Political institutions, governance, and consumption expenditure in developing countries: a panel data analysis.
Since the end of the Second World War in 1945, governments have become more influential, even in capitalist countries, as they provide social services and income supplements as well as produce foods, manage the economy, and invest in capital (Cameron 1978). In recent times, external shocks and internal structural failures have had an unavoidable effect on economies. Government intervention in such cases can play an important and effective role in adjusting adverse economic conditions. However, in choosing between monetary and fiscal policy, governments prefer the latter to ensure a confirmed quick recovery from the "recessionary condition" and to reestablish the confidence of the private sector (Kandil 2005, 269). In developing countries, in particular, fiscal policy has taken the lead in implementing and designing macroeconomic policies, while monetary policy has been subdued due to the need to stabilize a pegged exchange rate, or to accommodate budget deficits.
Distinction, however, has been made between "productive" and "unproductive" utilization and the composition of public expenditure. It is generally agreed that large government expenditure increases economic growth, which further improves social welfare and is found to be somewhat effective in reducing poverty (e.g., see Ram 1986). Such expenditure has a substantial impact on the aggregate productivity of an economy as demonstrated in the influential article by Aschauer (1989) in which he established a significant relationship between the aggregate productivity and the stock and flow of different government spending variables. He argued that nonmilitary public capital is more important for productivity and concluded that infrastructure spending (such as on streets, highways, mass transit, and sewerage) has the most significant association with productivity. However, Devarajan, Swaroop, and Zou (1996) in their seminal article, employing theoretical setup with empirical support, derived the conditions under which large government expenditure leads to economic growth. They showed that the current expenditure has positive and sizable growth effects whereas the capital expenditure when used excessively could become unproductive. The conclusions of these two articles were particularly important for developing countries in which right and adequate allocation of public expenditure symbolizes the "wheels"--if not the engine--of economic activity (World Bank 1994).
Addressing a taxonomic account of the size of government expenditure from the demand perspective, one needs to pay important attention to the role of political influence and institution in the distribution and size of government expenditure. The theoretical setup underpinning this demandside approach of government expenditure under the influence of political economic framework has been demonstrated in the literature by Persson and Tabellini (1999); Pesaran (2004), and Milesi-Ferretti, Perotti, and Rostagno (2002), to name a few. Similarly employing a political economic framework, Tridimas and Winer (2005) demonstrated that the demand for government initiates from voters with heterogeneous income and political influence, where voters demand public goods and redistribution into their favor. Under this setup, several testable hypotheses emerged which need further exploration to verify those comparative static properties with rigorous analytic framework, which has been initiated in the earlier works by Mueller and Stratmann (2003) and Sturm (2001).
The aim of this study is to identify how the broad specifications with economical, political, and institutional family of variables influence government expenditure of developing countries. In particular, this study examines the influence of political, governance, and institution setup on the government consumption expenditure (GCE) by employing indicators like electoral politics of a nation (e.g., the partisan composition and political structure of government), military dictatorship, the institutional structure, and the current state of governance. To facilitate this empirical investigation, categorized variables have been used, namely: (a) demographical variables, (b) fractionalization variables, (c) political variables, and (d) governance and institutional variables.
The main contribution of this article is twofold. First, the article employs a richer set of specifications while giving emphasis on the role of politics, governance, and institution on the government expenditure, which is a noteworthy contribution. Second, to show the robustness of the estimation, extreme bound analysis (EBA) has been conducted, which is a well-known practice in the growth literature but has not yet been applied in this literature.
Using panel dataset for 97 developing countries from 1984 to 2004, this study finds evidence that political and institutional variables as well as governance variables significantly influence the government expenditure of developing countries. Political institutional variables such as political ruling and political power in the parliament, and governance variables such as corruption and government effectiveness are found to have significant statistical association with government expenditure. Moreover, the study finds robust evidence that public expenditure significantly shrinks under military dictatorship during autocratic regime compared with other forms of governance.
We begin Section II with a selective review of the literature on the determinants of government expenditure. Section III talks about the leading hypothesis of government expenditure and the data used in the article. The methodology employed in the analysis has been described at length in Section IV. Section V provides the detailed explanation of the findings of the estimation. Sensitivity test of our estimation using EBA is discussed in Section VI and Section VII concludes.
The existing literature on this issue can be classified into two streams. One significant stream of literature focused on the economic factors that influence the government expenditure. Earlier works on this stream of literature were carried out by De Haan, Sturm, and Sikken (1996), Sturm (2001), Fan and Rao (2003), and Boix (2004). A recent influential contribution to this literature is that of Shelton (2007) who tested several leading hypotheses of government expenditure using data from the Global Financial Statistics of the International Monetary Fund (IMF) and various other sources. He tested separate sectors of government expenditure and different levels of government, and concluded that '"preference heterogeneity leads to decentralization rather than outright decreases in expenditure" (Shelton 2007, 2,230).
Another stream of literature looked at the response of governments to gross domestic product (GDP) variation over business cycles effects. By and large, the literature argues that there is a profound divergence between how fiscal policy is conducted in developing countries when comparing that to industrial countries. While fiscal policy in industrial countries is either acyclical or counter-cyclical, interestingly, fiscal policy in developing countries is mostly procyclical (Gavin and Perotti 1997; Ilzetzki and Vegh 2008; Talvi and Vegh 2005). The reason behind such finding is not surprising because developing countries suffer from imperfections in international credit markets that prevent these countries from borrowing in bad times (Caballero and Krishnamurthy 2004; Gavin and Perotti 1997; Riascos and Vegh 2003). Moreover explanations through the lens of political economy would also suggest that good times would likely encourage fiscal prodigality which may also increase the rent-seeking activities in developing countries, thus increase the fiscal expenditure (Alesina, Campante, and Tabellini 2008; Ilzetzki and Vegh 2008; Talvi and Vegh 2005; Tornell and Lane 1998, 1999).
III. DATA AND HYPOTHESIS
Numerous hypotheses have been proposed in various literatures on the determinants of government expenditure, which are mainly economic and demographical variables in nature. However, the influence of political, institutional, and governance on government expenditure has been rarely studied and only a handful of studies looked into this aspect of the determinants of government expenditure. To test the political economic influence on the government expenditure, we relied on the hypothesis derived by Tridimas and Winer (2005) where one important comparative statics property has shed light on how political influence could shape the size of the government. This particular property tells us that the more unequal the political distribution in relation to the distribution of income, the larger the public sector. To test this hypothesis, we used polity index to capture the political sphere for any country.
Political institutions play a pivotal role in deciding the shape and size of the government, and it is necessary to understand the determinants of government expenditure through the lens of political institutions. However, inadequate data on the political institutions of countries, especially for developing countries, have made the cross-country empirical work handicapped. In our article, we wanted to test whether different political regimes play a determinative role in explaining the cross-country variation in government expenditure. The existing literature on this mainly focuses on the public good provision under different forms of political system. Leading this literature, Persson and Tabellini (1999) and Persson, Roland, and Tabellini (2007) demonstrated that government spending significantly varies under different political systems and electoral rules (e.g., proportional vs. majoritarian political system or single party government vs. coalition). Similar work has been conducted by Milesi-Ferretti, Perotti, and Rostagno (2002) where they evaluated various categories of government spending under different electoral systems employing theoretical setup with empirics. Using a panel dataset, Mueller and Stratmann (2003) showed that more political participation leads to larger public sector in countries with democratic institutes. McGuire and Olson (1996) theoretically proved that democratic governments do more redistribution than autocratic governments because the latter maximize the welfare of an elite subset rather than the whole population. Niskanen (1997) showed that democratic governments produce substantially higher outcomes, income, and transfer payments by maximizing the welfare of the median income voter. Using a similar setup, Tridimas and Winer (2005) demonstrated that more unequal the political distribution in relation to the distribution of income, the larger is the public sector. Lake and Baum (2001) and Bueno de Mesquita et al. (2003) demonstrated empirical evidence in support of lower public good provision (in the case of public health and education) under a dictatorship.
There is a growing consensus among scholars, policymakers, and donors that good governance is one of the keys to achieve sustainable economic development. Evidence in the literature suggests that good governance contributes significantly to economic development (North 1981; Shleifer and Vishny 1993) as well as to economic growth (Easterly and Levine 1997; Mauro 1995); hence this appears to be a well-established economic proposition. However, empirical measures of governance are very difficult because such measures have to be comparable across countries and free from measurement errors. Only a handful of governance measurement indices are available in the literature and we selected variables from the Worldwide Governance Indicators (WGI) dataset of the World Bank for its wide coverage and comparability features across countries. (1) However, the direction of causality could be an issue, as one might wonder whether the variation of government expenditure drives the quality of governance, or the existing quality of governance affects the public expenditure. The direction of causality is perhaps more plausible from governance to government expenditure, that is, it seems reasonable to argue that the existing level of governance influences government expenditure rather than the current level of government expenditure being responsible for the quality of governance (Acemoglu, Johnson, and Robinson 2001; Mauro 1995).
Similarly, corruption, which is another very important indicator of the quality of governance, is a persistent feature of countries over time and space (Aidt 2003). Corruption is pervasive, consistent, and significant around the world, even in developed countries. Although some studies have concluded that some level of corruption might be desirable (Leff 1964), most studies suggest that corruption is quite harmful for the development process of any economy (Gould and AmaroReyes 1983; Klitgaard 1988), and is a particularly crucial issue for poor countries. Pioneering a systemic empirical analysis on corruption and composition of government expenditure, Mauro (1998) found that corrupt countries have been more frequently associated with low spending on public education and health. Corruption within countries has also been linked to poor quality roads and inadequate electric distribution (Tanzi and Davoodi 1998) and poor environmental protection outcomes (Welsch 2004).
To examine the role of institution in the government expenditure, we have used a series of variables. For instance, to test the influence of electoral rule as mentioned by Milesi-Ferretti, Perotti, and Rostagno (2002) and Persson, Roland, and Tabellini (2000), we used democracy variable. Similarly to test the impact of the structure of executive-legislative relationship characterized by the power in the parliament as mentioned by Persson and Tabellini (1999) and Persson, Roland, and Tabellini (2000), we used size of the ruling party in the parliament (number of government seats). Moreover, to test the common pool problem which arises when the legislators provide benefit to their allies at the expense of general taxpayers, as mentioned by Baqir (2002) and Pesaran (2004), we used autocracy variable to capture these impacts.
To summarize, we categorized four families of variables, other than economic ones, following the work of Sturm (2001), which are given below.
Base Variables. Real GDP per capita (e.g., see Gupta 1968; Musgrave ; Bird 1971; Cameron 1978), openness (in % of GDP) (e.g., see Cameron 1978; Rodrik 1998; Schmidt 1983; Saunders and Klau 1985), and log of population (e.g., see Alesina and Spolaore 1997; Alesina and Perotti 1997).
Demographical Variables. Elderly population, ages 65 and above (percentage of total population) (Remmer 2004; Sanz and Velazquez 2002), young population, ages 15 and below (percentage of total population), and urban population (percentage of total).
Fractionalization Variables. Ethnic fractionalization and linguistic fractionalization (Shleifer and Vishny 1993; Persson, Roland, and Tabellini 1997; Easterly and Levine 1997, to name a few).
Political-Institutional Variables. Number of government seats (by the ruling party) in the parliament, military officer (1 if the chief executive is a military officer), polity index (a standard measure of governance on a 21-point scale ranging from -10 [dictatorial] to +10 [consolidated democracy]) (Tridimas and Winer 2005).
Governance Variables. Control of corruption (varies from -2.5 to +2.5), corruption perception index (CPI) (varies from 0 to 145), and government effectiveness (varies from -2.5 to +2.5) (Kirchgassner 2002).
Pairwise correlation, summary statistics, and detailed description of these variables are available in Tables Al, A7, and A8, respectively in Appendix A.
B. Data Issues
All the base variables and demographical variables have been taken from the World Bank Development Indicators CD-ROM 2008 (WDI 2008) published by the World Bank. The fractionalization data have been obtained from the study of Alesina et al. (2003). In this study, new measures of ethnic, linguistic, and religious fractionalization for some 190 countries have been constructed. The set of political variables has been taken from the database of political institutions (DPI 2004) provided by the Development Research Group of the World Bank. This dataset is constructed by Beck (2001) and the index created in this series has been described in their appendix. In the case of institutional variables, the data were not available for the periods 1997, 1999, and 2001. For this set of variables, we have constructed the values of the missing years by using the means of the corresponding forward and backward years. Except for the CPI data, the institutional variables have been taken from the WGI project by the World Bank. This dataset is constructed by Kaufmann, Kraay, and Mastruzzi (2005), which was only available for the 1996-2005 period. The CPI has been taken from Transparency International's website. The "polity score" is a standard measure of governance on a 21-point scale ranging from -10 (dictatorial) to +10 (consolidated democracy).
Various literatures of panel data analysis have used random effect models rather than country-specific fixed effect models (e.g., see Shelton 2007). The reason for such practice is due to the problem of "unit centering" methodology of the fixed effect model (Beck 2008, 8). Under fixed effect models, because all cross-sectional effects have been eliminated, it is almost impossible to estimate the impact of any variables that do not change or seldom change (e.g., political regimes or institutional variables). (2) Also, when the error component of panel data has group-wise heteroskedasticity then fixed effect models cannot be efficiently estimated with ordinary least squares (OLS) (Yaffee 2003). Hence, in empirical and applied research, well-specified models do not require to use fixed effect models. As stated by Beck (2008, 8): "Ideally one would like to explain the effects by substantive variables, and not simply conclude that Germany grew faster because it was Germany." The advantage to random effect models is the ability to estimate time invariant variables as well as we can save a number of degrees of freedom, which will assist us in obtaining more efficient estimates of the regression parameters.
Initially, we tested for the pool-ability estimation and the result suggested that simple pooled OLS would be inappropriate. We then tested with the basic specification with base variables for the Hausman specification test (Hausman 1978), which could not reject the null hypothesis that a random effect model is consistent (p value .12); hence we could use the random effect model, which is also consistent with the recent work of Shelton (2007) and Freytag et al. (2011).
To test the hypothesis of cross-sectional independence in panel data models with small T and large N, we used semi-parametric tests proposed by Friedman (1937) and Frees (1995, 2004) as well as the parametric testing procedure proposed by Pesaran (2004). (3) In our study, we found evidence of contemporaneous correlation across the units using the above-mentioned tests. We also tested the group-wise heteroskedasticity and autocorrelation in the panel data with the help of a modified Wald test and Wooldridge test respectively. Both the tests showed evidence of heteroskedasticity and autocorrelation in the dataset. As mentioned by Baltagi (2005, 84) assuming homoskedasticity disturbances and ignoring serial correlation when heteroskedasticity and serial correlation are present will result in consistent but inefficient estimations and standard errors could be wrong. As a result, models needed to be corrected for such patterns of the error terms to obtain consistent and efficient estimates of the regressors.
Two standard methods used by the researchers to correct such problems in the data are the feasible generalized least square (FGLS) method and the Prais-Winsten transformation procedure. Both estimates will produce consistent estimates as long as the conditional mean is correctly specified. In our study, we chose to use the FGLS procedure for its power to produce estimates with time invariant variables. Following the suggestion of Beck and Katz (1995), we used panel-specific AR1 parameters in all regressions.
It is argued that current political, social, and economical institutions for many countries are largely determined by their past history, geography, religion, and climate (Acemoglu, Johnson, and Robinson 2001; Leonardi, Nanetti, and Putnam 2001). To capture such time-independent constant effect, we used continental dummies. All the regression estimations have year-specific dummies, which have accommodated the year-specific variation and business cycles in the model. We tested for panel unit root process in the dependent variable for both common and individual unit root processes, and five out of six tests rejected the null of having a unit root process in the dependent variable.
One could have reasonable doubts about the economic variables used in the regressions, mainly the per capita income and openness variable, which could be potentially endogenous and reverse causality could be an issue. One ideal way to tackle such issues is by employing instrumental variable techniques. However, finding right instruments to proxy such potentially endogenous variables is very difficult and we rarely have good instruments for such variables. One approach taken by this article is following the work of empirical economists like Hall (1988) and Yogo (2004) who used the past (lagged) economic variables as proxy for the contemporaneous level of economic variables. (4) Such technique controls for any two-way causality between dependent and independent variables and endogeneiety. The argument for the validity of such a technique lies in the theory of rational expectation and efficient market hypothesis; the current change in economic variables is uncorrelated with its all-lagged outcomes. As described by Hall (1988, 340): "Actual movements of consumption differ from planned movements by a completely unpredictable random variable that indexes all the information available next year that was not incorporated in the planning process the year before." Hence, one could moderately argue that the lagged variables are not systematically correlated with the unexpected current variables. Put it in a different way, the contemporaneous economic variables affect government consumption expenditure directly but lagged variables do not. In this article, we used 2-year lagged variables, which is even more systematically uncorrelated with the contemporaneous economic variable and tackles the issues like aggregation across a year (Hall 1988) and argued to be more credible (Murray 2006). Moreover, to check for the robustness, we also use alternative specifications like lagged 3-year averages of the economic variables and initial values of the economic variables, which are other standard techniques used in cross-country growth regressions. (5) Our results stay largely consistent throughout regressions using different measures of lag variables to proxy for the economic variables.
In order to test the robustness of the model, we imputed some missing variables of the countries to improve the degrees of freedom of the model and also to check the persistence of the estimations. There are some countries which have very good data but 1 or 2 years' data are missing for some variables. We have used linear trend imputation techniques to estimate the missing values for these countries. (6)
The basic specification for the model is
(1) [GCE.sub.i,t] = [alpha] + [beta] * [variables.sub.i,t] + [gamma] * year, + [delta] * continent [dummy.sub.i] + [[epsilon].sub.i,t].
where i denotes the country and t denotes the year. For an extended specification, we will keep the basic specification with an added set of new variables.
Finally, to test the robustness of our findings, we also used extended specification as well as tested the key variables with EBA suggested by Learner (1983) and Levine and Renelt (1992).
V. ESTIMATION RESULTS
Columns 1-6 in Table A2 report the specifications based on 2-year lagged values of the economic variables while columns 7 and 8 used initial values as an alternative. Here, all the regressions are based on FGLS estimation technique. The specifications differ mainly due to the combination of the governance and corruption indicators. Among all the specification estimated in Table A2, our richest specification is the one reported in column 6.
A. Base Variables
Our result suggests that real GDP per capita is highly significantly associated with government consumption expenditure. The point estimates from column 6 in Table A2, for example, suggest that one standard deviation increase in income per capita increases the GCE of the current year by almost 1.47% of GDP, suggesting the evidence in line with Wagner's law, a principle named after German economist Adolph Wagner (1835-1917) which states that for any economy public expenditure rises constantly with an upward sloping trend, which is more related to consumption spending in developing countries. This result indicates that with the increase in the per capita income of the population, developing economies appear to expand their public consumption spending plausibly due to the emerging pressure on the demand for publicly available goods and services. Such a finding is strongly consistent with all the other regressions estimated in this article and broadly in line with the recent literature of re-examining Wagner's law (e.g., the work of Durevall and Henrekson 2011 and Kirchner 2012).
We did not find a strong association between trade openness and government expenditure. Using 2-year lagged values of the current trade openness, the current government expenditure is positively influenced by the trade openness for the developing countries for a few specifications but such a result is not robust in all the estimations, which suggests that the variable might be subject to omitted variable bias and should be treated cautiously.
In our specification, we find that a one standard deviation increase in the log of population, as a measure for country size, leads to a decrease in GCE by 1.29% of GDP (using the specification of column 6). This result shows evidence of large preference heterogeneity-led reduction in government expenditure as hypothesized by Alesina and Wacziarg (1998). Among the continental dummies, we can observe that on average, GCE as a percentage of GDP is higher in European and African countries than Asia and Latin American ones, which is quite consistent with other extended specifications. This particular result confirms that European and African countries tend to accommodate a greater degree of publicly provided goods and services such as social security and health care than other continents, which has increased the relative size of their government expenditures.
B. Demographical Variables
To control for demographical variables, we used variables like population ages 0 to 14, population ages 65 and above, and urban population, as a percentage of total population. Our estimation did not find any robust statistically association with the government expenditure as a percentage of GDP with the share of older population. (7) The reason for this finding could be analyzed from the supply side. In most developing countries, it is very difficult to find adequate government-run social security or a well-established pension system for the aging population. Due to the resource limitation, these governments mostly prioritize their expenditure toward public education, public health spending (mostly focusing on mother and children) or infrastructure investment rather than dedicated spending on the rising older population share. This finding is in line with Razin, Sadka, and Swagel (2002) who argued that larger dependency ration leads to lower tax revenue thus less public transfer even in welfare states.
By contrast, a strong and positive association has been found between the population aged less than 15 and government expenditure as a percentage of GDP, and such trend continues even in our other specifications. This result reveals that developing countries on average allocate more expenditure to the growing fraction of younger population. A one standard deviation increase in the fraction of population less than 15 is associated with an increase in GCE by 1.22% of GDP (using the specification of column 6 of Table A2). Such a rise in expenditure can plausibly be directed toward the education and health sectors of the economy to fulfill the emerging demand for these services with the greater fraction of young population.
Similarly, strong positive association between the degree of urbanization and public expenditure has also been found in most specifications, showing the emerging demand for public utilities and services in urban areas as the fraction of population living in urban areas increases. Internal rural to urban migration is a common phenomenon in developing countries, because the expected income in urban areas is higher than in the rural areas. As the degree of urbanization increases, governments need to spend more on transportation, public utilities, and amenities to fulfill the rising demand for such services. However, this variable becomes statistically less powerful in the specifications with initial value of the economic variables, hence should be treated with caution.
C. Fractionalization Variables
Two different measures of fractionalization--ethnic and language--have been used to test the association of fractionalization with government expenditure in this article. The results of such regressions are reported in Table A2. The coefficient of ethnic fractionalization shows positive and highly statistically significant power in explaining the variation in government expenditure. This finding is in line with the work of Alesina, Baqir, and Easterly (2000) who demonstrated, using U.S. data, that greater ethnic fragmentation leads to higher public employment, because governments of an ethnically diverse economy tend to use public employment--led increase in government expenditure as an "implicit subsidy" (Alesina, Baqir, and Easterly 2000, 220) to ethnical interest groups who would otherwise receive transfer payments. Politicians are also interested in such strategies to disguise their redistributive policies, to avoid opposition of precise tax-transfer schemes.
However, Easterly and Levine (1997) find a strong negative relationship between ethnic fragmentation and some public goods (such as telecommunication, transportation, electricity grids, and education) in African countries. They conclude that, due to such high levels of ethnic division and conflict, African countries have largely adopted "growth-retarding" policies over the years which could be a principal reason for Africa's recent growth tragedy. The dataset used in this article also reveals that on average, ethnic fractionalization is remarkably higher in the African nations than in other continents. To be specific, the average probability that two randomly selected people do not belong to the same ethnic group in African countries is 0.25 whereas the average is only 0.06 in European nations. In our data, 18 out of 20 most ethnically heterogeneous countries belong to Africa, demonstrating the high degree of ethnic diversity in Africa.
Therefore, simply using ethnicity to explore the cross-country difference in government expenditure, it would be more sensible and interesting to explore the association between ethnic diversity and GCE in African nations. To explore this association, we employed an interaction term between ethnic fractionalization and African continental dummy in all the specifications. Our specifications reveal that the interaction term of ethnic fractionalization with Africa is significantly negatively correlated with government expenditure and has an economically large coefficient in all the specifications. Ethnic diversity influences the economic performance of any nation and has a direct influence on growth performance (Easterly and Levine 1997). The estimation confirms that with a greater degree of ethnic heterogeneity, nations in Africa tend to reduce the size of government expenditure. A high degree of ethnic fractionalization leads to under-provision of publicly available services like education, transportation, and infrastructure, which has a negative impact on the economic growth of the continent and could be used to explain the recent growth tragedy of Africa.
Linguistic fractionalization, on the other hand, is more or less a common phenomenon in any continent and as a result does not have a significant explanatory power to address the variation of government expenditure in cross-country regression. We did not use religious fractionalization in our regression because religious fractionalization is mostly endogenous in nature (Alesina et al. 2003). Individuals and families can convert to another religion quite easily, and a high degree of religious heterogeneity could be a sign of tolerance and harmony rather than conflict, which could also explain the reason for not employing religious fractionalization to explain cross-country government expenditure.
D. Political Institutional Variables
To capture the impact of different political regimes, we used the polity index which has been converted to regime categories as suggested by Marshall and Jaggers (2003). The categories used are basically dummy variable where the categorization of "autocracies" (-10 to -6), "anocracies" (-5 to +5), and "democracies" (+6 to +10) have been used. Our regression reveals strong association of autocracy and democracy with the variation of government expenditures when compared with anocracies, which is in line with the earlier theoretical and empirical findings. When a government moves from anocracy to democracy, or from anocracy to autocracy, the shift in political regime significantly increases the size of the government. However, the choice of public good provision under different political regimes could be completely different. Democratic governments mostly spend excess funds on providing better healthcare, education, environmental protection, and transfer payments (Deacon 2009). By contrast, autocratic government might focus expenditure on the expansion of law enforcement or providing better facilities to the elite to keep them satisfied. In a democratic regime, governments usually have a short-term fiscal horizon in contrast to autocratic governments, for which the policy horizon is mostly long term. As a result, autocratic governments can be aggressive in terms of expenditure and can continue to make expensive bad policy choices.
One extreme form of dictatorship is military dictatorship, in which the dictator comes from a military background. Our regressions suggest that public consumption expenditure shrinks significantly with military dictatorship under autocratic regime (this is captured with the interaction term of autocracy and military variable). This finding is not surprising because a military dictatorship historically has high entry and exit costs; entry may require the overthrowing of a powerful ruler, or mass killing through a military coup, or even civil war, whereas exit might involve the imprisonment, asylum, or death of the military dictator. Hence, military dictators under autocratic regime channelize a large share of government expenditure toward military investment, keeping in mind their own protection, which could possibly explain such negative impact on the current consumption of government expenditure. Moreover, military dictators would be unable to implement any effective fiscal policy because of the uncertain span of their government. This uncertainty might then influence military dictators under autocratic regime to cut down on government expenditure and could also make them even very reluctant to undertake ambitious projects which require further expenditure. In most cases, countries under the rule of a military dictatorship do not receive any foreign aid from international donors, and international trade with such countries becomes restricted, which may also reduce government expenditure in these countries. Other than political regime variables, we also used "number of government seats" variable, taken from the DPI, (8) which is found to be highly significant and positive in influencing government expenditure. "Number of government seats" variable shows how influential a government can be if one political party has an absolute majority in the parliament.
E. Governance Variables
Among the six different variables of WGI, to capture the quality of the governance in any nation, we used government effectiveness and control of corruption variable in our regressions. Table A2 reveals that both the WGI variables have highly significant and positive association with government expenditure. "Government effectiveness" variable captures the perception of the quality and degree of independence of the public and civil services. It also captures the quality of policy formulation and implementation, and the credibility of a government's commitment. As a result, a superior government effectiveness index means the civil and public services exercise a higher degree of independence and quality, and the government is also credible and effective in terms of policy implementation.
We used two variables to understand the impact of corruption on GCE. The first is the CPI, which measures the perceived level of public sector corruption based on 13 different expert and business surveys. (9) In our CPI variable, instead of corruption scores, we used the rank of countries to provide better variation in the variable. The higher the rank of a country in the CPI, the higher the perception of corruption for that country. Our regression confirms the prior assumption that corruption has considerable negative bearing on government consumption expenditure and the result is significant even with 1 % level of significance (columns 4 and 6 in Table A2) which is in line with the earlier finding of Mauro (1998). To test the robustness of our findings, we used a second corruption variable known as "control of corruption" from the WGI dataset where the variable measures the exercise of public power for private gain, including both petty and grand corruption and state corruption measured on a scale between +2.5 and -2.5. The lower the score for "control of corruption" in a country, the higher the corruption for that country. Similar to our previous findings, we see that the more corrupt countries spend less on current government consumption expenditure and this result is highly statistically significant. The reason for such finding could be that corrupt governments reduce the spending on current consumption and increase the spending on public investment and transfers where they have better scope for corruption. Our coefficient of "control of corruption" is positive and significant, suggesting that less corrupted countries spend more on salaries, wages, social securities, and safety net programs thus requiring larger GCE.
F. Robustness Check
To check for the robustness of our study, we estimated regressions with alternative specifications presented in Table A3. In columns 1 and 2, we dropped trade openness from the specifications and in columns 3-6 we added two additional control variables, total debt services, and aid per capita, to check the robustness of our estimations. All these economic controls are instrumented by 2-year lagged values of the respective variable. Even with the added control for economic variables, which are used to check for omitted control bias, the results are highly consistent and almost all the findings of the previous section hold with the same sign and statistical significance.
In Table A4, we estimated all our specifications of Table A2 with the lagged 3-year average of the economic variables as proxies for the economic variables. In doing so, we lost two-thirds of our observation; however, our major findings still remain consistent and statistically significant, showing credibility of our estimations. In addition, we reported simple pooled regression with country and year fixed effects for all our specifications in Table A5 which is largely consistent with our earlier findings.
VI. SENSITIVITY TEST: EBA
The purpose of this section is to examine the sensitivity of the regression results presented above with the aid of Learner's (1983) EBA, which is a common approach used in cross-country empirical work. EBA is useful in investigating the sensitivity of the empirical findings to the changes in the number of explanatory variables. The specification used in the article is similar to the technique discussed by Levine and Renelt (1992). Our specifications are:
(2) [GCE.sub.i,t] = [alpha] + [[beta].sub.M] + [[beta].sub.I]I + [[beta].sub.z]Z + [epsilon]
where M is the variable of interest, I is the base variables which are always included in the regression, and Z is a subset of variables chosen from a set of potentially important explanatory variables. EBA varies the subset of Z variables included in the regression to find the widest range of coefficient estimates (Levine and Renelt 1992). In such test, we need to find the lower extreme bound and the upper extreme bound for all the possible regressions with the same number of parameters. The extreme bound test suggests that the variable M is not robust if all the estimated [[beta].sub.M]'s are not statistically significant or changes the sign between lower and upper extreme values. This entails that as long as there exists one regression for which the sign of the coefficient is not significant or the coefficient changes the sign, the variable is not robustly influential.
The EBA procedure, proposed by Levine and Renelt (1992), is very restrictive, rigid, and it would be difficult to find any variable that could survive such an acid test. Even Levine and Renelt (1992) found that no variables are robust determinants of economic growth. Sala-i Martin (1997a, 1997b) argues that the decision mechanism used in EBA is too rigid and if the coefficient in consideration has some positive and negative supports then one is bound to have at least one regression for which the coefficient changes sign, if enough regressions are run. Hence, Sala-i Martin (1997b) suggested a more relaxed version of the EBA test, which uses the entire distribution of the estimated of [[beta].sub.M] and argued that if at least 95% of the density function of the estimated [[beta].sub.M] lies on either side of the zero, one could have reasonable confidence that the [[beta].sub.M] is robust. Following this suggestion, we have reported unweighted cumulative distribution function (CDF, which is the fraction of CDF lying on either side of the zero. Our conclusion of variable robustness is thus based on Sala-i-Martin's variant of the EBA.
In Table A4, the results of the robustness of the variables are reported. For our base variables, we used income, openness, population, and continental dummies in all of our regressions. In our investigation with EBA, only three variables, language fractionalization, openness, and urbanization, could not pass the test. On the basis of these results, one can assert a reasonable amount of confidence in the partial correlations of all the important variables of this study.
This study empirically investigated the determinants of government expenditures of developing countries by looking into a broad spectrum of specifications with emphasis on political and institutional variables. We used data from the WDI (2008), WGI, DPI (2004), and Transparency International for 97 developing countries from 1984 to 2004. However, the research has been affected by the unavailability of data of some economic variables over the period we examined. Some developing countries are still unable to provide important economic data as they have poor institutional facilities for providing up-to-date indexes. Despite out efforts to show the robustness of our study, some of the results should be interpreted with caution, due to multicollinearity issues among the family of variables, which is beyond our control.
Using the unbalanced panel dataset for developing countries, other than the economic variables such as per capita income and trade openness, we found robust evidence of political and institutional variables significantly influencing government expenditure. Among other results, corruption has been found to be negatively influencing the government consumption expenditure in the developing countries. Corrupt bureaucrats may have little rent-seeking opportunities from the consumption expenditure compared with other components of spending. As a result, corrupt officials may like to reduce the consumption portion of the government expenditure to channel some additional resources toward other forms of government spending which is more susceptible to corruption.
Demographic patterns of the population are found to have considerable impact on government consumption expenditure, having a positive association with younger population. We find confirming evidence that greater fractionalization leads to higher government expenditure; however, higher ethnic fractionalization in African countries leads to a sizable negative influence on government expenditure, which is aligned with previous studies. Furthermore, our results suggest that autocratic governments with military ruling are not particularly accommodative toward government consumption expenditures perhaps due to interest in the benefit and protection of their own as well as the welfare of the elites who will ensure their power in the politics. Hence, in developing countries, this form of governance is likely to hold little appeal for those who directly depend on publicly provided utilities and spending.
There are many issues that our analysis did not address and left for future research. It would be interesting to estimate the dynamic pattern of government expenditure which will enable us to estimate the level of persistence of government expenditure for developing nations. Furthermore, future research could also look into the supply side of government expenditure rather than the demand side of it. Such research will give us a better understanding of the overall determinants of the government expenditures of developing countries and could be an important area for future study.
CDF: Cumulative Distribution Function
CPI: Corruption Perception Index
DPI: Database of Political Institutions
EBA: Extreme Bound Analysis
FGLS: Feasible Generalized Least Square
GCE: Government Consumption Expenditure
GDP: Gross Domestic Product
IMF: International Monetary Fund
ODA: Official Development Assistance
OLS: Ordinary Least Squares
WDI: World Development Indicators
WGI: Worldwide Governance Indicators
TABLE A1 Pairwise Correlation Among Demographical and Fractionalization Variables, Unbalanced Panel (1984-2004) Income Openness Lpop Ethnic Language Income 1 Openness 0.131 1 Lpop -0.0866 -0.536 1 Ethnic -0.316 -0.117 0.0253 1 Language -0.441 -0.13 0.0817 0.718 1 Oldpop 0.416 0.186 -0.0016 -0.361 -0.32 Pop014 -0.502 -0.228 -0.118 0.418 0.388 Urbanpop 0.709 0.0768 0.0689 -0.285 -0.451 Govseat -0.0702 -0.17 0.432 -0.137 -0.102 Autocracy -0.194 -0.0946 -0.0373 0.0409 0.114 Democracy 0.362 0.11 0.028 -0.171 -0.269 Military -0.28 -0.205 0.0723 0.194 0.169 Goveffect 0.613 0.188 -0.0691 -0.238 -0.292 Concorp 0.581 0.189 -0.211 -0.261 -0.314 Oldpop Pop014 Urbanpop Govseat Autocracy Income Openness Lpop Ethnic Language Oldpop 1 Pop014 -0.899 1 Urbanpop 0.52 -0.607 1 Govseat 0.0209 -0.107 -0.0807 1 Autocracy -0.223 0.278 -0.293 0.188 1 Democracy 0.32 -0.389 0.398 -0.12 -0.576 Military -0.266 0.333 -0.272 -0.00145 0.323 Goveffect 0.281 -0.404 0.359 0.103 -0.181 Concorp 0.23 -0.328 0.304 0.0169 -0.0875 Democracy Military Goveffect Concorp Income Openness Lpop Ethnic Language Oldpop Pop014 Urbanpop Govseat Autocracy Democracy 1 Military -0.384 1 Goveffect 0.354 -0.195 1 Concorp 0.305 -0.169 0.837 1 TABLE A2 Government Consumption Expenditure: Unbalanced Panel (1996-2004) With Lagged Values of Economic Variable Independent Variable: GCE(% of GDP) (1) (2) (3) Real GDP per capita 0.001 *** 0.001 *** 0.001 *** (0.000) (0.000) (0.000) Openness 0.003 0.010 *** 0.010 *** (0.003) (0.003) (0.003) Log of population -1.684 *** -0.957 *** -0.828 *** (0.146) (0.129) (0.131) Africa 9.110 *** 9.076 *** 9.331 *** (0.655) (0.897) (0.787) Asia 4.064 *** 1.883 *** 2.052 *** (0.677) (0.436) (0.541) Europe 8.176 *** 9.070 *** 9.759 *** (0.736) (0.693) (0.743) Ethnic 6.781 *** 8.860 *** 9.726 *** (0.998) (0.951) (1.006) (Africa) x (ethnic) -10.993 *** -12.094 *** -13.179 *** (1.244) (1.354) (1.331) Language 0.727 0.097 1.245 (0.943) (0.743) (0.781) Population aged 65+ -0.048 -0.170 -0.030 (0.104) (0.104) (0.115) Population aged 0-14 0.171 *** 0.105 ** 0.158 *** (0.036) (0.049) (0.050) Urban population 0.049 *** 0.015 0.027 *** (0.011) (0.010) (0.010) Number of government seats 0.004 *** 0.003 *** 0.003 *** (0.000) (0.000) (0.000) Autocracy 0.594 *** 1.017 *** 1.165 *** (0.210) (0.304) (0.305) Democracy 0.594 *** 0.795 *** 0.650 *** (0.174) (0.186) (0.172) Military -0.123 -0.054 0.113 (0.186) (0.280) (0.287) (Autocracy) x (military) -0.496 * -0.747 * -0.994 ** (0.295) (0.424) (0.419) Government effectiveness 1.102 *** (0.213) Control of corruption 1.512 *** (0.205) CPI rank Observations 1,671 863 854 Number of countries 97 97 97 Year-specific dummy Yes Yes Yes With Lagged Values of Economic Variable Independent Variable: GCE(% of GDP) (4) (5) (6) Real GDP per capita 0.001 *** 0.001 *** 0.001 *** (0.000) (0.000) (0.000) Openness 0.006 0.010 *** 0.001 (0.005) (0.004) (0.005) Log of population 0.048 -0.799 *** -0.199 (0.183) (0.134) (0.173) Africa 9.204 *** 9.037 *** 8.170 *** (1.036) (0.873) (0.969) Asia 2.027 *** 1.838 *** 1.568 *** (0.484) (0.538) (0.521) Europe 11.663 *** 9.726 *** 11.455 *** (0.587) (0.731) (0.604) Ethnic 7.969 *** 9.702 *** 7.506 *** (0.906) (1.014) (0.960) (Africa) x (ethnic) -8.122 *** -12.986 *** -7.591 *** (1.914) (1.410) (1.795) Language -0.470 1.138 0.298 (0.732) (0.774) (0.759) Population aged 65+ -0.305 *** -0.061 -0.330 *** (0.105) (0.114) (0.107) Population aged 0-14 0.114 ** 0.160 *** 0.114 ** (0.052) (0.051) (0.052) Urban population 0.087 *** 0.026 *** 0.072 *** (0.013) (0.010) (0.012) Number of government seats 0.002 *** 0.003 *** 0.002 *** (0.000) (0.000) (0.000) Autocracy 0.618 ** 1.177 *** 0.721 ** (0.291) (0.308) (0.294) Democracy 1.075 *** 0.643 *** 0.977 *** (0.191) (0.177) (0.201) Military -0.447 0.114 -0.381 (0.365) (0.280) (0.348) (Autocracy) x (military) -1.299 *** -0.967 ** -1.207 *** (0.448) (0.421) (0.432) Government effectiveness 0.410 * 1.245 *** (0.224) (0.275) Control of corruption 1.376 *** (0.223) CPI rank -0.020 *** -0.015 *** (0.005) (0.005) Observations 451 854 451 Number of countries 70 97 70 Year-specific dummy Yes Yes Yes With Initial Values of Economic Variable Independent Variable: GCE(% of GDP) (7) (8) Real GDP per capita 0.001 *** 0.001 *** (0.000) (0.000) Openness 0.017 ** 0.012 (0.007) (0.008) Log of population -0.745 *** -0.790 *** (0.152) (0.158) Africa 5.369 *** 5.551 *** (0.886) (0.813) Asia 1.685 ** 1.060 (0.747) (0.730) Europe 3.088 4.159 ** (1.904) (1.744) Ethnic 4.901 *** 5.819 *** (1.443) (1.280) (Africa) x (ethnic) -7.917 *** -9.416 *** (1.590) (1.432) Language 0.565 0.534 (1.046) (1.103) Population aged 65+ -0.298 ** -0.290 ** (0.139) (0.140) Population aged 0-14 0.083 * 0.125 *** (0.048) (0.048) Urban population 0.011 0.008 (0.012) (0.012) Number of government seats 0.002 * 0.002 *** (0.001) (0.000) Autocracy 0.568 * 0.628 ** (0.321) (0.318) Democracy 0.324 0.330 * (0.203) (0.198) Military -0.268 -0.298 (0.294) (0.274) (Autocracy) x (military) -1.011 ** -0.918 ** (0.466) (0.452) Government effectiveness 0.790 *** (0.238) Control of corruption 1.586 *** 1.257 *** (0.213) (0.225) CPI rank Observations 665 665 Number of countries 75 75 Year-specific dummy Yes Yes Note: Values in parentheses are the reported standard errors of the estimation. Except for columns 6 and 7, all models used 2 years lag whereas for the model in columns 7 and 8, initial values have been used. *** Significant at 1%; ** significant at 5%; and * significant at 10%. Source: WDI (2008), Alesina et al. (2003), DPI (2004), Polity IV: Regime Authority Characteristics and Transitions Dataset (2008), CPI (2007), and WGI (2007). TABLE A3 Government Consumption Expenditure: Alternative Specifications, Unbalanced Panel (1996-2004) With Only Income Independent Variable: GCE (% of GDP) (1) (2) Real GDP per capita 0.001 *** 0.001 *** (0.000) (0.000) Openness Total debt service Aid per capita Log of population -0.859 *** -0.852 *** (0.110) (0.116) Africa 8.704 *** 8.421 *** (0.694) (0.741) Asia 2.379 *** 2.156 *** (0.522) (0.528) Europe 9.416 *** 9.470 *** (0.768) (0.758) Ethnic 7.794 *** 7.832 *** (1.054) (1.046) (Africa) x (ethnic) -10.854 *** -10.845 *** (1.298) (1.335) Language 1.182 0.960 (0.823) (0.827) Population aged 65+ -0.025 -0.051 (0.115) (0.114) Population aged 0-14 0.120 *** 0.132 *** (0.046) (0.047) Urban population 0.034 *** 0.030 *** (0.010) (0.010) Number of government seats 0.002 *** 0.002 *** (0.000) (0.000) Autocracy 1.121 *** 1.116 *** (0.293) (0.296) Democracy 0.704 *** 0.707 *** (0.175) (0.176) Military 0.121 0.071 (0.280) (0.279) (Autocracy) x (military) -1.129 *** -1.046 *** (0.396) (0.401) Control of corruption 1.284 *** 1.125 *** (0.198) (0.217) Government effectiveness 0.346 (0.220) Observations 854 854 Number of countries 97 97 Year-specific dummy Yes Yes With Additional Control of Aid and Debt Independent Variable: GCE (% of GDP) (3) (4) Real GDP per capita 0.001 *** 0.001 *** (0.000) (0.000) Openness 0.011 *** 0.011 *** (0.004) (0.004) Total debt service -0.012 -0.015 (0.014) (0.014) Aid per capita Log of population -0.806 *** -0.766 *** (0.132) (0.135) Africa 9.432 *** 9.039 *** (0.746) (0.843) Asia 2.220 *** 2.034 *** (0.547) (0.544) Europe 9.865 *** 9.818 *** (0.743) (0.729) Ethnic 9.041 *** 8.937 *** (0.979) (0.986) (Africa) x (ethnic) -12.924 *** -12.522 *** (1.316) (1.421) Language 1.247 1.068 (0.801) (0.797) Population aged 65+ 0.051 0.039 (0.110) (0.110) Population aged 0-14 0.203 *** 0.210 *** (0.044) (0.045) Urban population 0.034 *** 0.034 *** (0.010) (0.010) Number of government seats 0.003 *** 0.003 *** (0.000) (0.000) Autocracy 1.265 *** 1.320 *** (0.305) (0.308) Democracy 0.621 *** 0.619 *** (0.177) (0.182) Military 0.046 0.060 (0.285) (0.279) (Autocracy) x (military) -1.098 *** -1.085 ** (0.419) (0.424) Control of corruption 1.499 *** 1.376 *** (0.204) (0.224) Government effectiveness 0.401 * (0.227) Observations 854 854 Number of countries 97 97 Year-specific dummy Yes Yes With Additional Control of Aid and Debt Independent Variable: GCE (% of GDP) (5) (6) Real GDP per capita 0.001 *** 0.001 *** (0.000) (0.000) Openness 0.012 *** 0.011 *** (0.004) (0.004) Total debt service -0.019 -0.020 (0.014) (0.014) Aid per capita 0.001 0.000 (0.002) (0.002) Log of population -0.749 *** -0.798 *** (0.132) (0.131) Africa 8.582 *** 8.159 *** (0.671) (0.691) Asia 1.519 *** 1.422 *** (0.461) (0.449) Europe 9.751 *** 9.738 *** (0.690) (0.684) Ethnic 8.744 *** 8.602 *** (0.804) (0.800) (Africa) x (ethnic) -11.705 *** -11.341 *** (1.167) (1.170) Language 0.522 0.513 (0.758) (0.743) Population aged 65+ -0.023 -0.021 (0.104) (0.104) Population aged 0-14 0.171 *** 0.183 *** (0.040) (0.041) Urban population 0.035 *** 0.036 *** (0.010) (0.010) Number of government seats 0.002 *** 0.002 *** (0.000) (0.000) Autocracy 1.418 *** 1.481 *** (0.314) (0.318) Democracy 0.691 *** 0.647 *** (0.179) (0.183) Military 0.070 0.051 (0.285) (0.280) (Autocracy) x (military) -1.194 *** -1.281 *** (0.431) (0.432) Control of corruption 1.649 *** 1.459 *** (0.196) (0.218) Government effectiveness 0.484 ** (0.226) Observations 854 854 Number of countries 97 97 Year-specific dummy Yes Yes Note: Values in parentheses are the reported standard errors of the estimation. *** Significant at 1%; ** significant at 5%; * significant at 10%. Source: WDI (2008), Alesina et al. (2003), DPI (2004), Polity IV: Regime Authority Characteristics and Transitions Dataset (2008), CPI (2007), and WGI (2007). TABLE A4 Government Consumption Expenditure with Lagged 3-Year Average, Unbalanced Panel (1996-2004) With Lagged 3-Year Average Value Independent Variable: GCE (% of GDP) (1) (2) (3) Real GDP per capita 0.001 *** 0.001 *** 0.001 *** (0.000) (0.000) (0.000) Openness 0.022 *** 0.010 *** 0.011 *** (0.005) (0.002) (0.003) Log of population -1.353 *** -0.595 *** -0.666 *** (0.186) (0.130) (0.132) Africa 8.868 *** 9.458 *** 10.270 *** (0.863) (0.801) (0.852) Asia 3.068 *** 1.407 *** 0.935 ** (0.711) (0.458) (0.426) Europe 7.595 *** 6.962 *** 6.663 *** (0.717) (0.555) (0.577) Ethnic 6.424 *** 8.143 *** 7.573 *** (1.182) (0.597) (0.769) (Africa) x (ethnic) 1.625 0.079 0.244 (1.158) (0.554) (0.502) Language -10.816 *** -11.564 *** -12.280 *** (1.605) (1.130) (1.281) Population aged 65+ 0.019 0.139 * -0.053 (0.103) (0.073) (0.101) Population aged 0-14 0.201 *** 0.150 *** 0.091 ** (0.040) (0.025) (0.045) Urban population 0.056 *** 0.044 *** 0.039 *** (0.012) (0.009) (0.009) Number of government seats 0.003 *** 0.001 *** 0.001 *** (0.000) (0.000) (0.000) Autocracy 2.022 *** 3.939 *** 3.440 *** (0.415) (0.404) (0.460) Democracy 0.569 * 0.770 *** 0.762 *** (0.300) (0.258) (0.221) Military -0.092 -0.502 -0.648 (0.369) (0.439) (0.397) (Autocracy) x (military) -2.770 *** -4.841 *** -4.558 *** (0.565) (0.624) (0.697) Government effectiveness 1.966 *** (0.270) Control of corruption 2.183 *** (0.175) CPI Observations 538 290 290 Number of countries 97 97 97 Year-specific dummy Yes Yes Yes With Lagged 3-Year Average Value Independent Variable: GCE (% of GDP) (4) (5) (6) Real GDP per capita 0.000 *** 0.000 *** 0.000 (0.000) (0.000) (0.000) Openness 0.011 *** 0.019 *** 0.021 *** (0.002) (0.003) (0.004) Log of population -0.581 *** 0.302 * 0.516 *** (0.131) (0.162) (0.189) Africa 9.211 *** 6.410 *** 6.074 *** (0.818) (0.874) (1.040) Asia 1.098 ** 1.893 *** 1.030 * (0.481) (0.508) (0.567) Europe 6.850 *** 8.828 *** 8.483 *** (0.542) (0.495) (0.571) Ethnic 7.990 *** 8.132 *** 7.940 *** (0.581) (0.877) (0.962) (Africa) x (ethnic) 0.218 -2.551 *** -2.833 *** (0.554) (0.630) (0.768) Language -11.671 *** -2.089 -1.920 (1.116) (1.305) (1.606) Population aged 65+ 0.172 ** 0.048 0.084 (0.076) (0.086) (0.087) Population aged 0-14 0.171 *** 0.201 *** 0.223 *** (0.027) (0.031) (0.039) Urban population 0.040 *** 0.081 *** 0.081 *** (0.009) (0.006) (0.009) Number of government seats 0.001 *** 0.001 *** 0.001 (0.000) (0.000) (0.000) Autocracy 4.066 *** 3.674 *** 3.307 *** (0.407) (0.228) (0.304) Democracy 0.679 *** 2.859 *** 2.770 *** (0.255) (0.195) (0.251) Military -0.573 1.712 *** 0.897 (0.438) (0.558) (0.564) (Autocracy) x (military) -4.908 *** -9.885 *** -9.339 *** (0.617) (1.741) (1.587) Government effectiveness 1.649 *** (0.311) Control of corruption 0.543 * 1.057 *** (0.285) (0.300) CPI -0.056 *** -0.044 *** (0.006) (0.008) Observations 290 172 172 Number of countries 97 62 62 Year-specific dummy Yes Yes Yes Note: Values in parentheses are the reported standard errors of the estimation. All models used lagged 3-year average of economic variables. *** Significant at 1%; ** significant at 5%; * significant at 10%. Source: WDI (2008), Alesina et al. (2003), DPI (2004), Polity IV: Regime Authority Characteristics and Transitions Dataset (2008), CPI (2007), and WGI (2007). TABLE A5 Government Consumption Expenditure: Pooled Regression, Unbalanced Panel (1996-2004) Pooled Regression with Year and Country Fixed Effects Independent Variable: GCE (% of GDP) (1) (2) (3) Real GDP per capita 0.001 *** 0.000 ** 0.000 * (0.000) (0.000) (0.000) Openness -0.008 0.020 *** 0.022 *** (0.006) (0.006) (0.007) Log of population -13.491 *** -0.480 *** -0.294 (2.539) (0.176) (0.184) Africa 15.331 *** 8.450 *** 8.010 *** (5.346) (0.847) (0.847) Asia 98.116 *** -0.040 0.072 (24.157) (0.835) (0.832) Europe 26.295 *** 5.897 *** 5.956 *** (8.501) (0.861) (0.836) Ethnic 39.667 *** 8.345 *** 8.697 *** (10.006) (1.063) (1.091) (Africa) x (ethnic) 21.093 *** -0.647 -0.394 (4.022) (0.934) (0.972) Language -0.611 * -0.005 0.025 (0.359) (0.123) (0.121) Population aged 65+ 0.245 *** 0.131 ** 0.146 ** (0.067) (0.060) (0.060) Population aged 0-14 0.175 *** 0.032 ** 0.039 *** (0.056) (0.014) (0.014) Urban population 0.004 *** 0.000 0.000 (0.001) (0.000) (0.000) Number of government seats 0.569 5.800 *** 5.177 *** (0.390) (1.084) (1.032) Autocracy -0.192 0.926 * 0.760 (0.347) (0.478) (0.471) Democracy 0.561 -1.515 *** -1.195 ** (0.525) (0.502) (0.497) Military -1.880 *** -5.241 *** -5.115 *** (0.544) (1.070) (1.061) (Autocracy) x (military) -10.295 *** -10.379 *** (1.442) (1.455) Government effectiveness 2.403 *** (0.433) Control of corruption 3.002 *** (0.516) CPI Observations 1,671 863 854 Number of countries 0.764 0.314 0.322 Year-specific dummy Yes Yes Yes Country-specific dummy Yes Yes Yes Pooled Regression with Year and Country Fixed Effects Independent Variable: GCE (% of GDP) (4) (5) (6) Real GDP per capita 0.000 ** 0.000 0.000 (0.000) (0.000) (0.000) Openness -0.006 0.022 *** -0.006 (0.008) (0.007) (0.008) Log of population -0.186 -0.308 * -0.154 (0.254) (0.186) (0.255) Africa 7.483 *** 7.913 *** 7.314 *** (1.334) (0.877) (1.341) Asia 1.322 -0.021 0.863 (0.835) (0.846) (0.902) Europe 7.357 *** 5.920 *** 7.210 *** (1.133) (0.837) (1.107) Ethnic 7.242 *** 8.632 *** 7.437 *** (1.137) (1.091) (1.165) (Africa) x (ethnic) -0.067 -0.410 -0.050 (1.230) (0.966) (1.239) Language -0.021 0.027 0.005 (0.141) (0.121) (0.142) Population aged 65+ 0.173 *** 0.149 ** 0.186 *** (0.065) (0.060) (0.067) Population aged 0-14 0.071 *** 0.039 *** 0.069 *** (0.018) (0.014) (0.018) Urban population 0.001 0.000 0.001 (0.001) (0.000) (0.001) Number of government seats 3.782 *** 5.255 *** 3.990 *** (1.377) (1.041) (1.402) Autocracy 1.575 ** 0.737 1.436 ** (0.633) (0.475) (0.639) Democracy -0.586 -1.220 ** -0.870 (0.754) (0.499) (0.780) Military -4.338 ** -5.082 *** -3.943 ** (1.697) (1.062) (1.715) (Autocracy) x (military) -6.814 *** -10.277 *** -7.083 *** (2.278) (1.476) (2.333) Government effectiveness 0.447 0.916 (0.511) (0.610) Control of corruption 2.665 *** (0.642) CPI -0.052 *** -0.042 *** (0.012) (0.015) Observations 457 854 457 Number of countries 0.335 0.323 0.338 Year-specific dummy Yes Yes Yes Country-specific dummy Yes Yes Yes Note: Values in parentheses are the reported standard errors of the estimation. *** Significant at 1%; ** significant at 5%; * significant at 10%. Source: WDI (2008), Alesina et al. (2003), DPI (2004), Polity IV: Regime Authority Characteristics and Transitions Dataset (2008), CPI (2007), and WGI (2007). TABLE A6 Results of EBA, Unbalanced Panel (1984-2004) (1) (2) (3) Average Average Standard Variables Beta Error CDF Real GDP per capita 0.0007 0.0001 0.9999 Openness 0.0051 0.003 0.8974 Log of population -1.3676 0.1321 0.9999 Population aged 65+ -0.2478 0.0865 0.9871 Population aged 0-14 0.083 0.029 0.9871 Urban population 0.0156 0.0104 0.8974 Ethnic 1.6597 0.7126 0.9615 Language -1.9429 0.71139 0.923 (Africa) x (ethnic) -4.0671 0.999 0.9999 Government effectiveness 0.7185 0.1661 0.9999 Control of corruption 1.1268 0.18329 0.9999 Number of government seats 0.0017 0.0004 0.9999 Autocracy 0.2569 0.1496 0.9999 Democracy 0.2646 0.148 0.9999 Autocracy x military -0.2761 0.2256 0.9999 Military -0.2347 0.1646 0.9871 (4) (5) (6) Decision Based Lower Upper Variables on CDF Bound Bound Real GDP per capita Robust 0.0003 0.0012 Openness Fragile -0.0078 0.0207 Log of population Robust -2.0839 -0.4179 Population aged 65+ Robust -0.6173 0.3005 Population aged 0-14 Robust -0.117 0.1847 Urban population Fragile -0.0298 0.0685 Ethnic Robust -1.5756 7.7416 Language Fragile -5.9077 3.7163 (Africa) x (ethnic) Robust -12.1768 0.2971 Government effectiveness Robust -0.1576 1.3884 Control of corruption Robust 0.5096 1.7747 Number of government seats Robust -0.0003 0.0031 Autocracy Robust -0.1389 1.286 Democracy Robust -0.3389 0.7582 Autocracy x military Robust -1.4124 0.3614 Military Robust -1.2067 0.4175 Note: Results are based on FGLS model. The unweighted average of the estimated coefficients and standard errors are reported in columns 1 and 2. The decision criterion in column 3 has been used by following the studies of Learner (1983) and Levine and Renelt (1992). CDF reported in column 4 is the unweighted CDF of the coefficients at the 5% level of significance as suggested by Sala-i Martin, Doppelhofer, and Miller (2004). The "lower bound" and "upper bound" give the lowest and highest value of point estimates minus plus two standard deviation. Source: WDI (2008); Alesina et al. (2003); DPI (2004), Polity IV: Regime Authority Characteristics and Transitions Dataset (2008), CPI (2007), and WGI (2007). TABLE A7 Descriptive Statistics Standard Standard Deviation Deviation Variable Mean (Panel) (Between) Government expenditure 14.356 6.076 5.336 GDP per capita 1476.845 1572.844 1536.254 Openness 72.007 39.064 35.487 Total debt service 5.797 5.535 3.719 Aid per capita 40.196 46.569 38.502 Log of population 15.998 1.647 1.649 Africa 0.387 0.487 0.489 Asia 0.207 0.405 0.407 Europe 0.189 0.392 0.393 America 0.216 0.412 0.414 Ethnic fractionalization 0.497 0.239 0.240 Language fractionalization 0.440 0.298 0.300 Population ages 65+ 5.028 3.093 3.031 Population ages 0-14 37.696 8.626 8.349 Urban population 43.736 19.839 19.612 Number of government seats 126.745 301.587 291.355 Autocracy 0.352 0.478 0.338 Democracy 0.379 0.485 0.388 Military 0.249 0.432 0.333 Government effectiveness -0.382 0.610 0.582 Control of corruption -0.457 0.565 0.533 Corruption index 66.38831 27.56809 26.211897 Standard Deviation Variable (Within) Minimum Maximum N Government expenditure 3.419 1.375 66.028 2183 GDP per capita 318.091 81.009 9497.559 2221 Openness 16.850 10.831 280.361 2175 Total debt service 4.047 0.000 107.374 2104 Aid per capita 25.077 -133.78 421.675 2160 Log of population 0.130 11.547 20.983 2331 Africa 0.000 0.000 1.000 2331 Asia 0.000 0.000 1.000 2331 Europe 0.000 0.000 1.000 2331 America 0.000 0.000 1.000 2331 Ethnic fractionalization 0.000 0.000 0.930 2247 Language fractionalization 0.000 0.008 0.923 2184 Population ages 65+ 0.677 1.888 17.012 2331 Population ages 0-14 2.303 14.070 51.915 2331 Urban population 3.504 4.520 92.750 2331 Number of government seats 75.535 0.000 2978.000 2235 Autocracy 0.344 0.000 1.000 2,011 Democracy 0.297 0.000 1.000 2011 Military 0.272 0.000 1.000 2138 Government effectiveness 0.189 -1.960 1.310 963 Control of corruption 0.189 -2.130 1.510 948 Corruption index 15.25037 17 145 479 TABLE A8 Variable Description Variable Name Details Ethnic and linguistics Fractionalization is an indicator fractionalization variable whose value ranges from 0 to 1. Two component indices are: (1) index of ethnic fractionalization in 1960, which measures the probability that two randomly selected people from a given country will not belong to the same ethnic group (the index is based on the number and size of population groups as distinguished by their ethnic and linguistic status); and (2) probability of two randomly selected individuals speaking different languages. General government final General government final consumption consumption expenditure expenditure (formerly general (% of GDP) government consumption) includes all government current expenditures for purchases of goods and services (including compensation of employees). It also includes most expenditure on national defense and security, but excludes government military expenditures that are part of government capital formation. Per capita income Annual per capita income on current (current U.S.$) U.S.S. Per capita income is GDP divided by midyear population. GDP at purchaser's prices is the sum of gross value added by all resident producers in the economy plus any product taxes and minus any subsidies not included in the value of the products. It is calculated without making deductions for depreciation of fabricated assets or for depletion and degradation of natural resources. Aid per capita Aid per capita includes both official (current U.S.S) development assistance (ODA) and official aid, and is calculated by dividing total aid by the midyear population estimate. Total debt service Total debt service is the sum of (% of gross national principal repayments and interest income) actually paid in foreign currency, goods, or services on long-term debt, interest paid on short-term debt, and repayments (repurchases and charges) to the IMF. Openness (% of GDP) Trade is the sum of exports and imports of goods and services measured as a share of GDP. Population ages 65 and Population ages 65 and above is the above (% of total) percentage of the total population that is 65 or older. Population ages 0-14 Population ages 0 to 14 are the (% of total) percentage of the total population that is 14 or younger. Urban population Population living in the urban areas as (% of total) a percentage of total population. Control of corruption Capturing perceptions of the extent to which public power is exercised for private gain, including both petty and grand forms of corruption, as well as "capture" of the state by elites and private interests. Government effectiveness Capturing perceptions of the quality of public services, the quality of the 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. Polity index A standard measure of governance on a 21-point scale ranging from -10 (dictatorial) to +10 (consolidated democracy) Military If the chief of the state is from military/defense background Number of government seats Number of total parliament seats belonging to the ruling party CPI Ranking of each country based on corruption perception. The higher the perception of corruption for that country, higher the rank of a country. Variable Name Source Ethnic and linguistics Alesina et al. (2003) fractionalization General government final WDI (2008) consumption expenditure (% of GDP) Per capita income WDI (2008) (current U.S.$) Aid per capita WDI (2008) (current U.S.S) Total debt service WDI (2008) (% of gross national income) Openness (% of GDP) WDI (2008) Population ages 65 and WDI (2008) above (% of total) Population ages 0-14 WDI (2008) (% of total) Urban population WDI (2008) (% of total) Control of corruption WGI (2005) Government effectiveness WGI (2005) Polity index Polity Score (2008) Military DPI 2004 Number of government seats DPI (2004) CPI CPI (2007)
Acemoglu, D., S. Johnson, and J. Robinson. "The Colonial Origins of Comparative Development: An Empirical Investigation." American Economic Review, 91(5), 2001, 1369-401.
Aidt, T. S. "Economic Analysis of Corruption: A Survey." The Economic Journal, 113(491), 2003, F632-52.
Alesina, A., and R. Perotti. "Fiscal Adjustments in OECD Countries: Composition and Macroeconomic Effects." IMF Staff Papers, 44(2), 1997, 210-48.
Alesina, A., and E. Spolaore. "On the Number and Size of Nations." Quarterly Journal of Economics, 112(4), 1997, 1027-56.
Alesina, A., and R. Wacziarg. "Openness, Country Size and Government." Journal of Public Economics, 69(3), 1998, 305-21.
Alesina, A., R. Baqir, and W. Easterly. "Redistributive Public Employment." Journal of Urban Economics, 48(2), 2000, 219-41.
Alesina, A., D. Arnaud, E. William, K. Sergio, and W. Romain. "Fractionalization." Journal of Economic Growth, 8(2), 2003, 155.
Alesina, A., F. Campante, and G. Tabellini. "Why Is Fiscal Policy Often Procyclical?" Journal of the European Economic Association, 6(5), 2008, 1006-36.
Aschauer, D. A. "Is Public Expenditure Productive?" Journal of Monetary Economics, 23(2), 1989, 177-200.
Baltagi, B. H. Econometric Analysis of Panel Data. 3rd ed. Chichester, UK: John Wiley and Sons, 2005.
Baqir, R. "Districting and Government Overspending." Journal of Political Economy, 110(6), 2002, 1318-54.
Barro, R. "On the Determination of the Public Debt." The Journal of Political Economy, 87(5), 1979, 940.
--. "Government Spending in a Simple Model of Endo geneous Growth." The Journal of Political Economy, 98(5), 1990, S103-25.
--. "A Cross-Country Study of Growth, Saving, and Government," in National Saving and Economic Performance, edited by B. D. Bernheim and J. Shoven. Chicago: University of Chicago Press, 1991, 269-301.
Beck, N. "Time-Series-Cross-Section Data: What Have We Learned in the Past Few Years?" Annual Review of Political Science, 4(1), 2001, 271-93.
--. "Time-Series Cross-Section Methods," in The Oxford Handbook of Political Methodology edited by J. M. Box-Steffensmeier, H. E. Brady, and D. Collier. Oxford Handbooks Online. Oxford: Oxford University Press, 2008, 475-93.
Beck, N., and J. N. Katz. "What to Do (and Not to Do) with Time-Series Cross-Section Data." The American Political Science Review, 89(3), 1995, 634-47.
Bird, R. "Wagner's Law of Expanding State Activity." Public Finance, 26(1), 1971, 1-26.
Boix, C. "The Public Sector," in Global Change and East Asian Policy Initiatives, edited by S. Yusuf, M. A. Altai, and K. Nabeshima. Washington, DC: World Bank and Oxford: Oxford University Press, 2004, 239-92.
Bueno de Mesquita, B., A. Smith, R. Siverson, and J. Morrow. The Logic of Political Survival. Cambridge, MA: MIT Press, 2003.
Caballero, R., and A. Krishnamurthy. "Fiscal Policy and Financial Depth." NBER Working Paper No. 10532, 2004.
Cameron, D. R. "The Expansion of the Public Economy: A Comparative Analysis." American Political Science Review, 72(4), 1978, 1243-61.
De Haan, J., J. Sturm, and B. Sikken. "Government Capital Formation: Explaining the Decline." Review of World Economics, 132(1), 1996, 55-74.
De Hoyos, R., and V. Sarafidis. "Testing for Cross-Sectional Dependence in Panel-Data Models." Stata Journal, 6(4), 2006, 482.
Deacon, R. "Public Good Provision Under Dictatorship and Democracy." Public Choice, 139(1), 2009, 241-62.
Devarajan, S., V. Swaroop, and H.-f. Zou. "The Composition of Public Expenditure and Economic Growth." Journal of Monetary Economics, 37(2), 1996, 313-44.
Durevall, D., and M. Henrekson. "The Futile Quest for a Grand Explanation of Long-run Government Expenditure." Journal of Public Economics, 95(7), 2011, 708-22.
Easterly, W., and R. Levine. "Africa's Growth Tragedy: Policies and Ethnic Divisions." Quarterly Journal of Economics, 112(4), 1997, 1203-50.
Fan, S., and N. Rao. "Public Spending in Developing Countries: Trends, Determination and Impact." Working Paper. International Food Policy Research Institute, 2003.
Frees, E. "Assessing Cross-Sectional Correlation in Panel Data." Journal of Econometrics, 69(2), 1995, 393-414.
--. Longitudinal and Panel Data: Analysis and Applications in the Social Sciences. New York: Cambridge University Press, 2004.
Freytag, A., J. J. Kruger, D. Meierrieks, and F. Schneider. "The Origins of Terrorism: Cross-Country Estimates of Socio-Economic Determinants of Terrorism." European Journal of Political Economy, 27(suppl 1), 2011, S5-16.
Friedman, M. "The Use of Ranks to Avoid the Assumption of Normality Implicit in the Analysis of Variance." Journal of the American Statistical Association, 32(200), 1937, 675-701.
Gavin, M., and R. Perotti. "Fiscal Policy in Latin America." NBER Macroeconomics Annual, 12, 1997, 11-61.
Gould, D., and J. Amaro-Reyes. "The Effects of Corruption on Administrative Performance." World Bank Staff Working Paper No. 580, 1983.
Gupta, S. "Public Expenditure and Economic Development: A Cross-Section Analysis." FinanzArchiv, 28(1), 1968, 26-41.
Hall, R. "Intertemporal Substitution in Consumption." The Journal of Political Economy, 96(2), 1988, 339-57.
Hausman, J. A. "Specification Tests in Econometrics." Econometrica, 1978, 1251-71.
Ilzetzki, E., and C. Vegh. "Procyclical Fiscal Policy in Developing Countries: Truth or Fiction?" NBER Working Paper, 2008.
Kandil, M. "On the Effects of Government Spending Shocks in Developing Countries." Oxford Development Studies, 33(2), 2005, 269-304.
Kaufmann, D., A. Kraay, and M. Mastruzzi. "Governance Matters IV: Governance Indicators for 1996-2004." Policy Research Working Paper No. 3630, 2005.
--. "Governance Matters VIII: Governance Indicators for 1996-2008." World Bank Policy Research, 2009.
Kirchgassner, G. "The Effects of Fiscal Institutions on Public Finance: A Survey of the Empirical Evidence," in Political Economy and Public Finance, edited by S. L. Winer and H. Shibata. Cheltenham, UK: Edward Elgar, 2002, 145-77.
Kirchner, S. "Federal Legislative Activism in Australia: A New Approach to Testing Wagner's Law." Public Choice, 153(3-4), 2012, 375-92.
Klitgaard, R. Controlling Corruption. Berkeley: University of California Press, 1988.
Lake, D., and M. Baum. "The Invisible Hand of Democracy: Political Control and the Provision of Public Services." Comparative Political Studies, 34(6), 2001, 587-621.
Learner, E. "Let's Take the Con Out of Econometrics." American Economic Review, 73(1), 1983, 31-43.
Leff, N. "Economic Development Through Bureaucratic Corruption." American Behavioral Scientist, 8, 1964, 8-14.
Leonardi, R., R. Y. Nanetti, and R. D. Putnam. Making Democracy Work: Civic Traditions in Modern Italy. Princeton, NJ: Princeton University Press, 2001.
Levine, R., and D. Renelt. "A Sensitivity Analysis of Cross-Country Growth Regressions." American Economic Review, 82(4), 1992, 942-63.
Marshall, M., and K. Jaggers. "Polity IV Project: Political Regime Characteristics and Transitions, 1800-2001." Mimeo, University of Maryland, Center for International Development and Conflict Management (CIDCM), College Park, MD, 2003.
Mauro, P. "Corruption and Growth." Quarterly Journal of Economics, 110(3), 1995, 681-712.
--. "Corruption and the Composition of Government Expenditure." Journal of Public Economics, 69(2), 1998, 263-79.
McGuire, M. C., and M. Olson. "The Economics of Autocracy and Majority Rule: The Invisible Hand and the Use of Force." Journal of Economic Literature, 34(1), 1996, 72-96.
Milesi-Ferretti, G. M., R. Perotti, and M. Rostagno. "Electoral Systems and Public Spending." Quarterly Journal of Economics, 117(2), 2002, 609-57.
Mueller, D. C., and T. Stratmann. "The Economic Effects of Democratic Participation." Journal of Public Economics, 87, 2003, 2129-55.
Murray, M. "Avoiding Invalid Instruments and Coping with Weak Instruments." Journal of Economic Perspectives, 20(4), 2006, 111-32.
Musgrave, R. Fiscal Systems. New Haven, CT: Yale University Press.
Niskanen, W. A. "Autocratic, Democratic, and Optimal Government." Economic Inquiry, 35(3), 1997, 464-79.
North, D. Growth and Structural Change. New York: Norton, 1981.
Persson, T, and G. Tabellini. "The Size and Scope of Government: Comparative Politics with Rational Politicians." European Economic Review, 43(4-6), 1999, 699-735.
Persson, T., G. Roland, and G. Tabellini. "Separation of Powers and Political Accountability." Quarterly Journal of Economics, 112(4), 1997, 1163-202.
--. "Comparative Politics and Public Finance." Journal of Political Economy, 108, 2000, 1121-61.
--. "Electoral Rules and Government Spending in Parliamentary Democracies." Quarterly Journal of Political Science, 2(2), 2007, 155-88.
Pesaran, M. H. "General Diagnostic Tests for Cross Section Dependence in Panels." IZA Discussion Paper No. 1240, 2004, Bonn, Germany.
Ram, R. "Government Size and Economic Growth: A New Framework and Some Evidence from Cross-Section and Time-Series Data." American Economic Review, 76(1), 1986, 191-203.
Razin, A., E. Sadka, and P. Swagel. "The Aging Population and the Size of the Welfare State." Journal of Political Economy, 110(4), 2002, 900-18.
Remmer, K. "Does Foreign Aid Promote the Expansion of Government?" American Journal of Political Science, 48(1), 2004, 77-92.
Riascos, A., and C. A. Vegh. "Procyclical Government Spending in Developing Countries: The Role of Capital
Market Imperfections." Unpublished. Washington: International Monetary Fund, 2003.
Rodrik, D. "Why Do More Open Economies Have Bigger Governments?" Journal of Political Economy, 106(5), 1998,997-1032.
Sala-i Martin, X. X. "I Just Ran Four Million Regressions." Technical Report, National Bureau of Economic Research, 1997
--. "I Just Ran Two Million Regressions." American Economic Review, 87(2), 1997, 178-83.
Sala-i Martin, X., G. Doppelhofer, and R. Miller. "Determinants of Long-Term Growth: A Bayesian Averaging of Classical Estimates (Bace) Approach." American Economic Review, 94(4), 2004, 813-35.
Sanz, L, and F. Velazquez. "Determinants of the Composition of Government Expenditure by Functions." European Economy Group Working Paper No. 13/2002, 2002.
Saunders, P., and F. Klau. "The Role of the Public Sector: Causes and Consequences of the Growth of Government." OECD Journal Economic Studies, 4, 1985, 5-239.
Schmidt, M. G. "Why Governments Grow: Measuring Public Sector Size," in The Growth of the Tax State: The Industrial Democracies, 1950-1978, edited by C. L. Taylor. Beverly Hills, CA: Sage, 1983.
Shelton, C. "The Size and Composition of Government Expenditure." Journal of Public Economics, 91(11-12), 2007,2230-60.
Shleifer, A., and R. W. Vishny. "Corruption." Quarterly Journal of Economics, 108(3), 1993, 599-617.
Sturm, J.-E. "Determinants of Public Capital Spending in Less-Developed Countries." Working Paper, 2001.
Talvi, E., and C. Vegh. "Tax Base Variability and Procyclical Fiscal Policy in Developing Countries." Journal of Development Economics, 78(1), 2005, 156-90.
Tanzi, V., and H. Davoodi. "Corruption, Public Investment, and Growth," in The Welfare State, Public Investment, and Growth, edited by H. Shibata and T. Ihori. Tokyo: Springer Japan, 1998, 41-60.
Tornell, A., and P. Lane. "Are Windfalls a Curse? A Non-Representative Agent Model of the Current Account." Journal of International Economics, 44(1), 1998, 83-112.
--. "The Voracity Effect." American Economic Review, 89(1), 1999, 22-46.
Tridimas, G., and S. Winer. "The Political Economy of Government Size." European Journal of Political Economy, 21(3), 2005, 643-66.
Welsch, H. "Corruption, Growth, and the Environment: A Cross-Country Analysis." Environment and Development Economics, 9(5), 2004, 663-93.
World Bank. "Annual Report." 1994. Accessed December 28, 2015. http://documents.worldbank.org/curated/en/1994/ 08/698118/world-bank-annual-report-1994
Yaffee, R. "A Primer for Panel Data Analysis." Connect: Information Technology at NYU, New York, 2003. Accessed December 28, 2015. http://web.pdx.edu/~crkl/ec510/pda_yaffee.pdf
Yogo, M. "Estimating the Elasticity of Intertemporal Substitution When Instruments are Weak." Review of Economics and Statistics, 86(3), 2004, 797-810.
(1.) Details of WGI indicators as well as the disaggregated underlying indicators are available at Kaufmann, Kraay, and Mastruzzi (2009) and www.govindicators.org.
(2.) In political economy research, it is quite common to encounter a case where the dependent variable is affected by a variable which varies across units but remains constant over time.
(3.) We used xtcsd routine in STATA which is developed by De Hoyos and Sarafidis (2006) to check such assumptions in STATA.
(4.) It could be possible that some of the explanatory variables are correlated; however, our pairwise variables did not reveal such concern among the family of variables.
(5.) For example, see Barro (1979, 1990, 1991).
(6.) The maximum number of imputations done for any country for any variable is 3 years. If the data for more than 3 years are missing, we have dropped the country. Imputation has been done only for the government effectiveness and control of corruption variables. As a precautionary exercise and to check the robustness of our estimate, we dropped all the imputed missing data from our sample and re-estimated our main regression specifications which turned out to be largely consistent with our main findings.
(7.) Except for the specification in the column 7 of Table A3 which could be due to the lagged 3-year average data that decreases the observations to one-third.
(8.) Developed by the Development Research Group at the World Bank.
(9.) Details of CPI are available at www.transparency.org.
ABU S. SHONCHOY *
* I am extremely thankful to Denzil Fiebig, Kevin J. Fox, Raghbendra Jha, and Adrian Pagan for showing interests and for providing numerous ideas to fulfill this research. My heartfelt thanks go to them. I also benefited from insightful discussions with the participants of the Ajiken Power Lunch seminar at IDE-JETRO, Chiba, Japan and the GRIPS Seminar in Economics at National Graduate Institute for Policy Studies, Tokyo, Japan. I thank two anonymous referees for their extremely useful comments. Usual disclaimers apply. An earlier version of this article circulated under the title of "Determinants of Government Consumption Expenditure in Developing Countries: A Panel Data Analysis," available at http://www.ide.go.jp/English/Publish/Download/Dp/266.html.
Shonchoy: Research Fellow, Development Studies Center, Institute of Developing Economies, JETRO, 3-2-2 Wakaba, Mihama-ku, Chiba-shi, Chiba 261-8545, Japan. Phone 81-43-299-9695, Fax 81-43-299-9548, E-mail firstname.lastname@example.org
|Printer friendly Cite/link Email Feedback|
|Author:||Shonchoy, Abu S.|
|Publication:||Contemporary Economic Policy|
|Date:||Oct 1, 2016|
|Previous Article:||Effectiveness of monetary policy in Sweden.|
|Next Article:||Job reallocation, financial crises, and policy reforms: evidence from Korea.|