Relationship between government expenditure and economic growth in Visegrad Group

Szarowska, Irena Article Relationship between government expenditure and economic growth in Visegrad Group Financial Internet Quarterly Provided in Cooperation with: University of Information Technology and Management, Rzeszów Suggested Citation: Szarowska, Irena (2022) : Relationship between government expenditure and economic growth in Visegrad Group, Financial Internet Quarterly, ISSN 2719-3454, Sciendo, Warsaw, Vol. 18, Iss. 4, pp. 12-22, https://doi.org/10.2478/fiqf-2022-0024 This Version is available at: https://hdl.handle.net/10419/329825 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/3.0/ 10.2478/fiqf-2022-0024 Abstract The goal of the article is to examine the relationship between government expenditure and economic growth in the Visegrad Group in the period 2000-2020. Economic theory as well as published studies have suggested that government expenditure is an important factor in ensuring economic growth, which has been the subject of increased interest in recent years. The article focuses on development of government expenditure, changes in its composition in individual countries during the analyzed period, and also on the direction of influence between these variables. Real GDP time series were cyclically adjusted as well as annual data on government expenditure which were used in compliance with the COFOG international standard. The results suggest that government expenditure and their composition are similar in the Visegrad Group despite the existing differences in the size of the public sector. On the other hand, the cyclicality of government expenditure differs across the countries. Results suggest that countries of the Visegrad Group did not use government expenditure as a stabilizer in the monitored period. Government expenditure was acyclical in Slovakia and procyclical in other countries of the Visegrad Group. Applied Granger causality methodology provides mixed conclusions about the relationship between government expenditure and economic growth. Nevertheless, unidirectional Granger causality from GDP growth to government expenditure and its divisions predominates, which means that economic growth comes first, followed by government expenditure. JEL classification: C32, H50, E62 Keywords: government expenditure, COFOG classification, cyclicality Received: 25.08.2022 Accepted: 13.10.2022 Cite this: Szarowska I. (2022) Relationship between government expenditure and economic growth in Visegrad Group. Financial Internet Quarterly 18 (4), pp. 12-22. © 2022 Irena Szarowska, published by Sciendo This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License. 1 Silesian University in Opava, Czech Republic, e-mail: [email protected], ORCID: https://orcid.org/0000-0002-5197-6455. ture, there are two opposing views on the relationship between government expenditure and economic growth. The Keynesian view claims that government expenditure is an exogenous factor that affects economic growth and can be used as a policy instrument. On the other hand, Wagner's view states that government expenditure is an endogenous factor or result (not a cause) of economic growth (Romer, 1986; Nyasha & Odhiambo, 2019). From a Keynesian perspective, government expenditure should act as a stabilizer and move in a counter-cyclical direction. Procyclical fiscal policy is conversely policy expansionary in booms and contractionary in recessions. Servén (2013) draws attention to the fact that procyclical fiscal policy is generally regarded as potentially damaging for welfare. If a government respects the basic prescription that fiscal tools should function in a manner which is counter-cyclical, the optimal fiscal policy involves decreasing of government expenditure in “good times” and increasing of government expenditure in “bad times.” Contrary to the theory, a number of studies found evidence that government expenditure is procyclical (Fiorito & Kollintzas, 1994; Lane, 1998; Hercowitz & Strawczynski, 2004; Alesina et al., 2008; Rajkumar & Swaroop, 2008; Woo, 2009). Lane (2003) writes that the level of cyclicality varies across spending categories and across OECD countries. Later analysis of Talvi and Vegh (2005) shows that fiscal procyclicality is evident in a much wider sample of countries. Alesina et al. (2008) formulate the conclusion that the procyclicality of fiscal policy is more pronounced in more corrupt democracies. Abbott and Jones (2011) tested differences in the cyclicality of government expenditure across functional categories. Their evidence from 20 OECD countries suggests that procyclicality is more likely in smaller functional budgets, but capital spending is more likely to be procyclical for the larger spending categories Despite the fact that many studies have focused on the relationship between government expenditure and economic growth, it cannot be clearly stated that an increase in government expenditure has a positive effect on economic growth. Empirical studies provide mixed and unclear conclusions about the effect of government expenditure on economic growth, as studies with positive, negative, and even zero effects can be found. The first studies can be found in the 1980s when Romer (1986) confirmed a positive effect of a higher government expenditure creating higher economic growth as a result of expansionary fiscal policy. Alexiou (2009) empirically estimates the relationship between economic growth and government expenditure using Recently, the increasing interest of many governments is focused on the question of how to ensure economic growth. Government expenditure and factors behind its growth are a serious problem in many countries. The importance of this problem is related to exhausted public budgets after the covid pandemic and the need to ensure the development of economies in the future. It is important to note that government expenditure plays an important role in the fiscal policy of each country as a possible stabilizer as well as a tool for the implementation of individual government policies and the provision of public goods and services. In spite of the fact that many studies have been focused on the relationship between government expenditure and economic growth, it cannot be unequivocally claimed that increase of government expenditure has a positive effect on economic growth (for details look at Nyasha and Odhiambo, 2019). Previously published studies are weakly supported, especially by data from the Central European countries, therefore this analysis is focused on the Visegrad countries, in which the results may differ. The Visegrad Group, also called the Visegrad Four or V4, is an alliance of four Central European states – the Czech Republic (CZ), Hungary (HU), Poland (PL) and Slovakia (SK) – for the purposes of cooperation and furthering their European integration. All of the Visegrad countries have a developed free market economy and have enjoyed more or less steady economic growth since the revolutions of 1989. In 2009, Slovakia adopted the euro as official currency, the other countries have their national currencies. Despite their similar histories, the individual countries of the Visegrad Group have differently structured and oriented economies and social preferences, which are also reflected in the structure of government expenditure. The goal of the article is to examine the relationship between government expenditure and economic growth in the Visegrad Group in the period 2000-2020, including the direction of influence between the variables. The intention is also to analyze acyclicality and the structure of government expenditure according to COFOG classification. A country's economy is largely influenced by the level and the structure of government expenditure as government expenditure is an important tool for national governments for realizing individual government policies and to mitigate uneven economic development and economic shocks. In the theoretical scientific litera- ambiguous results across various countries focused on the relationship between public expenditure and economic growth and conclude the impact of government expenditure on economic growth is not clear cut. Alqadi and Ismail (2019) also assessed theoretical and empirical literature on the relationship between government expenditure and economic growth. The article finds that neither the theoretical literature nor the empirical literature provides conclusive evidence about the nature of this relationship. Many studies focus on African or South Asian economies, but these are not the subject of investigation and literature review due to very different characteristics. This article follows the study by Szarowská (2011, 2018) who analyzed the Czech Republic and V4 in a transition period which makes it possible to obtain information about the differences in the later development of these economies. The goal of the article is to examine the relationship between government expenditure and economic growth in the Visegrad Group in the period 2000-2020 which is why the methodology consists of several steps. Firstly, it is necessary to examine government expenditure, its composition and changes in the analyzed period. The dataset consists of annual data on GDP and government expenditure in compliance with the COFOG international standard in the period 2000-2020 (latest available data). All the data were collected from the Eurostat database. The time series for GDP, total government expenditure and its subcomponent are adjusted at constant prices. Most of the results were calculated in econometric program Eviews 12. Many studies point out that using a non-stationary macroeconomic variable in time series analysis causes superiority problems in regression. Thus, a unit root test should precede any empirical study employing such variables. We decided to make the decision on the existence of a unit root through the Augmented Dickey –Fuller test (ADF test). The equation (1) is formulated for the stationary testing: (1) ADF test is used to determine a unit root xt at all variables in the time t. Variable Δxt-i expresses the lagged first difference and ut estimate autocorrelation error. Coefficients δ0, δ1, δ2 and αi are estimated. Zero and the alternative hypothesis for the existence of a unit root in the xt variable are specified in (2). The result of the ADF test, which confirms the stationary of all time series on the first difference, is available on panel data model. The results proved that government expenditure had a strong positive impact on economic growth in the South Eastern European countries from 1995 to 2005. Afonso and Alves (2016) assess public expenditure and economic growth using data of 14 European countries over the period of 1996-2013. They found that some functions of government spending in Austria, France, the Netherlands and Portugal support the approach that an increase in national income causes more government spending. A contrary, negative impact appeared in a study of Lupu et al. (2018) who investigated the effect of public expenditure on economic growth in 10 selected Central and Eastern European countries. Their research found that public expenditures on defence, economic affairs, general public services, and social welfare had a negative impact on economic growth in the analyzed countries in the period 1995–2015. Pula and Elshani (2018a, 2018b) used a Granger causality test for identifying the direction of flow between variables in Kosovo (2002–2015) and discovered a unidirectional causality between government expenditures and economic growth. Sáez et al. (2017) also tested the relationship between government expenditure and economic growth using regression and panel data analysis. But they revealed mixed results for 15 selected European Union countries between 1994 and 2012. The effect of government expenditure on economic growth was found positive in Portugal, the United Kingdom, France, Greece and Luxembourg. On the other hand, the effect of government expenditure on economic growth was negative in Austria, Belgium, Denmark, Germany, Finland, Ireland, Italy, Netherlands, Spain and Sweden. Similarly, mixed conclusions were noted by Dudzevičiūtė et al. (2018) who indicated using correlation analysis and Granger causality testing, that eight European Union countries showed a significant relationship between 1995-2015. Specifically, a unidirectional causal relationship from economic growth to government expenditure was found in France, Belgium, Germany, Portugal and Cyprus; a uni-directional causal relationship from government expenditure to economic growth in Sweden and Slovakia, and no causality relationship was found between government expenditure and economic growth in Poland. Different results can be seen in a study of Roşoiu (2015) who found a bidirectional causality relation between government expenditure and economic growth using Granger causality methodology in Romania (1998 -2014). Nyasha and Odhiambo (2019) provided a comprehensive review of previous empirical evidence with 11() p t t j t j t j y t y Y      −− −  = + + +  +  empirical studies for testing relationships between economic growth and government expenditure (Zamanian et al., 2012; Roşoiu, 2015; Dudzevičiūtė et al., 2018; Lupu et al. 2018; Pula & Elshan, 2018a, 2018b; Nyasha & Odhiambo, 2019). That's why the Granger causality test is applied for modeling the relationship and for the estimation of causality between all COFOG government expenditure divisions and economic growth. The Granger causality test assumes that only time series data can expound the information needed for a relationship between variables (Gujarati 2003). A time series X is said to Granger-cause Y if it can be shown, usually through a series of t-tests and F-tests on lagged values of X (and with lagged values of Y also included), that those X values provide statistically significant information about future values of Y. Government expenditure can help in smoothing out cyclical fluctuations in the economy and influences the level of employment and price stability. Thus, government expenditure plays a crucial role in the economic growth of a country. Government expenditure can be classified into four categories: 1) Functional Classification or Budget Classification 2) Economic Classification 3) Cross Classification 4) Accounting Classification. Each classification of government expenditure serves one objective or other, i.e., financial control, economic growth, price stability, etc. Functional classification in compliance with the COFOG international standard (Classification of the Functions of Government) is used in this analysis. Government expenditure is divided into 10 basic divisions: C: Total function C10: General public services C20: Defense C30: Public order and safety C40: Economic affairs C50: Environment protection C60: Housing and community amenities C70: Health C80: Recreation; culture and religion C90: Education C100: Social protection Firstly, the total value and composition of government expenditure is analyzed in each individual country in the period 2000-2020. In Figure 1, data indicates a highly volatile development of government expenditure in individual countries in the monitored period. (2) The Hodrick-Prescott (HP) estimates an unobservable time trend for time series variables. Let yt denote an observable macroeconomic time series. The HP filter decomposes yt into a nonstationary trend gt and a stationary residual component ct, that is: (3) We note that gt and ct are unobservables. Given an adequately chosen, positive value of λ, there is a trend component that will minimize: (4) The first term of the equation is the sum of the squared deviations which penalizes the cyclical component. The second term is a multiple λ of the sum of the squares of the trend component’s second differences. This second term penalizes variations in the growth rate of the trend component. The larger the value of λ, the higher the penalty. Hodrick and Prescott advise that, for annual data, a value of λ = 100 is reasonable. Next, cross-correlation to all combinations of GDP – total value of government expenditure is applied. Cross correlation assesses how one reference time series correlates with another time series, or several other series, as a function of time shift (lag). Cross correlation does not yield a single correlation coefficient but rather a whole series of correlation values. Like all correlations, cross correlation only shows statistical associations not causation. Consider two financial series and, then the cross-correlation at lag (lead) k is defined as follows: (5) Where: p - correlation coefficient and mx, my - the means of corresponding series. The series can be related in three possible ways: 1) yt can lead xt(p(yt-k, xt) ≠ 0) 2) yt can lag xt(p(yt+k, xt) ≠ 0) 3) series can be related to each other (p(yt,xt)≠ 0) The standard Granger causality test developed by Granger (1980) is commonly used to test whether past changes in one variable help explain current changes in other variables as well as being used in the majority of empirical studies for testing relationships between eco- 0 2 2 : 0, : 0HH   = t t t y g c=+ 1 22 () ( )( ) ( , ) ( ) ( ) T t k y t x tk t k t TT Tk t k y t x t k t k T y m x m p y x y m x m + =− + ++ == −− = −−   1 22 11 12 ( ) (( ) ( )) TT t t t t t t tt y g g g g g  − +− == − + − − −  cation. Figure 2 briefly summarizes the structure of COFOG expenditures and its changes in each country at the beginning and end of the monitored period. The common trend is that Social protection (C100) has the highest average share of government expenditure across the V4 although the difference between the maximum (38.3% in Poland) and the minimum (30.1% in the Czechia) share of total government expenditure is more than eight percentage points. In the Czech Republic, the data demonstrates Social protection (C100) as the highest and the most stable item of government expenditure over the period (30% share in 2000 and 30.5% in 2020) though its share of total government expenditure is the lowest in the Visegrad Group. Economic affairs (C40) were second and Health expenditure (C70) in the third place till the year 2003. From 2004 the second and the third position of the largest share has changed (average share of C40 is 16.1% and C70 17.4%). Shares of percentage GDP demonstrate existing differences in the role of the public sector in the Visegrad countries. From this point of view, Hungary is the country with the highest role of government and its redistribution, although its role has also decreased (except in the pandemic period). Average share of government expenditure is 48.9% GDP. Contrarily, Slovakia is the country with the lowest share of government expenditure on GDP (average value is 42.1% GDP) and it means the smallest size of the public sector. The most volatile government expenditure (with average 42.7% GDP) was recorded in the Czech Republic in the selected period. Government expenditure is the most stable in Poland with average value of 43.7% GDP. Despite these variations, the final trend of a sharp increasing weight of government expenditure on GDP caused by the covid pandemic is common to the entire V4 Group. The intention is also to analyze the structure of government expenditure according to COFOG classifi- Figure 1: Development of total government expenditure in V4 (% GDP) Source: Author´s elaboration based on data from Eurostat. The results suggest that the composition of government expenditure is similar in the Visegrad Group despite the existing differences in the size of the public sector. Data confirms unstable and cyclical development of total government expenditure on GDP in all V4 members. Expenditure on Social protection C100 has the highest share of total government expenditure in all countries. Five of the highest COFOG expenditure divisions, on average, account for more than 80% of the total expenditure: Social protection (C100), Economic affairs (C40), Health (C70), General public services (C10) and Education (C90). As already mentioned, government expenditure is a possible stabilizer. From this point of view, government expenditure should move in a countercyclical direction. Time series are cyclically adjusted for the purpose of analysis. Firstly, logarithms of variables were calculated and then the cycle components were extracted using the Hodrick-Prescott filter. Figure 3 shows adjusted GDP and total government expenditure in each individual country of the Visegrad Group. Separate figures are available for each country on request. The order of the highest average shares on government expenditure is the following in Hungary: Social protection (C100 - average 31.7%), General public services (C10 - average 18.9%) and Economic affairs (C40 - average 14.3%). In Poland there can be seen a different structure of COFOG expenditures. The first position belongs also to Social protection (C100 41.1% in 2000 and 37.3% in 2020) and its share is higher than in the Czech Republic, Hungary and Slovakia. In 2000 and on average, the second position belongs to General public services (C10 – 14% in 2000, 12.4% on average) and the third to Education (C90 – 13.1% in 2000 and 12.4% on average). In 2020, the composition changed significantly due to the covid pandemic and the order was as follows: Social protection C100 (37.3%), Economic affairs C40 (18.7%) and Health C70 (11.1%). As already mentioned, Slovak total government expenditure is the lowest of all the analyzed countries. Social protection (C100 - average 34.4%) has the highest share of total government expenditure in all years, and its share increased from 27% in 2000 to 36.7% in 2020. General public services (C10 - 21.1% in 2000 and 13.5% in 2020), Economic affairs (C40 - 18% and 12.8%) and Health (C70 - 10.6% and 13.7%) alternately placed in the 2nd and the 3rd positions. Figure 2: Changes in structure of COFOG expenditure Source: Author´s calculations based on data from Eurostat. The results significantly vary across countries. In the Czech Republic, Hungary and Poland, correlation coefficients are positive and express procyclical development. Differences can be seen in the value of correlation coefficients – weak positive correlation is found in the Czech Republic, moderate procyclical development in Poland and strong procyclical government expenditure in Hungary. Only in Slovakia does a correlation coefficient describe a zero correlation between total government expenditure and GDP and it means that government expenditure is acyclical. Correlation is a statistical technique that can show whether and how strongly pairs of variables are related. The correlation coefficient can vary from -1 to +1. The correlation coefficient -1 indicates perfect negative correlation, and +1 indicates perfect positive correlation. Its value smaller than 0.1 means zero correlation, from 0.1 to 0.35 weak correlation, from 0.35 to 0.7 moderate correlation and higher than 0.7 expresses strong correlation. A positive correlation coefficient indicates the procyclicality of government expenditure, negative value means that variables are countercyclical, and values close to zero express acyclicality. Table 2 presents the summary of correlation analysis. Figure 3: Adjusted GDP and total government expenditure Source: Author´s calculations based on data from Eurostat. Country Correlation coefficient Correlation Cyclicality Czech 0.312 Weak positive Procyclical Hungary 0.884 Strong positive Procyclical Poland 0.645 Moderate positive Procyclical Slovakia 0.035 Zero correlation Acyclical Table 1: Cyclicality of GDP and total government expenditure Source: Author´s calculations. cyclicality has changed since the transition period, when a countercyclical movement was detected in Slovakia and as well as lower values of correlation coefficients in other countries. It can be stated that, compared to the expectations and recommendations of the theoretical literature, the countries of the Visegrad Group did not use government expenditure as a stabilizer in the monitored period. Moreover, compared to Szarowska´s papers (2011), does not by itself indicate causality in the more common use of the term. The null hypothesis should be rejected if probability is less than 0.05. Table 2 summarizes results for lags of one or two years (number of observations for each country is 20, resp. 19). GC means Granger cause; bold values indicate a hypothesis which should be rejected. The analysis presented in Table 2 shows the direction of causality between the analyzed macroeconomic variables. One can find two examples of unidirectional Granger causality from government expenditure to GDP growth, namely from Public order and safety (C30) and Recreation, culture and religion (C80) to GDP growth, both reported in Hungary for 1year lag, and so it provides support for the validity of the Keynesian view. Therefore, it can be stated that Public order and safety expenditure was a very important factor for the economic growth in Hungary in the period under review. The results of our evidence are parallel to findings of Alexiou (2009), Afonso and Alves (2016), Korkmaz and Güvenoğlu (2021) and Dudzevičiūtė et al. (2018) for Sweden and Slovakia, although our results of analysis differ in the case of Slovakia. The results of the empirical evidence are in line with the findings of other empirical studies which confirmed that government expenditure is procyclical (e.g. Fiorito & Kollintzas, 1994; Lane, 1998; Hercowitz & Strawczynski, 2004; Alesina et al., 2008; Rajkumar & Swaroop, 2008; Woo, 2009; Abbott & Jones, 2011; Servén, 2013). Results for Czech, Hungary and Poland confirm the statement of Talvi and Vegh (2005) that fiscal procyclicality is evident in a wider sample of countries. From this point of view, Slovak development is an exception compared to most of the previously analyzed countries. As mentioned above, the Granger causality refers to a specific notion of causality in time-series analysis. The Granger causality test is applied to a time-series data to determine the causality between total as well as individual COFOG government expenditure (C10…C) and economic growth (GDP). It is important to mention that the statement for example “C10 Granger causes GDP” does not imply that GDP is the effect or the result General public services (C10). Granger causality measures precedence and information content but Table 2: Results of Granger causality Null Hypothesis Czech (lag 2 years) Hungary (lag 1 year) Poland (lag 2 years) Slovakia (lag 1 year) F-Stat Prob. F-Stat Prob. F-Stat Prob. F-Stat Prob. GDP does not GC C10 3.412 0.062 8.766 0.009 4.351 0.034 17.691 0.001 C10 does not GC GDP 0.410 0.671 2.678 0.120 0.836 0.454 0.249 0.624 GDP does not GC C20 3.057 0.079 2.354 0.143 3.393 0.063 0.796 0.385 C20 does not GC GDP 0.419 0.666 0.038 0.847 0.828 0.457 0.078 0.783 GDP does not GC C30 3.934 0.044 0.010 0.921 6.310 0.011 14.896 0.001 C30 does not GC GDP 0.795 0.471 11.414 0.004 0.458 0.642 1.621 0.220 GDP does not GC C40 4.213 0.037 5.146 0.037 4.690 0.028 10.044 0.006 C40 does not GC GDP 0.131 0.879 0.138 0.715 3.958 0.043 0.073 0.790 GDP does not GC C50 4.142 0.039 4.567 0.047 2.900 0.088 18.248 0.001 C50 does not GC GDP 0.259 0.775 0.014 0.907 0.942 0.413 0.070 0.795 GDP does not GC C60 0.053 0.949 21.264 0.000 0.891 0.432 3.811 0.068 C60 does not GC GDP 1.874 0.190 3.043 0.099 1.261 0.314 0.151 0.703 GDP does not GC C70 2.596 0.110 7.540 0.014 5.423 0.018 3.899 0.065 C70 does not GC GDP 0.402 0.676 0.481 0.497 0.147 0.865 0.270 0.610 GDP does not GC C80 8.495 0.004 1.372 0.258 3.494 0.059 4.605 0.047 C80 does not GC GDP 4.372 0.034 16.118 0.001 2.090 0.161 1.664 0.214 GDP does not GC C90 7.274 0.007 2.827 0.111 6.643 0.009 8.068 0.011 C90 does not GC GDP 0.687 0.519 0.013 0.912 0.304 0.743 0.011 0.919 GDP does not GC C100 5.501 0.017 25.049 0.001 2.289 0.138 17.587 0.001 C100 does not GC GDP 0.226 0.801 1.878 0.188 0.802 0.468 0.085 0.774 GDP does not GC C 5.502 0.017 25.049 0.000 2.289 0.138 17.588 0.001 C does not GC GDP 0.226 0.801 1.878 0.188 0.802 0.468 0.085 0.774 Source: Author´s calculations.