The impact of corruption upon economic growth in the U.E. countries

The impact of corruption upon economic growth in the U.E. countries MIHAI DANIEL ROMAN mihai.roman@ase.ro MADALINA ECATERINA ANDREICA National Scientific Research Institute for Labour and Social Protection 6-8 Povernei Str., District 1, Bucharest madalina.andreica@gmail.com MADALINA ROXANA ONUTA madalina.onuta@gmail.com Abstract: The study consists in building a panel data model in order to explain the economic growth variations of the ten newest E.U. members based on macroeconomic variables, governance and corruption indicators for the period 2002-2011. One of the questions for which an answer was searched consisted in deciding whether corruption reduces the negative effect of poor governance and increases investment and economic growth successively, or if it actually increases the negative effect of a poor governance, leading to a decrease in investment and economic growth. Key-Words: corruption, economic growth, E.U. countries, panel data model 1 Introduction The purpose of this paper consists in studying the economic growth rate evolution of the newest ten European Union members that have joined the union during 1981 and 2007. In order to do that we built a panel data model which allowed us to explain the economic growth variations of the ten newest E.U. countries based on macroeconomic variables, governance and corruption indicators for the period 2002-2011. The corruption and governance indicators were especially used with the purpose of finding whether corruption reduces the negative effect of poor governance and increases investment and economic growth successively, or if it actually increases the negative effect of a poor governance, leading to a decrease in investment and economic growth. Even if the answer seems obvious, this issue is highly debated among economists. The corruption phenomenon proves to be a difficult field to study. This is caused by the obstacles in measurement and the ambiguities of its definition [1] Shleifer and Vishny [2] point out the fact that there are only few economic studies of corruption, mainly focusing on a principal-agent control problem, where the principal is represented by the top level of the public sector and the agent is represented by the officials who take bribes. Therefore, the economic theory of corruption is associated with the economics of crime, where the incentives for an official subordinate to agree and to participate to an illegal activity are studied. Generally, corruption is seen as an obstacle to economic development and growth, as it is supported by the recent results of the literature. For instance, Mauro [3] observed a significant negative relationship between corruption and investment. His results were also confirmed by Brunetti and Weder [4] and later by Mo [5]. As a result to these findings, the international organizations (e.g. the IMF, the World Bank, the UN or the OECD) had to set higher priorities in ISBN: 978-960-474-305-6 76

fighting corruption, by taking international initiatives, such as: the UN resolution (1998) or the OECD s Convention on combating bribery (1999). However, there are still researchers with opposite opinions. For example, Leys [6] had doubts about seeing corruption as a problem, while Bardhan [1] reminded some situations from the history of Europe and the U.S.A. where corruption seemed to have favoured development and therefore, growth, allowing entrepreneurs to grow out of bribes. Moreover, Lien [7] argued that corruption may improve the efficiency of investment, justifying that in countries where the governance is poor, bribes may act as a device of trouble-saving, because it may raise the efficiency of investments. This hypothesis was also put forward by Huntington [8] arguing that corruption may be beneficial for emerging countries. An inefficient bureaucracy represents an obstacle to investment, so he considers that some speed money may be a cure for this disease of bureaucracy. The empirical evidence regarding the negative impact of corruption upon growth is not in disagreement with the statement above. According to the hypothesis, corruption may be useful only in countries where other characteristics of governance are ineffective, otherwise it is still harmful. Corruption is generally related to lower growth and investment, but there is no proof of any correlation with the governance s quality. In fact, the existing evidence does not allow a rigorous rejection of the hypothesis. Mauro [3] tried to explain this issue by splitting his sample in two parts: first - countries where bureaucracy is excessive, and second - a sub-sample characterized by low level of bureaucracy. The result pointed out no significant difference between the two sub-samples regarding the negative impact of corruption. However, the size of the two subsamples was too small to admit the inclusion of control variables, because the threshold for splitting the sample was approximately arbitrary (score of 5 or 7 of the bureaucracy index). More recently, the corruption phenomenon became an interesting subject for economists, and is highly debated in various studies. For exemple, Guillaumemeon and Sekkat [9] studied the relationship between the impact of corruption on growth and investment and the quality of governance in a sample of 63 to 71 countries between 1970 and 1998. Like previous studies, they found a negative effect of corruption on both growth and investment. However, they also noticed that corruption has a negative impact on growth independently from its impact on investment. These impacts were different depending on the quality of governance and tend to worsen when indicators of the quality of governance deteriorate. This supports the sand the wheels view on corruption and contradicts the grease the wheels view, which postulates that corruption may help compensate bad governance. In our paper we raise the same issue concerning the impact of corruption on economic growth upon the latest countries that joined the E.U. The paper is structured as follows: section 2 offers a data description, section 3 describes the methodology required for building the panel data model, while the results of the study are presented in section 4. Section 5 concludes. 2 Data description For this study, annual data for the period 2002 2011 for ten E.U. countries were collected from the Growth Development Network of the World Bank s web site. The sample consisted in the latest ten countries that successively joined the European Union as follows: Greece (in 1981), Portugal and Spain (in 1986), Poland, Czech Republic, Hungary, Slovakia and Slovenia (in 2004), as well as Romania and Bulgaria (in 2007). The selection of the main set of macroeconomic variables for each E.U. country was conditioned by those variables that appeared in most empirical work, but also restricted to the availability of the macroeconomic data. Twelve variables were therefore used in the analysis as presented in table 1, in order to reflect not only macroeconomic performances of each country but also the corruption level and several governance policy aspects. All values of the macroeconomic indicators are reported in the prices of the year 2000 and are calculated as growth rates. When considering the corruption indicator, we decided upon the indicator of corruption perception (named CPI) that was available on the Transparency International s web site. ISBN: 978-960-474-305-6 77

Table 1. Macroeconomic variables CATEGORY CODE VARIABLES EXPLANATIONS Macroeconomic Corruption Governance rpib GDP per capita Used as a proxy for economic growth rate growth rate rvenit The growth rate of income per capita Used the growth rate of household consumption per capita as a proxy for income growth rate nivsc The enrolment rate The enrolment rate at the tertiary level was used as a proxy for human capital rpop The growth rate of Annual % population rinv The investment Calculated as percentage of GDP growth rate grpiata Market trade openness The ratio of exports plus imports to GDP, used as a proxy for the exposure of the economy to foreign markets CPI The Corruption Calculated as annual average of other indicators, with Perception Index values between zero and ten, where 10 corresponds to the vr Lvpol EG grregl Voice and accountability Lack of political violence Government effectiveness The degree of regulation of the economy absence of corruption Measures the degree to which a country's citizens are able to participate in selecting government and plays the role of a proxy for the openness of the political system Measures the perceptions of the likelihood that the government in power will be destabilized or overthrown by unconstitutional/or violent means. It provides an assessment of the political risk of a country. Measures the perceptions of quality of service delivery, quality of bureaucracy, the competence of civil servants, the independence of public service to political pressures and the credibility of the government. Measures the incidence of unfriendly market policies such as price controls or inadequate bank supervision, as well as the perceptions upon fiscal burden caused by excessive regulations statdr Rule of law Measures the extent to which agents have confidence in law and abide by the rules of the society 3 Models and methodologies The econometric study is based on panel data estimation, using STATA software. A panel data regression has the following general form: y x ' it i it it (1) where i=1...n, t=1... T. In a panel data estimation, the individual effects can be either assumed to be correlated with explanatory variables as in the case of fixed effects model (FEM) or to be incorporated into the error term for the case of random effects model (REM) and assumed uncorrelated with the explanatory variables [10-11]. In our estimation, we assumed the presence of fixed effects between the E.U. countries and tested the assumption that a FEM is more appropriate than a REM using the Hausman test. For the FEM the most used estimator is the within estimator. Moreover, when estimating a panel data model, one must check the validity of the assumptions concerning the absence of both heteroskedasticity and serial correlation of the idiosyncratic error term [12]. When heteroskedasticity is present the standard errors of the estimates will be biased and one needs to compute robust standard errors. Another problem is the serial correlation of the idiosyncratic error term which leads to biased estimates, but Wooldridge proposed a simple test to check the autocorrelation of the residuals [13]. In order to overcome these problems, one should estimate the regression model using robust standard errors [11, 14-16] which is implemented in the STATA software. 4 Results of the analysis The paper aimed to study the economic growth rate evolution of the newest European Union members in all stages of its expansion between 1981 and 2007. ISBN: 978-960-474-305-6 78

The econometric analysis allowed explaining the economic growth rate variations of the E.U. countries based on macroeconomic variables, governance indicators and the current level of corruption registered in each country. The general form of the GDP per capita growth rate equation estimated using a panel data model is the following: rpib it = β 0 + β 1 rvenit it + β 2 nivsc it + +β 3 rpop it + +β 4 rinv it + +β 5 grpiata it + β 6 CPI it + +β 7 Vr it + β 8 Lvpol it +β 9 EG it + +β 10 grregl it + β 11 statdr it Out of the all explanatory variables used in the panel data model estimation, only four turned out to be statistically significant. More precisely, the Perception Index (CPI) and the indicator of the growth rate of income per capita (rvenit), the investment growth rate (rinv), the Corruption Rule of law (statdr) which measures the extent to which agents have confidence in law and abide by the rules of the society, were the only variables that can explain the variation of the GDP per capita growth rate for the 10 newest European Union members.when running the Hausman test in order to decide whether a RE model is more appropriate than a FE model, the probability was less than 5%. Concluding that we are dealing with fixed-effects, we estimated the model using the within estimator. When performing both the modified Wald test for groupwise heteroskedasticity in the FE model, implemented in Stata by Baum [11] and the serial correlation test proposed by Drukker [14], it resulted that the errors were both autocorrelated and heteroskedastic. That is why, in order to ensure the validity of the statistical results, we had to estimate a robust fixed-effects (within) regression with Driscoll and Kraay standard errors (see fig. 1). Regression with Driscoll-Kraay standard errors Number of obs = 91 Method: Fixed-effects regression Number of groups = 10 Group variable (i): Col1 F( 4, 9) = 9.19 maximum lag: 2 Prob > F = 0.0031 within R-squared = 0.2503 Drisc/Kraay Rpib Coef. Std. Err. t P> t [95% Conf. Interval] Fig. 1. The robust panel data estimation output (2) Rvenit.2231944.0638588 3.50 0.007.0787357.3676532 rinv.0062549.0022415 2.79 0.021.0011843.0113256 cpi 2.384594.6720236 3.55 0.006.8643709 3.904817 StatDr -14.39015 5.780557-2.49 0.034-27.46668-1.313627 _cons -.6906764 2.019888-0.34 0.740-5.25998 3.878628 The new GDP per capita growth rate equation obtained after the robust fixed-effects estimation with Driscoll-Kraay standard errors is the following: rpib it = 0. 691 + 0. 2 rvenit it +0. 0063 rinv it + 2. 385 CPI it 14. 39 statdr it (3) As expected, according to economic theory, the investment rate has a positive impact on the economic growth rate, even if, for the countries selected in the study, the impact is quite small (the magnitude of the coefficient of investment rate is approximately 0.0063). Therefore, on average, an increase of one percentage of the investment rate leads to an increase of only 0.006 percentage points on the economic growth rate. Regarding the income rate, there is also a positive impact upon economic growth rate. The coefficient is significant, having a value of 0.223. Thus, a 1% increase of the income rate leads on average to an increment of 0.223 % in the economic growth rate, keeping all other variables constant. The corruption perception index is a qualitative variable and has a reduced value (closer to 0). The selected sample of countries that we analysed is characterised by a high level of corruption, as 2.385 is not even close to the average value of corruption index (which is 5). Therefore, corruption directly influences economic growth. This is consistent with Mo s research, who found a significant relationship between corruption and growth when controlling for the investment ratio, but the coefficient of corruption becomes insignificant when he takes into account the human capital. Our results contradicts, however, Mauro s results [3], who found no significant relationship between corruption and growth, when investment was included in the regression as an explanatory variable. Another qualitative variable used in our study represents the governance indicator regarding the rule of law. The negative value suggests that, for our sample of selected countries, corruption has a strong and negative impact on growth. Hence, corruption does not act as a troublesaving device, fighting with the inferences of a poor governance (ineffective administration or cumbersome bureaucracy) and it shows that it tends to worsen the negative impact on investment and ISBN: 978-960-474-305-6 79

growth. In other words, out regression s results tend to reject the hypothesis that corruption reduces the negative effect of a bad governance, thereby raising investment and eventually, economic growth. Thus, reducing corruption is most advisable to countries experiencing a weak rule of law. 5 Conclusion In this paper we tried to clarify the interaction between corruption, investment and economic growth and other dimensions of governance in order to test whether corruption reduces the negative effect of poor governance and increases investment and economic growth successively, or if it actually increases the negative effect of a poor governance, leading to a decrease in investment and economic growth. The econometric results reject the first hypothesis in favour of the assumption that a reduction of corruption could bring more benefits in countries where other aspects of governance are weak. Our results contradict the view of those who see corruption as a way to minimize the effect of these deficiencies. Therefore, corruption does not appear as a way to counterbalance a poor governance (such as ineffective administration or cumbersome bureaucracy) but rather as a way to make it more harmful. Since the investments positively affect the economic growth, the economic policies should include in their strategy incentives for investment, subsidies and tax reductions for entrepreneurs. Moreover, since the governance indicator has a negative impact on economic growth, the analyzed states should promote policies that would earn the trust of their citizens. In future research the study of the main channels through which corruption has an impact on economic growth may become a challenging subject of interest. different uncertainty measures Weltwirtschaftliches Archiv 134: 513 533, 1998. [5] Mo, P.H. Corruption and economic growth, Journal of Comparative Economics 29:66 79, 2001 [6] Leys, C. What is the problem about corruption? Journal of Modern African Studies 3: 215 230, 1965. Reprint in A.J. Heidenheimer, M. Johnston and V.T. LeVine (Eds.), Political corruption: A handbook, 51 66, 1989. Oxford: Transaction Books. [7] Lien, D.H.D. A note on competitive bribery games Economics Letters 22: 337 341, 1986. [8] Huntington, S.P. Political order in changing societies, New Haven: Yale University Press, 1968 [9] Guillaumemeom și Khalid Sekkat, Does corruption grease or sand the wheels of growth?, Public Choice 122: 69 97, 2005 [10] B. H. Baltagi, Econometric Analysis of Panel Data, John Wiley & Sons Ltd, 2008. [11] C. F. Baum, Residual diagnostics for crosssection time series regression models, The Stata Journal, vol. 1, pp. 101 104, 2001. [12] A.C.Cameron,P.K.Trivedi, Microeconometric s Using Stata, Stata Press, 2009. [13] J.M. Wooldridge, Introductory econometrics A modern approach, Second Edition, pp. 501-528, 2002 [14] D. M. Drukker, Testing for serial correlation in linear panel-data models, The Stata Journal, vol. 3, pp. 168 177, 2003. [15] D. Hoechle, Robust Standard Errors for Panel Regressions with Cross-Sectional Dependence, The Stata Journal, vol. 7, pp. 281-312, 2007. [16] R. M. Kunst, Econometric Methods for Panel Data Part II, 2009, Available: http://homepage.univie.ac.at/robert.kunst/panels2e.pdf http://www.transparency.org/ http://data.worldbank.org References: [1] Bardhan, P., Corruption and development: A review of issues, Journal of Economic Literature 35: 1320 1346, 1997. [2] Shleifer, A. and Vishny, R.W. Corruption, Quarterly Journal of Economics 108: 599 617, 1993 [3] Mauro, P., Corruption and growth, Quarterly Journal of Economics 110: 681 712, 1995. [4] Brunetti, A. and Weder, B., Investment and institutional uncertainty: A comparative study of ISBN: 978-960-474-305-6 80