotti, 1996). Unfortunately, such (rich) data are not available for most ECOWAS countries for a considerable number of years. Moreover, the objective indices are not without shortcomings. For instance, frequent changes in government, which is an indication of political instability (Edwards, 1996), may give the wrong information about the political conditions in a country. Mauro (1995) noted that although Italy had more than fifty changes in government between 1945 and 1995, the country remained relatively stable during the period. Thus, we captured political instability by using the ICRG political risk rating (index).
Although the ICRG index is subjective, it has been increasingly used in empirical research and has been found be highly associated with economic variables (see Erb, Harvey, & Viskanta, 1996;Hayakawa, Kimura, & Lee, 2013). The index measures the extent of political instability or uncertainty in a country, and its components include political leadership, military role in politics, external conflicts, the role of organized religion in politics, racial and national tension, law and order, political terrorism, civil war, and political party development. The index ranges from 0% (indicating higher political instability) to 100% (indicating higher political stability), and it has been employed in previous studies (see Abu et al., 2013;Erb et al., 1996;Hayakawa et al., 2013;Heaney & Hooper, 1999;Linder & Santiso, 2002).
Corruption is difficult to measure/quantify, and what is perceived as a norm in one country at one point in time may be considered corruption in other countries. Also, given that most corrupt practices are regarded as unlawful activities, they occur in secrecy.
Thus, it is difficult to measure/quantify them. Moreover, the only objective measure of corruption that is available is the number of individuals who have been convicted of engaging in corrupt practices. However, higher conviction rates (as in the case of Singapore and Hong Kong) do not necessarily imply that corruption is higher but may indicate the effectiveness of the judiciary and anti-corruption agencies in detecting and prosecuting offenders (Lambsdorff, 1999;. Given the weakness of such objective data, corruption perception indices (subjective data) have been widely used. Furthermore, because of the lack of adequate measurements of corruption, one may resort to using corruption perception indices (Gyimah-Brempong, 2002). In this study, therefore, we employed the TI corruption perception index (CPI) that has been employed in many empirical studies as a measure of corruption. The CPI reveals the extent to which a country is perceived to be corrupt. This index is also a reflec-

Results
Having specified the respective models, we conducted a unit root test to ascertain whether the series used in this study are stationary. Standard economic theory requires series to be stationary prior to estimating their relationship to avoid generating spurious results. Fisher aug- The table clearly indicates that the series have a unit root at level but are stationary at the first difference. This outcome supports the claim that many macroeconomic variables are non-stationary at level but stationary after the first difference (Nelson & Plosser, 1982). Our next task is to investigate if there is a long-term equilibrium relationship (cointegration) between the series using the Pedroni residual cointegration test (Pedroni, 1999). The The existence of cointegration suggests that the estimated relationship is not spurious. In addition, if the tests reveal the presence of cointegration, then causality will exist in at least one direction (Granger, 1986). The results of the cointegration test are presented in table 3.
Given that the variables are cointegrated, we took another step to determine the direction of causality between them. Granger (1969)

Results of the Pedroni Residual Cointegration Test
Note: POL refers to political instability, COR refers to corruption, and PCY refers to economic development. ** and *** indicate a rejection of the null hypothesis of no cointegration at the 5% and 1% significance levels, respectively.

Forecast Error Variance Decomposition Analysis
The Granger causality analysis conducted above is limited to the 1996-2012 period, but it does not consider the dynamic interaction of the variables beyond that period. In an attempt to understand the dynamic relationship among corruption, political instability and economic development outside of the sample period of 1996-2012, we performed a forecast error variance decomposition analysis (FEVD) (Sims, 1980). The FEVD is useful in assessing the amount of variation in a variable caused by its own shock and by shocks to other variables. In the short term, a larger percentage of the variation in a variable results from its own shock, while in the long term, the impact of shocks on other variables increases. Each of the variables in the system is disturbed by one standard deviation.
The results of the variance analysis presented in

Impulse Response Function Analysis
The In sum, the empirical results indicate that there is positive causality running from political instability to economic development in the short term and from political instability and economic development to corruption in the long term in ECOWAS countries. In separate studies, Fosu (1992Fosu ( , 2001Fosu ( , 2002aFosu ( , 2002b confirmed that political instability is deleterious to economic growth in Sub-Saharan African (SSA) countries. For instance, Fosu (1992)  Moreover, military rulers in the ECOWAS region have been blamed to a greater extent for the institutionalization of corruption as evident in some countries.
The various military regimes alluded to corruption as one of the reasons that they seized power (Edi, 2006).

Conclusion and recommendations
Given that less developed ECOWAS countries are corrupt and politically unstable, it is important to examine the interaction among economic development, corruption and political instability in these countries. This paper examines the causal relationship among corruption, political instability, and economic development in ECOWAS countries within a multivariate cointegration and error-correction framework. The Pedroni cointegration test reveals that the variables are cointegrated, indicating the existence of a long-term equilibrium relationship among corruption, political instability, and economic development. Having confirmed the existence of cointegration, we investigated the direction of causality between the variables using the VECM. The results illustrate that there is short-term unidirectional causality from political instability to economic development, while in the long term, causality runs from economic development and political instability to corruption in ECOWAS countries.
Moreover, we employed the FEVD and IRF to examine the dynamic interaction among corruption, political instability and economic development in ECOWAS outside the sample period of 1996-2012. The FEVD confirmed that corruption, political instability and economic development are endogenous. Political instability is the most important variable accounting for shocks in corruption, while corruption is the most important variable accounting for shocks in political instability and economic development. Furthermore, the IRF illustrated that a shock to political instability and economic development has a positive effect on corruption. Additionally, a shock to corruption has a positive impact on political instability, while a shock to economic development has a negative effect on political instability. In addition, a shock to political instability has a positive effect on economic development, whereas a shock to corruption has a negative impact on economic development. Thus, there is positive unidirectional causality from political instability and economic development to corruption in the long term and positive unidirectional causality from political instability to economic development in the short term in ECOWAS countries.
The findings of this study suggest that years of political instability have contributed to the high rate of corruption and underdevelopment in ECOWAS countries. Researchers have confirmed that corruption increases with political instability (see Lederman et al., 2005;Leite & Weidmann, 1999;Park, 2003). Likewise, other proxies of political instability have also been found to be correlated with corruption. For instance, it has been observed that democracy (as measured by press freedom) is negatively related to corruption (Brunetti & Weder, 2003;Lederman et al., 2005). Similarly, improvements in civil liberty reduce corruption (Lederman et al., 2005). In addition, higher levels of decentralization reduce corruption by bringing government closer to the people and ensuring that government officials can be held accountable when the need arises (Fisman & Gatti, 2002). Furthermore, political instability destroys physical capital and displaces human capital (Le, 2004), disrupts production activ-Corruption, Political Instability and Economic Development ity (Aisen & Veiga, 2013), and encourages the violation of property rights, including lack of guarantee for contracts (Svensson, 1998). All of these consequences adversely affect the economy (Aisen & Veiga, 2013;. High levels of corruption and underdevelopment in ECOWAS have been blamed on political instability primarily resulting from many years of military rule (including ethno-religious crises). For instance, Edi (2006) posited that the failure to improve socioeconomic conditions and high corruption, among other factors, led to reoccurring military takeovers in ECOWAS countries. Based on our findings, ECOWAS governments should employ policies to promote political stability to improve economic development and reduce corruption in the region. However, as we stated in the introduction, the use of a single policy option of political stability may not be sufficient to achieve higher levels of economic development and to eliminate corruption from ECOWAS countries.