Unequal results: improving and reconciling the UK’s household income statistics

Published on Incomes and Inequality

Although GDP, employment and average earnings are (rightly) key economic barometers, household incomes and inequalities are indispensable measures of living standards. Indeed, there has been increased recognition across the world of the need for economic statistics to give more emphasis to the distribution of growth as well as averages. Household survey data has proven invaluable for assessing the challenges facing the UK and for measuring the impact of policy changes.

But the recognition of the importance of distribution; heightened concerns about inequality; and increased potential for big data and data sharing all give cause to think about whether the UK’s current data is as good as it can be. It is not. Worse, perhaps, two official sources have published quite different inequality figures.

This briefing note explores why this is, some of the improvements already planned, and what else needs to be done to give the UK world-leading household finance data to continue to inform debate and public policy.

  1. The Office for National Statistics (ONS) should extend its inequality statistics back to 1961 or earlier – rather than 1977 – and it and the Department for Work and Pensions (DWP) should aim for these to receive National Statistics status
  2. The 1960 Family Expenditure Survey should be digitised, with funding sought from research councils and trusts
  3. DWP should seek to make its annual Households Below Average Income (HBAI) publication more prominent and user-friendly, recognising that these results are the number one source for household income and inequality trends and learning lessons from the ONS’s statistical bulletins
  4. ONS should include additional variables in its datasets to facilitate comparison – and combination – with HBAI, such as HBAI-consistent definitions of income and pensioners
  5. ONS and DWP should explore whether their survey designs and microdata could eventually be formally linked to produce a single, very large source of household income data
  6. ONS and DWP should reconsider their treatment of tuition fee loans in household surveys (HBAI currently records receipt of these loans as disposable income) – and whether inheritances and capital gains could be captured as supplementary information (they are currently not included at all)
  7. ONS should include in its household income survey a measure of income that includes imputed rent
  8. ONS should switch from household weighting to person weighting for its equivalised income statistics
  9. ONS and others should consistently equivalise incomes relative to those of couples, rather than the single adult reference point still used in some releases and datasets
  10. ONS, DWP and other stakeholders should regularly explore and outline the judgements and empirical evidence underlying equivalisation factors
  11. ONS and UKSA should make clear that its Effects of Taxes and Benefits on Household Income (ETB) data is known to underestimate top incomes and that HBAI is therefore currently the superior source for measures such as the disposable income Gini
  12. ONS, in reforming its surveys and using administrative data, should attach high importance to the accuracy of top income statistics, using Real Time Information and other sources and building on the DWP’s experience of top income adjustment
  13. HMRC should finally release unit record Survey of Personal Incomes (SPI) data for 2008-09, 2011-12 and 2012-13; and documentation on its SPI methodologies
  14. DWP should improve its HBAI top income adjustment to more comprehensively correct for underestimation and inequality within the top income group, and periodically revise earlier figures using outturn data and a consistent method
  15. Given that the tax data is an imperfect measure of incomes due to tax evasion, researchers and government should aim to quantify the scale of income missing from the UK tax data across the distribution and across time
  16. ONS and DWP should use administrative benefit data including the Universal Credit system to improve the quality of data for lower income households, where there are very significant inaccuracies
  17. Government should provide the resources required for ONS, HMRC, and DWP to attain by 2020 the high standards the public would expect for household income statistics