Today HMRC releases new figures on the tax gap – the share of tax due that wasn't collected, for example because of mistakes or deliberate under-reporting (evasion). Around a fifth of this comes from taxpayers filing income tax self assessment.

New IFS research published today uses data from HMRC’s random audit programme to show which types of people are more likely to be under-reporting taxes and how their behaviour changes after a tax audit. The results are based on data from audits covering tax returns for the years 1999–2009.

The analysis shows that:

  • 36% of self-assessment taxpayers have some under-reporting on their taxes. This rises to almost 60% among the self-employed.
  • Most under-payments are less than £1,000. But, a small minority of taxpayers (less than 4%) owe more than £10,000 and account for nearly half of the missing tax revenue from self-assessment.
  • Audits can recover significant revenue but the self-employed tend to return to under-reporting within a few years. Audits have bigger long-run effects if they uncover errors on income sources that are more stable over time, such as pensions or income from letting property.

Other headline results from our new research include:

  • The likelihood of under-reporting doesn’t vary much with income, but the cash amount of tax not reported is highest for those with the highest incomes.
  • Men are more likely to under-report than women (40% vs 27%). However, among both men and women who under-report, 32% of total tax owed was not declared.
  • More than half of taxpayers in the construction, hospitality, and transport industries were found to be under-reporting. In hospitality and transport more than half of total tax owed was not reported. These are typically bed and breakfast owners and taxi drivers.
  • Each random self-assessment audit recovers for HMRC an initial £830 on average. An additional £1,230 is raised in the following five years because taxpayers’ change their reporting behaviour. The amounts recovered are higher for certain groups. If HMRC were to target those with the top 20% of reported incomes, for example, they could expect a total yield of £10,260 per audit after 5 years.
  • Audits have a lasting impact if they reveal a source of income that varies little over time, such as pension or income from letting properties. Once HMRC uncover these income sources, it is hard for taxpayers to hide them in future years. In contrast knowing about today’s self-employment income provides relatively limited information about future self-employment income.

There are three things to note about all of the above. First, the figures don’t account for avoidance – by definition this is legal (unlike evasion) and therefore does not lead to extra tax payments after audit. Second, some under-reporting will result from genuine mistakes. Third, audits won’t capture all evasion. For example, it may be very difficult for HMRC to capture the full extent of cash-in-hand working and those who fail to declare any of their income will not be in the audit programme at all. As such, the figures above are likely to underestimate the true tax gap for those filing self-assessment.

Arun Advani, Assistant Professor at the University of Warwick, Research Fellow at the Institute for Fiscal Studies, and author of the study said “Between errors and deliberate under-reporting, a significant share of self-assessment tax goes unpaid. Audits bring in tax directly, but also change taxpayers’ behaviour. Audits work not because they scare people into complying in future years, but because they give HMRC more information about people’s incomes. The change in behaviour actually brings in more than the original audit. HMRC have got better at targeting their audits and spotting under-reporting. But, this didn’t translate into more revenue from audits because they did fewer of them.”

Helen Miller, Associate Director at the Institute for Fiscal Studies said “The self-assessment tax gap is significant. This new research fills in some of the details about where the revenues are being lost. Most revenue is lost to a relatively small proportion of people who evade large amounts of tax. Evasion is highest for the types of income which are easiest to under-report.”

Notes to Editors

Who does and doesn't pay taxes? (IFS Briefing Note BN218), by Arun Advani (IFS and warwick), is available to view on the IFS website

The research was funded by the Economic and Social Research Council (ESRC) through the Centre for Microeconomic Analysis of Public Policy (CPP) at IFS (grant number ES/M010147/1), the Tax Administration Research Centre (TARC, grant number ES/K005944/1), and a University of Warwick Impact Acceleration Award (grant number ES/M500434/1).

The use of HMRC statistical data in this work does not imply the endorsement of HMRC in relation to the interpretation or analysis of the information.