One of the biggest misconceptions about the benefits system is that we split neatly and permanently into two groups: ‘benefit claimants’ and ‘everyone else’. As soon as you take a long view, though, you realise how wrong this is: many people move in-and-out of struggles at different times in their lives, and one of the key roles of the welfare state is to help us smooth out the bad times. John Hills calls this ‘the welfare myth of them and us’ (the sub-title of his recent book), and has dedicated many years to challenging it.
Now in a great new piece of research, the Institute of Fiscal Studies has put together the data to show us the long view – and this tells us that the majority of people are ‘benefit claimants’ at some point in their lives (even ignoring pensions and child benefits…).
What we already knew
My version of this story comes from my time at LSE in the late 2000s, where I was taught (and later had my PhD supervised) by John Hills. I think John had long-known that the welfare state uses money from people during their good years, to give it back to them in their bad years – but getting the data to show this was an almighty challenge, because it requires following people over long periods of time, and trying to look at today’s welfare state even though people live through a constantly changing one.
In a seminal piece of work in the 1990s (which I’ve previously mentioned on the blog), Jane Falkingham and John Hills took 4000 representative people and simulated the lifetime effect of the 1991 welfare state. As summarised in John’s 2004 book Inequality and the State (#8.4), “most benefits are self-financed over people’s lifetimes rather than being paid for by others… nearly three-quarters of what the welfare state was doing in the late 1980s and early 1990s was like a ‘savings bank’, and only a quarter was ‘Robin
Hood’ redistribution between different people”. John’s recent book (p67) presents the figures for the 1991 welfare state updated to 2010 prices:
Benefits recipiency over the long run
This is where the IFS study by Barra Roantree and Jonathan Shaw comes in (beind a paywall in a journal (DOI), but see also their free 2014 report). They use a 20-year survey of a representative sample of the British population (BHPS 1991-2008), and start by showing that about 20% claimed one means-tested benefit in each wave (see table below). But if you look across the whole 18-year period, you can see that 54% of Britons – over half – claimed a means-tested benefit at some point.
(As a quick aside, Table 2.3 of their 2014 report came to a total figure of just under half (47.8%). As far as I can tell, this is because their earlier figure for ‘tax credits’ ignored ‘family credits’ – the precursor to tax credits – before 1999).
What is more, they say – very plausibly – that this is likely to be an underestimate. This is partly because the survey only asks people about their claims once per year, so some people may start and stop a claim without BHPS knowing. But it’s also because we know that survey respondents under-report their benefit claims – indeed, one IFS study shows that BHPS sometimes captures less than half the families that claim income support, council tax benefit or housing benefit.
And there’s more in the IFS study…
This is perhaps the most headline-grabbing message that comes out of the research – but they go on to build on the insights of the Falkingham-Hills research, which can be cut one of three ways:
- The tax and benefits system is more redistributive between people in a single year than it is in the long-run (because much of it is redistributing between good and bad times within people’s own lives);
- Income inequality is lower when you average people’s income over several years than if you look at single year, but the difference made by the welfare state is lower too;
- In total across 18 years, looking only at the tax and benefits system (and not the rest of the welfare state), about 12% of all redistribution is within the same person rather than between different people. As they say, 18 years is some way short of a full lifetime – they cite Falkingham and Harding (1996)’s book and Bovenberg et al. (2008) in Denmark showing that this rises to about 60-75% when looking over the whole lifetime.
This is one of those times that a single number conceals a wealth of careful analysis (and indeed, of careful and expensive data collection in BHPS) – it’s great to see the IFS doing this kind of thing, building on the earlier work of Hills, Falkingham and many others.