The Department for Work and Pensions (DWP) has today published its annual statistics on the distribution of household income in the UK. The latest release covers the years up to and including 2013–14.
On the 16th July, IFS researchers will publish a detailed report, funded by the Joseph Rowntree Foundation, analysing what these data tell us about recent changes in living standards, inequality and poverty. In this Observation article, we briefly outline some of the key information contained in the DWP publication.
One difficulty in summarising the headline figures is that the official statistics continue to use the discredited Retail Price Index (RPI) to adjust for inflation. Since the RPI is known to significantly and systematically overstate the rate of inflation due to defects in its formula, the official figures are based on a methodology which makes changes in living standards look worse than they really are.
The use of the RPI in these figures is sensibly under review. Fortunately, the report also contains an annex with figures adjusting for inflation using the more appropriate RPIJ measure. In what follows, we discuss the RPIJ-adjusted figures rather than the more prominent (but less appropriate) headline figures.
After accounting for RPIJ inflation, median (middle) income grew by just under 1% in 2013–14, following a similarly small rise in 2012–13. (The HBAI headline figures, based on RPI inflation, suggest that median income was unchanged in 2013–14.)
This represents a slow recovery in average incomes, which follows the sharp decline between 2009–10 and 2011–12 when workers’ real earnings fell rapidly. It did mean that real median income had crept back to within about 1% of its pre-recession (2007–08) level, though it was still almost 3% below its 2009–10 peak (household incomes actually continued to rise in real terms during the recession, partly because of temporary stimulus measures such as the temporary VAT cut).
Mean income grew by almost 3% in 2013–14, rather faster than the median. However, this is highly likely to reflect increases right at the top of the income distribution which were, at least in part, somewhat artificial. The cut to the top rate of income tax in April 2013 from 50% to 45% not only had the direct effect of increasing the incomes of the richest, but will also have resulted in significant amounts of income being taken in 2013–14 rather than in 2012–13 – in other words, a distortion to the timing of income. As a result the measured change in mean income will overstate the growth in underlying living standards.
In 2013–14 incomes grew at a similar rate across almost all of the income distribution, resulting in little change in income inequality. An exception seems to have been at the very top of the distribution – which as discussed above is likely to have been artificially distorted – driving a small but statistically insignificant 1ppt rise in the oft-cited Gini coefficient measure of inequality.
At 0.34, the Gini in 2013–14 remains at around the same level as in the early 1990s, but lower than before the Great Recession (the Gini was 0.36 in 2007–08). This is largely explained by the fact that while real earnings fell sharply between 2009–10 and 2011–12, benefit incomes were more stable. Since poorer households get a greater share of their income from benefits, their incomes have risen relative to higher-income households (who tend to get most of their income from earnings).
Once you take account of falling mortgage payments, however, overall inequality fell less than those numbers suggest since it is largely better off households who benefited from this reduced cost. This is illustrated in the after housing costs figures in today’s release, which show smaller falls in inequality since the recession.
The latest data show little or no change in poverty rates, for the population as a whole and for the major demographic groups (pensioners, working-age adults and children).
There was a small and statistically insignificant fall in absolute income poverty (with income measured before housing costs (BHC) and a poverty line fixed over time at 60% of 2010–11 median income), from 10.1 million individuals in 2012–13 to 9.8 million in 2013–14 (16% of the population). This leaves it at a similar level to that seen in 2009–10 and 2010–11, and below levels seen prior to the recession during which income poverty actually fell – particularly for pensioners and children – as some real benefit rates increased.
Using a relative measure of poverty, which instead uses a poverty line of 60% of median income in each year (and hence measures how low-income households fare not in absolute terms, but relative to middle-income households), the figures show 9.6 million (15% of) individuals were in poverty in 2013–14 with incomes measured BHC. This is essentially unchanged from the previous year and down from 18% in 2007–08, reflecting the fact that median income (and hence the relative poverty line) is still a little below its 2007–08 level. The rate of relative child poverty was also unchanged in 2013–14 and remained below its pre-recession level.
The lack of a measured increase in poverty in 2013–14 may come as a surprise. IFS researchers had estimated that known changes to tax and benefit policy (and in particular a number of substantial working-age benefit cuts), combined with the labour market trends recorded by the Labour Force Survey (LFS), would be likely to result in an increase in income poverty in 2013–14. The DWP document mentions employment increases (which were larger in the data underlying today’s statistics than in the LFS) and direct tax cuts as explanatory factors, and in our detailed report next month we will analyse the underlying data to assess the extent to which these do explain the changes observed, and the extent to which they are consistent with other data sources.
The most important point, though, as we emphasise every year, is that limited weight should be placed on a single year of data, which is subject to some margin of error, based as it is on a sample of the population. This margin of error means that a stable poverty rate in the data is perfectly consistent with a small increase or decrease in reality. It is broader trends measured over a number of years that tend to be the most reliable and upon which most weight should be placed.
As we argued last year, when looking at recent changes it is better to measure poverty after housing costs (AHC) due to the variation in housing cost trends across the income distribution. Unfortunately, the official document does not attempt to use a non-RPI based inflation measure to uprate the AHC absolute poverty line (i.e. the real line with which after-housing-cost incomes are compared to assess whether households are in poverty). But the change in AHC absolute poverty in 2013–14 is likely to have been very similar to the BHC change. However, looking at the period since 2007–08 (just before the recession), we know that AHC poverty trends have looked less favourable, because lower-income groups have not benefitted to the same extent from falls in mortgage interest rates.