CHAPTER TWO
LITERATURE REVIEW
2.1Conceptual and Theoretical Review
This chapter presents a review of related literature on the governance and growth in the ECOWAS member states, and extended globally.
2.1.1Conceptual Review
Beyond governance which is used to capture economic growth in this study, there are other factors that affect the economic growth of a country or region, this study will dwell on governance and the description of the chosen variables is given. This explanation will help us to choose the appropriate methodology to examine the relationship between governance and economic growth. Therefore, the focus of this section is to establish a theoretical and conceptual framework for the quantitative research.
2.1.1.1Governance
In recent time in the study of economics, more attention is now been given to the institutional arrangement in the quest of growth and development. Specialists in the Medieval English Society were the first to use the concept of governance as referred to in the growth and development economics literature. They characterized governance as cooperation between the different sources of powers. The term governance today is widely used in different international development literature and there are numerous interpretations given to describe the concepts (Haq, 2009). Governance means process of decision making and the process by which decisions are implemented. The quality of governance is determined by the impact of this exercise of power on the quality of life enjoyed by the citizens. Governance can be used as several contexts such as international governance, corporate governance, national governance and local governance. Government is one of the actors in governance. Although no consensus exists in terms of defining governance, a common idea among scholars is that governance means more participation in the political and decision-making process by non-governmental institutions (Agere, 2000; de Ferranti et al 2009; Lovan et al 2004). Thus, under governance, government is one of several players—rather than the only player—managing a nation’s affairs (Frahm and Martin, 2009; Kettl, 2002; Lovan et al., 2004).
Asian Development Bank (1995) identified four basic elements of good governance such as accountability, participation, predictability and transparency. McCawley (2005) categorizes governance issues at the macro and micro level. The macro level includes constitution, the overall rule of size and resources of the government, and relationship between legislators, the judiciary and the military, while micro issues of governance includes commercial firms, social institutions and civil society affairs. According to de Ferranti et al. (2009), “governance describes the overall manner in which public officials and institutions acquire and exercise their authority to shape public policy and provide public goods and services” (p. 8). Governance “represents the overall quality of relationship between citizens and government, which includes responsiveness, efficiency, honesty, and quality” (p. 8). Similarly, the United Nations defined governance as “the process of decision-making and the process by which decisions are implemented (or not implemented)” (UNESCAP, 2009, p. 1). United Nation Development Programme (1997) defines governance as the exercise of economic, political and administrative authority to manage a country’s affairs at all levels. It comprises mechanisms, processes and institutions through which citizens and groups articulate their interests, exercise their legal rights, meet their obligations and mediate their differences. The United Nations also introduced characteristics of good governance practices as a global standard to be adopted by governments that receive their aid. According to the United Nations, “good governance has 8 major characteristics; it is participatory, consensus oriented, accountable, transparent, responsive, effective and efficient, equitable and inclusive, and follows the rule of law” (UNESCAP, 2009). International Country Risk Guide (ICRG) covering 140 countries from 1980 to the present analyses and forecast risk for international investors. It includes 22 components that are grouped into three categories of risk: political, financial and economic. The political risk assessments are made on the basis of subjective analysis of the available information, while the financial and economic risk assessments are made solely on the basis of objective data. In determining the component rating, political risk contributes 50 percent to the rating while the other two categories contribute 25 percent each.
The current use of the governance concept may be traced to a World Bank study in the year 1989 on Africa that defined governance as “the exercise of political power to manage a nation’s affairs”. Later, the World Bank (1992) defined governance as “the manner in which power is exercised in the management of a country’s economic and social resources for development”. The Organization for Economic Co-operation and Development (OECD), on the other hand, defined governance as “the exercise of authority in government and the political arena”. In line with this definition, “Good public governance helps to strengthen democracy and human rights, promote economic prosperity and social cohesion, reduce poverty, enhance environmental protection and the sustainable use of natural resources, and deepen confidence in government and public administration” (Tarschys 2001, 28).
World Bank aggregate governance indicators data set developed by Kaufmann, et al. (2005) hereafter called the KK Data sets, is a set of worldwide measures of six composite dimensions of governance perception indicators for 105 countries. These indicators are oriented so that higher value correspond to better outcomes, on a scale refers to the point estimates range from –2.5 to 2.5. These estimates are also rescaled and ranked in percentile (0–100). The lower percentile is ranked as worse off governance indicators whereas upper percentile is ranked as best governance for any given country. These perceptions may often be more meaningful than objective data, especially when it measures public faith in institutions. These averages of governance indicators are considered to capture institutional quality. These dimensions can be classified into three clusters with two indicators in each group is given as:
2.1.1.1Governance from Economic Dimension
Under economic governance the two indicators representing this indicator are Government effectiveness and Regulatory quality. These two indicators summarize various indicators that include the government’s effectiveness which shows the quality of public services, the quality of civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility of the government’s commitment to such policies. The thrust of this index is on the input required government to be able to produce or implement good policies and quality delivery of public good. The ‘regulatory quality’ governance indicator covers the ability of the government to formulate and implement sound policies and regulations that permit and promote private sector development and this includes measures of the incidence of market unfriendly policies such as price control or inadequate bank supervision, as well as the perceptions of the burdens imposed by excessive regulation in area such as foreign trade and business development.
2.1.1.2 Governance from Political Dimension
Political governance covers these two; voice and accountability and political instability and violence. The political governance indicator is intended to capture the process by which government is selected, monitored and replaced. First indicator which is voice and accountability measures political, civil and human rights and independence of the media. It includes a number of indicators measuring diverse aspects of political process, political rights and civil liberties. It expresses to which extent a country’s citizens are able to participate in selecting their government, as well as freedom of expression, freedom of association and free media while the ‘political instability’ indicator captures the likelihood that the government in power will be destabilized or overthrown by possibly unconstitutional or violent means, including military cop, terrorism etc.
2.1.1.3. Institutional Dimension of Governance
Institutional dimension of governance is captured by Rule of law and Control of corruption. The final dimensions of governance indicators are summarized in broad terms as the respect of citizen and the state of institutions that govern their interactions. Rule of law, summarizes several indicators that measure the extent to which agents have confidence in and abide by the rules of the society. It measures the quality of contract enforcement, the police and the courts as well as likelihood of crime and violence. These indicators also measure a society’s success in developing environment in which fair and predictable rules form basis for economic and social interactions. Control of corruption measures its perceptions conventionally defined as the exercise of public power for private gains. This aspect of corruption differs somewhat, ranging from the occurrence of additional payment to get things done, to assess the effect of corruption on business environment, to measure grand corruption in political arena or in the tendency of elites to engage in state capture. The presence of corruption is often a manifestation of a lack of respect on the part of both the corruptor and the corrupted for the rules that govern their interaction, thus represents a failure of governance (Haq, 2009).
2.1.2Economic Growth
Over the years, there has been no universally agreed- upon definition of economic development, however, there is a unionism in the view of development researchers that economic development brings about improved living standards for people and is necessary for a strong long-term national economy. Economic development implies both the improvement of people’s health, education, and general well-being and the presence of positive economic indicators, such as economic growth and low unemployment rates (Adams ;Mengistu, 2008). Sustainable development is another issue related to economic development because, without strong long-term economic growth, an economy will be in danger of collapsing during any economic or political crisis (Blair and Carroll, 2008; Mayer-Foulkes, 2009; Nafziger, 2006). Economic development is important because it has implications on people’s lives (Adams and Mengistu, 2008; Chong and Calderon, 2000). With economic development, people will have better education and healthcare and be more productive. Economic development also affects crime rates and political stability as better-developed nations tend to have lower crime rates and greater political stability than less-developed countries (Kaufmann and Kraay, 2002; Abdellatif, 2003; Adams and Mengistu, 2008). Consequently, economic growth concerns all nations trying to increase their GDP per capita in order to increase their citizens’ well-being (Adams and Mengistu, 2008; Mankiw, 2009; UNDP, 2010). Although scholars continue to debate whether it is a consequence of human development or a precondition for human development, economic growth is considered an important component of economic and human development. Smith (2007) found that human development and economic development need each other; as such, countries cannot concentrate on one and ignore the other. According to Smith (2007), “there is in effect a virtuous circle of human development and economic development, each enhancing the other” (p. 14). In addition, the United Nations Development Program (UNDP) demonstrated that economic growth, education, and health are the key parts of human development, with each part dependent on the others. According to the UNDP (2000), “resources generated by economic growth have financed human development and created employment while human development has contributed to economic growth” (p. 7).
Economic growth is the increase of real gross domestic product (GDP) or other measurements of aggregate income (AlBassam, 2013). According to the World Bank (2004), economic growth is “quantitative change or expansion in a country’s economy” (par. 10). In addition, the World Bank (2004) contended that “economic growth is conventionally measured as the percentage increase in gross domestic product (GDP) or gross national product (GNP) during one year” (par.10). After acknowledging the existence of the relationship between economic growth and human development, the Human Development Reports (HDRs, 2010) indicated that the direction of the relationship is not clear cut. According to the HDRs (2010), “even if there is a causal relation, the direction is unknown: higher incomes could improve quality of life, or improvements in health and education could make societies more productive” (p. 48). In addition, both the income distribution among citizens and the quality of goods and services produced are as important for any nation as increasing income levels. According to Petrovskiy (2009), “from a human development perspective, the quality of economic growth is just as important as its quantity” (p.134).
Furthermore, economic growth has been linked to governance improvement (Albassam, 2012; Furubotn and Richter, 2005; Kaufman and Kraay, 2008; Mantzavinos, 2001). Kaufmann and Kraay (2002) argued that governance quality and economic growth are positively related. In their evaluation of the worldwide governance indicators (WGI) from 1996 to 2002, they found that “per capita incomes and the quality of governance are strongly positively correlated across countries” (p. 1). Accordingly, the relationship between economic growth and quality of governance impacts international aid assistance from countries such as the U.S. and the U.K. and from international organizations such as the World Bank and the IMF. According to Mehanna et al (2010), “the issue of causality between governance and economic development is crucial and has many implications from an international agency perspective; resolving this issue would assist international organizations in their choices between prioritizing pro-growth or institutional policies” (p. 123). Therefore, the power and direction of the relationship between economic growth and governance have been and will continue to be the subject of disagreement among policymakers and those in academia (Acemoglu et al., 2001; Alkire, 2010).
2.1.3Governance and Economic Growth
Different literature has linked several theories especially the institutionalist’s theories proposed by the School of Compatibility among others, to the study of the relationship of governance and economic growth (Porras and Vázquez, 2015). These theories argue that development depends on the ability to secure property rights and enforce contracts in the economies. Governance, and eventually democracy, promotes growth because they ensure property rights, business transactions, social rights and the provision of public goods. Studies of the relationship between governance and growth cannot be separated from the debate about the role of government in the economy. Particularly, Bevir (2007) and Aguilar (2010) argue that the concept of governance reflects the transformation in the vision of the State that occurred after the reforms of the public sector of the years 1980 and 1990. As it is known, these reforms reduced the importance of bureaucracy and introduced market criteria in decision-making. For this reason, studies on governance have normative connotations. Governance has changed the way of formulating and implementing public policies; these are no longer considered as unilateral processes. Public decisions, from the perspective of governance, involve coordination, discussions, understandings, negotiations, agreements and public, private and social commitments. In this context, Bevir (2011) highlights the importance of markets, social networks and non-state actors for public policy.
In line with the above view, Bevir emphasizes that governance defines rules that go beyond those defined by the formal powers of the State. Traditionally it is assumed that an improvement in governance and institutions can promote economic growth (Acemoglu, Johnson and Robinson, 2005; Acemoglu and Robinson, 2006 and 2012). So, some argue that consolidated governance and democratic institutions can enhance the growth of developing economies. However, it should be recognized that this regulatory requirement has been questioned by Glaeser et. al. (2004) and by advocates of the “School of Conflict” and “skeptical school”. Therefore, despite popular belief, there are still theoretical debates on the relationship between governance and economic growth.
2.2Theoretical Review
2.2.1Growth-Governance Hypothesis
The Gross Domestic Product (GDP) has been the most frequently used measure of economic growth in the academic literature and most studies have revealed that there exist a strong positive association between high-quality governance institution and good economic performance (Wilson, 2015). Change in the quality of a country’s quality of governance can be brought about by improvements in the economic performance through different mechanisms. The first mechanisms are the relative payoff to investments in the formal governance, rather than resting on informal mechanisms, which will in turn increase with a country’s level of economic activity. Personal ties and repeated interactions is the drive of economic exchange and activities when it is localized and grooming stage (Dixit, 2004). As the volume of trade and economic activities grows and the economic develops, it may enhance the relative efficiency of formal governance mechanisms, creating stronger incentives for public investments in improved governance institutions (Li, 2003).
Furthermore on growth to governance hypothesis, the cost and technical requirement of quality governance reforms are lacking in many countries (Rodrik, 2007). Therefore, the advocates of a program of good enough governance argue that growth can often be sparked by relatively minor reforms that encourage investment and that such growth can allow developing countries the time and resources to establish higher quality governance institutions at a later stage of economic development (Grindle, 2007). There are some studies where the linkage between economic growth and good governance is contradicted. For instance, some authors (Stojanovic et al, 2016) argue this conclusion with the examples of certain countries such as Cambodia, China and Vietnam in which the economic growth is clearly demonstrated despite of the lack of good governance. There are few studies that investigated the governance-to-economic performance causal relationship and contradict the conclusions resulted from most of the previous papers which argue the positive influence of good governance on economic development. The author of this study confirms a positive correlation between governance and economic growth, but he concludes that is not necessary to consider the quality of governance as a key factor, determinant for economic performance. Furthermore, they demonstrated a positive influence of economic growth on subsequent quality of governance; therefore he suggests that more attention should be paid to analyze the possible reverse causality between governance and economic growth, particularly in cross-country analyses (Wilson, 2015).
Another channel which have been identified that growth may cause governance is through creating a constituency of business and customers with the interest and ability to demand such improvements. In many developing countries, especially those experiencing transition from a centrally planned economy to a market oriented economy, state-owned firms are the one occupying the central stage in the economy despite being inefficient (Li and Xia, 2008). In other to still maintain their position, these firms have strong incentive to resist reforms that would tighten their soft budget constraints and expose them to competition from new, more efficient private enterprises. In such a setting, growth in the non-state sector – largely supported by informal, network-based forms of governance in place of missing or ineffective formal institutions – may be required to give private economic interests the economic and political power to effectively advocate the governance reform. A substantial number of historical case studies support the hypothesis of causality from economic growth to improvements in governance. For instance, Chang (2003) investigated the historical development of several aspects of governance in the now developed countries of Western Europe and North America. Many of these features of good governance – including a professional bureaucracy, effective corporate regulation, an impartial and independent judiciary, consistent and impersonal enforcement of contracts and protection of property rights, efficient broad-based tax collection, and modern social welfare institutions were shown to have been implemented in the most advanced countries only in the late 19th century (or in many cases well into the 20th century), by which time these countries had already enjoyed half a century or more of industrialization and sustained economic growth. Goldsmith (2007) acknowledges the same trend in his finding for Jamaica, Mauritius, Argentina and United States. His finding suggests that quality of governance at least in the modern sense, was not required for economic growth to take off at the early stages, pointing to the fact that improvement in governance may have been as a result of economic growth.
2.2.2Governance to Growth Hypothesis
Causality may operate between governance and growth through several potential ways. The first is through the professionalization of the bureaucracy and this goes a long way to give the bureaucrats with the assurance of merit-based career and predictable paths in the civil service which bring about a definite stable stay in office which enhances investment in public infrastructure with long-term payoffs rather than short-term which is in form of consumption (Wilson, 2015). A bureaucracy that one can bank upon is capable of motivating private businesses to engage in long-term investment since the perceived risk that comes along change in power or government policies will be highly reduced.
The correlation between the quality of governance and economic growth has been the objective of some cross-country empirical studies (Wilson, 2015; Sharma, 2001) while only few paper are addressing the impact of bureaucratic professionalization and effectiveness of government on economic growth (Cingolani et al, 2015). Economists and policymakers have common ground about the role of institutions in influencing economic development, while several authors in their individual research finds that “There is increasing recognition that corruption and other aspects of poor governance and weak institutions have substantial, adverse effects on economic growth” (Sharma, 2007). In the attempt to answer the question of how important is good governance for economic growth, another author admits the answer is best highlighted by the “oft-cited aphorism that good governance promotes growth and that growth further improves governance”. Sharma (2007) also admits there are a lot of econometric papers whose findings emphasize a strong correlation between long-term economic development and good governance, thus the quality of governance definitely influences long-run economic performance outcomes. Also, the bureaucratic consistency accomplished through organized rule-based decision-making should also increase the success of major infrastructure projects that involve partnership between different levels of governments. Bureaucratic professionalization is of many advantages such as encouraging investment that are productive and also reduces the chances of corrupt practices. The strong positive correlation across countries between GDP per capita and the quality of governance was also admitted by the World Bank (Kaufmann and Kraay, 2002). Even more, researchers from the World Bank proposed in their working papers an empirical strategy to analyze this correlation from two directions, namely: (a) a strong positive causal effect running from a better quality of governance to GDP per capita; and (b) a weak and even negative causal effect running in the opposite direction from GDP per capita to the quality of governance. Their results provide sufficient evidence on the significance of good governance for economic performance and the same study also points out the non-existence of “virtuous circles in which a higher GDP per capita determines further amelioration in the quality of governance (Bota-Avram et al, 2018).
Governance to growth relationship can also be seen from the standpoint of institutional and policy perspective, economic growth can thrive well in the society when laws and regulations are enforced effectively with an impartial judicial system. In examining China (Zhoung, 2001), role played by educational attainment in selecting and assessing public officials have been on the increase which is as a result of the significant legislative and bureaucratic changes taking place over the past few decades thereby improving China’s quality of governance. In line with the movement towards the global benchmark, room is given to independent regulatory agencies to oversee key industries yet they are still subject to political interference (Wilson, 2015).
2.2.3Endogenous Growth Theory
Convergence was predicted by the Neo-Classical using Solow’s view of 1956, and the view is that economies will move toward their steady state growth which implies that in the long run, there will be convergence in their per capita income but to the lack of empirical backup for convergence has been a challenge to their models. Endogenous growth theory came up with some conditional factors that can bring about convergence and these factors are related to institutions. Knowledge spillovers as it is referred to by the new growth theories is the based on the assumption that any sectors in the developing countries can catch-up with the latest trend of technology whenever in innovates and the adaptation of this new innovation rest on the institutional arrangement which governance is key part. Technological changes when seen as endogenous its innovation can be driven by the governance based on the incentive and transaction cost which goes a long way to determine how fast technological changes will actually progress (Siddiqui and Ahmed, 2009).
The quality of governance brings about growth by minimizing the risk of doing business and this in turn brings about channeling of resources to innovation by prevent diversion of resources and preventing predatory rent seeing activities. When governance can give an enabling environment which is free of diversion, the productive units are assured of being rewarded fully to the amount of what they have inputted into the production and this reduces diversion of resources by individual and increases investment in knowledge.
2.3Empirical Review
Growth has been seen as the foundation for development and the role of good governance has been attributed a major role in achieving a sustainable growth. Different academic literature has attempted to explore the relationship between governance and growth. In the empirical literature there are several studies that have examined the relationship between governance and economic growth. Among these studies are found the Knack and Keefer (1995), Mauro (1995) and Alesina (1997). These studies were pioneers in evaluating the mentioned relationships focusing on the role of institutions and property rights. They pioneered the use of cross-sectional indicators for statistical validation. Furthermore, they were the first to consider governance as a multidimensional phenomenon and that its analysis could be separated from the political regimes. Contemporary studies on the relationship of governance and economic growth emphasize the role of property rights and the quality of institutions. These studies tend to find positive relationships between the development of institutions and growth. Among these studies are those of Knack and Keefer (1995) and Ndulu and O’Connell (1999), who evaluate how risk and violence are linked to economic growth. Rivera-Batiz (2002), meanwhile, analyzes the relationships between democracy, good governance and economic growth. The mentioned studies are important because they tend to support the institutionalist theories (see North, 1990 and 2005; Olson, 1996). Furthermore, they are important because they complement other studies on economic development. In particular, they complement studies that emphasize the role of political rights and civil liberties (Grier and Tullock, 1989), democracy and social conflicts (Keefer and Knack, 2002) and the importance of social capital (Gutierrez-Banegas and Ruiz-Porras, 2014).
Alesina et al (1996) studied the effects of political instability on per capita GDP growth of a sample of 113 countries over the period 1950-1982. The major result of their paper is that political instability has negative effects on economic growth. That is, political instability lessens the growth. Moreover, their results suggest that regime changes affect growth adversely. The same findings were reported by Feng’s (1997) study. Using the three stages least-squares estimation for data covering 96 countries covering the period of 1960-80, the author’s findings demonstrated that political instability has significant and negative effects on economic growth.
Kaufmann, Kraay, and Zoido-Lobatón (1999) studied more than 150 countries, provides empirical evidence that good governance matters a great deal for economic outcomes. Kaufmann and Kraay (2002) conducted another study of 175 countries for the period 2000/01, asserting that good governance is necessary for high per capita income and economic development. The same result was concluded by Knack (2002). It is worth mentioning that Kaufmann (with other authors) has examined the impact of governance on economic outcomes in many studies. In each one of them, he comes to the same conclusion stated above.
Moreover, the results of a study by Calderoân and Chong (2000) have confirmed that there is strong causality from institutional quality to economic growth. The authors’ results have also shown that economic growth causes institutional quality. Although their findings indicate that policies that attempt to improve the state’s institutional quality by securing propriety rights, controlling corruption, and limiting uncertainty need considerable time to achieve the desired goal, these policies are important for economic growth. In addition, such a study has shown that institutional reforms have high influence on economic growth, especially for the very poor countries. Furthermore, by answering the question: why do some countries produce so much more output per worker than others? The results of Hall and Jones (1998) have revealed that a country’s long-run productivity, capital accumulation, and thereby productivity per worker are influenced the most by institutions and government policies.
Using the PRASH Model, Campos and Nugent (2000) analyzed the relationship between the growth volatility and political stability of Argentina over the period of 1896-2000. The authors’ findings suggest that “informal” political instability, such as assassinations, directly and negatively affects economic growth; and “formal” political instability affects economic growth indirectly.
In a cross-sectional analysis of all developing countries, Chauvet and Collier (2004) found that those countries suffering from poor governance, on average, experience 2.3 percentage points less GDP growth per year relative to other developing countries. There are also other recent findings that suggest a strong causal effect running from better governance to better development outcomes. In spite of such a broad array of support for the positive impact of good governance on economic growth, there are only few studies that show results to the contrary. For example, an important challenge to the significance of good governance for the economic growth of African countries comes from Sachs et al. (2004). In an empirical analysis, they show that the differences in performance among African countries cannot be explained by differences in the quality of their governance once differences in their levels of development have been accounted for and thus conclude that a focus on governance reforms is misguided.
Acemoglu et al (2005) examined the link between institutions and long-run growth. They argue that when political power is allocated to groups that enforce propriety rights, when there are few rents that can be sought by the groups in power, and when there are effective constraints on power-holders, there will surely be a causality from economic institutions to economic growth.
Drury et al (2006) studied the connection between corruption, democracy, and growth in more than 100 countries for the period 1982-97. The authors’ findings show that corruption does not have any significant impact on growth in democracies, whereas corruption has strong negative effects on growth in non-democratic countries.
Dam (2006) reviewed the relationship between the rule of law and the economic growth of China. The author argues that China is currently facing the same type of governance issues that Asian Tigers have experienced. Asian Tigers’ economic growth has been negatively affected by such governance issues. The author contends that these issues may affect China’s economic growth as they have affected Asian Tigers’ growth. Dam avers that China’s governance weaknesses are associated with many problems, such as a weak judiciary. Additionally, the author concludes that there is nothing in China’s experience from which one can conclude that institutions and rule of law are not important for economic development.
Morita and Zaelke (2007) have studied the link between the rule of law, good governance, and economic development. The authors argue that rule of law and good governance are important to achieve sustainable development. They also emphasize that good governance and sustainable development goals will not be achieved just by making laws and regulations, rather by enforcing those regulations and laws by governments.
In sharing his ideas on governance with World Bank economists, Rodrik (2008) said that governance is an important tool for development. He suggests that it is a good instrument to achieve better economic outcomes and enhance a country’s policy making. Rodrik also distinguishes between governance as a means and as an end. The author advises economists not to try to address governance as an end because it is the political scientists’ task. For governance as a means, however, he argues that only countries having governance as binding constraint can give a governance reform the priority to boost their economic growth.
Amirkhalkhali and Dar’s study investigates the connection between regulatory quality and economic growth of the 23 OECD countries over the period 1996-2008. They use a generalized version of the production function model of Solow. Their findings suggest that regulatory quality and economic growth are positively correlated. That is, a better regulatory quality leads to a high growth rate. The authors argue that regulatory quality has an impact on economic growth through its effects on total factor productivity.
Guisan (2009) examined the link between government effectiveness, education and economic development by comparing European countries to the U.S. and Canada over the period of 2000-07. The author’s results have shown the importance of government effectiveness to economic development.
Many studies have been done to determine the relationship between corruption and growth at the macro-level. One such study has been conducted by Hodge et al, (2009) where the authors used an econometric methodology that can take into account the multidimensional nature of, as well as the inherent endogeneity in, the relationship corruption-growth. Overall, their results have shown that corruption has negative impacts on investment in human capital, physical capital, and political stability, which means that corruption indirectly impedes growth.
Huynh and Jacho-Chavez (2009) have used a nonparametric method to analyze the relationship between governance and growth. Their findings indicate that three of the six indicators of governance: voice and accountability, political stability, and rule of law are economically and statistically significant, while regulatory control, control of corruption, and government effectiveness are insignificant. The authors state that their empirical results support the findings of Glaeser, La Porta, de Silva, and Shleifer (2004) that poor countries get out of poverty and grow through good policies pursued by a dictator.
Using the studies by Knack, Stephen, Baliamoune-Lutz and Stefan Lutz, and alongside the study of Transparency International in Cameroon, Sikod and Teke (2012) established that there is a direct relationship between governance and economic performance, and that Cameroon has lagged behind in development in a major part because of bad governance which led them to give a policy implication that as governance indicators improve, the economic performance will also improve.
Another study was conducted by Cebula and Foley (2011) to test three hypotheses, one of which is about how quality government regulation affects per capita real GDP. By using panel data and PLS estimation for OECD countries over the period of 2003-06, the authors conclude that better regulatory quality is positively associated with economic growth because it has a positive effect on the way market functions, and it allows for the avoidance of unnecessary costs of managing businesses in the marketplace.
Ahmad et al (2012) used panel data over the period of 1984-2009 for 71 developed and developing countries to test whether corruption affects growth. Their study demonstrates that the relationship between corruption and long-run economic growth is hump-shaped. Their results also suggest that the quality of public institution has a crucial impact on any country’s growth performance. They conclude that there are many ways though which corruption can lessen economic growth, such as lowering domestic and foreign direct investment, and overblown government expenditure.
Another study was done by Aisen and Veiga (2013) to determine the impact of political instability on the growth. The authors used the system-GMM estimator for linear dynamic panel data models on a sample covering 169 countries for the period of 1960-2004. Their results have proved that political instability and lower GDP per capita are strongly associated. Political instability has negative effects on economic growth by reducing the rates of productivity growth, and lowering capital and human accumulation.
The study by Han et al (2014) determines whether countries with below-average governance grow slower than countries with above-average governance. Their results show that government effectiveness, political stability, control of corruption, and regulatory quality are more significantly positively correlated with economic growth than rule of law and voice and accountability. The results also indicate that the studied Asian countries’ above-average governance grow faster than those with below-average governance.
Emara and Jhonsa (2014) used the Two-Stage Least Square method for a cross-sectional dataset of 197 countries to investigate the interrelationship between the improvement in the quality of governance and the increase in per capita income. Their findings show that there is a strongly positive and statistically significant causation from the quality of governance to per capita income. The results also prove a positive causation in the opposite direction. The authors used their results to interpret the relationship between the studied variables for 22 MENA countries. They contend that one of their surprising results is that even though most of the studied MENA countries had low performance on all six indicators of governance, these MENA countries’ income per capita is relatively higher than the rest of the countries in the sample.
Wilson (2015) tested the casual relationship between quality of governance and economic growth in China at the provincial level and found out that under some certain circumstances, successful economic growth can be achieved without reliance on the improvements in formal governance institutions and that such economic growth can in turn support subsequent governance improvements.
2.4Implications of the Review for the Current Study
A major question that has been raised repeatedly with respect to the applicability of industrial policies that were so successful in East Asia to Africa economies is “governance.” Therefore, from the above literatures one can conclude that the effects of governance on economic growth might be positive or neutral. This study draws from the case study of Wilson (2015) which examined the bi-causal relationship between governance and growth in China and from a similar work conducted by Bota-Avram (2018) in studying some countries from the Sub-Saharan African countries. Both authors acknowledge the importance of the double feedback relationship between institutions and economic performance in their informal analyses. The motivation for this paper is initiated by the gap of applying this study to the region of West African countries in line with the important policy questions from the recent empirical evidence that was summarized above in which question such as: Does governance improve the economic growth of a country? Can appropriate policies make a contribution of governance more efficient? Why would some countries benefit more from a governance reform than other countries? These are challenging questions to answer because there are many interrelated factors that affect the long-term economic growth of a country. In spite of that, one empirical fact is that nations can improve economic growth by adopting appropriate policies. Therefore the goal of this study is to measure association and causality between governance and economic growth in the Economic Community of West African State countries homogenously and heterogeneously.
CHAPTER THREE
THEORETICAL FRAMEWORK AND METHODOLOGY
3.1Theoretical Framework
This study is based on the governance-growth hypothesis. Professionalization of the bureaucracy is an important factor that drives economic growth according to the study not leaving out institutional and policy perspective (Wilson, 2015). The hypothesis presented that long-term investment that in turn after growth can be promoted by a stable and trusted bureaucracy. Also, the effective enforcement by an impartial system of governance will bring about a conducive environment that can spur innovation and investment for economic growth. Also by growth-governance hypothesis, economic growth can spur increase in the quality of governance. When an economic starts at thriving and begins to growth, there is a high demand for more of property rights, rule of law and other necessary policy to sustain growth. Following the hypothesis, the model is specified using the Toda-Yamamoto’s approach following the VAR system:
Governancei,t = a1,i+ k-1ki?1,ikGovernancei, t-k+ k-1ki?1,i(k)(GDP)i,t-k+ ?1,i,t (3.1)
GDPi,t = a2,i+ k-1ki?2,ikGDPi, t-k+ k-1ki?2,i(k)(Governance)i,t-k+ ?2,i,t(3.2)
3.2Methodology
To test for the causality of quality of governance and economic growth in the countries of the region of ECOWAS, Granger Causality test will be used to test relations from both governance to economic growth and from economic growth to governance. Vector Autoregressive (VAR) model is used for this study and reason been that to test the potential for causal relationships to differ across countries in the ECOWAS region which may be as a result of difference in their private-state interaction and level of government participation at the onset of the country’s economy. Taking this heterogeneity into consideration, heterogeneous panel vector autoregressive (VAR) model is adopted to examine the focus of the study, governance and growth nexus, a case of ECOWAS.
3.2.1Model Specification
In achieving the objectives of this research by carrying a robust analysis on determining whether governance causes growth and vice-versa, this study adopts the model of Wilson and made necessary adjustment to the model. i = 1,….,N for countries and t = 1,….T for time to test for the cross country level heterogeneity.
Govi,t = a1,i+ k-1ki?1,ikGovi, t-k+ k-1ki?1,i(k)(GDP)i,t-k+ ?1,i,t (3.3)
GDPi,t = a2,i+ k-1ki?2,ikGDPi, t-k+ k-1ki?2,i(k)(Gov)i,t-k+ ?2,i,t (3.4)
Where ai are the country level effects, (GDP)i,t which represent economic growth and (Gov)i,t capturing quality of governance. Both GDP and Governance are stationary variables and ?i,t are normally distributed error terms with mean zero.
The heterogeneity of the cross-country is accounted for in the model by allowing the coefficients ?(k)and ?(k), and the lag length Ki, to vary across the countries. Finding a significant effect in the model when the two null hypotheses are tested and the coefficients ?1,i=(?1,i(1),..…,?1,iki) and ?2,i=(?2,i1,….,?2,1ki) are found to be zero for all country against the alternative will be considered as evidence for the presence of the corresponding casual relationship in at least one country in the sample.
3.2.2Research Design
The research design for the study is based on bi-variate, which is estimated by employing the VAR/ VECM model. This study made use of secondary data covering the period of 1996 to 2016. The Panel Unit Root test summary consisting of Levin-Lin-Chu (to check for the common unit-root among the variables), Im-Pesaran-Shin and Augmented Dickey-Fuller (ADF-Fisher) will be used to check that variables are stationary and will be used to determine the order of integration for each of the variables that has the alternative hypothesis that at least one cross-section unit is stationary. The Pedroni Residual Co-integration test will be used to check if there is long-run equilibrium relationship among the variable. For absence of co-integration, we run the VAR model. We will also use Wald coefficient test to check for the joint significance or short run effect. For the heterogeneous causality tests, block-bootstrapped p-values will be calculated to test for cross-sectional dependency and will also use heterogeneous panel Granger causality tests introduced by Dumitreseu and Hurlin (2012) to take account of the individual coefficient of each country in the region.
3.2.3Sources and Measurements of Data
The 15 countries under the region of ECOWAS shall be the focus for this study, using only secondary data which span across the period of twenty-one years i.e. 1996 to 2016. To capture economic growth this study uses GDP and in measuring quality of governance, data were sourced from The Worldwide Governance Indicators which is based on the information provided by various organizations worldwide. The quality of governance used will be the un-weighted average of the six governance indicators which are: Voice and accountability; Political stability and absence of violence; Government effectiveness; Regulatory quality; Rule of law; Control of corruption
Table 3.1Description of Variables and Data Sources
S/N VARIABLES MEASUREMENT(S) SOURCES
1 Governance Aggregated governance measure which consist of an average value of six governance indicators World Bank
2 Gross Domestic Product Growth, measured by GDP which is the market value of all goods and services produced within a country World Bank
Source: Author’s Computation
CHAPTER FOUR
PRESENTATION AND ANALYSIS OF RESULT
4.1Presentation of Results
Since it is never known aprior whether the selected variables for the study are integrated, co-integrated or stationary in its trend, the empirical result presented in this analysis starts with the test of the time-series in order to examine the stationarity problem and the possible presence of unit roots in series with Panel Unit-root test which will be followed by Pedroni Residual Co-integration Test.
4.1.1Panel Unit-root test for time series
Testing the time-series of GOV and GDP is the first step of the Toda-Yamamoto approach on Granger causality in order to investigate the presence of unit root in series and to determine their order of integration.
Table 4.1 Panel Unit-root Test
Variables Method At level At First Difference Comment
Statistics Prob. Statistics Prob. I(1)
GOV Levin, Lin ; Chu t* -1.13175 0.1289 -11.3861 0.0000 IPS W-Stat -0.24576 0.4029 -8.92861 0.0000 ADF – Fisher Chi-Square 29.2187 0.5061 129.803 0.0000 PP-Fisher Chi-Square 27.8819 0.5767 131.371 0.0000 GDP Levin, Lin ; Chu t* -1.57985 0.0571 -8.29686 0.0000 I(1)
IPS W-Stat 3.18975 0.9993 -10.1800 0.0000 ADF – Fisher Chi-Square 40.2207 0.1007 151.787 0.0000 PP-Fisher Chi-Square 28.4673 0.5457 182.115 0.0000 Source: Author (Using Eviews 9)
Table 4.1 gives the summary of the panel unit root test; the null hypothesis is that there is unit root which means the variables are not stationary. Rejecting the null hypothesis means validation of an alternative hypothesis that there is no unit root which is based on the significance level of 5%. In the GOV time-series, the Levin-Lin Chu t* statistics for common unit root p-value (0.1289) ; 0.05, we fail to reject the null hypothesis of the presence of unit root. Also the individual unit root test conducted by the Im-Pesaran-Shin, Augmented Dickey-Fuller (ADF-Fisher) and PP-Fisher statistics at the p-value of 0.4029, 0.5061 and 0.5767 respectively which is greater than the specified significance level of 5%, we fail to reject the null hypothesis. Since it is not stationary at first level, we apply first order difference to make it stationary at d=1. With the result from first difference, all the method of the test statistical p-value ; 0.05 and we therefore fail to accept the null hypothesis that there is presence of unit root
Also in the GDP time-series, the Levin-Lin Chu t* statistics for common unit root p-value (0.0571) ; 0.05, we fail to reject the null hypothesis of the presence of unit root. Also the individual unit root test conducted by the Im-Pesaran-Shin, Augmented Dickey-Fuller (ADF-Fisher) and PP-Fisher statistics at the p-value of 0.9993, 0.1007 and 0.5457 respectively which is greater than the specified significance level of 5%, we fail to reject the null hypothesis. Since the GDP is not stationary at first level, we apply first order difference to make it stationary at d=1. With the result from first difference, all the method of the test statistical p-value ; 0.05 and we therefore fail to accept the null hypothesis that there is presence of unit root.
4.1.2Panel Co-integration Test
With all the variables stationary of the same order, we conduct the Pedroni Residual Co-integration Test to ascertain if they have inner long-run relationship as this will determine the direction of whether to use VAR or VECM. One of the conditions for running the panel con-integration test is that the variables be stationary at the same level which the variables used for this study satisfied.
Table 4.2: Pedroni Residual Co-integration Test
Methods Common AR Coefficient
Statistic Prob. Weighted
Statistic Prob.
Panel v-Statistic -1.779134 0.9624 -2.211266 0.9865
Panel rho-Statistic 1.901582 0.9714 2.517670 0.9941
Panel PP-Statistic 1.122842 0.8692 2.525334 0.9942
Panel ADF-Statistic 2.560657 0.9948 2.900267 0.9981
Individual AR Coefficients
Group rho-Statistic 3.439916 0.9997 Group PP-Statistic 3.356592 0.9996 Group ADF-Statistic 4.557396 1.0000 Source: Author (Using Eviews 9)
The null hypothesis of Pedroni Residual Co-integration test is that there is no co-integration among the variables and the decision rule is that the probability value that conforms more to either the null or alternative hypothesis out the 11 probability values in the test. From Table 4.2, the result shows that co-integration does not exist. The p-value ; 0.05 which is not significant, we will fail to reject the null hypothesis that there is no co-integration among GOV and GDP which means that there is no long-run relationship between GOV and GDP of the 15 ECOWAS countries between 1996 – 2016; therefore, we will be using VAR for the system of analysis.
Due to the interest in the Granger non-causality tests, we will first establish the number of correct lags because the number of lags has a significant influence on the results of the Granger non-causality test. Using the lag order selection criteria, the LR test statistic (LR), Final Prediction Error (FPE), Akaike Information Criterion (AIC) and Hannan-Quinn Information Criterion (HQ) techniques reveals a maximum lag length of 2 for each of the variables.
4.2Discussion of Results
Table 4.3 VAR with GOV as the dependent variable
Variable Coefficient Std. Error t-Statistic Prob.
GDP(-1) -0.057485 0.072589 -0.791929 0.4291
GDP(-2) 0.054997 0.072620 0.757332 0.4495
GOV(-1) 1.196444 0.058109 20.58953 0.0000
GOV(-2) -0.22757 0.057019 -3991219 0.0001
C 0.041829 0.073417 0.569744 0.5693
R-Squared 0.973088 F-statistic 2531.055
Adj. R-Squared 0.972703 Prob(F-statistic) 0.000000
Durbin-Watson stat 2.1008888 Source: Author’s Compilation with eviews 9
The result presented in table 4.3 shows the vector autoregression model with GOV as the dependent variable. The result shows that the lag of GDP has both negative and positive impact on GOV but however, their impact is not significant on governance and this conforms with recent research conducted (Bota-Avram, et al, 2018). Also in the result, the quality of governance has both positive and negative impact on itself. In the first period, 1% increase in the quality of governance will bring about 119.6% increase in GOV while in the second period, 1% increase in GOV will bring about 22.7% decrease in GOV. The R-square suggests that the model is very good as about 97% variation in the dependent variable is explained by the regressors in the model.
Table 4.4Wald Coefficient Test for GDP on GOV
Test Statistic Value Df Probability
Chi-square 1.188988 2 0.5518
Source: Author’s Compilation with eviews 9
The Wald Test for joint significance of both lag of the GDP to see if the lags combined together can cause the GOV, the null hypothesis is that the both lag of GDP cannot cause improvement in the quality of governance. From Table 4.4, the p-value ; 0.05, we will therefore accept the null hypothesis which means that GDP in the first and second period can jointly bring a significant effect on the GOV.
Table 4.5 VAR with GDP as the dependent variable
Variable Coefficient Std. Error t-Statistic Prob.
GDP(-1) 1.202070 0.045663 26.32466 0.0000
GDP(-2) -0.199530 0.045683 -4.367748 0.0000
GOV(-1) 0.116578 0.036555 3.189140 0.0016
GOV(-2) -0.106533 0.035869 -2.970106 0.0032
C -0.017138 0.046184 -0.371078 0.7109
R-Squared 0.998826 F-statistic 59567.90
Adj. R-Squared 0.998809 Prob(F-statistic) 0.000000
Durbin-Watson Stat 2.073161 Source: Author’s Compilation with eviews 9
After analyzing the result that gives table 4.5, and it shows the vector autoregression model with GDP as the dependent variable. The result shows that the lag of GDP has a positive impact on the current year GDP. In the first period, one percent increase in the GDP will bring about 120% increases in the following year while holding constant other explanatory variables. Furthermore, the second lag of GDP shows a negative relationship and significant. Also in the result, the quality of governance has both positive and negative impact on the GDP. In the first period, 1% increase in the quality of governance will bring about 11.6% increase in GDP while in the second period, 1% increase in GOV will bring about 1.7% decrease in GDP. The R-square suggests that the model is very good as about 99% variation in the dependent variable is explained by the regressors in the model.
Table 4.6Wald Coefficient Test for GOV on GDP
Test Statistic Value Df Probability
Chi-square 11.27789 2 0.0036
Source: Author’s Compilation with eviews 9
The Wald Test is to test for joint significance of both lag of the governance to see the lags combined together can cause the GDP, the null hypothesis is that the both lag of governance cannot cause growth. From Table 4.6, the p-value ; 0.05, we will therefore reject the null hypothesis which means that governance in the first and second period can jointly bring a significant effect on the GDP
Table 4.7: Granger Causality/Block Exogeneity Wald Tests
Dependent variable: GDP
Excluded Chi-Square Df Prob.
GOV 11.27789 2 0.0036
Dependent Variable: GOV
LGDP 1.188988 2 0.5518
Source: Author’s Compilation with eviews 9
The Table 4.7 shows the result of the VAR Granger Causality/Block Exogeneity Wald tests and it reveals the a tangible evidence of Granger causality runs from country-level to GDP with the p-value < 0.05 which is significant but the causality from GDP to quality of governance is not confirmed since the p-value is > 0.05. From this, it is established that it is a unidirectional casual relationship that exist and it is from the quality of governance to growth.
Table 4.8 Heterogeneous Panel Causality Test
Null Hypothesis W-Stat. Prob.
GOV does not homogenously cause GDP 5.06080 0.0001
GDP does not homogenously cause GOV 3.15239 0.2506
Source: Author’s Compilation with eviews 9
Allowing for difference in the coefficient of each nation in the region Table 4.8 shows the result of heterogeneous panel granger causality for each cross-section unit independently and the result still tally along with the result of the VAR Granger Causality/Block Exogeneity Wald tests. It shows evidence of Granger causality from country-level to GDP for p-value ; 0.05 but the causality from GDP to quality of governance is not confirmed since the p-value is ; 0.05. Unidirectional casual relationship is what still exist and it is from the quality of governance to growth.
This result is mainly as a result of nations in the ECOWAS such as Cape Verde that has a high level of quality of governance yet they have a very low GDP. While some countries such as Cote d’Ivoire and Nigeria with low level of quality of governance and yet maintain a high GDP, accounting for over 70% of the GDP of the Region and being referred to as the economic powerhouse of West Africa (ECOWAS Annual Report, 2016). So there is causality when we homogeneously test between the quality of governance and growth because of this happening highlighted as regarding countries with low quality of governance and high growth that contradicts economic theory or it could be a pointer that the yardstick in capturing governance as given by Kaufmann et al (2005) cannot be fully relied upon to explain growth for the ECOWAS region community.
4.3Comparison of Results with Previous Findings
The result of this study conforms partly to economic literature and it has confirmed some of the results previous studies such as Bota-Avram et al (2018), Calderoân and Chong (2000), Chauvet and Collier (2004) which provide evidence of Granger causality from quality of country level governance to economic growth and so also the causality from economic growth to country-level governance is not confirmed. The result of this study, same as their own result follows the empirical findings of the World Bank researcher that concludes there is no significant influence of economic growth on the quality of governance establishing unidirectional causality because in the opposite direction there is a strong positive causal effect from governance to economic growth. It is noted also that the result of this study is at variance with Sachs et al. (2004) which find out an important challenge to the significance of good governance for the economic growth of African countries. He further asserts that the differences in performance among African countries cannot be explained by differences in the quality of their governance once differences in their levels of development have been accounted for and thus concludes that a focus on governance reforms is misguided. Wilson (2015) which uses heterogeneous Granger causality introduced by Dumitreseu and Hurlin (2012) allowing for potential differences in the causal relations across countries showing a significant and positive effect of quality of governance on economic growth and also confirms a bi-directional causality between the two variables which is at a variance with this study.
APPENDIX A
Data on Quality of Governance and Economic Growth
APPENDIX B
Panel Unit Root Test for GOV
Panel unit root test: Summary Series: GOV Date: 08/11/18 Time: 16:00 Sample: 1996 2016 Exogenous variables: Individual effects
Automatic selection of maximum lags Automatic lag length selection based on SIC: 0 to 1
Newey-West automatic bandwidth selection and Bartlett kernel
Cross- Method Statistic Prob.** sections Obs
Null: Unit root (assumes common unit root process)
Levin, Lin & Chu t* -1.13175 0.1289 15 297
Null: Unit root (assumes individual unit root process)
Im, Pesaran and Shin W-stat -0.24576 0.4029 15 297
ADF – Fisher Chi-square 29.2187 0.5061 15 297
PP – Fisher Chi-square 27.8819 0.5767 15 300
** Probabilities for Fisher tests are computed using an asymptotic Chi
-square distribution. All other tests assume asymptotic normality.
Panel unit root test: Summary Series: D(GOV) Date: 08/11/18 Time: 16:07 Sample: 1996 2016 Exogenous variables: Individual effects
Automatic selection of maximum lags Automatic lag length selection based on SIC: 0 to 1
Newey-West automatic bandwidth selection and Bartlett kernel
Cross- Method Statistic Prob.** sections Obs
Null: Unit root (assumes common unit root process)
Levin, Lin & Chu t* -11.3861 0.0000 15 283
Null: Unit root (assumes individual unit root process)
Im, Pesaran and Shin W-stat -8.92861 0.0000 15 283
ADF – Fisher Chi-square 129.803 0.0000 15 283
PP – Fisher Chi-square 131.371 0.0000 15 285
** Probabilities for Fisher tests are computed using an asymptotic Chi
-square distribution. All other tests assume asymptotic normality.
Panel Unit Root test for LGDP
Panel unit root test: Summary Series: LGDP Date: 08/11/18 Time: 16:10 Sample: 1996 2016 Exogenous variables: Individual effects
Automatic selection of maximum lags Automatic lag length selection based on SIC: 0 to 4
Newey-West automatic bandwidth selection and Bartlett kernel
Cross- Method Statistic Prob.** sections Obs
Null: Unit root (assumes common unit root process)
Levin, Lin & Chu t* -1.57985 0.0571 15 289
Null: Unit root (assumes individual unit root process)
Im, Pesaran and Shin W-stat 3.18975 0.9993 15 289
ADF – Fisher Chi-square 40.2207 0.1007 15 289
PP – Fisher Chi-square 28.4673 0.5457 15 300
** Probabilities for Fisher tests are computed using an asymptotic Chi
-square distribution. All other tests assume asymptotic normality.
Panel unit root test: Summary Series: D(LGDP) Date: 08/11/18 Time: 16:14 Sample: 1996 2016 Exogenous variables: Individual effects
Automatic selection of maximum lags Automatic lag length selection based on SIC: 0 to 4
Newey-West automatic bandwidth selection and Bartlett kernel
Cross- Method Statistic Prob.** sections Obs
Null: Unit root (assumes common unit root process)
Levin, Lin & Chu t* -8.29686 0.0000 15 277
Null: Unit root (assumes individual unit root process)
Im, Pesaran and Shin W-stat -10.1800 0.0000 15 277
ADF – Fisher Chi-square 151.787 0.0000 15 277
PP – Fisher Chi-square 182.115 0.0000 15 285
** Probabilities for Fisher tests are computed using an asymptotic Chi
-square distribution. All other tests assume asymptotic normality.
Co-integration test
Pedroni Residual Cointegration Test Series: LGDP GOV Date: 08/13/18 Time: 14:19 Sample: 1996 2016 Included observations: 315 Cross-sections included: 15 Null Hypothesis: No cointegration Trend assumption: No deterministic trend Automatic lag length selection based on SIC with a max lag of 4
Newey-West automatic bandwidth selection and Bartlett kernel
Alternative hypothesis: common AR coefs. (within-dimension)
Weighted Statistic Prob. Statistic Prob.
Panel v-Statistic -1.779134 0.9624 -2.211266 0.9865
Panel rho-Statistic 1.901582 0.9714 2.517670 0.9941
Panel PP-Statistic 1.122842 0.8692 2.525334 0.9942
Panel ADF-Statistic 2.560657 0.9948 2.900267 0.9981
Alternative hypothesis: individual AR coefs. (between-dimension)
Statistic Prob. Group rho-Statistic 3.439916 0.9997 Group PP-Statistic 3.356592 0.9996 Group ADF-Statistic 4.557396 1.0000 Cross section specific results Phillips-Peron results (non-parametric) Cross ID AR(1) Variance HAC Bandwidth Obs
1 0.671 0.009299 0.010115 1.00 20
2 0.939 0.004551 0.012744 3.00 20
3 0.910 0.012537 0.016111 1.00 20
4 1.051 0.001927 0.003776 2.00 20
5 0.520 0.004243 0.004243 0.00 20
6 0.904 0.025090 0.022147 2.00 20
7 1.007 0.001995 0.005780 3.00 20
8 0.827 0.009170 0.007400 3.00 20
9 0.508 0.021728 0.052674 3.00 20
10 0.806 0.008084 0.009150 1.00 20
11 1.025 0.003905 0.006522 2.00 20
12 0.972 0.007498 0.015807 2.00 20
13 0.989 0.003199 0.008497 3.00 20
14 0.695 0.014132 0.014600 1.00 20
15 0.998 0.002069 0.004379 3.00 20
Augmented Dickey-Fuller results (parametric) Cross ID AR(1) Variance Lag Max lag Obs
1 0.671 0.009299 0 4 20
2 1.007 0.001965 2 4 18
3 0.910 0.012537 0 4 20
4 1.051 0.001927 0 4 20
5 0.520 0.004243 0 4 20
6 0.904 0.025090 0 4 20
7 1.028 0.000515 3 4 17
8 0.827 0.009170 0 4 20
9 0.632 0.009363 2 4 18
10 0.806 0.008084 0 4 20
11 0.901 0.001106 4 4 16
12 0.967 0.004645 1 4 19
13 0.968 0.001902 1 4 19
14 0.695 0.014132 0 4 20
15 1.071 0.000787 1 4 19
Test for selecting the number of lags
VAR Lag Order Selection Criteria Endogenous variables: GDP GOV Exogenous variables: C Date: 08/11/18 Time: 18:24 Sample: 1996 2016 Included observations: 195 Lag LogL LR FPE AIC SC HQ
0 -5325.270 NA 1.84e+21 54.63866 54.67223 54.65226
1 -4417.648 1787.317 1.74e+17 45.37075 45.47146* 45.41152
2 -4410.158 14.59628* 1.67e+17* 45.33495* 45.50280 45.40291*
3 -4409.827 0.638802 1.74e+17 45.37258 45.60756 45.46772
4 -4407.181 5.047105 1.76e+17 45.38647 45.68859 45.50880
5 -4404.985 4.143395 1.80e+17 45.40498 45.77424 45.55449
6 -4404.153 1.553970 1.86e+17 45.43746 45.87386 45.61416
7 -4402.627 2.817190 1.90e+17 45.46284 45.96638 45.66672
8 -4401.632 1.815859 1.96e+17 45.49366 46.06434 45.72472
* indicates lag order selected by the criterion LR: sequential modified LR test statistic (each test at 5% level) FPE: Final prediction error AIC: Akaike information criterion SC: Schwarz information criterion HQ: Hannan-Quinn information criterion Appendix C
VAR Result
Vector Autoregression Estimates
Date: 08/13/18 Time: 15:16
Sample (adjusted): 1998 2016
Included observations: 285 after adjustments
Standard errors in ( ) & t-statistics in
LGDP GOV
LGDP(-1) 1.202070 -0.057485
(0.04566) (0.07259)
26.3247 -0.79193
LGDP(-2) -0.199530 0.054997
(0.04568) (0.07262)
-4.36775 0.75733
GOV(-1) 0.116578 1.196444
(0.03655) (0.05811)
3.18914 20.5895
GOV(-2) -0.106533 -0.227574
(0.03587) (0.05702)
-2.97011 -3.99122
C -0.017138 0.041829
(0.04618) (0.07342)
-0.37108 0.56974
R-squared 0.998826 0.973088
Adj. R-squared 0.998809 0.972703
Sum sq. resids 0.785713 1.985497
S.E. equation 0.052973 0.084208
F-statistic 59567.90 2531.055
Log likelihood 435.4481 303.3459
Akaike AIC -3.020689 -2.093655
Schwarz SC -2.956610 -2.029577
Mean dependent 22.46003 -0.598788
S.D. dependent 1.535270 0.509684
Determinant resid covariance (dof adj.) 1.97E-05
Determinant resid covariance 1.90E-05
Log likelihood 740.0036
Akaike information criterion -5.122832
Schwarz criterion -4.994675
Dependent Variable: LGDP Method: Panel Least Squares Date: 08/13/18 Time: 15:18 Sample (adjusted): 1998 2016 Periods included: 19 Cross-sections included: 15 Total panel (balanced) observations: 285 LGDP = C(1)*LGDP(-1) + C(2)*LGDP(-2) + C(3)*GOV(-1) + C(4)*GOV(-2) +
C(5) Coefficient Std. Error t-Statistic Prob.
C(1) 1.202070 0.045663 26.32466 0.0000
C(2) -0.199530 0.045683 -4.367748 0.0000
C(3) 0.116578 0.036555 3.189140 0.0016
C(4) -0.106533 0.035869 -2.970106 0.0032
C(5) -0.017138 0.046184 -0.371078 0.7109
R-squared 0.998826 Mean dependent var 22.46003
Adjusted R-squared 0.998809 S.D. dependent var 1.535270
S.E. of regression 0.052973 Akaike info criterion -3.020689
Sum squared resid 0.785713 Schwarz criterion -2.956610
Log likelihood 435.4481 Hannan-Quinn criter. -2.995001
F-statistic 59567.90 Durbin-Watson stat 2.073161
Prob(F-statistic) 0.000000 Wald Test: System: sys01 Test Statistic Value df Probability
Chi-square 11.27789 2 0.0036
Null Hypothesis: C(3)=C(4)=0 Null Hypothesis Summary: Normalized Restriction (= 0) Value Std. Err.
C(3) 0.116578 0.036555
C(4) -0.106533 0.035869
Restrictions are linear in coefficients.
Dependent Variable: GOV Method: Panel Least Squares Date: 08/13/18 Time: 15:20 Sample (adjusted): 1998 2016 Periods included: 19 Cross-sections included: 15 Total panel (balanced) observations: 285 GOV = C(6)*LGDP(-1) + C(7)*LGDP(-2) + C(8)*GOV(-1) + C(9)*GOV(-2) +
C(10) Coefficient Std. Error t-Statistic Prob.
C(6) -0.057485 0.072589 -0.791929 0.4291
C(7) 0.054997 0.072620 0.757332 0.4495
C(8) 1.196444 0.058109 20.58953 0.0000
C(9) -0.227574 0.057019 -3.991219 0.0001
C(10) 0.041829 0.073417 0.569744 0.5693
R-squared 0.973088 Mean dependent var -0.598788
Adjusted R-squared 0.972703 S.D. dependent var 0.509684
S.E. of regression 0.084208 Akaike info criterion -2.093655
Sum squared resid 1.985497 Schwarz criterion -2.029577
Log likelihood 303.3459 Hannan-Quinn criter. -2.067968
F-statistic 2531.055 Durbin-Watson stat 2.100888
Prob(F-statistic) 0.000000
Wald Test: System: sys01 Test Statistic Value df Probability
Chi-square 1.188988 2 0.5518
Null Hypothesis: C(6)=C(7)=0 Null Hypothesis Summary: Normalized Restriction (= 0) Value Std. Err.
C(6) -0.057485 0.072589
C(7) 0.054997 0.072620
Restrictions are linear in coefficients.
Causality Test
VAR Granger Causality/Block Exogeneity Wald Tests
Date: 08/13/18 Time: 16:53 Sample: 1996 2016 Included observations: 285 Dependent variable: LGDP Excluded Chi-sq df Prob.
GOV 11.27789 2 0.0036
All 11.27789 2 0.0036
Dependent variable: GOV Excluded Chi-sq df Prob.
LGDP 1.188988 2 0.5518
All 1.188988 2 0.5518
Pairwise Dumitrescu Hurlin Panel Causality Tests
Date: 08/14/18 Time: 05:53
Sample: 1996 2016 Lags: 2 Null Hypothesis: W-Stat. Zbar-Stat. Prob.
GOV does not homogeneously cause LGDP 5.06080 3.82617 0.0001
LGDP does not homogeneously cause GOV 3.15239 1.14900 0.2506
