Abstract

This exploratory study analyzed the taxpayer compliance function at the Kenya Revenue Authority (KRA). The objectives of the study were three-fold. The first was to develop a regression model for taxpayer compliance risk management. The second objective was to identify determinants of taxpayer compliance risk while the third objective was to assess the effectiveness of the current risk profiling frameworks used by the revenue departments. The literature reviewed pointed at six categories of factors that determine taxpayer compliance. These are economic, social, psychological, demographic, institutional and political factors. These factors were further categorized into five areas that could be easily captured for regression purposes. These are cultural and behavioral factors, control process for complex transactions, financial performance, history of taxpayer compliance and erratic factors. A binary logistic regression model was specified where the dependent variable was a measure of payment compliance. The independent variables were also categorical in nature. The data used were obtained from the Large Taxpayers Office database covering the period 2009/10 to 2011/12. The results showed that the model was well specified and had an overall percentage of prediction of 77%. This was further supported by the goodness of fit as measured by the Hosmer and Lemeshow statistic and the pseudo R-squared by Nagelkerke. The variable selection procedure was stepwise forward selection (Wald). The results show that nine variables were significant namely: business ownership, characteristics of tax agents, performance targets, erratic performance of the sector, nature of business, financial performance of the taxpayer in terms of profitability and liquidity, company structure and frequency of investment deduction claims. This implies that cultural and behavioral factors, control of complex transactions, financial performance, history of taxpayer compliance and erratic factors are significant determinants of payment compliance. The model is recommended for taxpayer profiling based on the characteristics found to be significant determinants of taxpayer compliance. Whereas the study does not recommend a replacement of the current risk profiling frameworks used by revenue departments, it is recommended that more weight should be placed on the areas that were found to be significant in determining taxpayer compliance. It is further recommended that KRA needs to maintain a more extensive and systematic taxpayer database that permits a more effective risk assessment. This data base would go a long way in ensuring a more effective and transparent tax administration.