Pat Rabbitte (Dublin South West, Labour)
Link to this: Individually | In context

Question 93: To ask the Tánaiste and Minister for Finance the number of random audits carried out by the Revenue Commissioners in 2006 and to date in 2007; the way this compares with each year from 2003 to date in 2007; the number expected to be undertaken during 2007; and if he will make a statement on the matter. [17173/07]

Brian Cowen (Laois-Offaly, Fianna Fail)
Link to this: Individually | In context

I have been advised by the Revenue Commissioners that 402 cases were selected for random auditing in 2006. As of 31 March 2007, 187 of these audits have been concluded. 215 cases are ongoing. I am advised that 410 cases were selected for random audit in 2005. 6 cases are still ongoing. Revenue's approach to random compliance testing changed significantly for their 2005 programme. The new approach introduced both better design and improved statistical vigour.

In 2004, the traditional random audit programme was not carried out due to the change in Revenue's approach following a review of the programme. However, 25 cases selected under the programme for previous years were completed in 2004. In 2003, 274 random audits were completed in 2003. Finally, in relation to 2007, I am advised that the random audit programme commenced at the end of March and 401 cases have been randomly selected. Internationally a sample in the order of 400 cases is regarded as an adequate sample for the purposes of an exercise such as this.

There are two primary objectives to this programme:

Compliance measurement

To provide some measure of the effect that Revenue's compliance programmes have on taxpayer compliance over time.

Deterrence

Random case selection ensures that no taxpayer is exempt (or believes himself or herself to be exempt) from the possibility of having his or her compliance position checked.

Q.What was wrong with the previous random audit programme? Why was it changed?

I am informed by the Revenue Commissioners that the previous programme was criticised by the Comptroller and Auditor General on the grounds that the cases selected by Revenue were not actually selected at random. Typically a certain amount of stratification was carried out before the cases were selected and there was subsequent screening of cases at Regional/District level. In other words, cases were not selected on a purely random basis and results could not be used to validate Revenue's other programmes or regarded as representative of the taxpayer population.

Q.How are cases selected under the Taxpayer Compliance Testing (Random) Programme?

For the 2005, 2006 and 2007 programmes a total of 400 cases approximately were selected at random on a pro-rata basis for each of Revenue's four regions and Large Cases Division. The cases were selected from those who had a live registration for any of the following taxes — Income Tax, Corporation Tax, VAT, Employer's PAYE/PRSI and Relevant Contracts Tax.

Q.What does the programme test?

The programme tests taxpayer compliance. Compliance measurement is a measure of the extent to which taxpayers fulfil their tax obligations. There are three basic tax obligations:

Taxpayers are obliged to file their returns on time (filing obligation);

Taxpayers are obliged to pay the liability on time (payment obligation), and

Taxpayers are obliged to report accurately on the return whatever is required for the correct liability to be established (reporting obligation).

Q.How do the Revenue Commissioners use the results of the new random audit programmes?

I am informed by the Revenue Commissioners that there are a number of secondary objectives associated with the new random programme and results will be used to evaluate these objectives. These are:

1. To validate the risk criteria being used to select cases for audit: This is one of the most frequently mentioned uses for random audit programmes as expressed by jurisdictions that operate a random audit programme. Typically this implies auditing a random sample of cases and comparing the results with those of cases that have been selected using risk-scoring systems. The expectation is that the adjustments, which arise on the risk selected cases, should be considerably higher than those from the randomly selected cases.

With the advent of Revenue's computerised risk scoring system, REAP, which is use throughout Revenue, this validation exercise will be more important in the future. The random programme will also have a role to play in the future identification of risk and the development of the risk assessment model.

2. Identifying emerging issues

Changes in the nature of economic activity and in tax legislation can give rise to new sources of tax evasion activity. Random case selection offers Revenue a means of gathering information on new forms of tax evasion activity that may not come to light using the existing risk assessment model.

3. Policy Changes

Random selection may provide data to allow the Revenue Commissioners to identify policy changes for improving tax compliance, including requests for staffing and resources.

Q.When will the results be available?

I have already given you the results for 2005 and prior years as well as an update for 2006. I am advised by the Revenue Commissioners that they expect to have a comprehensive analysis of the 2006 programme available by October 2007. However, the finalisation of audits depends on the nature and complexity of the cases selected, as some difficult cases will be expected to have a longer timescale for completion and this may have an impact on the result evaluation target date. I am advised that the Revenue Commissioners expect to have a comprehensive analysis of the 2007 programme available by the Summer of 2008.

Q.Why confine the programme to 400 cases?

Internationally a sample in the order of 400 cases is regarded as an adequate sample for the purposes of an exercise such as this. With a sample size of 400 we can be 95% confident that the average results are broadly representative of the taxpayer population sampled. However we will keep this figure under review.

While the random audit programme is important therefore, its significance should not be overplayed. The main focus of Revenue will continue to be on selecting cases for audit based on the presence of various risk indicators and other information available. This is the type of targeted audit that gets best results and that is most likely to change the behaviour of the taxpayer into the future. The targeted approach is now greatly enhanced by the new computerised Risk Evaluation Analysis and Profiling System (REAP) developed by Revenue. This system, which categorises taxpayers in accordance with defined risk criteria, has been successfully piloted and has been introduced nationwide during 2006. The system allows for the screening of all tax returns against sectoral and business norms and provides a selection basis for check or audits. This effectively means that 100% of self-assessed taxpayers will be risk assessed at least once a year. This approach allows valuable audit resources to be assigned to tackling those cases featuring in the higher end of risk ranking and analysis of the 2005 and 2006 random results to date shows that there is strong evidence to suggest that this risk-based approach is the correct one.

Q.Are there any big settlements arising from the 2005 and 2006 random audits closed to date?

Q.How does the average yield from Random Audits compare with the yield from targeted audits?

Q.What is the detailed breakdown of the €1.2m yield in 2005?