Application of Markov Chain Model to Provide the Appropriate Model of Tax Discount with Dynamic Programming Approach

Since at the present time, one of the major revenues of the government are from taxes, realizing the amount of tax revenues and making decisions of policies likely to enhance earnings in proper times with considering the tax current value is very important. According to the Markov chain method capabilities in understanding the possible random processes and dynamic programming methods, in this article, combination of these two methods with using information from income tax firms in the country in 1384-88 were used to represent a suitable pattern for discount in tax policy. Results showed that according to the prior data periods, the amount of non-received tax revenue was 3926 billion Rials. With 2% discount to customers and the tax received current value, it proceeded to 4511 billion Rials. In this way, the discount- index was determined for each year. As the process of tax revenues is also a transition period, paying taxes during the 12- month- period and the discount based on the tax current value were taking into consideration. The current value of future tax -cash flows during this period was equivalent to 16,034 billion Rials. Finally, using dynamic programming, the break event point of tax discount index was determined. Accordingly, in this manner, if the value of discount was less than 22%, the optimal policy for the government to increase the value of the received tax would be not to give discount to the customers and if the discount factor was over 22%, the favorable government policy would be providing discount to customers.