Evaluation of the optimal rate of value-added tax in oil and non-oil sectors in Azerbaijan

Transcription

1 Evaluation of the optimal rate of value-added tax in oil and non-oil sectors in Azerbaijan Yadulla Hasanli 1,2, Nazim Hajiyev 3,4, Aynur Suleymanova 2 Abstract The optimal tax rate maximizes Gross Domestic Product and tax revenues to the budget, minimizing the level of tax debts and hidden economy. This theory is associated with the name and was first applied in the United States. It has since become a focus of politicians and academics in the world. It is true that such researches may not include many factors influencing the tax policy. That is why opponents of this type of researches have also emerged. However, determining the optimal tax rate on the way to improving the effectiveness of tax policy has not lost its relevance in the countries with high level of hidden economy and tax debts, also in the countries with rich natural resourses. In the study, the I and II types of the s in the non-oil sector were assessed based on shortterm and long-term econometric models. Note that the I type of the is the at which the volume of production receives the maximum estimate, and the II type of the is the that maximizes tax revenues. The following results were obtained from the study: - The optimal VAT rates in the oil and non-oil sectors in the long term are lower than those in the short term; - Due to the fact that the norm of value added in the oil sector is higher, in this area, different results were obtained. Thus, the optimal VAT rate in the oil sector should be higher than that in the nonoil sector. Thus, the results, obtained during the econometric analyzes give us opportunity to propose the application of different VAT rates in the oil and non-oil sectors. Key words: taxes, optimal VAT rate, oil and non-oil sectors,. JEL classification: H21 1 Scientific Research Institute of Economic Studies within Azerbaijan State University of Economics (UNEC), Baku, Republic of Azerbaijan 2 Head of Laboratory of Modeling the socio-economic processes in Institute of Control Systems of ANAS, Baku, Azerbaijan 3 Department of Economics and Business management in Azerbaijan State University of Economics (UNEC), Baku, Republic of Azerbaijan 4 Visiting Scholar, Davis Center for Russian and Eurasian Studies, Harvard University ( ), Boston, USA yadulla_hasanli@yahoo.com, n.hajiyev@unec.edu.az, nazimxx@yahoo.com, suleymanova.aynur@mail.ru

2 1. Introduction As it is known, taxes have 5 main functions: fiscal, distributing, regulating, integrating and controlling functions [Mammadov F., Mammadov A., Rzayev P., Xalilov S., Huseynov E., Seyfullayev I., Balakishiyeva Y. (2014)]. All types of taxes perform one or more of these functions in one or another way. The value added tax (hereinafter referred to as VAT) was first implemented in France by Maurice Lauré in In the beginning it was tested in Cote d'ivoire as a recognized province of France, but since 1958 it has been applied throughout France. Along with the fact that VAT plays an important role in ensuring budget revenues, one of its main functions is to control the economy. Thus, the use of VAT at all stages of the production process by calculating and reimbursing, and ultimately levying a tax on the final consumer by including it in the price of the product, makes it possible to control all areas of the economy. Nevertheless, in the conditions that the legislation and tax administration are not perfect the high level of VAT can lead to tax evasion and the growth of the hidden economy. In Azerbaijan, VAT began to be applied since 1992 as a result of the entry into force of the Law of the Republic of Azerbaijan on the Value Added Tax of December 31, When the VAT began to be applieng in Azerbaijan, its rate was 28 percent of the value added, in subsequent years this rate gradually decreased and now it is 18 percent of the value added. Since 2006, the use of sub-account VAT accounts helped to reduce the level of tax evasion and the volume of tax arrears. It should be noted that the share of VAT revenues in total tax revenues has gradually increased from 14.4 percent in 2006 to 33.0 percent in In 2016 and 2017, this indicator was equal to 29.8 and 26.7 percent, respectively. In addition, although the share of VAT revenues in the state budget revenues gradually decreased from 2006 to 2012, it tended to increase from 13.7 percent in 2012 to 20.7 percent in The relevance of research The optimal tax rate maximizes the volume of tax revenues to the budget and the Gross Domestic Product, minimizing the level of tax debts and hidden economy. This theory is associated with the name of and was first introduced in the US. It has become a focus of politicians and scholars in the world since then. It is true that such research can leave many factors that affect the effectiveness of tax policy. That is why opponents of this type of research have also emerged. But in countries where there are tax debts and the level of shadow economy is high, and also in countries with rich natural resources, determining the optimal tax rate on the way to improving the efficiency of tax policy has not lost its relevance. As mentioned above, one of the main functions of taxes is to control the economy. This function is mainly implemented by VAT. The imposition of an optimal level of VAT improves the effectiveness of fiscal policy, including tax policy. In Azerbaijan, rich in oil and gas resources, there are sharp differences in the value added of the oil and non-oil sectors. If one reason for this is the excess value received as a result of the exploitation of natural resources, then another reason is the application of various tax regimes [Hasanli Y. 2007, 2012, 2014]. Since on the basis of international oil contracts (Production Sharing Agreements) signed in the Republic of Azerbaijan, foreign oil companies belonging to the Consortium are exempt from all taxes except for profit tax. Each Contractor Party, the Operating Company and their Sub-contractors shall be exempt with credit (zero (0%) percent rate) from VAT in connection with Hydrocarbon Activities. However, the State Oil Company of Azerbaijan Republic (SOCAR) is a payer of all state taxes, including VAT. Therefore, the results of determining

3 the optimal VAT rate for the whole country may be incorrect. Thus, it is more expedient to calculate the optimal VAT rate, or rather s for the oil and non-oil sectors separately. Note that the I type of is the that maximizes the volume of the product, and the II type is the that maximizes the tax revenues. Implementation of effective tax policy depends largely on the optimal level of tax rates, including VAT rates [Balatsky E.V. 2005]. Evaluation of the optimal VAT rate makes it necessary to conduct certain studies. In a number of countries, including Azerbaijan, certain studies have been carried out to calculate the optimal tax burden (I and II type s), as well as to calculate the optimal rates of various taxes [Balatsky E.V. 1997, 1999; Hasanli Y. 2007, 2008; Musayev A., Amirov N., Hasanli Y. 2006; Kalbiyev Y. 2005; Movshovich S.M., Sokolovsky L.E.1994]. Considering that the optimal VAT rate for the country was not studied on a scientific basis and taking into account the need for regular research, a study was conducted to determine the optimal VAT rate using econometric modeling [Dougherty C. 2004; Magnus Y.R., Katyshev P.K., Peresetsky A.A. 2004; Tikhomirov N.P., Dorokhina E.Y. 2003; Valentinov V.A. 2006] and the results were analyzed. It should be noted that with the change in the economic situation and crises in the world economy due to the fall in oil prices in world markets, it became necessary to determine the optimal VAT rate for the current time in terms of a flexible and efficient tax policy. It also should be noted that there are certain differences in taxation in the oil and non-oil sector of the Republic of Azerbaijan. Thus, enterprises operating in the oil sector calculate and pay taxes in accordance with the Production Sharing Agreements. According to these Agreements, persons engaged in hydrocarbon activities pay only profit tax and personal income tax. Since Contractor Party, the Operating Company and their Subcontractors shall be exempt with credit (zero (0%) percent rate) from VAT in connection with Hydrocarbon Activities [Agreement on the joint development and production sharing for the extraction on oil field in Azerbaijan sector of Caspian Sea]. Therefore, the results of determining the optimal VAT rate for the whole country may be incorrect. That is why in this research we have used the information of tax revenues of SOCAR and non-oil sectr separately. As it is known in economic studies in the direction of determining the optimal rate for each type of taxes and increasing fiscal efficiency, the s of the I and II type are distinguished. Since the of the I type determines the tax rate at which GDP reaches its maximum, and at the of the II type, the tax rate at which tax revenues reach their maximum is determined [Musayev A., Amirov N., Hasanli Y. 2006]. 3. Theoretical and methodological bases for evaluating the optimal level of VAT It should be noted that despite the fact that the idea of determining the optimal tax burden, including the curve, which reflects the dependence between tax burden and tax revenues for the first time occured in the USA in the 80s of the last century, in subsequent periods, Western economists themselves began to critically evaluate this concept. Russian scientists, on the other hand, have used this idea to study the geometrical features of 's curve. It also should be noted that the concept of, as well as the differences between the s of the I and II type were also included in the scientific works of Russian researchers for the first time [Balatsky E.V. 2003]. When we get acquainted with the studies carried out to determine the optimal tax rates, we note that if in the early times such studies were aimed at determining the optimal tax burden in general, then recently these studies began to be conducted for each type of taxes.

4 4. Realization of models: Evaluation of optimal level of VAT in oil and non-oil sectors First of all, try to determine the VAT rate that maximizes the value added of the non-oil sector in other words, we will try to determine the I type for VAT on the non-oil sector. For this we estimate the following regression equation. Q_N_UDM(T) = C(1)*EDV_DER + C(2)*EDV_DER^2 + e(t) (1) Here: Q_N_UDM shows the value added of the non-oil sector. The following result was obtained from the econometric evaluation of the regression equation (1): Q_N_UDM = *EDV_DER *EDV_DER^2 + [AR(3)= ] (2) (t-statistic (8,385527) (-7,646605) R-squared = 0,750158; Adjusted R-squared = 0,678774; Durbin-Watson stat = 1, The statistical indicators of the regression equation (2) give the reason to say that the model is adequate. The optimal rate of VAT received from this equilibrium was 10,46 percent. Now we will try to determine the VAT rate that maximizes revenues of VAT (the II type ). For this we will estimate regression equation (3). Q_N_EDV_D(T)= C(1)*EDV_DER(t) + C(2)*EDV_DER^2(t) + e(t) (3) Here, e(t) is a random variable. We should note that equations 1 and 2 are dynamic models for a long period since they involve the time factor and the volume factors of GDP and VAT revenues. In other words, the models reflect the characteristics of the long-term period. Economic patterns reflecting the non-linear dependence of value added of non-oil sector on VAT rate and dependence of VAT revenues on VAT rate are studied on the statistics covering the years in the models (1) and (2) respectively. So, we can say that the optimal VAT rate obtained from this model is reliable in the longterm period. Q_N_EDV = *EDV_DER *EDV_DER^2 (4) (t-statistica) ( ) ( ) R-squared = ; Adjusted R-squared = ; Durbin-Watson stat = The statistical indicators of the regression equation (4) and statistical tests give the reason to say that the model is adequate. The actual, fitted and residual dynamics of VAT of non-oil sector are shown in Graph 1. 2,500,000 2,000,000 1,500, , , , ,000, , , , , Residual Actual Fitted Graph 1. The actual, fitted and residual dynamics of VAT of non-oil sector.

5 As can be seen from the Graph 1, there is a similarity (correlation) between the values of the VAT revenues obtained from the model and the actual values. From model (4) we have got that, the optimal VAT rate that maximizes VAT revenues is 10,6 percent. Summarizing the abovementioned, it is possible to say that, in the long run the volume of GDP is maximized at a 10,46 percent VAT rate, and tax revenues are maximized at a 10,6 percent VAT rate. As we have noted, this result characterizes a long-term period. That is, the volume of GDP and tax revenues reach their optimal level when the VAT rate drops to 10,46 percent and 10,6 percent, respectively. Since the stability of the tax policy is the main indicator of the effectiveness of tax policy, it is not advisable to reduce the VAT rate to the optimal level found suddenly. It should also be added that the optimal level was obtained on the condition that other factors affecting VAT revenues (for example, GDP, tax administration, changes in management, inflation levels, etc.) remain unchanged. But in the long-term period, these factors do not remain stable. Therefore, in order to determine the optimal level of the current VAT, it is necessary to study short-term models. In short-term models, the increase in the indicators of volume-time series and the rate of increase are observed, and the volume indicators and at the same time the indicators of time are removed. This is achieved by using the logarithm (LOG) or changing the logarithm (DLOG) in econometric models. The logarithm of the variable (LOG) reflects the percentage of its change, and the change in the logarithm (DLOG) reflects the rate of increase in the change. Note that with the inclusion in the model of changing variables or the rate of change of variables, we create an important condition - stationarity of the time series, and this in turn removes the problems from the model created by the nonstationarity of the time series. LOG(Q_N_UDM) = C(1)*EDV_DER + C(2)*EDV_DER^2 (5) The model obtained from the econometric estimation of this regression equation is as follows: LOG(Q_N_UDM) = *EDV_DER *EDV_DER^2 (6) (t-statistica) (24,04318) (-14,24709) R-squared = 0,717848; Adjusted R-squared = 0,682580; Durbin-Watson stat = 1, The statistical indicators of the regression equation (6) and statistical tests give the reason to say that the model is adequate. The maximum (the where the first-order derivative equals zero) of the function expressed by the regression equation (6), in other words, the VAT rate that maximizes the value added of the non-oil sector in the short-term period is equal to 15,97 percent. To determine the VAT rate, optimizing the VAT revenues in the non-oil sector in the short-term period, we will estimate the following regression equation: LOG(Q_N_EDV_D) = C(1)*EDV_DER + C(2)*EDV_DER^2 (7) Here, EDV_DER- is the VAT rate. Due to the fact that not the VAT revenues but their increase (LOG(EDV_D)) are involved in the equation (7), the time factor loses its significance. Thus, in the econometric estimation (7), the increase in VAT revenues can be 12 percent in any year, and in evaluation it does not matter whether it was in 2000 or in In other words, the influence of current factors influencing the change in VAT receipts for a short period of time is studied. Therefore, a fixed coefficient reflecting the influence of factors influencing the result, but whose values remain stable in the long run, usually does not participate in short-term models.

6 LOG(Q_N_EDV) = *EDV_DER *EDV_DER^2 +e(t) (8) (t-statistica) ( ) ( ) R-squared = ; Adjusted R-squared = ; Durbin-Watson stat = The statistical indicators of the regression equation (8) and statistical tests show that the model is adequate. The dynamics of the actual, calculated from the model (fitted) logarithmic values of the VAT revenues and their difference (residual) is shown in the graph Residual Actual Fitted Graph 2. The dynamics of the Actual, Fitted (8) and Residual logarithmic values of the VAT revenues The study of the regression function of the econometric model (8) constructed for a short period shows that the extremum is 0, In other words, the VAT rate, which gives the highest level of VAT revenues in the short-term period, is about 15,2 percent. This result has once again proved the theoretical propositions we have mentioned above. Thus, the optimal VAT rate for the short-term period is not 10,6 percent, but 15 percent. That is, if in subsequent years the VAT rate is reduced to the optimal rate of 15 percent, this will ensure an increase in VAT revenues to the maximum level. If other factors affecting VAT revenues remain stable, then the reduction of the VAT rate should take a regular character and stabilize at about 10,6 percent. From real life it is known that in the long-term period, factors affecting VAT revenues are subject to certain changes. Therefore, every year it is necessary to conduct regular researches to determine the optimal VAT rate for both long-term and short-term periods. It should be noted that the results of the macroeconomic model of "Azerimodel" for the long-term period, carried out by the UK OED-Oxford Economics Forecasting experts at State Oil Company of Azerbaijan Republic, show that the effective rate of VAT is percent (look at: http: And now we will begin to determine the optimal VAT rate for the oil sector. First, we find the I type for VAT for the oil sector in the long-term period. For this we will estimate regression equation below: EDV_NEFT_DOVRIYYE = C(1)*EDV_DER + C(2)*EDV_DER^2 (9) Here: EDV_NEFT_DOVRIYYE- the value added tax declared by taxpayers on the oil sector as well as the value added tax determined and collected by the tax authorities. In spite of the fact that the added value created by the enterprises operating under the Production Sharing Agreement in the Azerbaijan Republic is included in the GDP of the oil sector, the taxation of these enterprises by the zero (0) VAT rate was the reason that when finding the I type, the model did not include information on value added in the oil sector, but information on the turnover and receipts of VAT of

7 enterprises operating in the oil sector and taxed in accordance with the Tax Code of the Republic of Azerbaijan. The result of the econometric evaluation of the regression equation (9) was as follows: EDV_NEFT_DOVRIYYE = *EDV_DER *EDV_DER^2 (10) (t-statistica) ( ) ( ) R-squared = ; Adjusted R-squared = ; Durbin-Watson stat = According to the results obtained from the equation (10) in the long-term period I type on VAT equals to 13,33 percent in the oil sector. Now to calculate the VAT rate that maximizes VAT revenues in the oil sector in the long run, in other words, to find the II type, we will estimate the following regression equation: EDVNEFT = C(1)*EDV_DER + C(2)*EDV_DER^2 (11) Here: EDVNEFT- is VAT revenues of enterprises operating in the oil sector and taxed in accordance with the Tax Code of the Republic of Azerbaijan. The model obtained from the econometric estimation of regression equation (11) is as follows: EDVNEFT = *EDV_DER *EDV_DER^2 (12) (t-statistica) ( ) ( ) R-squared = ; Adjusted R-squared = ; Durbin-Watson stat = According to the results obtained from the equation (12) in the long-term period II type on VAT equals to 11,57 percent in the oil sector. And now we find I type on VAT for short-term period in oil sector. For this we have to evaluate the following regression equation: LOG(EDV_NEFT_DOVRIYYE) = C(1)*EDV_DER + C(2)*EDV_DER^2 (13) The result of the econometric evaluation of the regression equation (13) was as follows: LOG(EDV_NEFT_DOVRIYYE)= *EDV_DER *EDV_DER^2 (14) (t-statistica) ( ) ( ) R-squared = ; Adjusted R-squared = ; Durbin-Watson stat = According to the results obtained from the equation (14) in the short-term period I type on VAT equals to 17,32 percent in the oil sector. And in the end we will try to find a VAT rate that maximizes VAT revenues in the oil sector in the short term period. For this we will estimate the following regression equation: LOG(EDVNEFT) = C(1)*EDV_DER + C(2)*EDV_DER^2 (15) The following results were obtained from evaluation of the regression equation (15): LOG(EDVNEFT) = *EDV_DER *EDV_DER^2 (16) (t-statistica) ( ) ( ) R-squared = ; Adjusted R-squared = ; Durbin-Watson stat = According to the results obtained from the model (16) in the short-term period VAT rate that maximizes VAT revenues equals to 16,89 percent in the oil sector.

8 The results of the researches carried out to determine the optimal rate of VAT in a generalized form are given in the table below: Table 1. Information on optimal levels of VAT rates in short and long term oil and non-oil sectors Non-oil sector Oil sector Long-term period Short-term period Long-term period Short-term period I type II type I type II type I type II type I type II type 10,46 10,6 15,97 15,2 13,33 11,57 17,32 16,89 source: calculated by authors There is a practice of widespread use of the elasticity coefficient in the analysis and forecasting of tax revenues. The elasticity coefficient answers the question of how many percent one indicator changes when another indicator changes by one percent. The following regression equation was evaluated for this purpose. LOG(Q_N_EDV_D) = C(1) + C(2)*LOG(EDV_DER) (17) Here, coefficient C (2) is the coefficient of elasticity of tax revenues (EDV_D) at the tax rate (EDV_DER). LOG(Q_N_EDV) = *LOG(EDV_DER) (18) (t-statistica) ( ) ( ) R-squared = ; Durbin-Watson stat = The statistical indicators of the regression equation (18) and statistical tests show that the model is adequate. It should be noted that this model is more adequate because the statistical characteristics of this model are better than the statistical characteristics of all previous models. More precisely, the results of this model have a higher percentage of reliability. Graph 3 shows the dynamics of the logarithmic values of the actual VAT revenues (Actual), calculated from the model (18) (Fitted) and the difference between them(residual) Residual Actual Fitted Graph 3. the dynamics of the logarithmic values of the actual VAT revenues (Actual), calculated from the model (18) (Fitted) and the difference between them(residual). Model (18) shows that coefficient of elasticity of VAT revenues (EDV_D) at the rate of VAT equals to In other words, an increase in the VAT rate by 1 percent reduces the VAT revenues by 9,7 percent.

9 Results So the results are shown below: - Revenues from VAT reach their maximum when the VAT rate in the non-oil sector in the long-term period is approximately 10,6 percent. As it was noted this result characterizes the long-term period. That is, when the VAT rate reaches 10,6 percent as the last limit, VAT revenues reach the optimal level. Since the stability of the tax policy is the main indicator of the effectiveness of tax policy, it is not advisable to reduce the VAT rate to the optimal level found suddenly. It should also be added that the optimal level was obtained on the condition that other factors affecting VAT revenues (for example, GDP, tax administration, changes in management, inflation levels, etc.) remain unchanged. But in the long-term period, these factors do not remain stable. Therefore, to determine the optimal level of the current VAT rate, we need to review the short-term models. - The VAT rate in the short-term period, which gives the highest level of VAT revenues in the non-oil sector, is approximately 15,2 percent. This result has once again proved the theoretical propositions we have mentioned above. Thus, the optimal VAT rate for the short-term period is not 10,6 percent, but 15 percent. That is, if in subsequent years the VAT rate is reduced to the optimal rate of 15 percent, this will ensure an increase in VAT revenues to the maximum level. If other factors affecting VAT revenues remain stable, then the reduction of the VAT rate should take a regular character and stabilize at about 10,6 percent. From real life it is known that in the long-term period, factors affecting VAT revenues are subject to certain changes. Therefore, every year it is necessary to conduct regular researches to determine the optimal VAT rate for both long-term and short-term periods. It should be noted that the results of the macroeconomic model of "Azerimodel" for the long-term period, carried out by the UK OED-Oxford Economics Forecasting experts at State Oil Company of Azerbaijan Republic, show that the effective rate of VAT is percent (look at: http: - Due to the fact that the added value of the oil sector is high, we have obtained different results in this area. As in the oil sector, the VAT rate that provides the maximum level of VAT revenues in the longterm period is 11,57 percent, and in the short-term period 16,89 percent. As you can see, the optimal VAT rates obtained in this research for the oil sector are higher than that in the non-oil sector both in the long-term period and in the short-term period. And the main reason for obtaining such a result is that the specific weight of the added value of the oil sector in GDP is higher than that in the non-oil sector. The results obtained during the econometric analysis give us the opportunity to propose the application of different VAT rates in oil and non-oil sectors. - If the VAT rate increases (decreases) by 1 percent, the VAT income will decrease (increase) by 9,7 percent. References Agreement on the joint development and production sharing for the extraction on oil field in Azerbaijan sector of Caspian Sea Balatsky E.V., (2005), Stability of the tax system as a factor of economic growth. Society and economics, 2, pp Balatsky E.V., (1997), effects and financial criteria of economic activity, World Economy and International Relations, 11.

10 Balatsky E.V., (1999), The s and their quantification, World Economy and International Relations, 12. Balatsky E.V., (2000), Effectiveness of fiscal policy of the state, Problems of forecasting, IEF RAS, N0 5, p.32. Balatsky E.V., (2003), Analysis of the impact of the tax burden on economic growth through production-institutional functions, Problems of forecasting, 2. Christopher Dougherty, (2004), Introduction to Econometrics. EViews-7 User s guide. Hasanli Y., Suleymanov N., (2007), Statistical analysis and forecasting of VAT in the Republic of Azerbaijan. Statistical issues, Moscow, 11, pp Hasanli Y., (2007), Analysis and forecasting of VAT. Azerbaijan tax news, 10, pp Hasanli Y., (2008), Evaluation of the optimal rate of VAT. Azerbaijan tax news, 2, pp Hasanli Y., Suleymanov N., (2007), Analysis of intersectoral relations in the economy of Azerbaijan based on the Input-Output model, Statistical issues, Moscow, 1, pp Hasanli Y., (2012), Evaluation of final products impact of oil and non-oil sectors on key economic indicators, including tax revenues to the state budget, Reports of ANAS, 2, pp Hasanli Y., (2014), Research of impact of natural resources wealth on economic development and human capital using Marx s reproduction scheme, EcoMod 2014, İnternational Conference on Economic Modeling, Bali, Indonezia, pp Kalbiyev Y., (2005), Fiscal policy and problems of regulation of national economy. Baku, Elm, 468 p. Kontorovich G.G., (2002), Time Series Analysis, Economic Journal, HSE, 1, pp Magnus Y.R., Katyshev P.K., Peresetsky A.A., (2004), Econometrics. The initial course., 6-th edition, Moscow. Mammadov F., Mammadov A., Rzayev P., Xalilov S., Huseynov E., Seyfullayev I., Balakishiyeva Y., (2014), Introduction to taxes, Shamakhi. Movshovich S.M., Sokolovsky L.E., (1994), Output, taxes and the curve, Economics and Mathematical Methods, 3. Musayev A., Amirov N., Hasanli Y., (2006), Evaluation of the effectiveness of fiscal policy. Finding I and II type Lafer s, Azerbaijan tax news, 3, pp The Tax Code of the Republic of Azerbaijan. Tikhomirov N.P., Dorokhina E.Y., (2003), Econometrics, Moscow. Valentinov V.A., (2006), Econometrics, Moscow.