LABOUR TAXATION IMPACT ON ECONOMIC GROWTH: DYNAMIC PANEL REGRESSION Rudolf Macek 1.

Transcription

1 LABOUR TAXATION IMPACT ON ECONOMIC GROWTH: DYNAMIC PANEL REGRESSION Rudolf Macek 1 1 VŠB-TU Ostrava, Economic Faculty, Sokolská třída Ostrava rudolf.macek@vsb.cz Abstract: Aim of the article is to verify the impact of labour taxation on economic growth in OECD for time period with application of dynamic panel regression model and implicit tax rates. From the methodological point of view the methods and tests of dynamic panel regression with Arellano-Bond estimator were used. From the empirical results it is evident (1) there exists a non-linear relationship between tax revenues (implicit tax rates) and tax burden (tax rates) (2) labour taxation negatively affects economic growth. There was also proven the hypothesis that labour taxation impact on economic growth is enhanced by its connection with labour market, because from the quantitative point of view it was the most negative. It is possible to state that labour taxation in OECD should be reduced in an effort to stimulate economic growth. Keywords: Labour Taxation, Economic Growth, Implicit Tax Rates, Laffer Curve, Dynamic Panel Regression, OECD JEL classification: E20, H20, H24, O47 1. Introduction Generally, it is important to say that redistributive processes can be considered as necessity in current globalized society due to fulfilment of elementary state s functions. Therefore the existence of taxes and government expenditures can be considered as justified. But on the other hand the tax system structure itself, tax mix or other tax characteristics are the subject of many polemics among economists and other experts. That is the reason why individual tax systems are noticeably heterogeneous. Also almost all of developed countries integrated in OECD are currently affected by significant crisis of public finance. Therefore on the one hand the individual governments have to reduce their government spending and to increase tax revenues on the other hand. In case of individual types of taxes impact on economic growth it is necessary to say that impact of labour taxation can be enhanced by its connection with labour market. Because the individual tax systems are significantly heterogeneous it is appropriate to work with suitable approximator of taxation which includes factors determining the final or real taxation rate. Therefore it is not desirable to work with traditional tax quota which is characteristics by many shortages and application of implicit tax rates can be considered as justified. Aim of this article is to verify the impact of labour taxation on economic growth in OECD for time period with application of dynamic panel regression model and implicit tax rates. The analysis is based on modified neoclassical growth model of Mankiw, Romer and Weil (1992). 2. Theoretical background of taxation impact on economic growth and literature review At introduction it is necessary to realize that taxation has potential to influence long run economic growth through its influence on physical and human capital, resp. on individual growth variables (Kotlán, Machová and Macek, 2014). Also the growth theories must be considered as fundamental elements of fiscal variables impact on economic growth explanation. These theories must be understood as the complex theories where the individual types of taxes are integrated simultaneously. Barro and Sala-i-Martin (2004) state that the basic growth model can derived from the Cobb-Douglas production function:. (1) -327-

2 The individual accumulation of physical capital in economy is possible to express as: In both these expressions (1) and (2) Y represents the total production; C presents the private consumption; δ is the depreciation rate; v expressed the part of the physical capital K which is dedicated to the production; w expressed the part of the human capital H which is dedicated to the production; A is the technological level and finally the coefficient α is the rate of diminishing returns to the physical capital. Furthermore, by the gradual derivation it is possible to derive the expression: (2). (3) It is evident that in equation (3) there are implemented all common taxes which are generally used in developed countries. Concretely, those are the labour taxes, property taxes, corporate taxes and consumption taxes. The whole process of derivation is possible to see in Kotlán and Machová (2014a) or Kotlán and Machová (2014b). Taxation influence economic growth through its influence on individual growth variables which are physical accumulation and human capital accumulation and therefore these channels will be described now. Pfaffermayr, Stockl and Winner (2008) or Daniel and Jeffery (2013) state that corporate taxation reduces not only the return of invested capital but also the structure of capital. Corporate taxation represents very important factor for decision making of investors and determines the extent of foreign direct investment (Mooj and Ederveen, 2001). Overesch and Wamser (2010) state that corporate taxation impacts has two basic aspects of supranational activities. First one is the localized decision and the second one is decision about making business. They argue that corporate taxation has negative impact on both of these decisions. The impact of corporate taxation can be different in developed and less developed countries, see Macek (2013) and also its impact is strongest one with two-year lag (Machová, Kliková and Kotlán, 2013). For horizon of fiscal policy see Kotlán and Macek (2014). Feld and Kirchgässner (2001) state the negative effect of labour taxation on economic growth can be visible in reduction of entrepreneurial activities and also in decline of savings which represent the key factor of economic growth and key restriction of investment realization (Leibfritz, Thornton and Bibbee, 1997). For completeness, the capital accumulation can be also influenced by consumption taxes which have same impact on investment as labour taxation (Salanié, 2003). Taxation can influence also human capital in case of widened neo-classical growth model. It is necessary to realize that positive dependency between taxation and economic growth exists if the tax revenues are used only for the human capital accumulation (Lin, 2001). 3. Empirical analysis: methodology, data and results Aim of this article is to verify the impact of labour taxation on economic growth in OECD for time period with application of dynamic panel regression model and implicit tax rates. In harmony with Barro and Sala-i Martin (2004), there will be analyzed the homogenous group of countries which can be countries with e. g. similar production functions, institutional parameters etc. This approach is fulfilled in our analysis, where the basic criterion for homogeneity is the country s membership in OECD. It is evident that OECD can be understood as a divergent group of countries. However in case of the e.g. European Union countries the problem can be caused by a limited number -328-

3 of observations and also the fact that the tax systems are harmonized and coordinated to some extent. It is necessary to realize that other more homogenous grouping of countries is not available. Also Barro and Sala-i-Martin (2004) state that groups of countries where the growth theory is valid (so called conditional convergence), exist within the growth theory; such group of countries can be considered the OECD countries. The data about the GDP amount per capita in purchasing power parity and government spending were drawn from OECD database National Accounts Statistics. 1 The information about the amount of ratio of investments on GDP is acquired from database Penn World Table (Penn World Table database ). 2 Information about human capital was acquired from the new database of Penn World Table (Feenstra, Inklaar and Timmer, 2013). Data about implicit tax rates are from database of Eurostat (2015). 3 The analysis is based on dynamic panel regression model, where cross section is represented by individual countries of OECD and time aspect is period Resulting number of observation is therefore equal to the product of number of analyzed periods and number of cross-sectional units. Dynamic panel is characteristic by the endogeneity existence which is caused by lagged value of dependent variable. This endogeneity means that lagged value of dependent variable can be correlated with independent variables and residuals. Therefore the model cannot be estimated with OLS (ordinary least squared) method but special differenced form of GMM (generalized method of moment) with instrumental variables (standard IV estimator, resp. Arellano-Bond estimator) has to be used (Arellano a Bond, 1991). The possible existence of autocorrelation and heteroskedasticity was eliminated with White period method. Main econometric program was E-views, version (8). Because the empiric analysis is based on implicit tax rates, now it is necessary to introduce them. Implicit tax rates represent an appropriate measure for effective tax burden comparison. These tax rates do not relate their tax collection to the GDP as a base, but to the activity with the tax is directly affected. Therefore it is possible to express the real tax burden of labour (ITRL), capital (ITRC) and consumption (ITRc). Furthermore, this approximator does not take into account only the level of statutory rates but also other aspects of tax systems e. g. tax benefits, tax reliefs or tax surcharges. Vogel (2009) states that if tax deductible items exist then no statutory tax rate will be at the same level as implicit tax rate. Mutual comparison of statutory and implicit tax rates reflects the tax incentives which are provided by state organs in individual countries (Bach, 2009). These rates are calculated according to harmonized systems of national and regional accounts ESA95. As it is stated above econometric analysis is based on Mankiw, Romer and Weil s model (1992). This modified model belongs among the most used ones because it widens the basic neoclassical growth model by human capital. According to this model and in regard with the dynamic panel regression, it is possible to write the explored mathematical formula shown in equation (1): i = ; t = (4.5) 1 OECD National Accounts Statistics available from: 2 Penn World Table available from: 3 Eurostat Statistics available from:

4 Individual variables of the analyzed model can be written down as: - GDP - gross domestic product growth per resident expressed by the amount of real GDP per capita in purchasing power parity in USD (dependent variable) [USD/resident.]; - CAP - capital accumulation approximated by indicator of proportion of real investments to GDP, expressed in purchasing power parity per one resident [%]; - HC - human capital expressed with the index of human capital per one resident based on the number of schooling years and return on investment in education; - GOV - total government spending as a % GDP [%]; - TAX - taxation rate approximated by tax implicit tax rates (ITR) [%]. At first, the individual variables entering the analysis were transformed into logarithms. Due to that, it is possible to interpret the resulting coefficients, if some independent variable changes by 1% this fact will lead to the growth or decrease of GDP growth rate by the amount of estimated coefficient. Then the stationarity of time series was tested, where it is tested the hypothesis of existence of single root. Time series stationarity of individual variables was explored by tests of Levin, Lin and Chu (2002); Im, Pesaran and Shin (2003) and ADF and PP tests according to Maddalu and Wu (1999). The existence of unit root was confirmed in case of gross domestic product, human capital, government spending and implicit tax rates of labour. Their stochastic instability were removed with using first differences which have already shown the stationarity. With regard to analyzed issue it is necessary to realize that all variables should have same base and form. Therefore the first differences were used in case of all variables so they entered into the model as a first difference of logarithmic value. Before the results of analysis will be shown it is appropriate to explore the potential connection between tax revenues expressed by implicit tax rates on labour and size of tax burden expressed by tax rate. In considering the representative type of tax rate it has to be taken account that tax rate differs in case of married couple or single persons and also in case of children in family etc. Therefore the tax rate related to single no-child person was chosen as the representative tax rate. Correlation between chosen variables was made for year 2012 for all OECD countries same as e.g. Kotlán and Machová (2012). In identifying and evaluating this potential connection numerous models were estimated where individual values were gradually transferred into line and parabola. Table no. 1 show the values of achieved coefficients of determination in individual cases. It is obvious that the coefficients of determination are low, but from the view of statistical verification the use of parabola is more appropriate. Tab. 1. Characteristics of Laffer curve in case of implicit tax rates on laobur ITRL Function R 2 Linear 0,226 Quadratic 0,264 Source: own calculations -330-

5 Pic.1. Laffer curve for implicit tax rates on labour Source: own calculations Picture 1 graphically illustrates the Laffer curve for implicit tax rate on labour. It is obvious that between tax burden (tax rate) and tax revenues (implicit tax rates on laobur) there exists non-liner relationship. This means that in case of low tax burden the positive correlation to the implicit tax rates exists and from higher tax burden about 40% the relationship is negative. Finally it is possible to verify the impact of labour taxation on economic growth. The results of empirical analysis are shown in table 2. Tab. 2. The impact of labour taxation on economic growth in OECD ( , dynamic panel regression) Dependent variable D(LOG(HDP)) Number of observations 131 J-statistics 7,86 Probability J-statistics 0,249 Coefficient t-statistics D(LOG(GDP(-1))) 0, ,519861*** D(LOG(CAP)) 0, ,80138*** D(LOG(HC)) 0, , D(LOG(GOV(-1))) 0, ,572674*** D(LOG(ITRC(-2))) -0, ,300786*** D(LOG(ITRL(-1))) -0, ,799963*** D(LOG(ITRC01(-2))) -0, , Note: *, **, *** represent the significance level at 10 %, 5 % and 1 %. Source: own computation From the table no.2 it is obvious that within the model as a whole two essential conditions for dynamics panel are fulfilled. Number of instruments is larger than J-statistics and probability of J-statistics (Sargan test) is at the level 24,9% which means that instruments are chosen properly. With high probability it is possible to say that endogeneity was removed from the model

6 It is obvious that lagged value of dependent variable is significant at 1% significance level which means that using of dynamic panel regression is justified. Capital accumulation is also statistically significant at 1% significance level and there was confirmed the positive relationship with the economic growth. This fact can be connected with conclusions of basic neoclassical growth model. In this model the growth of capital accumulation (represented by increased savings or investment activity) is the basic source of economic growth up to the achievement of stable state. Based on this and due to its statistical significance it is possible to assume that OECD countries have not reached the stable state yet. Human capital is not statistical significant, but this variable was kept in the model as a basic growth controlled variable. It is also clear that human capital is positively connected to the economic growth. Now it is possible to focus on fiscal variables (government spending and taxes). It is necessary to realize that econometric significance of fiscal variables time lag is justified because the change of government spending (taxes) level or structure is one of the basic economic-policy maker s decision. This decision is accompanied by certain time-lag. In this case there was proven the positive theoretical assumption which means that the positive effect of positive part of government spending prevails, see e.g. Kneller, Bleaney and Gemmell (1999) or Machová (2013). Corporate taxation is statistically significant at 1% significance level with negative relationship with economic growth. It is obvious that this type of taxation reduces the return of capital, FDI inflow and capital accumulation. In case of individual income taxes it is evident that these taxes reduces the economic growth and then gradually also welfare and the standard of living of working individuals. It is necessary to realize that this type of taxes reduces the creation of savings as the basic source of investments in the neoclassical growth model. Subsequently there is also a decrease of disposable sources which finance investments and by this the number of realized investments also falls. The next impact of this taxation is through its connectedness to the labour market. The growth of labour taxation causes the decrease of labour supply which can be problematic in case of high specialized work force. The increase of labour taxation also causes labour costs rise which creates pressure to lower the enterprises profits. Due to this, the enterprises abandon the investments localization and also the structure of capital accumulation changes. Also the increase in labour cost causes the substitution of labour by capital and decreases marginal product of capital. Therefore from the long-term view it comes to the growth of unemployment, which is also connected with a low rate of product growth. Consumption taxes are negatively connected to the economic growth, but this variable is statistically insignificant. 4. Conclusion Aim of the article was to verify the impact of labour taxation on economic growth in OECD for time period with application of dynamic panel regression model and implicit tax rates. Before the analysis was made there was explored the potential connection between tax revenues expressed by implicit tax rates on labour and size of tax burden expressed by tax rate. These relationship was non-linear which means that there exists Laffer curve between these variables. The empirical analysis was based on dynamic panel regression with special differenced form of GMM (generalized method of moment) with instrumental variables (standard IV estimator, resp. Arellano-Bond estimator). The lagged value of dependent variable (economic growth) was statistically significant so using of dynamic panel was justified

7 Capital accumulation has positive impact on economic growth and this variable was statistically significant. Therefore it is possible to say that that the OECD countries have not reached the stable state yet. In case of human capital there was proven the positive theoretical assumption but this variable was statistically insignificant. Government spending impact on economic growth is positive which means that the positive effect of positive part of government spending prevails Corporate taxation was statistically significant at 1% significance level. This type of taxation reduces return of capital, FDI inflow and finally also the capital accumulation. Personal income taxes were also statistically significant at 1% significance level with negative relation with economic growth. There was proven the hypothesis that the impact of labour taxation is enhanced by its connection with labour market. From the quantitative point of view its impact was the most negative. Consumption taxes were statistically insignificant but its impact on economic growth was negative. In case of tax quota use the impact was positive see previous author s work (Macek, 2014) and therefore it is possible to say that implicit tax rates are more appropriate approximator of taxation than tax quota. According to higher stated it is possible to say that labour taxation in OECD should be reduced to stimulate economic growth. Also the final tax system should be a result of society-wide consensus because tax uncerainty can be more harmful to economic growth than taxation rate itself (Macek, Machová and Kotlán, 2014). This paper was financially supported within the VŠB - Technical University SGS grant project No. SP2015/110 The Influence of Fiscal Deficit on Economic Growth and the Possibilities of its Decreasing in the Czech Republic. References [1] Arellano, M. and S. Bond Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Applicaiton to Employment Equations. The Review of Economic Studies, vol. 58, issue 2, p [2] Bach, S The Impact of Loses on Income Tax Revenue and Implicit Tax Rates of Different Income Sources. DIW Wochenbericht, Working Paper No [3] Barro, R. and Sala-i-Martin, X Economic Growth (2 nd ed). London: The MIT Press. [4] Daniel, B. and P. Jeffrey Investment Taxation and Portfolio Performance. Journal of Public Economics, vol. 97, issue 1, pp [5] Feld, L. and G. Kirchgässner The Impact of Corporate and Personal Income Taxes on the Location of Firms and on Employment: Some Panel Evidence for Swiss Cantons. Center for Economic Studies, Working Paper No [6] Feenstra, R., Inklaar, R. and M. Timmer The Next Generation of the Penn World Table. National Bureau of Economic Research, Working Paper No [7] Im, K., Pesaran, H. and Y. Shin Testing for Unit Roots in Heterogeneous Panels. Journal of Econometrics, vol. 115, issue 1, pp [8] Leibfritz, W., Thornton, J. and A. Bibbee Taxation and Economic Performance. OECD Economics Department, Working Paper No

8 [9] Lin, S Taxation, Human Capital Accumulation and Economic Growth. Japanese Economic Review, vol. 52, issue 2, pp [10] Kotlán, I. and Z. Machová. 2014a. Modern Thought and Integration of Taxation into Economic Growth Models. Pensee Journal, vol. 76, issue 6, pp [11] Kotlán, I. and Z. Machová. 2014b. Horizont daňové politky v zemích OECD. Politická ekonomie, vol. 62, issue 2, pp [12] Kotlán, I. and R. Macek. What is the Horizon of Fiscal Policy in OECD Countries? Pp in Proceedings of the 23rd International-Business-Information-Management-Association Conference on Visio 2020: Sustainable Growth, Economic Development, and Global Competitiveness. NORRISTOWN: IBIMA Publishing. [13] Kotlán, I. and Z. Machová Vliv zdanění korporací na ekonomický růst: selhání daňové kvóty? Politická ekonomie, vol. 60, issue 6, p [14] Kotlán, I., Machová, Z. and R. Macek Influence of Tax Uncertainty and Government Expenditures Volatility on Economic Growth Case Study of OECD Countries. Pp in Proceedings of the 7 th Annual Conference of the EuroMed Academy of Business THE FUTURE OF ENTREPRENEURSHIP. Kristiansand: EuroMed Press. [15] Kneller, R., Bleaney, M. and N. Gemmell Fiscal Policy and Growth: Evidence from OECD Countries. Journal of Public Economics, vol. 74, issue 2, pp [16] Levin, A., Lin, Ch. and Ch. Chu Unit Root Tests in Panel Data: Asymptotic and Finite-Sample Properties. Journal of Econometrics, vol. 108, issue 1, pp [17] Macek, R The Impact of Taxation on Economic Growth: Case Study of OECD Countries. Review of Economic Perspectives, vol. 14, issue 4, pp [18] Macek, R., Kotlán, I. and Z. Machová How does Tax Uncertainty Influence Economic Growth? Pp Proceedings of the 8th International Days of Statistics and Economics. Praha: MELANDRIUM. [19] Macek, R Is the Level of Corporate Taxation Equally Important in all Countries? A Case Study of Developed and Less Developed EU Countries. Pp in Proceedings of the 3 rd International Scientific Conference: TAXES IN THE WORLD. Ostrava: VŠB-TUO. [20] Maddala, G. and S. Wu A Comparative Study of Unit Root Tests with Panel Data and a New Simple Test. Oxford Bulletin of Economics and Statistics, vol. 61, issue 0, pp [21] Machová, Z Pro-growth Effects of (Un)productive Government Spending in the OECD. Pp in: Proceedings of the 3 rd International Scientific Conference: TAXES IN THE WORLD. Ostrava: VŠB-TUO. [22] Machová, Z., Kliková, Ch. and I. Kotlán, Corporate Tax Changes: When is Fiscal Policy Effective? Pp in Proceedings of the 3 rd International Scientific Conference: TAXES IN THE WORLD. Ostrava: VŠB-TUO. [23] Mankiw, G., Romer, D. and D. Weil A Contribution to the Empirics of Economic Growth. The Quarterly Journal of Economics, vol. 107, issue 2, pp [24] Mooij, R. and S. Ederveen Taxation and Foreign Direct Invesment: A Synthesis of Empirical Research. CESifo, Working Paper No

9 [25] Overesch, M. and G. Wamser The Effects of Company Taxation in EU Accession Countries of German FDI. Economics of Transition, vol. 18, issue 3, pp [26] Pfaffermayer, M., Stockl, M. and H. Winner Capital Structure, Corporate Taxation and Firm Age. University of Innsbruck, Working Paper No [27] Salanié, B The Economics of Taxation. Cambridge: MIT Press. [28] Vogel, L Tax Avoidance and Fiscal Limits: Laffer Curves in an Economy with Informal Sector. European Comission, Working Paper