Michael McGrath (Cork South Central, Fianna Fail)
Link to this: Individually | In context | Oireachtas source

323. To ask the Minister for Finance if he will provide details of the Revenue Commissioners compliance project in respect of the self-employed; the scope of the project and the number of tax cases that will be examined; the results of the pilot project; and if he will make a statement on the matter. [1859/17]

Michael Noonan (Limerick City, Fine Gael)
Link to this: Individually | In context | Oireachtas source

It is assumed that the Deputy refers to work that the Revenue Commissioners have been engaged in during the course of 2016 to use 3rd party information and data analytics to identify income-tax payers who have understated their income to Revenue.

I am advised that this work is one of a number of projects designed to deliver on a specific objective of the Commissioners' published Statement of Strategy 2017-2019, which is to use data analytics to identify sectoral and behavioural risks and detect non-compliance. To this end, Revenue has established an in-house analytics team of eight permanent staff, supported by a technical services team of an additional four permanent staff. The role of this unit is to use statistical and machine-learning methods to support Revenue's risk assessment and operational decision-making. The focus of the team's work is on detecting potentially suspect anomalies in Revenue data, and on learning from the outcomes of past Revenue interventions in order to target non-compliance more accurately in future.   

The particular project referred to falls into the former category: it aims to identify non-compliant income-tax payers by highlighting those whose declared income is unusually low compared to their peers. To identify such taxpayers, the project draws on a wide range of data sources, including third-party information returns made by financial institutions, insurance providers, government bodies, and other tax administrations.  A predictive model built by Revenue's data analysts then compares each income-tax payer's declared income against the average declared income of other taxpayers with a similar expenditure profile and demographic characteristics. Taxpayers whose declared income is substantially lower than that of their peer group are flagged for further review (and possible intervention) by Revenue staff.  I am advised that similar programmes are used by a number of other tax administrations to better target non-compliance.

In a small ongoing pilot study, Revenue is examining an initial sample of 72 taxpayers identified through the means described above. When available, the results of this pilot will be used to evaluate the effectiveness of the approach and to refine the case-selection methodology used.  Depending on the effectiveness of the resulting methodology, this approach may be used more widely in future to select cases for Revenue audit or other compliance intervention.  The pilot study may also highlight additional sources of third-party and other data that are required to optimise the effectiveness of the predictive model.