The Joseph Rowntree Foundation has just published a new report, Counting the cost of UK poverty by Heriot Watt and Loughborough Universities [1]. This finds that £69 billion of all spending on public services, £1 in every £5, goes on services which are needed to pick up the pieces from the damage and cost poverty has on the lives of the 13 million people living in it in the UK today. The total £78 billion also includes £9 billion lost tax revenue and additional benefits spending resulting from the experience of poverty. One of the report’s authors, Glen Bramley, muses on these findings.

So now we know. The cost of poverty in the UK is £78 billion. That is, £69bn cost to the government in extra costs of services, plus an extra £9bn ‘knock on’ costs of past child and adult poverty. This is not to be confused with the £70bn  the government also spends on income-related benefits and tax credits for poorer people – we spend this to prevent poverty turning into destitution and as one of the main ways of alleviating poverty. Nor also to be confused with the subjective value of the pain, and suffering experienced by the poor, doubtless also a very big number but harder to calculate and even more difficult to agree on. We are talking hard cash from the Government’s coffers here.

So what? This may be the response, from most of us who have some difficulty taking in large numbers. Perhaps that question may be broken down into some more practical and searching questions, were we to be looking at this in the popular BBC Radio 4 programme, ‘More or Less?’.

Is it a big number? How does it compare with some other big numbers the Chancellor and the financial/political press get worried about – e.g. the deficit, the welfare cuts – as well as the overall scale of public spending, or the  economy?

Is it right? How convincing is the evidence in the report? Can it be cross-checked against different sources? Are there other interpretations of the evident associations of spending with poverty?

Is it cashable? If we were successful in reducing poverty by a significant amount, would we see commensurate savings, which could be reinvested in some other public activity, or given back to the grateful taxpayers?

How long would it take to fully realise such a poverty-reduction dividend? One term of parliament?  A generation, or more?

What other fruit are sitting on the fiscal tree, waiting to be picked? But are there other big bills and nasty surprises which are going to turn up and gobble up this or any other fiscal dividends?

A big number? Well it is a moderately big number, in the same league as some other moderately big numbers which Chancellors worry about. For example, it is about the same size as the government’s deficit in 2015/16, a pesky £72bn which seems to be reluctant to actually go down despite this being the dominant strategy of government since 2010. It is about half the total spending on the NHS, which everybody loves and the government claims to protect. It is about five times larger than the welfare cuts proposed in 2015, which caused a lot of political difficulty, while being about the same size as the current spend on means tested benefits and tax credits. On the other hand, it is only about 4% of GDP, although it is a fifth of total public spending on services.

Is it right?  Well, that is a big question; how long have you got? You can guess I am going to say it is about right, more or less. Or, to put it more academically, it is a prudent, middle figure drawing on a range of estimates (particularly for the most important items), making reasonable assumptions. It reflects very clear evidence on the association between poverty and higher levels of public service activity and spend (measured across small areas or individuals/households). It tries to control for (i.e. exclude) the possibly confounding effects of other measurable causes of higher service costs, as well as factors like differential policies between local authorities. For example, in the case of healthcare we take account of factors like age, household types, housing, past job mix, institutions, urban density, greenspace, pollution. However, it cannot effectively separate and exclude the effects of some factors/conditions which are very closely associated with poverty – they come as a package. We are saying, in effect – to end or reduce poverty you have to also end/reduce these closely associated conditions (e.g. damaging health behaviours, addictions, etc.). Also, importantly, costs today reflect poverty experienced in the past, which has obvious implications for the timescale to turn this around. It can be the case that ill-health leads to or reinforces poverty, but we assume (with some evidence [2]) that this is not the predominant reason for the association. We have put more effort, including looking at different strands of evidence and analysis techniques, into the most important/expensive services, notably health. In general, the more ways we look at it, the more this reinforces our central estimate of the poverty-related component of costs [3].

Is it cashable? –  good question for a prudent Chancellor, and a difficult one. The policy measures that would reduce poverty have different timescales. Some, such as raising the rates of certain benefits, can be done quickly. Radically improving childcare could be done at moderate net cost quite quickly, and would yield fiscal savings in the medium term [4]. Some, such as proposals to tackle complex multiple disadvantages through new strategies like ‘Housing First’, involve a change in the mix of spending and a retraining and reorientation of services, that may take several years [5]. Some, such as raising educational attainment among lagging groups, will not see their benefits fully realised for another 10-20 years or so. Getting adults to adopt healthier lifestyles will bring savings in healthcare costs building up over decades – however, the potential demand for better healthcare and new treatments is insatiable, so that the ability to claw back cash may be limited. More generally, the cashability of savings depends partly on the way resources are allocated and managed, particularly on the extent to which spending budgets are tied to sets of outputs and outcomes, but also on consistent long term strategies and monitoring. Politicians are not very good at maintaining the latter in the face of electoral cycles.

What timescale? Overall, as the preceding examples illustrate, quite a lot of the costs of poverty seem to be attributable to past experiences of poverty, as well as current experience, although these are often correlated, individually or geographically. In the study one piece of analysis illustrates this particularly well. Using the UK Poverty and Social Exclusion Survey 2012 data, this showed that one could separately quantify the impacts on poor health of (a) direct effects of current poverty (b) indirect effects of poverty, through recent unemployment experience and neighbourhood effects; (c) the personal experience of poverty in the past, including through indirect routes such as associated involvement in crime. As percentages of the total volume of ill health and healthcare demand, these three components accounted for 15%, 14% and 10% respectively – and our central estimate of 29% only included the first two of these.  This evidence suggests that there is plenty that can be done on a shorter time horizon, for example on the income, work and neighbourhood fronts, while still recognising the longer term factors.

Fiscal pickings (and bear traps)  The idea that ‘there isn’t any money’, or that fiscal limits are immovable, is a political construct, whether you are talking about the deficit or the share of the economy devoted to public spending. Also, if you examine some of the ideas for tackling poverty, you find that they may generate a positive fiscal impact, for example the Living Wage or Child Care enhancements, because they generate more tax and NI contributions from higher workforce participation or higher levels of pay, even after allowing for the modest negative effects on economic growth likely to result (based on OBR assessments). Arguably, there are opportunities to adjust tax and NI parameters in ways which make some sense from an equity point of view, which would not significantly add to poverty, and which would generate a lot more revenue; the obvious examples concern the retirement age population, who are now officially better off than the working age population [6]. Thirdly, there is the whole area of tax avoidance and evasion, both corporate and in respect of the wealthy, which it is now respectable to mention in polite company [7].

However, these have to be set against several bear traps which are going to make the fiscal position for Chancellors more difficult in the coming period. These include Brexit itself (clearly a net negative once allowance is made for the short and medium term hit on GDP growth [8]), the ‘productivity puzzle’ [9] (if it doesn’t fix itself very quickly), the NHS as a bottomless pit (illustrated by current deficits issue), local government having to be resuscitated after a period of extreme austerity (particularly in the Brexit-voting north of England), and the generally recognised need to spend a lot more on infrastructure and an industrial strategy, in most of the country other than London.

See also

Drawing on this research, Julia Unwin of the Joseph Rowntree Foundation argues in Inside Housing that investing in affordable housing is vital if we are to reduce the amount we spend dealing with the fallout from poverty.

Notes and References

[1] Glen Bramley, Donald Hirsch, Mandy Littlewood and David Watkins (2016) Counting the Cost of UK Poverty. York: Joseph Rowntree Foundation. July 2016. https://www.jrf.org.uk/report/counting-cost-uk-poverty

[2] I specifically tested this proposition, using the British Household Panel Survey which is longitudinal. I compared models to predict poor health, including past health and poverty measures and many other socio-demographic control factors, with comparable models to predict poverty. This shows clearly that the effects from poverty onto health are much larger than the effects from health onto poverty.

[3] One special dataset, which links the ‘Children in Need’ social services data for England with the National Pupil Database, was delivered too late for inclusion in the published report. Analysis of these data indicate that the share of social services spending on children attributable to poverty is 66%, higher than the 58% ‘average’ of several simpler/older figures used in the report. This would add another £0.8bn to the total.

[4] See Adam Butler and Jill Rutter (2016) Creating An Anti-Poverty Childcare System. JRF Anti Poverty Strategy Paper. York: Joseph Rowntree Foundation, www.jrf.org.uk , and also Glen Bramley with others (2016 forthcoming) What Would Make a Difference: Modelling scenarios for tackling poverty in the UK. Edinburgh: Heriot-Watt University. When published the report will be available here.

[5] For discussion and analysis of ‘Housing First’ and similar initiatives, see Appendix G of Bramley and others (2016), note 4 above.

[6] See Chris Belfield, Jonathan Cribb, Andrew Hood and Robert Joyce (2016) Living Standards, Poverty and Inequality in the UK: 2016. ISBN 978-1-911102-17-5 London: Institute for Fiscal Studies. Pages 58-60.

[7]  See Nicholas Shaxson (2011) Treasure Islands: Tax havens and the men who stole the world. London: The Bodley Head.

[8] See Angus Armstrong and Jonathan Portes (2016), ‘Commentary: the economic consequences of leaving the EU’,  National Institute Economic Review , May 2016, and other contributions to that special edition; and Paul Johnson (2016) ‘We are in uncharted waters without a compass’, The Times, 25 June 2016, reproduced in IFS Commentaries http://www.ifs.org.uk/publications/8337 .

[9] see Andrew Goodwin and Martin Beck (2016) ‘The UK Economic Outlook’ in The IFS Green Budget 2016, London: Institute for Fiscal Studies,  esp pp. 45-46; and Richard Disney, Wenchao Jin and Helen Miller (2013) ‘The productivity puzzles’, in The IFS Green Budget 2013, London: Institute for Fiscal Studies, pp.53-90.