Abstract:
This study aims to examine the effectiveness of the tax presence in combating tax evasion and promoting tax compliance in the segment of small and medium-sized companies, domiciled in the 5th Tax Region of Brazil, states of Bahia and Sergipe, opting for the tax regime based on presumed profit. These companies were notified by the Brazilian Federal Revenue Service in 2021, due to inconsistencies found in the declared income of these legal entities, when compared with income reported in other databases such as financial transactions, electronic invoices, credit card sales, payments received and bookkeeping of sales transactions. Strategies with intelligent use of advanced techniques for analysing data from a large number of taxpayers make it possible to optimise public resources and expand tax presence, promoting a greater perception of the risk of a tax assessment. This is a descriptive study with a quantitative approach, using documentary sources of secondary data, using machine learning, association and regression functions, the aim of which was to estimate income evasion in the small and medium-sized enterprise segment. The data was collected from a Federal Revenue Service operation called ECF zerada, which extracted 1,144 registrations of legal entities with tax domiciles in the states of Bahia and Sergipe that incurred discrepancies in the revenue reported to the tax authorities, in the 2018 calendar year, when compared to other databases. The results show that medium-sized companies accounted for only 0.3 per cent of those notified, and that small companies accounted for a large proportion of those that did not respond positively to IRS notifications. With 42 per cent of the companies notified having rectified their declared income, this situation may be related to the lack of any idea of what audit probabilities these companies actually face. With an estimate of income evasion in the order of R$877 million in 2018, for small and medium-sized companies opting for presumed profit and based in the 5th Tax Region, the research advanced by calculating the taxes potentially due on the estimated evasion, resulting in a tax non-compliance indicator of around 45%. The intelligent use of advanced data analysis techniques broadens the base of taxpayers affected by non-compliance with their tax obligations, strengthens control and increases the perception of audit risk.
Frequent monitoring is shown to drive changes in taxpayer behaviour that result in greater tax
compliance. Improving the methodology for estimating tax gaps in the various segments will
help to devise inspection strategies with efficient use of resources, pursuing tax justice and
preserving the integrity of the Brazilian tax system, a primary source of funds to finance the
State and develop public policies.
