Analysis of the relationship between tax revenue and gross value added in the Romanian economy

Ogneru, Victor Article Analysis of the relationship between tax revenue and gross value added in the Romanian economy Financial Studies Provided in Cooperation with: "Victor Slăvescu" Centre for Financial and Monetary Research, National Institute of Economic Research (INCE), Romanian Academy Suggested Citation: Ogneru, Victor (2019) : Analysis of the relationship between tax revenue and gross value added in the Romanian economy, Financial Studies, ISSN 2066-6071, Romanian Academy, National Institute of Economic Research (INCE), "Victor Slăvescu" Centre for Financial and Monetary Research, Bucharest, Vol. 23, Iss. 2 (84), pp. 37-55 This Version is available at: https://hdl.handle.net/10419/231676 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/4.0/ ANALYSIS OF THE RELATIONSHIP BETWEEN TAX REVENUE AND GROSS VALUE ADDED IN THE ROMANIAN ECONOMY Victor OGNERU, PhD Student Abstract Any government is interested in knowing, to a certain extent degree, the level of tax revenue at a given time in order to design public expenditures. On the other hand, this level of budget revenue is desirable to be sustainable, i.e. to be supported by the existing economic conditions at a given moment. One way to estimate the expected revenues is the relationship of the tax bases with the main macroeconomic indicators. It is assumed that the main source of tax is gross value added in the economy. This article examines the nature of gross value added links with the tax revenue, on the one hand, and with the tax bases of each category of tax, on the other hand, in order to identify the best predictors of tax revenue for Romania. The analysis was carried out using multiple time series regressions in the cases of Romania and the standard (benchmark) states (Germany, France, the United Kingdom and Italy), respectively regressions on cross-sectional data in the case of Member States of the European Union. Keywords: tax revenue, macroeconomic analysis, public finance, econometric modeling, estimation JEL Classification: H20, C51 1. Introduction Securing budget revenues in a sustainable manner is the main concern of any responsible government. In the literature, the sustainability dimension is studied in complex macroeconomic models and is analyzed in relation with the most important factors defining an economy. From an operational perspective, economists are interested in the type of relationships that exist between tax revenues and macroeconomic or social indicators, in order to identify the best ways  The Faculty of Economic Cybernetics, Statistics and Informatics, Bucharest University of Economic Studies, Bucharest, Romania. Financial Studies – 2/2019 38 of evaluating and forecasting the level of public funds, within the timeframe for which they are based, and accordingly design fit for purpose public policies. Moreover, establishing the nature of a type of relationship between these macroeconomic aggregates may allow future development of top-down methodologies to assess the risk of not achieving a certain level of income. This study aims to evaluate the relationship between tax or budget revenue and gross value added in the economy (GVA). Why GVA? For a neutral observer, this relationship might appear to be a common one, and for a macroeconomist this kind of connection is easy to understand. Perhaps the lack of "evidence" is the reason why this relationship has not been studied at all in literature. The present study shows that things must be judged with caution, despite some existing empirical evidence. In every economic doctrine on taxation, ranging from the two extremes, neoliberalism and statism, any state, however liberal or centralized it is, is interested in consolidating its revenue. And beyond fiscal policy changes, as a lever of intervention, it is interesting to what extent a state can secure a steady level of income. While the issue has been extensively dealt with in several papers, literature remains poor, primarily because of the very high interest in assessing the public spending mechanism and the presumption of (partial) neutrality of the level of taxation. GVA is defined in national accounts as the production account's balance, measured as the difference between the value of goods and services (valued at basic prices) and intermediate consumption (measured at purchaser's prices) and is calculated before determining the consumption of fixed capital. Therefore, after deducting the depreciation, the net value is obtained. Therefore, GVA represents the newly created value in the production process of an economy. The very definition of this indicator reveals that there should be a causal link between GVA and tax or budget revenue and, moreover, an economically meaningful one. Firstly, the newly created value implies the accumulation of income at the level of the economic agents, which is the basis of taxation for the profit tax. Secondly, generating new value in an economy involves the payment of wage costs or other forms of remuneration for labor or capital, which in turn constitute the tax base for income tax and mandatory social contributions. Thirdly, as added value is created, money can be released in the economy in the form of wages or other income that go, even in part, into consumption, which in turn supports the creation of new value. Therefore, there should be Financial Studies – 2/2019 39 mutual influences between consumption and GVA, the first being a tax base for value added tax (VAT) and excise duties. If we also take into account the way microeconomic value added is calculated, we can deduce that there should be a close link between the two variables (tax revenues and GVA). It is worth mentioning that the macroeconomic analysis mainly studies the relationship between GVA and investments, the latter being considered as a generator of new value. From a strictly fiscal point of view, this type of relationship can be omitted. In a comprehensive analysis, however, the influence of interplay between investment (capital formation) and value added and the level of taxation on the achievement of a certain level of budgetary or fiscal revenues should provide an interested topic for study. At microeconomic level, determining and studying added value is an important step as this indicator reflects the ability of an economic entity to generate new value and to resume its work at higher levels of performance. Two methods of calculating the added value are known and used in corporate finance: the subtractive method and the additive method. By subtracting, the added value (VA) is determined by the formula: VA = trade margin + production value of the exercise - value of consumption This relationship shows that there should be a significant relationship at least between GVA and tax revenue from corporate taxation. The additive method consists in applying the following formula: VA = staff costs (including social contributions) + taxes (excluding VAT) + dividends + reinvested earnings + depreciation This formula results, also, in a relationship between VAB and income from taxes and contributions, less VAT. Taking into account both the macroeconomic and the microeconomic perspectives presented above, it may be assumed that there are causal and/or direct relationships between the two variables, GVA and tax revenues (as increases or decreases in the level of one generates increases or decreases in the level of the other). 2. Literature review Perhaps the most extensive research on the determinants of tax revenue is the study of A.S. Gupta, Determinants of Tax Revenue Efforts in Developing Countries (Gupta, 2007), and is based on Financial Studies – 2/2019 40 regression models applied to panel data. The analysis is focused on identifying the main factors that can influence the tax collection capacity of a country. One of the most important findings of Gupta's study is that there are no "universal" recipes to assess a state's tax performance. The analyzed factors are: GDP per capita in current prices, share of agriculture in GDP, weight of imports in GDP, weight of state aid in gross national income, share of public debt in gross national income, tax revenues from goods and services (as a share in total revenues), tax revenues from taxes on income, profit and capital gains (as a share of total income), tax revenue from commercial activities, as well as tax revenue from exports (as a weight in total revenues), the highest marginal tax rate for individuals, the highest marginal corporate tax rate, the average tariff applied to trade between countries and institutional factors such as political stability, economic stability, corruption, law enforcement and governmental stability. Concerns for identifying tax revenue determinants and understanding tax patterns and fiscal potential, especially in emerging economies, are not new (see Chelliah, 1971). At the beginning of the 1970s, the research effort was focused on explaining the differences in tax regimes (especially the differences in quotas and the differences in the weights that fiscal revenues hold in GDP in different countries). One of the strongest explanatory factors is the share of agriculture in GDP, and it is expected that as the share of value added in the agricultural sector increases, tax revenue will decrease due to the narrower tax base existing in this sector (Tanzi, 1992). Vito Tanzi and Parthasarathi Shome (1992) find that the structure of the tax system is not relevant in an unstable macroeconomic environment; therefore, a certain tax regime will not be able to explain the variation in tax revenue under conditions of macroeconomic instability. From the point of view of these authors, but also of others (Gupta, 2007), the key factor in achieving a comfortable level of tax revenue is the level of corruption. Other authors, also using econometric modeling, find a direct relationship between gross value added in mining and tax revenue and an indirect relationship between gross added value in agriculture and these incomes, as well (Bahl, 1971). The direct relationship between the mining industry and tax revenue must be attributed to the very high share that this sector held in the seventh decade of the last century. More recently, attention was directed to assessing performance in terms of tax collection by inspecting the elasticity of tax bases (Sobel Financial Studies – 2/2019 41 and Holcombe, 1996; Bruce et al., 2006; Fricke and Suessmuth, 2014; Koester and Priesmeier, 2017). Davoodi and Grigorian (2007) highlight the impact of institutional factors on tax revenues, the only macroeconomic factor being GDP per capita. Institutional factors are also analyzed in most of the papers on the relationship between tax revenue and GDP (Gupta, 2007; Tanzi, 1992; Le et al., 2008; Javid and Arif, 2012). Other authors study the reverse relationship between tax revenue and economic growth (Ofoegbu et al., 2016), respectively, between fiscal policy and economic output (Baum and Koester, 2011). The study of Gobachew et al. (2018), one of the most recent in the field, considers macroeconomic factors only, in the case of Ethiopia. Based on the econometric modeling they found that structural factors, macroeconomic conditions (such as inflation rates) and external trade are levers that can improve tax revenue. It is also found that the inflation rate and the share of agriculture in GDP negatively affects tax revenue, while openness to foreign trade, the share of manufacturing and per capita income positively influence the level of tax revenue. Another recent study exclusively addresses the influence of structural and conjectural factors (Yi and Suyono, 2014), assessing the effect of the fiscal multiplier on both tax revenue and GDP, under the general hypothesis on the incompatibility between maximizing of tax revenue and consecutively maximizing the GDP. Finally, Aizenman and Jinjarak (2009) analyze government expenditures as a potential determinant of tax bases. These studies primarily focus on the relationship between GDP and tax revenue, even when contextual factors such as institutional ones are present. Thus, the link between gross value added in the economy and tax revenue is analyzed either indirectly (through economic output) or directly (by economic sectors), but we do not have an explicit picture of this type of relationship. This study aims to inspect the evolution of fiscal and GVA variables at different time points and across the EU as a whole, but in some selected countries. 3. Data and methodology 3.1. Description of the data The following macro-variables have been taken into account, the values of which have been extracted from national accounts: gross value added (GVA), tax revenue, indirect tax revenues, direct income, gross operating surplus, compensation of employees, mixed income Financial Studies – 2/2019 42 and final consumption. In the case of Romania, these variables are tracked in time series, consisting in annual values of the indicators, for the period 1995-2017. As a "blank sample", values of those variables were used for each Member State of the European Union in 2017 (cross-sectional data). In addition, for the comparative analysis, four EU-level benchmarks have been considered, at which the highest level of GVA is recorded. The data source was the European Commission's AMECO database. The extracted values were reconciled with the values in the national accounts published by Eurostat and the information regarding the tax revenue collected in Romania was confronted with the existing data at the level of the Romanian National Agency for Fiscal Administration and adjusted, where necessary. Together with the primary variables, GVA and aggregate tax revenue, variables reflecting or approximating the tax bases of the main taxes were also considered, in order to assess their relationships with the GVA. At the same time, the link between GVA and the main categories of taxes (direct and indirect) was also assessed. The macroeconomic aggregates considered as proxy tax bases are: • Gross operating surplus (GOS) approximates the tax base for corporate tax; • Compensation of employees approximates the tax base for wage tax and compulsory social contributions; • Mixed Income approximates the tax base for income tax obtained by individuals in self-employment arrangements; • Final consumption approximates the tax base for VAT and excises (generally for indirect/consumption taxes). Total revenue from taxes is by far the main category of public revenue. 3.2. Method Definitions and conceptualizations of Gross Value Added allow us to formulate hypotheses about the influence it may have on tax revenue. If we also take into account the neo-Keynesian theory regarding the relationship between output and tax, we can admit that the level of taxation does not significantly influence the achievement of a certain level of tax revenue. If the tax burden is increased, we expect an “eviction” effect and the compression of economic activity, due to higher potential tax bills. In the case of a reduction in the tax burden, Financial Studies – 2/2019 43 economic activity is expanding, leading to higher tax base, as economic expansion is achieved. Consequently, we can neglect the possible effect of fiscal policy changes in the analysis, considering only the link between GVA and these revenues. In the light of the above, the following working hypotheses are formulated: • There is a direct (positive) link between GVA and tax revenue; • The level of taxation is neutral with respect to this relationship; • GVA significantly determines the achievement of a certain level of tax revenue, ceteris paribus; • Considering the above assumptions, we expect revenue to be higher as GVA increases. If the hypotheses are confirmed, it is possible to identify, using the GVA criterion, what economic sectors or categories of taxpayers can provide or generate higher tax revenue or, in correlation with the level of tax compliance, may pose a lower or higher risk in terms of budget revenue. For studying the nature and significance of some relationships, the most appropriate method is classical regression modeling and Granger causality analysis. 4. Results and discussions 4.1. The link between GVA and tax revenue The central hypothesis of this study is that there is a relationship of determination between GVA and tax revenue, a hypothesis based on economic theory. At the level of Romania, the two variables co-vary, but the difference between the values of the two tends to increase over time (Figure 1). For the past years (2015-2017), a quasi-exponential increase in GVA may be observed, contrasting with the steep slowdown in tax revenue growth. Financial Studies – 2/2019 44 Figure 1 Evolution of GVA and tax revenue over 1995-2017 (annual data) 0 40 80 120 160 200 96 98 00 02 04 06 08 10 12 14 16 VAB pc VF Note: VAB pc = GVA in current prices; VF = tax revenue Source: AMECO database and own computations These most recent developments indicate a possible rupture of economic logic or, more precisely, a possible refutation of the hypothesis. Figure 2 Correlation between GVA and tax revenue (Romania, 1995-2017 and EU, year 2017) Source: AMECO database and own computations Variables GVA and tax revenue are strongly correlated, and a positive and linear relationship exists between the two (Figure 2). 0.00 20.00 40.00 60.00 80.00 0.00 50.00 100.00150.00200.00 Relationship between GVA and tax revenue in Romania (time series) 0.00 500.00 1000.00 1500.00 2000.00 0.00 1500.00 3000.00 4500.00 Relationship between GVA and tax revenue in EU (cross-sectional data) Financial Studies – 2/2019 51 The two models show that the only predictor for both GVA and tax revenue is consumption (statistically significant coefficient), and in the case of tax revenue, mixed income may be accepted as a predictor for a significance threshold of 5%. Although the regression models indicate the final consumption variable as the only predictor for GVA (regression coefficient is statistically significant for a significance threshold below 1%), the Granger causality is not verified. The test was performed for lags 2, 3 and 4, with only the relevant results being selected. There is causality between the GOS (EBE) and GVA variables, as well as between the employee compensation variable and GVA, the latter at lag 2. The relationship GVA - compensation of employees is also bidirectional for a 10% significance threshold (Table 9, Appendix). Broadly, it can be admitted that GOS (EBE) and compensation of employees significantly influence GVA. The link between GOS (EBE) and GVA seems to be rather negative (Table 9, Appendix). Consumption remains a good predictor for GVA. In short, the tax bases, with the exception of mixed income, influence the achievement of gross value added in the economy. Therefore, tax revenue can be assessed through GVA, but in the case of Romania some adjustments are needed to mitigate or, on the contrary, to highlight the impact of circumstantial or institutional factors. 5. Conclusions This article looks at the evolution of tax revenue for the first time in terms of Gross Value Added (GVA), which is considered a proxy variable for tax bases rather than GDP, unlike most approaches that analyze the link between tax revenue and GDP. GVA simultaneously captures both the influence of existing economic conditions at one time and the variation of the basis on which different tax regimes are applied. Furthermore, the relationship between tax revenues and GVA has been assessed with the tax bases of each tax category, so that variation patterns and reliable predictors of tax revenue can be identified. Although the GVA and tax revenues (as a whole) variables are almost perfectly correlated, and the regression analysis indicates a causal link between the two, where tax revenue is the dependent variable (adjusted coefficient of determination is 68%), the Granger causality is not verified in any sense of the implication between the two Financial Studies – 2/2019 52 variables in the case of Romania. However, it should be noted that for Romania the coefficient of regression is 0.3579, very close to the value of regression coefficient (cross-section data) at European level - 0.3686. Both models of regression confirm the existence of a positive causal relationship between GVA and tax revenue. Under these circumstances, we can expect that the EUR 1000 increase in gross added value will increase tax revenue by 360 euros. However, level estimates and Granger's causality check after staging variables, signal a possible skew of the model. In the first step, attention was focused on identifying patterns of variation according to tax categories, taking as a benchmark the links established between GVA and tax revenue of different tax categories in the countries with the highest gross added value from the EU, namely Germany, France, the United Kingdom and Italy. In all these countries except Italy, the link between the variables considered is perfectly linear, which corresponds to the economic model. In the case of Romania, the links between the GVA and the categories of taxes (direct and indirect) are not linear. This situation reveals that, in reality, atypical evolution in Romania does not correspond to an economic reality, but rather to other causes. One of the causes can be tax evasion. In the second step, the relationship between the GVA and the tax bases was assessed. The GVA link with the tax bases is perfectly linear, with the exception of mixed income, but this variation pattern is similar to the one existing at the level of the European Union. Therefore, from an economic point of view, Romania does not register atypical developments from the perspective of this type of relationship. Verifying the Granger causality for each GVA relationship with the tax bases shows that there is only a determination between GVA and GOS (EBE), and between GVA and compensation for employees. The causal relationship GVA - compensation for employees is even bidirectional for a materiality threshold of 10%. In the case of mixed income, an absence of direct causality was expected, but no causal relationship between GVA and consumption was verified, although the linkage is perfectly linear. Based on two multiple regression models, the link between GVA and tax bases was assessed, on the one hand, and the link between tax revenues and tax bases, on the other. From these models it was found that GVA variation is only explained by final consumption, in circumstances where the Granger causality was not verified in this case. Also, the variation in tax revenue is explained only by the change in final consumption, which can be attributed to the fact Financial Studies – 2/2019 53 that more than 50% of the state budget revenue comes from indirect taxes. In short, atypical variations in Romania are not economic in nature, but are due to other factors, one of which may be tax evasion. The causal link between GOS (EBE) and compensation of employees on the one hand and GVA on the other hand is due to the fact that the tax bases mentioned are, by definition, included in the GVA but at the same time these tax bases do not explain variation in GVA and tax revenue as a whole. Final consumption is a very good predictor for both GVA and tax revenues as a whole, but no strict causality between these two variables is identified, which can be attributed to the very large share of imports in final consumption. References Aizenman, J. and Jinjarak, Y. (2009) ‘Globalisation and Developing Countries—A Shrinking Tax Base?’, Journal of Development Studies 45:5, 653–671. Bahl, R. (1971) A Regression Approach to Tax Effort and Tax Ratio Analysis, International Monetary Fund, Staff Paper 18, 570–612. Baum, A. and Koester, G. (2011) The impact of fiscal policy on economic activity over the business cycle - evidence from a threshold VAR analysis, Deutsche Bundesbank Discussion Paper Series 1: Economic Studies, No. 03. Bruce, D., Fox W. and Tuttle M. (2006) ‘Tax Base Elasticities: A Multi-State Analysis of Long-Run and Short-Run Dynamics’, Southern Economic Journal, Vol. 73(2), pp. 315–341. Chelliah, R. J. (1971) Trends in Taxation in Developing Countries, International Monetary Fund, Staff Papers 18, 254–0331. Davoodi, H. and Grigorian D. (2007) Tax Potential vs. Tax Effort: A Cross-Country Analysis of Armenia’s Stubbornly Low Tax Collection, International Monetary Fund, Washington, DC. (IMF Working Paper No. 106). Fricke, H. and B. Suessmuth (2014) ‘Growth and Volatility of Tax Revenues in Latin America’, World Development, Vol. 54, pp. 114-138. Financial Studies – 2/2019 54 Gobachew, N., Debela, K.L. and Shibiru, W. (2018) ‘Determinants of Tax Revenue in Ethiopia’, Economics, Vol. 6, No. 1, 2018, pp. 58-64. doi: 10.11648/j.eco.20170606.11. Gupta, A.S. (2007) Determinants of Tax Revenue Efforts in Developing Countries, IMF Working Paper No: WP/07/184, Washington: International Monetary Fund. Javid, Y.A. and Arif, U. (2012) ‘Analysis of Revenue Potential and Revenue Effort in Developing Asian Countries’, The Pakistan Development Review 51:4 Part II (Winter 2012) pp. 51:4, 365–380. Koester, G. and Priesmeier, C. (2017) Revenue elasticities in euro area countries: An analysis of long-run and short-run dynamics, ECB Working Paper Series No 1989 / January 2017. Le, T.M., Moreno-Dodson, B. and Rojchaichaninthorn, J. (2008) Expanding Taxable Capacity and Reaching Revenue Potential: Cross-Country Analysis, The World Bank, Policy Research Working Paper 4559, March 2008. Ofoegbu, G.N., Akwu, D.O. and Oliver O. (2016) ‘Empirical Analysis of Effect of Tax Revenue on Economic Development of Nigeria’, International Journal of Asian Social Science, 2016, 6(10): 604-613. Sobel, R., and Holcombe, R. (1996) ‘Measuring the Growth and Variability of Tax Bases over the Business Cycle’, National Tax Journal, Vol. 49(4), pp. 535–552. Tanzi, Vito (1992) ‘Structural Factors and Tax Revenue in Developing Countries: A Decade of Evidence’ in Tanzi, V. and Davoodi, H.R. (ed.) Open Economies: Structural Adjustment and Agriculture Corruption, Growth and Public Finances, Washington, DC: International Monetary Fund (IMF Working Paper No. 182). Tanzi, V. and Shome, P. (1992) ‘The Role of Taxation in the Development of East Asian Economies’ in The Political Economy of Tax Reform, NBER-EASE Volume 1, University of Chicago Press: 31-65. Financial Studies – 2/2019 55 Yi, Feng and Suyono, E. (2014) ‘The Relationship between Tax Revenue and Economic Growth of Hebei Province Based on The Tax Multiplier Effect’, Global Economy and Finance Journal Vol. 7, No. 2, September 2014, pp. 1 – 18. APPENDIX Table 1 ADF test for the variables GVA and tax revenue Null Hypothesis: VAB_PC has a unit root Exogenous: Constant, Linear Trend Lag Length: 1 (Automatic - based on SIC, maxlag=6) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -2.704987 0.2443 Test critical values: 1% level -4.467895 5% level -3.644963 10% level -3.261452 *MacKinnon (1996) one-sided p-values. Null Hypothesis: VF has a unit root Exogenous: Constant, Linear Trend Lag Length: 0 (Automatic - based on SIC, maxlag=6) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -2.353094 0.3912 Test critical values: 1% level -4.440739 5% level -3.632896 10% level -3.254671 *MacKinnon (1996) one-sided p-values. Table 2 Assesment of the relationship, in level, between GVA and tax revenue Dependent Variable: VF Method: Least Squares Sample: 1995 2017 Included observations: 23 Variable Coefficient Std. Error t-Statistic Prob. C -0.168491 0.887527 -0.189843 0.8513 VAB_PC 0.368649 0.009166 40.21828 0.0000 R-squared 0.987183 Mean dependent var 31.17778 Adjusted R-squared 0.986573 S.D. dependent var 17.57095 S.E. of regression 2.036020 Akaike info criterion 4.342812 Sum squared resid 87.05292 Schwarz criterion 4.441551 Log likelihood -47.94234 Hannan-Quinn criter. 4.367645 F-statistic 1617.510 Durbin-Watson stat 1.136489 Prob(F-statistic) 0.000000 Table 3 Assesment of the relationship, in log, between GVA and tax revenue Dependent Variable: LOG(VF) Method: Least Squares Sample: 1995 2017 Included observations: 23 Variable Coefficient Std. Error t-Statistic Prob. C -1.169449 0.083089 -14.07465 0.0000 LOG(VAB_PC) 1.036336 0.019285 53.73675 0.0000 R-squared 0.992780 Mean dependent var 3.247132 Adjusted R-squared 0.992436 S.D. dependent var 0.672497 S.E. of regression 0.058487 Akaike info criterion -2.757094 Sum squared resid 0.071835 Schwarz criterion -2.658355 Log likelihood 33.70658 Hannan-Quinn criter. -2.732261 F-statistic 2887.639 Durbin-Watson stat 0.993960 Prob(F-statistic) 0.000000 Table 4 Assesment of the relationship, in differential, between GVA and tax revenue Dependent Variable: D(VF) Method: Least Squares Sample (adjusted): 1996 2017 Included observations: 22 after adjustments Variable Coefficient Std. Error t-Statistic Prob. C -0.147010 0.581001 -0.253029 0.8028 D(VAB_PC) 0.357970 0.052254 6.850628 0.0000 R-squared 0.701185 Mean dependent var 2.181371 Adjusted R-squared 0.686245 S.D. dependent var 3.945811 S.E. of regression 2.210202 Akaike info criterion 4.510553 Sum squared resid 97.69986 Schwarz criterion 4.609739 Log likelihood -47.61608 Hannan-Quinn criter. 4.533918 F-statistic 46.93110 Durbin-Watson stat 1.825343 Prob(F-statistic) 0.000001 Table 5 Assessment of the relationship between GVA and tax revenue at EU level, in 2017 (cross-sectional data) Regression Statistics Multiple R 0.993571074 R Square 0.987183479 Adjusted R Square 0.986573168 Standard Error 2.036019887 Observations 23 ANOVA df SS MS F Significance F Regression 1 6705.189311 6705.189311 1617.510142 2.34318E-21 Residual 21 87.05291662 4.145376982 Total 22 6792.242228 Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Intercept - 0.168490599 0.887526782 -0.189842834 0.851255333 -2.014203584 1.677222385 X Variable VAB 0.368648679 0.009166197 40.21828119 2.34318E-21 0.349586529 0.387710828 Table 6 Granger causality between GVA and tax revenue (in differential) Lags: 1 Null Hypothesis: Obs F-Statistic Prob. DVAB_PC does not Granger Cause DVF 21 0.17576 0.6800 DVF does not Granger Cause DVAB_PC 0.19065 0.6676 Lags: 2 Null Hypothesis: Obs F-Statistic Prob. DVAB_PC does not Granger Cause DVF 20 0.48719 0.6237 DVF does not Granger Cause DVAB_PC 0.11747 0.8900 Lags: 3 Null Hypothesis: Obs F-Statistic Prob. DVAB_PC does not Granger Cause DVF 19 0.86066 0.4879 DVF does not Granger Cause DVAB_PC 0.24189 0.8655 Lags: 4 Null Hypothesis: Obs F-Statistic Prob. DVAB_PC does not Granger Cause DVF 18 1.58694 0.2591 DVF does not Granger Cause DVAB_PC 1.32039 0.3336 Lags: 5 Null Hypothesis: Obs F-Statistic Prob. DVAB_PC does not Granger Cause DVF 17 1.02557 0.4782 DVF does not Granger Cause DVAB_PC 1.06421 0.4618 Lags: 6 Null Hypothesis: Obs F-Statistic Prob. DVAB_PC does not Granger Cause DVF 16 1.11345 0.5051 DVF does not Granger Cause DVAB_PC 0.56442 0.7478 Table 7 Multiple regression model between GVA and tax bases in Romania (stationary series) Dependent Variable: D(VAB_PC) Method: Least Squares Sample (adjusted): 1996 2017 Included observations: 22 after adjustments Variable Coefficient Std. Error t-Statistic Prob. C 0.259021 0.545378 0.474938 0.6409 D(COMP) 0.237391 0.221131 1.073529 0.2980 D(CONS) 0.961199 0.158557 6.062166 0.0000 D(EBE) -0.025263 0.071172 -0.354960 0.7270 D(VENIT_MIXT) 0.281491 0.223228 1.261003 0.2243 R-squared 0.959100 Mean dependent var 6.504395 Adjusted R-squared 0.949476 S.D. dependent var 9.230083 S.E. of regression 2.074693 Akaike info criterion 4.494220 Sum squared resid 73.17399 Schwarz criterion 4.742184 Log likelihood -44.43642 Hannan-Quinn criter. 4.552633 F-statistic 99.66128 Durbin-Watson stat 1.500032 Prob(F-statistic) 0.000000 Table 8 Multiple regression model between tax revenue and tax bases in Romania (stationary series) Dependent Variable: D(VF) Method: Least Squares Sample (adjusted): 1996 2017 Included observations: 22 after adjustments Variable Coefficient Std. Error t-Statistic Prob. C -0.311272 0.473541 -0.657328 0.5198 D(COMP) -0.328644 0.192004 -1.711655 0.1051 D(CONS) 0.485610 0.137672 3.527299 0.0026 D(EBE) -0.008950 0.061797 -0.144827 0.8866 D(VENIT_MIXT) 0.486571 0.193824 2.510369 0.0225 R-squared 0.831273 Mean dependent var 2.181371 Adjusted R-squared 0.791573 S.D. dependent var 3.945811 S.E. of regression 1.801415 Akaike info criterion 4.211738 Sum squared resid 55.16662 Schwarz criterion 4.459702 Log likelihood -41.32912 Hannan-Quinn criter. 4.270151 F-statistic 20.93862 Durbin-Watson stat 1.984141 Prob(F-statistic) 0.000002 Table 9 Granger causality between GVA and tax bases in Romania Lags: 4 Null Hypothesis: Obs F-Statistic Prob. DVENITMIXT does not Granger Cause DVAB_PC 18 0.23088 0.9141 DVAB_PC does not Granger Cause DVENITMIXT 0.59324 0.6764 Null Hypothesis: Obs F-Statistic Prob. DEBE does not Granger Cause DVAB_PC 18 8.08257 0.0047 DVAB_PC does not Granger Cause DEBE 0.72890 0.5943 Null Hypothesis: Obs F-Statistic Prob. DCONS does not Granger Cause DVAB_PC 18 2.33980 0.1332 DVAB_PC does not Granger Cause DCONS 1.55981 0.2657 Lags: 2 Null Hypothesis: Obs F-Statistic Prob. DCOMP does not Granger Cause DVAB_PC 20 3.39906 0.0606 DVAB_PC does not Granger Cause DCOMP 5.16655 0.0196