Evidence of rising incomes is difficult to square with the lived experience of people on low incomes.
Last week the Office for National Statistics (ONS) released new figures on household disposable income and inequality, which said that the incomes of the bottom fifth of households – more or less equivalent to the poverty line - have risen by £1,800 since 2007.
This shows a very different picture to other Government datasets, such as the Households Below Average Income (HBAI) series, which the ONS said in their release is a more reliable data source, and to reports on rising poverty and inequality by JRF. Evidence of rising incomes is also difficult to square with the lived experience of people on low incomes, many of whom are finding it harder and harder to make their incomes strentch far enough to cover the basics. So, which figure is most accurate?
To look at what’s going on we’ve used the HBAI dataset produced by the Department for Work and Pensions, as this uses a much larger sample size and so is likely to be more robust. The key points are:
- Real income growth, which takes into account rising inflation, for the poorest fifth of households between 2007/08 and 2015/16 was £600 a year (£1,200 lower than the £1,800 suggested by the ONS on 10 January).
- If you ignore housing costs, over the whole period from 2007/08 to 2015/16, the bottom fifth of households did see real income growth whilst the incomes of the top fifth were lower in real terms in 2015/16 than in 2007/08.
- However there is a difference in the period up to 2010/11 and after: between 2007/08 and 2010/11, the incomes of the poorest were protected and the incomes of the better-off fell rapidly. Since 2010/11, better-off households have experienced faster income growth.
- When you take into account housing costs, the pattern of growth is reversed: increasing rents mean that the increase in real incomes in the eight years to 2015/16 is mostly wiped out.In contrast, lower housing costs for the top fifth saw their real income after housing costs grow by £1,600 a year.
Real income growth before housing costs
HBAI shows that median equivalised net income for the poorest fifth of households increased by £600 per year in real terms (2015/16 prices) between 2007/08 and 2015/16. This is likely to be more accurate than the £1,800 reported by the ONS on 10 January because the sample size of Households Below Average Income is 19,000 households compared to only 5,000 in the Living Costs and Food Study used in the ONS release. As the ONS release itself says (Section 3, paragraph 4), the differences between the two surveys “make HBAI a better source for looking at income-based analysis that does not need a very long time series” (see the methodological note below for more details).
2010/11 is a turning point
However, the picture between 2007/08 and 2015/16 looks very different when you split the analysis in 2010/11. During and immediately after the recession, up to 2010/11, the incomes of the poorest were protected by benefit increases, whilst the incomes of the better-off fell (see the purple bars in the chart below).
However, from 2010/11 to 2015/16, the pattern reversed, and the bottom fifth saw slower income growth than everyone else (see the green bars in the chart below). So looking at the change over the whole period from 2007/08 to 2015/16 masks the fact that very different trends were going on in the two periods before and after 2010/11.
Real change in annual net income before housing costs by quintile
Housing costs reverse the pattern
The ONS analysis does not take into account housing costs, which have seen significant change in the last decade. As we saw above, when you put the whole period of 2007/08 to 2015/16 together, we do see real income growth amongst the bottom fifth and a small fall in real income amongst the top fifth.
However, when housing costs are taken into account, this pattern reverses. For the bottom fifth, the £612 a year gain is mostly wiped out whilst the gains for the top fifth jump to £1,600 a year. This is because rents have been rising in the last decade whilst very low interests have held down the costs of owner-occupation.
Net change in real median income 2007/08 to 2015/16, before and after housing costs
The data that the ONS used for their publication – the Living Costs and Food Study (LCFS) – is the only such source that allows comparison back to 1977. However, with a sample size of only 5,000 households, it is relatively small. For any analysis for periods from 1994/95 onwards, it is highly preferable to use the Households Below Average Income (HBAI) – a National Statistics dataset produced by the Department for Work and Pensions for two main reasons:
- HBAI is based on a sample of 19,000 households, compared to 5,000 for the LCFS. This means that results from that source are less likely than LCFS to be the result of luck based on who got chosen for the survey in that year.
- As surveys are not very good at getting accurate details of incomes for those with high incomes, HBAI data is adjusted to be consistent with data from Her Majesty’s Revenue and Customs (HMRC). This means that estimates of inequality using this source are more likely to be accurate than the LCFS.
The ONS themselves acknowledge this in Section 3, paragraph 4 of their 10 January bulletin:
These differences make HBAI a better source for looking at income-based analysis that does not need a very long time series (the FRS data are available from financial year ending (FYE) 1995) and when looking at smaller sub-groups of the population, particularly at the upper end of the income distribution.
At present the HBAI data is only available until 2015/16. The 2016/17 data will be released in March 2018.