Trends in income taxation: are taxes converging in Central and Eastern European countries?

The concept of convergence, used in many fields, means the process of gradual similarity (“getting closer”) to each other or to the adopted pattern. It should also be remembered that convergence does not have to mean a qualitative improvement in the situation of the studied entities, because the similarity can be both “upward convergence” and “downward convergence”, which means approaching toward the worst or the weakest [see, e.g., Eurofound, 2020].

When researching convergence, we must always confine ourselves to a certain, relatively limited set of indicators that measure a selected aspect (or aspects) of the economy covered by the study. Based on the selection of variables, many types of convergence can be distinguished. There is, e.g., nominal convergence, which refers to the requirements for joining the euro area. These requirements, known as the Maastricht convergence criteria, focus on inflation, interest rates, volatility of exchange rates, and deficit and public debt. The EU legislation requires also the laws of candidate countries to comply with the Treaty on European Union and with the Statute of the European System of Central Banks and of the European Central Bank (legal convergence). Economic literature provides also many other types of convergence, such as the real ones focusing on the gross domestic product (GDP) (or GDP-related measures), productivity, competitiveness, labor market outcomes [e.g., Barro and Sala-i-Martin, 1992; Boyle and McCarthy, 1999; Borys et al., 2008; Halmai and Vásáry, 2010; Marelli and Signorelli, 2010; Szeles and Marinescu, 2010; Alexe, 2012; Monfort et al., 2013, Diaz del Hoyo et al., 2017], or structural convergence [Palan and Schmiedeberg, 2010; Vandenbroucke, 2017; Raleva, Marikina, 2021].

The studies also address tax convergence defined by Bénassy-Quéré et al. [2014] as a reduction of tax rate differentials and base differentials motivated by tax coordination and competition. Delgado and Presno [2017] stress that tax convergence may limit, control, or identify tax competition; but it may also be used to check the prediction from the theory of international tax competition, while the achieved results may contribute to a better understanding and design of tax policy.

There exist many methodological approaches to convergence in economics [Bernardelli et al., 2017]: e.g., beta-convergence based on the growth theory [Solow, 1956], which occurs when poor countries/regions grow faster than rich ones (and catch up with them); and sigma-convergence, which refers to the reduction of disparities over time [Barro and Sala-i-Martin, 1992; Quah, 1993; Sala-i-Martin, 1996; later, e.g., Monfort, 2008; Young et al., 2008; Lessmann and Seidel, 2015], broken down into subgroups (club convergence – introduced by Baumol, 1986, developed by, e.g., Durlauf and Johnson, 1995; Galor, 1996; Canova, 2004; Phillips and Sul, 2007; Alexiadis, 2012; Maynou et al, 2021; Ripollés, Vidal-Meliá, 2022).

The theoretical background of the concept of convergence in terms of the macroeconomic variables is well known for decades in the form of models of economic growth. A range of approaches have been developed to date, beginning with simple linear regression models, through the econometrics of panel data and Bayesian averaging, ending with very advanced and complicated mathematical constructions. Due to the changes in economic and political conditions, new approaches to the issue of convergence are constantly being developed.

The most common in the literature on economic growth models, namely, neoclassical models and endogenous models, have one common flaw in the form of the restrictive assumptions of application. Compared to these conventional econometric methods, an approach based on HMMs, which was used in the analysis presented in this article, is much more flexible in terms of the assumptions of a statistical or asymptotic nature and gives a wider possibility of applicability. Moreover, it can be adapted to a variety of input data (not only GDP), which makes it a perfect tool for tax convergence analysis. This field of application of the HMM is relatively new in scientific research and was used for the first time by Felis et al. [2020]. The authors have attempted to analyze and assess the directions of evolution of income taxation in EU countries. The concept of HMMs, as well as the description of the explored procedure of convergence analysis based on the HMM and Viterbi paths, is presented in the Appendix.

Studies related to tax convergence have a different scope with respect to time and geography, and their authors follow diverse methodological approaches. Wang [2007] used time series and cluster analyses to examine tax burden and economic growth convergence among China, Taiwan, and the OECD countries in the 1970s, 1980s, and 1990s. His research indicated tax burden convergence in the group of China, Taiwan, and Korea; however, this was not pairwise. The methodology of Phillips and Sul [2007] was used by Apergis and Cooray [2014] to investigate tax policy convergence across the Association of South-East Asian Nations (ASEAN) countries and the Asia–Pacific and Oceania countries in the years 1990–2012. Empirical evidence has revealed that there was no full convergence across the analyzed countries and that the tax policies remained national. The Phillips-and-Sul methodology [2007, 2009], beta-convergence, and sigma-convergence were also adopted by Chen et al. [2016] in the study of corporate tax rate convergence across 15 countries from the Asia–Pacific region. Their research showed significant dynamic tax convergence and identified three convergence clubs within the period 1980–2014. While studying worldwide corporate tax rate convergence, Slemrod [2004] found a strong link between the top individual rate and the top statutory corporate rate in the years 1980–1995.

Fiscal pressure and fiscal policy convergence across the EU countries were investigated by Esteve et al. [2000], who focused on convergence in fiscal pressure, defined as the total fiscal revenue/GDP ratio, during the years 1967–1994. They used cross-sectional convergence indicators and time series analysis to find, inter alia, both convergence (1979–1994) and divergence (1967–1979) periods. Similar research was also conducted by, among others, Sosvilla-Rivero et al. [2001], Blot and Serranito [2006] (Economic and Monetary Union [EMU]) countries), and Delgado and Presno [2010, 2011]. According to Delgado and Presno [2011] – who applied the time series method using unit root and stationarity tests with an endogenous break – over the period 1965–2004, only the United Kingdom and Germany exhibited signs of long-term convergence although there were some other countries for which convergence could be observed. Tax convergence in the EU member states was also addressed by Tibulcă [2015], who studied the sigma-convergence using the coefficient of variation and the Gini coefficient. The author identified several convergence periods for the years 1965–2012, of which the last one began in 1997 and ended in 2008. Regis et al. [2015], while investigating convergence in statutory corporate tax rates between 1980 and 2014, report that four clubs of converging European countries can be observed.

Tax convergence measuring the dispersion of tax-to-GDP ratio (sigma-convergence) was investigated in a study carried out for the OECD member countries in the years 1995–2016 [OECD, 2018]. This investigation also measured the convergence of tax structures taking into account the share of diverse income categories in total tax revenue. The D-index [Delgado, 2013], an indicator showing the extent to which a country’s tax structure differs from the OECD average structure, was used. It was found that, over the period covered by the study, both the distribution of countries around the OECD average tax-to-GDP ratio and the average tax structure had become more similar. The OECD countries were getting closer to a higher level of taxation and since the dispersion of tax-to-GDP levels had decreased, the level of the tax-to-GDP ratio was increasing. Nevertheless, a clear heterogeneity was observed for the tax-to-GDP ratio in different countries; thus, they were divided into four groups (with low, low-mid, high-mid, and high tax-to-GDP ratios). In each group, convergence unrolled along different lines. In the OECD member countries, convergence of tax structures was also reported for value-added tax (VAT), and quasi-fiscal charges (social security contributions) and CIT increased their share, however to a lesser extent. Over the examined period, eight countries moved away from the OECD average structure (Lithuania, Poland, the United States, Italy, Mexico, New Zealand, Hungary, and Estonia).

The OECD countries⁴ were also studied by Godbout and Robert-Angers [2022], who showed, among other things, that in the years 1981–2018, there was a reduction in the general dispersion around the average for the tax burden and the analyzed components of the tax structure, such as PIT, CIT, social security contributions and payroll tax, tax on property, consumption tax, and customs and import duties. They also noted that majority of the countries were closer to the OECD average in 2000–2018 than in 1981–1999.

Delgado et al. [2019] investigated the convergence of effective CIT rates in the EU (EU-27) over the years 2000–2016 using the club approach. They argue that since the emphasis is primarily on the harmonization of indirect taxes, the EU member states deploy different patterns of corporate taxation. Member states covered by the study were divided into three clubs: (1) France, Malta, Spain, and Portugal – countries with high effective tax rates, where poor relative convergence was observed (convergence speed: 0.007); (2) Belgium, Germany, Greece, Luxembourg, Italy, Austria, the Slovak Republic, Hungary, Poland, and the Czech Republic (average-level tax rates; convergence speed: 0.146); and (3) the 13 remaining member states with low effective tax rates, showing average convergence speed equal to 16.99.

Systemic change and the adoption of market economy principles were the major factors that triggered the reform of tax systems in the CEECs. For instance, the split of Czechoslovakia into the Slovak Republic and the Czech Republic was linked with the transition from a centrally planned economy to a market economy, which would have not been possible without fundamental tax reforms. Reforms in these countries started already in the early 1990s. The timing of the launch of the first fundamental change in the tax system in January 1993 accidentally coincided with the birth of the Slovak Republic. The establishment of an independent state, however, did not have any direct impact on the shape of the system. The new system incorporated relevant components of the tax system in the EU countries. Its basic pillar was a VAT with two rates, which were later changed a few times. The PIT was progressive, with six income brackets and rates ranging from 15% to 47%. However, the system also included five other tax rates applied to specific sources of income. The CIT was set with a linear rate of 45% [Miklos et al., 2005].

According to Ebrill [1999, p. 8] and Stepanyan [2003, p. 13], the Baltic countries managed very quickly to dismantle the Soviet Union tax models and successfully implement new solutions consistent with internationally approved practices. In 1994, Estonia introduced a PIT flat rate and reduced the CIT rate, both at the level of 26%; since 2000, the CIT taxpayers are allowed to defer payment of that tax until profit distribution. Latvia replaced the tax on profits with a corporate tax in 1995. Initially, the Latvian CIT rate was set at 29%; the country then reduced it to 24% in 2000 and to 15% since 2002 for several consecutive years, which significantly increased the competitiveness of this country.

Undergoing reforms also included stimulating foreign direct investment in order to modernize the postsocialist economy and to meet enormous capital needs. The goal was to adapt the tax systems of these countries to models operating in the other member states of the EU. In most cases, reforms embraced PITs and CITs, in particular, the reduction of tax rates or tax exemptions. Tax incentives were the main factor to attract foreign direct investments in the CEECs. For example, in the 1990s, Poland introduced income tax holidays for companies with foreign capital. Investment incentives took different forms and were offered within the framework of targeted programs. One of the most significant examples of supporting economic and social goals was the setting up of special economic zones (SEZs), accompanied often by subsequent exemption from property taxes. For example, 14 SEZs were created in Poland as a regional aid instrument. Investors in SEZs were allowed tax exemptions for 10 years and partial exemptions for the following 10 years. In turn, in 1992 and 1993, Hungary introduced a 10-year tax holiday for foreign investors [Krajewska, 2006].

When the CEECs joined the EU, the principal objective was to enhance the competitiveness of their tax systems in relation to the systems of the “old” EU member states. Reforms at that time introduced fundamental changes in many tax structures to ensure stable and long-lasting economic growth. The race to the bottom could have been observed to make the tax system more competitive, manifested to a large extent in the reduction of income tax rates and tax reliefs, as well as in the simplification of the entire taxation system. For example, the rate of 19% for businesses in Poland was introduced. The reduction of direct taxes shifted also the source of tax revenues to indirect taxes. This was happening with the introduction or broadening of the scope of VAT and excise duties. In 2006, an important priority for the Czech government was to simplify tax law and strengthen incentives for investment. At that time, shortened depreciation periods for movable property and research-and-development (R&D) tax credits were introduced.

The second and the most fundamental reconstruction of the tax system in the Slovak Republic (effective from January 2004) was to no longer copy the standard tax systems of Western Europe. It was rather based on a fundamental change of some taxation principles. A larger emphasis on the competitiveness of the economy made the tax system more attractive and motivating. The implemented reform consisted primarily in the introduction of a flat rate of 19% in direct taxes and the reduction of a number of exemptions and special tax regimes, although allowance for children in the PIT was left untouched. The reduction in the amount of direct taxes shifted the source of budget revenues to indirect taxes. Increase in certain rates of excise duty on selected goods (e.g., mineral oils, beer, and tobacco) and unification of VAT rates were implemented. Instead of the previous two rates of 20% and 14%, a single VAT rate of 19% was introduced [Miklos et al., 2005].

The global financial crisis of 2008–2009 affected the performance of public finances in most European countries and, as a consequence, a necessity to revisit fiscal policy tools emerged. For example, in 2009, the GDP of the EU member states decreased by 4.2% on average (Latvia suffered the most, with a decrease of 18.0%, and Poland, on the contrary, reported a growth of 1.7%).⁵

The actions of the governments of the member states were geared toward ensuring sufficient revenues to the budget, e.g., by limiting the income and VAT tax preferences and exemptions and by increasing the tax rates of indirect taxes. The biggest change was observed in Hungary, where the VAT rate – which was 20% in 2008 – was first increased to 25% in 2009 and then to 27% in 2012.

As of January 1, 2009, almost 40% of Lithuanian tax legislation was changed, and the changes concerned the key elements of the system and applied to the vast majority of taxpayers. In the PIT, a lower uniform rate of 15% was introduced, while in the CIT, the rate was raised to 20%, the tax base was extended, and new preferences were implemented [Skačkauskienė and Tunčikienė, 2012]. The increase in the CIT rate was a one-off event: since 2010, the rate is 15%, same as before.

Countries that increased the tax rates declared that the change was only temporary and to mitigate the effects of the crisis. However, for instance, Poland introduced a higher VAT rate (23%), which was supposed to be binding only between 2011 and 2013 but, in fact, still remains in force.

The coronavirus disease-2019 (COVID-19) pandemic crisis also entailed adjustments to tax systems. Different measures have already been adopted (e.g., tax payment deferrals, accelerated depreciation, and double deduction of expenses incurred to alleviate the pandemic from the taxable base or from the tax liability) to reduce the tax burden to entrepreneurs locked down. The ongoing pandemic prompts further research in the future.

The tax systems of the EU member states are influenced by globalization and its key elements, such as tax harmonization, tax competition, and combating tax avoidance or evasion.

The tax systems of the member states are similar to some extent in their tax constructions due to the harmonization process that had been going on, in particular, in the 1990s and in the first decade of this century. This is especially true with respect to indirect taxes as a result of the free movement of goods and services in the internal EU market. Therefore, the member states must ensure uniform taxation of supplies of goods and services. Harmonization of direct taxes is much more difficult, which is partly related to the reluctance of the member states to give up their sovereignty in the area of tax policy and largely is influenced by the different design of direct taxes in the member states. The concepts emerging around tax base harmonization in corporate taxes, ranging from the least intrusive with the tax-raising powers of the member states (e.g., Home State Taxation) to the most radical one (European Union Corporate Income Tax), which assumes the existence of a single EU-wide tax ensuring revenue flows to the EU budget, are rather unrealistic in the current political situation.

Since the 1990s, progressing trade liberalization, no control over the movements of capital and foreign exchange convertibility at a global level, regional integration programs, the EU Single Market, the North American Free Trade Agreement (NAFTA), and new communication and transportation technologies have reduced transaction costs related to the movement of goods, services, capital, and people. All these factors have provided fertile ground for crawling tax competition or even to some extent unfair/harmful tax competition. The main reason for tax competition was economic growth by creating optimum tax conditions for business and attracting foreign capital. Differences in taxation systems of individual member states may thus affect the investment decisions of enterprises [Genschel and Schwarz, 2011].

Increasing tax competition between the EU member states, especially after the accession of the CEECs to the EU, has led to further differences in tax rates, reliefs, and ways of collecting taxes. In order to become competitive for attracting international capital and retaining domestic investment capital, the CEECs have reduced the nominal tax rates. However, tax competitiveness is influenced not only by tax rates but also by reliefs and exemptions, tax base, tax preferences, or favorable tax procedures. The Czech Republic and the Slovak Republic are examples of countries that attract foreign capital through taxes. Favorable rules of taxation for CIT and the lack of restrictions in VAT deduction make entrepreneurs choose precisely these countries when looking for a place to do business. Tax competition between countries takes the form of a “race to the bottom,” which – in extreme cases – may result in the tax rate being reduced to zero. An example is Estonia, where the CIT on retained earnings is 0%, which is a form of tax deferral until its payment. A similar construct has been implemented in Poland in 2021.

Parallelly, the trend to secure tax revenues is visible in all the EU member states whose tax base has been/is being eroded. The Anti-Tax Avoidance Directive (ATAD) contains legally binding antiabuse measures. In particular, ATAD-I includes rules on hybrid mismatches, controlled foreign companies (CFCs), switchover rule, exit taxation, interest limitation, and general antiabuse rule. ATAD-II adds rules on mismatches with third countries, which apply to all taxpayers subject to corporate tax in one or more member states, including permanent establishments (PEs) in one or more member states of entities resident for tax purposes in a third country. Rules on reverse hybrid mismatches also apply to entities treated as transparent for tax purposes.

Broadly sketched measures with regard to the above have also been undertaken under the auspices of the OECD/G20 within the framework of the project Base Erosion Profit Shifting (BEPS) 1.0 and, currently, BEPS 2.0.⁶ The OECD/G20 Inclusive Framework on BEPS has agreed on a two-pillar solution to address the tax challenges arising from the digitalization of the economy.⁷ Under BEPS 2.0 Pillar 1, automated digital services (ADSs) and consumer-facing businesses (CFBs) would be taxed in the jurisdiction that the revenues are sourced from. BEPS 2.0 Pillar 2 introduces new rules granting jurisdictions additional taxing rights in cases where either another jurisdiction has not exercised its primary taxing right or the income is subject to a “low tax rate” of 15%.

Income taxes (on company or personal income) are, along with sales taxes, one of the main pillars of the tax systems in the CEECs. It was demonstrated that contemporary models of taxation in the analyzed countries were determined by several factors.

Income taxes (PIT, CIT) may have a negative impact on the economic situation of taxpayers as they deprive them of some (determined by tax base and tax rates) of the financial resources that they have earned, in addition to distorting the market mechanism. Therefore, it is desired to reduce the negative effects of taxation on economic growth by tax reforms while maintaining an appropriate level of public goods and services. However, the impact of these factors on the implemented solutions differed across individual countries. In the CEECs, significant differences in the construction of income taxes can be observed (Table 1), as follows:

various system principles regarding tax transparency of taxpayers, in particular, in respect to partnerships;

Most CEECs have reduced tax rates for both CIT and PIT. Low CIT rates are in force, inter alia, in Bulgaria (10%) and Hungary (9%). Many countries have carried out radical tax reforms and introduced flat PIT rates, e.g., Bulgaria, the Czech Republic, Lithuania, Romania, the Slovak Republic, and Hungary. Importantly, this clear downward trend in CIT was coupled with the broadening of the tax base. Among the solutions implemented in Poland were, inter alia, preferences for small taxpayers (e.g., lower CIT rate, accelerated depreciation, quarterly frequency of advance payments). Estonia and Latvia brought tax reliefs and exemptions down to a minimum level. However, in Latvia, tools that support the desired socioeconomic objectives or encourage taxpayers to engage in specific economic activities (e.g., in the field of innovation, in an investment project, or the acquisition or production of new technologies) have not been abandoned straight away. Many countries use R&D tax relief or accelerated depreciation, offer preferential tax treatment to the SME sector, allow tax losses to be carried forward (often for extended periods),⁸ or put regional policy instruments in place (SEZs).

Tax system assumptions in Croatia and Estonia may deserve special attention. Croatia has been a precursor in this part of Europe of a consumption-oriented approach to income taxation. The CIT has been initially designed to offer allowances for corporate equity. Estonia, on the other hand, exempted income retained in the company from corporate tax. Estonian CIT is levied only when profit earned is distributed. The deferred moment of taxation can cause the desired lock-in effects. A similar solution was also implemented in Latvia (in 2018).

To find out how much similar are countries when it comes to the tax burden, we selected four indicators and carried out convergence analysis for each of them separately. The study was conducted for 11 countries (Bulgaria, Croatia, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, the Slovak Republic, and Slovenia) and covered the period 1995–2018. We used the following indicators as variables that describe taxation systems⁹: 1)

IND.1: current taxes on income, wealth, etc. as percentage of GDP;

Therefore, a differentiated approach was adopted in the selection of indicators – with reference to both the total tax revenue of the state, as well as the income resulting from taxation of corporations and separately of households.

Basic descriptive statistics for the selected indicators for the CEECs and EU-15 countries are presented in Table 2. For each of the four indicators describing the taxation systems, the corresponding values (minimums, maximums, quartiles, and arithmetic averages) for the EU-15 countries are much higher than for the CEECs. This could be explained by the tax reforms that were introduced in most CEECs since the mid-1990s, mainly in the area of income taxes attracting the interest of this article. In the case of IND.1 and IND.4, the values are approximately twice as large. For IND.2 and IND.3, the differences are not that big, but still distinct. In fact, only in two cases, the values of the taxes on the income or profits of corporations as a percentage of the total (IND.3) are greater for the CEECs compared to those for the EU-15 countries. The first case is during the financial crisis of 2007–2009, when incomes were strongly determined by macroeconomic factors. In 1995 also, IND.2 for CEECs exceed the value recorded for EU-15 countries (2.7 versus 2.6). For all other years and the other two indicators, the values for EU-15 are always greater than those for CEECs, in most cases significantly. Such an observation is a quantitative argument for the diversity of these groups of countries and an additional premise for a more detailed analysis of the CEECs.

Usually, the convergence analysis is conducted on pairs of entities [Pesaran, 2007] or – in the case of a group of entities – with regard to some benchmark. For the considered group of countries and set of indicators, it is difficult to provide an adequate benchmark; therefore, some kind of leader was selected. The country for which the value of a given indicator was the highest in the last year of the analyzed period was identified with the reference time series. The reasons for using this selection criterion were – on the one hand – in line with the concept of the most recent leader and, on the other hand, this criterion is the solution of the problem in which, over the years under consideration, the highest values of indicators were observed for different countries. In order to ensure that the reference values in each year were not lower than those reported for other countries, the values of the analyzed indicator for the selected reference country were previously recalculated using an appropriate multiplier. The multiplier is the minimal real number such that, after multiplication, the resulting reference time series is not smaller than the indicators for the other countries in the entire period. This approach allows for easy observation of changes in the distances between the values of the indicators for different countries for each year consecutively, and the convergence or discrepancy is counted against a hypothetical maximum (changing over the years). This technical adjustment does not affect the convergence dynamics but facilitates the interpretation of the results, whereby changes in distances or speed of those changes are read in terms of convergence. Changing the reference time series is likely to change the results and the related interpretation, which informs about the convergence of the selected country toward the group of other countries. However, such a change would not allow for the analysis of the convergence defined as a determination of the degree of similarity between the CEECs over time. Therefore, as a reference, a leader in recent years was selected.

The classic convergence analysis commonly uses econometric models¹⁰ that require the meeting of many assumptions of a statistical or asymptotic nature (particularly important for the short time series). In this study, this is one of the restrictions that have to be taken into consideration. In addition, this convergence is analyzed as indicator by indicator, and in this case, multifactorial relationship study is not needed. Indeed, the comparison of the indicators in the considered group of countries should be treated as a similarity analysis between time series. However, there are relatively few adequate measures of the dependence of the time series, and some of them require meeting certain complex statistical assumptions. Simplified and inadequate measures, such as the linear correlation coefficient, have not been used. One of the latest approaches to this issue is the procedure of using HMMs.¹¹ It has been successfully used to assess the degree of cyclical convergence and the leveling of income levels between Poland and the EU-27 on the basis of GDP per capita according to purchasing power parity (PPP) or the real GDP growth rate and related macroeconomic variables [Bernardelli et al., 2017, pp. 59-80]. The same method has also been used for the convergence analysis in this study. Compared to the conventional econometric methods, an approach based on HMM does not require many assumptions, because it works rather in terms of automatic pattern recognition. The advantage of using this approach is the ease of interpretation and wide applicability in terms of input data (including missing data), while the disadvantage is the difficulty to precisely quantify the speed of convergence, which is obvious in the case of econometric modeling.

Among the group of 11 countries considered in the paper, subgroups representing specific features may be selected. However, for such short time series and under rapidly changing tax reality, it is not possible to identify them unequivocally. When dividing countries into groups, automatic grouping into three groups using the k-means algorithm has been applied to avoid subjectivity, where the Euclidean norm has been adopted as a measure of the distance [Hartigan and Wong, 1979, pp. 100–108]. The grouping has been applied on the basis of observations placed in multidimensional space, where each of the coordinates was related to a specific indicator in a specific year.

Indicators for each country are presented in Figure 1. Before the latest pandemic period, many CEECs exhibited a clear tendency to reduce income taxation. Reduction of revenue from income tax stemmed to a large extent from the reduction of tax rates. Revenue generated by direct taxes and other levies in the CEECs ranged on average between 6% and 8% of GDP. These amounts are important for the equilibrium of general government, although they are clearly smaller than the revenue generated by other categories of taxes: consumption taxes and social insurance contributions. For example, on analyzing direct taxes in the Czech Republic, there is a downward trend with regard to the share of the CIT in the GDP. Its share in GDP in 1995 was +4.6% and, in 2010, –3.2% of GDP. Undoubtedly, it is related to the reduction of the rate of this tax. It should be noted that the CIT rate has been lowered many times in the Czech Republic over the past 27 years. Currently, it is 19%, while in 1993, it was 45%.

Figure 1.

Examples of countries where IND.1 clearly dropped include Estonia, Lithuania, Poland, Romania, and the Slovak Republic. In the CEECs, the fiscal relevance of taxes imposed on corporate income is rather small; they represent on average 3%–4% of the GDP and approximately 8%–10% of all revenue from taxes. Reasons can be sought in the negative impact of this category of taxes on taxpayers’ economic activity. Considering that taxes on capital (and CIT is one of them) generate the highest losses in prosperity, the decreasing levels of IND.2 and IND. 3 should be assessed positively in countries such as Bulgaria, Romania, and the Slovak Republic.¹² For example, the reason for positive assessment in Bulgaria is that this country focused, in the long-term policy, on additional revenues from indirect taxes to compensate the revenue losses caused by lowering the CIT rate. However, there are countries such as Croatia and Slovenia in which revenue from these levies have been growing in recent years. In other countries, fiscal relevance of tax measured with IND.2 and IND.3 varies. For PIT, the situation is more complicated (IND.4) as the tax base must be divided into two components: income from work (e.g., part of the income from business activities attributed to an entrepreneur as remuneration for her/his work) and income from capital gains (including part of the income from business activities attributed to an entrepreneur as return on invested capital).¹³ In most countries, this indicator was either gradually decreasing or remained at a rather unchanged level. It needs to be underlined that over the period 2008–2009, the above-mentioned indicators were on the rise, meaning that most of the countries covered by the study decided to change their previous tax policies. The global crisis drove government revenue down; however, on the other hand, it triggered measures designed to avoid any further economic slowdown and to stimulate the economy. It also turned governments’ attention to measures intended to increase tax revenue (increased tax rates, broadening of the tax base, and sealing the tax system).

The convergence analysis was conducted for each indicator according to the adopted methodology for the group of 11 countries covered by the study. A collective representation of the examined indicators is presented in Figure 2. Values (left Y-axis) vary between “0” and “1”, where “0” indicates perfect similarity between countries and “1” indicates divergence from the reference country. For the comparison on the right Y-axis, there are the original time series of the considered indicators: IND.1, IND.2, IND.3, and IND.4, respectively.

Figure 2.

The examination of graphs reveals strong similarity in the course of convergence for indicators IND.1, IND.2, and IND.3. In other words, the CITs as a percentage of the total and as a percentage of the GDP, as well as current taxes on income as a percentage of the GDP, indicate analogous behavior in the analyzed countries: they are more similar or more dissimilar in relation to the Czech Republic (reference country) at approximately in the same time period. Divergence growth since the end of the 20th century was stopped at the outbreak of the global financial crisis in 2007, when convergence processes intensified. At that time, the countries covered by our study, taken together as a group, became very similar and stabilized at a relatively permanent level. Only the IND.1 indicator informs about slightly more powerful processes that would suggest some divergence between these countries in terms of the direct taxes-to-GDP ratio.

While the results of the studies demonstrate that tax policies pursued by these countries in corporate taxes over the recent 10 years were similar, a totally different conclusion can be drawn for PIT (IND.4). Since 1998, a rather systematic divergence from Latvia (reference country) can be observed in the PIT. As for Latvia, the PIT rates have remained at the level of 25% or lower for many years. Only once – in 2010 – the rate was increased to 26%. In 2018–2020, the standard flat rate was replaced by a progressive PIT, with rates of 20%, 23%, and 31.4%. Our results do not mean that the countries concerned have completely different and dissimilar tax policies in this area. It should be noted that the adopted indicator of the level of fiscalism contains data on taxation of capital and labor. Income taxes on work and income from capital have many common features, but they are also distinct. In the case of income tax on work, it is not only about solutions that may lead to the inflow of capital and migration of wealthy and educated people, but also about personalization and social circumstances and, hence, the greater or lesser progressiveness of PIT (the tax-free amount, number of tax thresholds, and range of tax reliefs).

Valuable data can be obtained from the convergence analysis conducted for subgroups of countries identified using k-means algorithm, whereby, after incomplete data had been removed, grouping was carried out in a 68-dimensional space, one dimension for each indicator in a given year (four indicators, 17 years from 2002 to 2018; the shorter time interval was due to the lack of data available for Croatia for the years preceding 2002). The resultant division into groups, with the arithmetic means for indicators in the individual groups, is given in Table 3.

Clustering produces groups of the most similar countries in terms of tax policy during the first 17 years of the 20th century. The countries are not grouped by a detailed analysis of tax systems, but rather based on the contribution of individual tax entities to the total taxes or GDP. It is done year by year in a completely algorithmic way. Therefore, it is not easy to give the ideal description of the groups obtained by automatic division; but, based on the averages, we can interpret the relative ordering of individual groups. In accordance with the clustering, Group 3 is made up of countries for which the values of indicators IND.1 and IND.4 were on average the highest, while values of indicators IND.2 and IND.3 were the lowest. Group 1 brings together countries with the lowest values of IND.1 and IND.4 and moderate values of the two remaining indicators that inform about corporate taxes (taxes on corporate income or profits). It should be emphasized that the resulting clustering applies to the selected indicators for a given group of countries in a specific period. Changing the indicators or study period will potentially affect the groups’ composition.

A research procedure that identifies the speed of convergence with the reference country in a given group was carried out for all three groups of countries included in Table 3. The comparison of results of analyses for the four indicators is presented in Figures 3–5. The value “0” indicates the period of convergence between the countries and “1” indicates the period of divergence of the considered group of countries from the reference country. Compared to the global convergence for the total group of 11 countries, the behavior in the subgroups shows some differentiation, which means that some groups are more similar than others in terms of the tax policy (at least in the considered time period).

Figure 3.

Figure 4.

Figure 5.

In Group 1, due to the lack of data available for Croatia for the years preceding 2002, the analysis was confined to the period 2002–2018. For all indicators in this group, except for IND.3, Croatia turned out to be the reference country. For IND.3, convergence was researched with reference to Romania. It could be understood that, in Romania, the relation between tax revenues received on the basis of CIT and other taxes was the highest in comparison to the same in other countries. Group 2 consists of only two countries, and every time, we studied the speed with which the Slovak Republic converged with the Czech Republic. Finally, in Group 3, the reference countries for the individual indicators were Poland (IND.1 and IND.2), Estonia (IND.3), and Latvia (IND.4).

From the group perspective, conclusions about the course of convergence processes may differ significantly from those drawn collectively for all countries. One reason is the change of the reference country and its tax policy (changes in CIT and PIT), and the other lies in the homogeneity of groups of countries with regard to all the indicators under consideration. These differences can be observed in the course of the IND.4 indicator, which clearly shows a divergent behavior pattern for all the countries covered by the study (when applied to Group 1 only, it reveals a clear convergence). Attention should also be paid to changes in the rate of convergence over the period under consideration, e.g., for taxes on corporate income or profits (IND.2), especially for countries from Groups 2 and 3, in which a significant change in the speed of convergence occurred with the outbreak of the global financial crisis 2007–2009. At that time, some countries became similar in terms of the rate of CITs (the Czech Republic and the Slovak Republic, Group 2), but some countries began to differ in some degree when hit by the crisis (Group 3). Concerning the characteristics of individual tax systems, this can be explained by similar tax solutions applied by the countries in Group 2 (flat tax in PIT, lump-sum tax-deductible costs, and a similar range of tax preferences), i.e., a common source of tax regulation and the legal legacy of a common state. For example, an instrument that only occurs in the Czech Republic and the Slovak Republic is the lump-sum tax-deductible costs of running a business. It can therefore be said that the structural elements of CIT in both these countries are at a very similar level.

Group 3 countries, due to their heterogeneity and diversity of tax solutions, do not satisfy this condition. Let us recall, e.g., the so-called Estonian income tax, i.e., tax deferral until the profits are distributed (Estonia and Latvia); different CIT rates (e.g., 9% in Hungary, 15% in Lithuania, and 19% in Poland and Slovenia); and different tax preferences (reliefs for innovation, SEZs).

Despite differences in the tax policies of the CEECs manifesting themselves in the original method that the authors applied (by grouping CEECs), some regularities can be observed over the period under examination. First of all, income taxes paid on corporate income, which are far from the theoretically effective models, do not prevail in the structure of tax revenues of these countries. Moreover, in most countries, a reduction of the income tax burden has been noticed. However, the process of reducing the tax burden should be seen not only in the context of globalization-driven tax competition but also as an exploration of tax rates that would be consistent with long-term economic growth.