Does political orientation affect economic indicators in the Czech Republic?

Available online at www.sciencedirect.com Journal of Policy Modeling 44 (2022) 1219–1231 www.elsevier.com/locate/jpm Does political orientation affect economic indicators in the Czech Republic? Michal Krajňák a,⁎,1 , Kateřina Krzikallová a,2 , Václav Friedrich b,3 a VSB – Technical University of Ostrava, Faculty of Economics, Department of Accounting and Taxes, 17. Listopadu 15/2172, 708 00 Ostrava, Czech Republic b VSB – Technical University of Ostrava, Faculty of Economics, Department of Mathematical Methods in Economics, 17. Listopadu 15/2172, 708 00 Ostrava, Czech Republic Received 3 May 2022; Received in revised form 22 August 2022; Accepted 24 September 2022 Available online 10 October 2022 Abstract The article evaluates the relationship between the tax burden on labor and magic quadrangle indicators in the Czech Republic in the years 1993 through 2020. The article examines whether indicators such as the effective rate or tax rate on labor affect the macro-economic indicators of the magic quadrangle. The originality of this study lies in the fact that it deals with the influence of political factors. The analysis shows the strongest correlation between the growth of gross domestic product and the implicit tax rate on labor. Moreover, the study finds that the factor with the most significant – and surprising – bearing on the findings is that fact that right-wing Parliament behaved like left-wing parties. The conclusions reached by this study further underline the significance of the tax burden on labor on the selected magic quadrangle indicators. © 2022 The Authors. Published by Elsevier Inc. on behalf of The Society for Policy Modeling. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Subject classification codes: C01; E62; H24 Keywords: Personal income tax; Effective tax rate; Magical quadrangle; Implicit tax rate on labor; Political orientation https://doi.org/10.1016/j.jpolmod.2022.10.001 0161-8938/© 2022 The Authors. Published by Elsevier Inc. on behalf of The Society for Policy Modeling. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). ]]]] ]]]]]] ⁎ Corresponding author. E-mail address: [email protected] (M. Krajňák). 1 ORCID 0000–0003-4924–3583 2 ORCID 0000–0001-6863–4630 3 ORCID 0000–0003-0076–5233 1. Introduction A country's economic policy is determined by the government. This affects the level of inflation, the unemployment rate, gross domestic product and the share of the current account balance on gross domestic product. These indicators make up the so-called ‘magic quadrangle’. The country’s political direction influences not only these indicators but also the tax burden, which is most often expressed as the personal income tax rate. According to OECD methodology the tax rate includes all payments to the state – not only personal income tax, but also health insurance and social security contributions. To enable standardized international comparison this OECD-defined tax rate is calculated from the average income of unmarried childless taxpayers (OECD, 2021). The overall personal tax rate – the implicit tax rate – describes the proportion of payroll expenditure on personal income tax and social security contributions. The overall tax burden is defined as the rate of personal income tax and mandatory social security contributions. Tax and social levies increase employer costs (Bingley & Lanot, 2002). As these levies increase, the competitiveness of the domestic labor market decreases (Nielsen & Smyth, 2008). High labor taxation may affect macroeconomic indicators such as the unemployment rate or economic growth rate. As stated by Surugiu et al. (2012), high marginal taxes reduce output; the substitution effect results in the taxpayer prioritizing their leisure time at the expense of working time which has an impact on the unemployment rate. At high levels of unemployment, one cannot expect economic output to be high, which negatively affects gross domestic product and therefore economic growth. Low domestic output affects imports thus influencing the current account of the balance of payments. The inflation rate, whose relationship with unemployment is shown by the Philips curve (Nedomlelová, 2008) also gains significance. Therefore, all macroeconomic indicators (inflation, unemployment, economic growth rate and the ratio of the current account balance to gross domestic product) can be affected by the overall personal tax burden. At the same time, it follows (from the ‘magic’ in the magic quadrangle) that meeting one objective often results in failure to achieve other objectives. These indicators can thus be influenced by the country's political orientation. Is this the case in the Czech Republic? Is it true that governments implement economic policy according to their political leanings, or are their situations where a right-wing government can be characterized by left-oriented economic policy and vice versa? The main aim of this article is to evaluate the relationship between political orientation and the above-mentioned economic indicators (the effective tax rate, implicit tax rate on labor, and magic quadrangle indicators) in the Czech Republic in the years 1993 through 2020. Comprehensive study evaluating the interrelationship between personal income tax rates, magic quadrangle indicators and political orientation has not been carried out in the Czech Republic, therein lies the value of this paper. Selected studies, for example, Bokrošová (2005) and Kuric (2015), only evaluated the economy based on the relationship between the macroeconomic indicators of the quadrangle. This article specifically deals with the Czech Republic, where fiscal policy is often implemented inconsistently: the differing agendas of successive ruling parties lead to frequent changes to the tax code concerning income tax rates, personal allowances and tax relief, with implications for the taxpayer's net wage. Payroll costs, including income tax, have an impact on the employer's total costs and the entire production process (Bingley & Lanot, 2002). Tax rates thus have an effect on economic growth, unemployment, inflation, and foreign trade. M. Krajňák, K. Krzikallová and V. Friedrich Journal of Policy Modeling 44 (2022) 1219–1231 1220 The paper is structured as follows: the introduction is followed by an outline of the theoretical background (Section 2) with focus on personal income tax. Section 3 includes the data we used for the analysis. Section 4 presents the research methodology. The main part of the article is Section 5 where we present the results of our analysis. The last section of the paper summarizes the results of our research topic. 2. Theoretical background In the field of economic policy, the direction and thus the program of leftand right-wing political parties often differ greatly from one another. The Czech Republic was established in 1993 after the division of the Czechoslovak Republic. ODS, a right-wing party, was the first to come to power following democratic elections in this new state. In his analysis, Šimoník (1996) found that in the first years of the independent Czech Republic there was a slight shift in voter preference away from right-wing politics, mainly due to centrist voters moving to the left, which was the reason why a left-wing party won the next election. Overall however, the centreright prevails in Central and Western European politics. According to a study by Choma et al. (2010), left-wing parties do not always have to pursue purely left-wing politics and right-wing parties do not necessarily follow a purely right-wing agenda. On the contrary, there are two dimensions: social left-right and economic left-right. Political orientation in Central European countries is partially changing, as evidenced by certain steps in economic policy as shown in the studies of Hlousek and Kaniok (2021), and Hooghe et al. (2002), for example. The state's fiscal policy aims to ensure economic growth, full employment, price stability and external economic balance by influencing aggregate supply and aggregate demand. According to Buettnerr and Krause (2021), fiscal policy instruments include a progressive income tax. The general principles of personal income tax are stated by du Prezz and Stiglingh (2018), among others. The main goal of progressive personal income tax is to ensure the redistribution of income in the economy. The level of government revenue and expenditure impacts aggregate demand and supply which affects economic growth, employment, price stability and external economic balance. The amount of government revenue is also affected by the extent of the tax gap. (For more about the definition and methodology concerning the tax gap and its determination see, for example, Moravec et al., 2018). Personal income tax can be defined as progressive in most countries (Sanz-Sanz, 2020). Tax progressivity ensures an automatic countercyclical effect: in a period of economic growth, income and tax increases, but because of the progressiveness the tax growth is higher than the growth of income; in a period of recession the opposite is true: the tax burden decreases faster than incomes do (Chen, 2019; Oddou, 2020). The effects of political orientation on tax policy, economic growth and employment have not been discussed in specialist studies. In the cited studies, only individual magic quadrangle indicators were examined. The fact that this topic has not been discussed in the context of the Czech Republic, and only minimally regarding other countries, contributes to the uniqueness of the research included in this article. The relationship between the tax burden and economic growth in 35 countries of the Organization for Economic Cooperation and Development (OECD) was the subject of a study by Andrasic et al. (2018). It determined that if tax revenues increase by 1%, gross domestic product increases by 0.29%. Similar positive correlation between these variables is confirmed by Nguyen (2019), and Bleaney et al. (2001). A combination of appropriate fiscal policy tools can therefore affect economic growth (Surugiu et al., 2012). Labor taxation in relation to M. Krajňák, K. Krzikallová and V. Friedrich Journal of Policy Modeling 44 (2022) 1219–1231 1221 economic growth does not always positively correlate, especially between medianand highincome households, as mentioned by Biswas and Chakraborty (2017). If labor taxation is high, a further increase in the tax burden can, on the contrary, reduce the degree of economic growth. Kotlán and Machová (2013), and Izák (2011) found that taxation has a negative impact on economic growth. An increase in unemployment and thus the slowdown of economic growth are the consequences of the excessively rapid growth of labor costs. The reason for the increase in labor costs in the European Union in particular are higher labor taxes. An increase in the tax rate on labor by 14% would result in an increase in unemployment by 4% and the reduction of economic growth by 0.4% per year in EU countries (Daveri & Tabellini, 2000). The growth of labor costs led companies to replace labor with capital, which in turn slows down an economic growth. Tax policy can be an effective instrument for reducing unemployment and increasing economic growth (Michaelis & Birk, 2006), while Bohringer et al. (2005) states that taxation as a tool for reducing unemployment does not have a significant effect. In the long term, an increase in the tax rate leads to a rise in the natural rate of unemployment (Neverauskiene Okuneviciute et al., 2017). The increase in prices puts increased pressure on employees to increase wages. If the tax rate is progressive, there is an increase in the tax burden. Unless the range of particular bands of personal income tax base area are adjusted according to prices, the taxpayer’s higher income is subject to a higher tax rate. Therefore, it is possible that, despite the growth of gross wages, the real net wage is lower after taxation (Nam & Zeiner, 2016). The tax burden therefore becomes more progressive (Gerber et al., 2020). For more about tax progressivity and the method of measuring see, for example, Wisniewska-Kuzma (2020), and Kristjansson and Lambert (2015). 3. Data used for the analysis For the purposes of analysing the relationship between tax burden and magical quadrangle indicators and political orientation in the Czech Republic (see Table 1) the following data has been gathered: •the unemployment rate (UNEM), inflation rate (INFL), gross domestic product growth (GDP) and the ratio of the current account balance of payments to gross domestic product (CA/GDP) have been sourced from the Czech Statistical Office (Czech Statistical Office, 2020), •the effective tax rate (ETR) determined in general terms (1), •the implicit tax rate on labor (ITR L ) determined in general terms (2), •data on political orientation and election results were obtained from the portal Volby operated by the Czech Statistical Office (Volby, 2022). In the Czech Republic the employer’s payroll costs consist of gross wages and social security contributions. Average gross wages in the analysed period are obtained from the database of the Czech Statistical Office (Czech Statistical Office, 2020). Social security contributions are calculated in accordance with the legislation governing this insurance. The overall personal tax rate is defined as the rate of personal income tax and social security contributions as a percentage of gross wages. M. Krajňák, K. Krzikallová and V. Friedrich Journal of Policy Modeling 44 (2022) 1219–1231 1222 4. Research methodology In order to analyse the data mentioned in the previous chapter, we specify the methodology used to conduct the analysis. Firstly, we define in general the method of calculating the personal tax rate. Subsequently, we formalize methods for assessing the relationship between selected indicators of the magic quadrangle, formalized tax burden indicators and political orientation. 4.1. Tax burden on labor indicators The effective tax rate E TR is calculated according to the formula (1), =ETR IT Y 100% (1) where IT is personal income tax, and Yis total gross salary. The implicit tax rate on labor ITR L is determined according to the formula (2), =ITR LT LC 100% L (2) Table 1 The political orientation in the Czech Republic. Year Orientation Prime Minister Predominance 1993 right Václav Klaus (ODS) 0.25 1994 right Václav Klaus (ODS) 0.25 1995 right Václav Klaus (ODS) 0.25 1996 right Václav Klaus (ODS) 0.08 1997 right Václav Klaus (ODS) 0.08 1998 left Josef Tošovský (non-party) -0.08 1999 left Miloš Zeman (ČSSD) -0.08 2000 left Miloš Zeman (ČSSD) -0.08 2001 left Miloš Zeman (ČSSD) -0.08 2002 left Vladimír Špidla (ČSSD) -0.23 2003 left Vladimír Špidla (ČSSD) -0.23 2004 left Stanislav Gross (ČSSD) -0.23 2005 left Jiří Paroubek (ČSSD) -0.23 2006 left Mirek Topolánek (ODS) -0.09 2007 left Mirek Topolánek (ODS) -0.09 2008 left Mirek Topolánek (ODS) -0.09 2009 left Jan Fischer (non-party) -0.09 2010 right Petr Nečas (ODS) 0.06 2011 right Petr Nečas (ODS) 0.06 2012 right Petr Nečas (ODS) 0.06 2013 left Jiří Rusnok (non-party) -0.14 2014 left Bohuslav Sobotka (ČSSD) -0.14 2015 left Bohuslav Sobotka (ČSSD) -0.14 2016 left Bohuslav Sobotka (ČSSD) -0.14 2017 center Andrej Babiš (ANO) 0.12 2018 center Andrej Babiš (ANO) 0.12 2019 center Andrej Babiš (ANO) 0.12 2020 center Andrej Babiš (ANO) 0.12 Source: own calculations. M. Krajňák, K. Krzikallová and V. Friedrich Journal of Policy Modeling 44 (2022) 1219–1231 1223 where LT represents labor taxes, and LC labor costs (Friedrich et al., 2012). 4.2. Methods Correlation analysis was used to evaluate the dependence of quantitative variables. The Pearson correlation coefficient ris a measure of the strength of a linear association between two variables, for example Xand Y. It is calculated using the formula (3), =rS S S XY X Y (3) where s XY is covariance between Xand Y,s X is standard deviation of X, and s Y is standard deviation of Y. The formula returns a value between − 1 and 1, where r= 1 indicates a maximal (linear) positive relationship, and r= −1 indicates a maximal negative relationship. A result of r= 0 indicates no relationship at all. If the absolute value of ris higher than 0,5, the association between Xand Yis considered strong (Field, 2013). The statistical significance of the correlation coefficient is determined by t-test. The test statistic (tvalue) is calculated as (4), =t r r n 1 2 2 (4) where nis the number of values of variables Xand Y. Statistics thas a Student's distribution with df =n2 degrees of freedom (Pearson, 1931). Linear discriminant analysis is a method of multidimensional statistical classification. It is used to classify a statistical unit into one of the groups based on the values of several quantitative variables based on the similarity of their values (Klecka et al., 1980). A test statistic called Wilks’s lambda is used to evaluate the differences in the values of the variables in the individual classes. It is calculated as (5), = + W W B Adet det( ) (5) where Wis the matrix of variability within classes and Bis the matrix of variability between classes. The value of Wilks’s lambda is in the range 0–1, where Λ= 0 represents absolute discrimination of objects and Λ= 1 means that the objects cannot be distinguished based on the observed variables. The output of discrimination analysis can be expressed using discrimination functions and scores, which allows the probability that a given statistical unit belongs to a given class to be expressed. Thus, for each unit, the sum of the probabilities is equal to 1 (Everitt & Dunn, 2001). The graphical representation of the results of discrimination analysis are biplots called (canonical) discriminant charts (Lipkovich & Smith, 2002). In these graphs, the individual units are displayed in two-dimensional space, with the distance between the units representing their similarity. Units belonging to the same group form clusters in the graph. In the discrimination chart, boundaries between individual groups represented by lines and half-lines can be created. Each group is then represented by a given polygonal area. Such graphs are also called territorial maps. M. Krajňák, K. Krzikallová and V. Friedrich Journal of Policy Modeling 44 (2022) 1219–1231 1224 4.3. Political data To determine political orientation in the Czech Republic, members of the Chamber of Deputies of the Parliament of the Czech Republic were monitored according to the political parties and movements they belonged to and subsequently classified as left, center, or right. Two variables were created for statistical analysis. The first variable was a nominal variable with values of left, center and right that determined the prevailing political orientation in Parliament (modal value). The second variable Pis a quantitative variable that expresses the predominance of the left or right political wing in the Chamber. The value of this variable is calculated as (6), = + PL R N (6) where Pis predominance, Lis number of left-oriented deputies, Ris number of right-oriented deputies and Nis the number of members of the Chamber. The value of Pranges from − 1 (100% predominance of the left wing) to + 1 (100% predominance of the right wing). It is a weighted average where left-wing members of the Parliament have a value of − 1, center 0 and right-wing + 1. The values of both variables for the period 1993–2020, supplemented by the names and political parties of the Prime Ministers, are given in Table 1. Predominance was calculated according to (6). Since the members of the Chamber of Deputies are elected (with a few exceptions) based on proportional representation, there is concordance between parliamentary predominance and the political orientation of the Prime Minister (head of government). In other words, there is consistency between the executive and the legislature. The exceptions are the period 2006–2008, when, although the Chamber of Deputies was predominantly left oriented, the right-wing government of Mirek Topolánek was in power, but for most of its term without confidence; and the period 2016–2020, when Andrej Babiš's central-oriented government operated in a predominantly right-wing Parliament. 5. Results and discussion In table 2, the selected characteristics of the analysed variables are listed first. For each variable, the minimum and maximum value, the mean value, median, and standard deviation is stated. 5.1. Research results The influence of political orientation in the Czech Republic on macroeconomic indicators of the magic quadrangle (UNEM, GDP, INFL, and CA/GDP) and selected tax rates (ETR and ITR L ), was evaluated by correlating these indicators with political predominance in the Chamber of Deputies. The results are shown in Table 3. The tax progressivity indicators (ETR and ITR L ) are not significantly dependent on the prevailing political orientation in the Chamber of Deputies. Of the magic quadrangle indicators, a statistically significant and strong dependence was found between the unemployment rate (UNEM) and inflation (INFL). M. Krajňák, K. Krzikallová and V. Friedrich Journal of Policy Modeling 44 (2022) 1219–1231 1225 As the correlation coefficients show, in periods when the right wing in Parliament is stronger, unemployment in the Czech Republic is lower and inflation is higher; in years when the political left has a stronger influence, higher unemployment and lower inflation are more likely. These seemingly surprising results can have two causes. Firstly, macroeconomic indicators tend to have higher inertia, reacting to changes with a delay (sometimes several years). It is therefore not uncommon for governments to be criticized for negative outcomes whose roots need to be traced back to their predecessors. Secondly, the political scene in the Czech Republic (and historically in the Czech lands) tends to be more left-oriented. This means that even parties that present themselves as rightwing are rather more center-right and have many left-wing elements in their manifestos. The exception is the ODS party led by Václav Klaus in the 1990 s, which strongly espoused Reaganomics and Thatcherism. At present, such ‘far right’ parties (as labeled by political scientists and journalists) are in the minority and outside the political mainstream. In addition, the political movement ANO (originally ANO 2011) has a specific character in the Czech political scene. It was founded by businessman and multimillionaire Andrej Babiš as a centrist, catch-all protest movement in opposition to the established left and right parties. ANO’s agenda features elements of both leftand right-wing politics so the movement has been willing to cooperate with other parties regardless of their political orientation. It participated in government from 2013 to 2021. In 2017 it won the election and Babiš became Prime Minister. Linear discriminant analysis was used to verify the stability of the perception of the political left and right in the Czech Republic. The previously-mentioned magic quadrangle indicators were again used as inputs and the state policy in the Czech Republic was evaluated as left-wing, central, or right-wing for each year of the period from 1993 to 2020. Wilks’s lambda reached a value of 0.238 (p-value < 0.001), indicating that the resulting discrimination of units into three groups called left, center and right is significant. Table 2 Descriptive statistics of indicators for analysis. Variable Minimum Maximum Mean Median Std. deviation Y 1 - ETR 8.64 14.00 11.04 11.08 1.38 Y 2 - ITR L 41.96 43.95 42.83 42.75 0.52 X 1 - UNEM 2.00 8.80 5.87 6.50 1.99 X 2 - GDP -4.80 6.85 2.65 2.68 2.60 X 3 - INFL 0.10 20.80 4.25 2.50 4.56 X 4 - CA/GDP -6.15 1.56 -2.25 -2.11 2.25 Source: own calculations. Table 3 Correlation between economic indicators and political predominance. Indicator Correlation p-value UNEM -0.643 < 0.001 GDP -0.153 0.436 INFL 0.597 < 0.001 CA/GDP 0.307 0.112 ETR -0.124 0.538 ITR L -0.042 0.836 Source: own calculations. M. Krajňák, K. Krzikallová and V. Friedrich Journal of Policy Modeling 44 (2022) 1219–1231 1226 The results of the discriminant analysis are shown in Table 4. It turns out that in most years the government's economic policy, as expressed by macroeconomic indicators (last column in Table 4), coincides with the prevailing political orientation of the Chamber (column 2). Over the entire 1993–2020 study period there are only two periods when there is significant deviation from this trend. The first of these is 2007–2008 when the right-wing government of Mirek Topolánek ruled without confidence, despite a left-wing majority in Parliament. As the discriminatory analysis shows, this period nevertheless displays the characteristics of a right-wing government. This government introduced the flat tax in the Czech Republic. Paulus and Peichl (2009) have pointed out the success of flat tax implementation in Eastern Europe, and Magnani and Piccoli (2020) document the positive effects of income tax reform in France stemming from a shift from progressive taxation and existing benefits to universal basic income scheme with a flat tax. On the other hand, the results of a study made in Russia by Duncan (2014) show that in developing economies, where tax evasion is widespread, this kind of reform is failing. Secondly, in the years 2010–2012, the right-wing government of Prime Minister Petr Nečas had the support of Parliament, in which there was a slight right-wing dominance. Nevertheless, Table 4 The discriminant analysis results. Year Orientation Probability Result Left Center Right 1993 right 0.07 0.00 0.93 right 1994 right 0.11 0.00 0.89 right 1995 right 0.09 0.00 0.91 right 1996 right 0.10 0.00 0.90 right 1997 right 0.18 0.00 0.82 right 1998 left 0.45 0.00 0.55 right 1999 left 0.82 0.00 0.18 left 2000 left 0.84 0.00 0.16 left 2001 left 0.77 0.00 0.23 left 2002 left 0.71 0.00 0.29 left 2003 left 0.82 0.00 0.18 left 2004 left 0.90 0.00 0.10 left 2005 left 0.89 0.00 0.11 left 2006 left 0.60 0.00 0.40 left 2007 left 0.33 0.00 0.67 right 2008 left 0.26 0.02 0.72 right 2009 left 0.71 0.00 0.29 left 2010 right 0.81 0.00 0.18 left 2011 right 0.80 0.00 0.20 left 2012 right 0.82 0.00 0.18 left 2013 left 0.82 0.00 0.18 left 2014 left 0.71 0.01 0.28 left 2015 left 0.44 0.23 0.33 left 2016 left 0.09 0.77 0.14 center 2017 center 0.00 0.99 0.01 center 2018 center 0.00 1,00 0.00 center 2019 center 0.00 1,00 0.00 center 2020 center 0.00 1,00 0.00 center Source: own calculations. M. Krajňák, K. Krzikallová and V. Friedrich Journal of Policy Modeling 44 (2022) 1219–1231 1227