Document Type : Research Paper

Authors

• Roghaye Shojaeddin ¹
• Majid Sameti ²
• Zahra Dehghan Shabani ³

¹ Ph.D. Candidate, Department of Economics, University of Isfahan, Isfahan, Iran

² Professor, Department of Economics, University of Isfahan, Isfahan, Iran

³ Associate Professor, Department of Economics, University of Shiraz, Shiraz, Iran

Abstract

The significance of tax revenue as the primary source of government finance underscores the importance of accurately measuring tax efforts using unbiased methodologies. This study employs a state-space model and the Kalman filter algorithm to estimate tax effort as an unobservable variable within the tax revenue equation in Iran from 1970 to 2021.The findings reveal a nuanced relationship between various factors and the tax ratio. Per capita income exhibits a positive impact, while the agriculture share in GDP exerts a negative influence. Interestingly, the coefficients of openness and monetization initially have negative elasticity but transition to positive after reaching a certain threshold, indicating a dynamic relationship with the tax ratio. Conversely, the services and industry share in GDP demonstrate a positive effect on the tax ratio before reaching a peak, after which their squared coefficients turn negative.Tax effort in Iran, throughout the studied period, has never been more than 0.25 highlighting a significant disparity between actual and potential tax revenue and underscores inefficiencies within the tax system.

Introduction

Due to dependence on oil and structural problems, attention to important tax indexes has been neglected in Iran’s Economy. Therefore, in order to achieve more accurate results, a new approach has been taken by the research to estimate tax effort as an indicator that shows the ability of the government to enhance tax revenues.
Traditionally, tax effort is calculated by dividing actual tax revenue by potential tax revenue. However, this method is inherently biased as it fails to account for the influence of economic, social, and political factors on tax revenue collection, alongside tax effort itself. To address this limitation, the study employs the Kalman Filter estimation technique, which treats tax effort as an unobservable variable within the tax revenue function, alongside other economic variables.

Methods and Material

In literature, tax effort is calculated by estimating the following equation:




 
(1)


F₁ =  =   =




where T is tax share, F is tax effort, Z is a vector of other factors affecting tax share and ω is an error term. It is evident that the estimator F₁ is a biased estimator for tax effort F. Considering the effect of tax effort on tax revenue, the index should be included as a dependent variable in the tax revenue function. Referring to the study of Kim (2007) and in order to overcome the bias, the research suggests a state-space approach and Kalman Filter Algorithm. The structural time series method allows tax effort to be taken into consideration in the tax revenue function as an unobservable variable.
In this context, the tax revenue function considered a linear form incorporating variables such as per capita income, the share of agriculture, services and industry, openness, and monetization . However due to the low coefficient of determination and the results of Ramsey Reset test, adopting a quadratic function  became imperative. Consequently,the final equation was changed as follows:




 


(3)





Results and Discussion

In econometric analysis, the  stationary test of data typically examined. However, According to Harvey, the stationary test holds less signifcance in the structural time series model.. The analysis of variables confirms that all variables exibite statistical normality. The results of estimating equation 3 are reported in Table 1, which shows that the variables are significant at one percent level.
Table 1. The results of estimating the square function of tax revenue using the STSM method




Prob


t-statistic


RMSE


Coefficients


Variables




0.0090


-2.7836


0.0647


-0.1801


Level break 1998




0.0000


5.3930


0.0586


0.3159


Level break 2005




0.0000


-5.8246


0.0569


-0.3314


Level break 2000




0.0000


6.4744


0.9254


5.9920


LPY




0.0000


-6.0820


0.0611


-0.3719


LPY^2




0.0031


-3.2082


2.4578


-7.8851


LMO




0.0030


3.2130


0.2998


0.9631


LMO^2




0.0449


-2.0900


0.5520


-1.1536


LOP




0.0626


1.9312


0.0766


0.1480


LOP^2




0.0000


5.9359


3.2214


19.1223


LIND




0.0000


-6.1448


0.4668


-2.8682


LIND^2




0.0007


3.7434


7.1427


26.7385


LSEV




0.0004


-3.9563


0.8990


-3.5566


LSEV^2




0.0106


-2.7203


0.1102


-0.2997


LAGR




Reference: Research calculations and software output
There were breaks in 1998, 2000, and 2005. The cause of these breaks can be attributed to the Asian financial crisis, the dot-com bubble, and oil fever, respectively.
"In addition to examining elasticities, the squared coefficients of variables hold significance in the analysis. Despite the elasticity of the agricultural sector share being -0.29, its squared coefficient was omitted from the model due to its low explanatory power. Notably, the tax exemption status of the agricultural sector in Iran contributes to a negative impact on tax revenues.
Regarding per capita income, its elasticity is positive, yet its squared coefficient is negative. Initially, an increase in per capita income enhances tax revenues, but subsequently leads to a decline in the tax ratio. This phenomenon arises because governments can only collect a specific portion of per capita income as taxes. Continued taxation may result in taxpayer resistance, consequently leading to a reduction in tax revenue.".
The share of industry and services, both, have negative elasticity and squared coefficient. Initially, an increase in these variables leads to a rise in tax revenue, followed by a subsequent decrease where the negative effect predominates. Notably, only in cases where production is efficient, the industry can generate a significant taxable surplus.  Therefore, the inefficient industry sector will not result in higher tax revenue in Iran.
Due to the lack of a full database and since some economic activities in Iran are unregistered and consequently untraceable, a significant percentage of tax evasion occurs in the services sector. Hereupon, the increase of this sector in Iran will not lead to more tax revenue.
Both openness and monetization exibit positive quadratic coefficients. The effect of these variables on the tax ratio is negative at first and becomes positive after the minimum point. The negative elasticity of these two variables is respectively caused by the government's policies such as lower tariffs for essential goods and the adverse effects of inflation on monetization and tax revenue as a result.
The tax effort trend is shown in Figure 1. The unevenness of the trend is caused by the fluctuation in oil revenues in Iran. Tax effort in the last 50 years has always been lower than 0.25, which indicates the misutilization of tax capacities.
Figure 1. Tax Effort Trend in Iran during the years 1970-2021
 
Reference: Research Findings

Conclusion

Recognizing the significance of taxes as a primary source of government revenue, this research employs the Kalman Filter algorithm and a state-space model to calculate tax effort in a novel manner. In this approach, tax effort, treated as an unobservable variable, is incorporated into the tax revenue function alongside six other variables. Given the low coefficient of determination and the outcomes of the Ramsey reset test, the linear model was deemed unsuitable. Consequently, the quadratic form of the function was adopted to better capture the complex relationship between tax effort and tax revenue."The estimations showed that the effect of per capita income on tax ratio is positive due to the increase in the potential of citizens to pay taxes, and the impact of agriculture share is negative due to tax exemptions.
The elasticity of monetization is negative owing to high inflation in Iran and its adverse effects on sales tax. The effect of this variable on tax ratio is initially negative, but after the minimum point, it becomes positive due to the compliance of taxpayers with inflationary conditions. Openness also has a negative elasticity due to the negative effect of import promotion policies. The share of industry and services have a positive effect on the dependent variable before reaching the maximum point. However, due to a high rate of tax evasion within this sector and production inefficiency in manufacturing, this effect reverses after surpassing the maximum point..
The discrepancy in the signs of the elasticities for some variables can be attributed to utilization of  different approaches in estimating the tax effort. The low tax effort in Iran reveals the necessity to make changes in government tax policies to make the most of tax capacities.
For the purpose of enhancing tax revenue, some measures should be taken to reduce tax evasion and increase the tax potential of economic sectors. Eliminating unnecessary tax exemptions can also improve tax performance. The exemptions should be gradually phased out until they are completely eliminated,as long-term tax exemptions in Iran create non-competitive structures.

Keywords

• Tax Effort
• Tax Revenue
• Underlying Trend
• Kalman Filter Algorithm
• State-Space Model