1. Introduction

The United Kingdom no longer has an income safety net worth its name. As a result, millions of households have too little to meet their basic needs. Food insecurity, the use of food banks and destitution have risen sharply, affecting millions of households. Without decisive action to guarantee at least a basic subsistence, these unacceptably high levels of hardship will continue. 

Evidence of the breakdown in the operation of an income safety net is clear-cut, and was set out in an earlier briefing paper, which demonstrated four aspects of this breakdown. First, benefits for working age people which had previously been deemed just enough to live on have declined in real terms. Second, the basic level of benefits can be shown to be inadequate for covering even the most basic of human needs, in some cases too little even to pay for food and home energy. Third, different guarantees for different groups cover vastly different amounts relative to need:  a working age adult gets minimum benefits worth only a third as much as a pensioner’s, relative to their minimum requirements. Finally, there are huge holes in the safety-net, now affecting the majority of working age claimants, whether because of shortfalls in coverage of rent, because of caps in benefit levels or number of children covered or because of repayments of loans taken out while awaiting the initial payment.

In March 2024, the House of Commons Work and Pensions Select Committee took the unprecedented step of publishing a report highlighting not just evidence of inadequate benefit levels but also the lack of any evidence-based system to ensure that benefits do meet needs, and a recommendation that this should be remedied. Couched in language designed to be acceptable to the cross-party committee members, it was nevertheless clear cut:

“…we set out a wide range of evidence which suggests that benefit levels are too low, and that claimants are often not able to afford daily living costs…”

“… a key challenge when evaluating the adequacy of benefit levels is that the Government has not set objectives for what benefit levels ought to achieve or prevent…”

The committee went on to recommend that the government should make changes informed by:  “a benchmark and objectives linked to living costs to measure the effectiveness of benefit levels” alongside explicit guarantees that these levels will be uprated annually on a stated basis.

Coming up with an evidence-based, consistent and politically feasible basis for setting benefit levels is immensely challenging, not least because the present structure and level of entitlements is so distant from such an outcome. But there is a powerful case for ensuring, at the very least, that the system be moved in the right direction, starting to reduce the large gaps between income and need for the families affected. That process can involve identifying both long-term objectives influencing direction of travel and how annual changes in benefit levels can systematically move in such a direction.

This report suggests a new framework for addressing the adequacy of safety net benefits and how this could be used to tackle the general inadequacy of such benefits. over time. A key argument is that uprating systems are immensely important in the long term, to each group’s benefit levels relative both to meeting their own needs and to other groups’ entitlements. In the next section, the past impact of upratings according to different systems is set out, before Section 3 suggests a way forward for improving the adequacy of the safety net in future years.

2. The shifting sands of inadequacy: past changes in benefit rates, relative to inflation and to each other

In 1999, a single person aged over 25 and below pension age could get £51.40 a week to cover their day-to-day living costs, through Income Support/Jobseekers’ Allowance. On the other hand, a pensioner with a low income had a minimum income guarantee of £75 a week, representing a safety-net increase of 46% once they reached state pension age.

Today, 25 years later, the working age benefit, via Universal Credit, is £90.50, while the pensioner guarantee, via Pension Credit, is more than double this, at £218.15. The “pensioner supplement” has grown from 46% to an astonishing 141% (i.e. pensioner benefits are 2.41 times working age benefits[1]).

Our safety net benefits system has always allocated more to a low-income pensioner having to survive on benefits indefinitely than to an unemployed person who might have to live on benefits only for a relatively short period.[2] This addition had grown from about a quarter up until the 1980s, to a third at the start of the 1990s, and to about half later in that decade. At that stage it could still have been seen as linked to additional long-term costs of people living permanently on a safety-net income. Yet today, with safety-net income nearly two and a half times as high for pensioners than non-pensioners, it can only be interpreted as the state guaranteeing very different kind of living standards at different stages of life. Part of this difference can be linked to the state’s desire to limit working-age benefits, but not pensions, to improve work incentives, but the implications for the different living standards of those remaining outside work, before and after pension age, are huge.

This fundamental change is mainly a result not of policy shift explicitly changing the purposes of different safety-net benefits, but rather of different ways in which benefits have been uprated over time. Differences in uprating regimes are partly a consequence of the political popularity of treating some “deserving” groups more favourably than others. Yet over the long term, they are an inherently illogical way of expressing society’s priorities in terms of whom to help the most. Different uprating regimes, with no end point, continue indefinitely to widen the relative generosity of benefits to different groups, rather than fixing this at a ratio reflecting what levels of income are needed or considered fair for different groups. Under the present regime, there is nothing to stop pensioner benefits eventually growing from 2.4 to five or ten times the level of working age benefits.

One reason why upratings have been so important in recent years is because, notwithstanding the introduction of Universal Credit (UC), there has not been significant structural change in entitlements since 2003, when the second generation of tax credits, and Pension Credit, were introduced. Standard allowances in UC were imported directly from “legacy benefit” rates, and indeed continue to run alongside them at the same level.[3] Thus, it has been incremental change over time, not explicit assessments of where it is fair to set entitlements, that have had the biggest influence on today’s rates.

Figures 1 to 5 show how different benefits have risen since the 2003 system was introduced, and how this relates to two measures of inflation [4]and to average earnings. As well as the Consumer Prices Index (CPI), it shows a previous inflation measure known as the “Rossi” index [5] – essentially a version of the Retail Prices Index (RPI) that excludes items such as rent that are not meant to be covered by basic benefits. Importantly, the RPI and Rossi index use a formula for averaging prices that consistently shows a higher rate than CPI. There remains controversy over which is more appropriate, but from 2011, CPI rather than Rossi has been used for inflation upratings.

Looking first at Figure 1, the benefit rate for working age adults without children has declined substantially relative to both prices and earnings in the past 21 years. While up to 2010, it was uprated by the Rossi index, since then this benefit has been increased by CPI at best, and between 2013 and 2019 by less than CPI. It has thus fallen in real terms on all measures.

It is worth noting here that over this period as a whole, average earnings have risen only modestly relative to CPI (11% real-terms growth over 21 years), and adjusted by this index are no higher than in 2008. Even though earnings grew more healthily in the early 2000s, they have grown by slightly less than the Rossi inflation rate since 2003. Thus, while earnings indexation is sometimes seen as the “gold standard” of earnings uprating, a switch to earnings uprating at the start of this period would have left working age benefits no higher than if there had not been the switch in the inflation index used from 2011. This also means that going forward, an earnings link may not be sufficient to avoid periods of declining real benefit levels, and an additional guarantee related to inflation may be needed for this purpose.

 Figures 2 to 4 look at three aspects of family benefits. Out-of-work families with children have seen very different uprating rules applied to various aspects of their benefits. The adult element of their benefits is the same as for non-parents, which as Figure 1 shows has been erratically uprated and fallen in real terms. In contrast, the main means-tested child element, the per-child part of Child Tax Credit and now of Universal Credit, was raised sharply in real terms between 2003 and 2011, as shown in Figure 2. The default was to raise it in line with earnings, but successive budgets in the 2000s gave substantial ad hoc additional increases, as a key tool for pursuing ambitious targets to bring child poverty down. Even though this element was subsequently subject to a benefits freeze, it remains well above its 2003 level adjusted for prices and earnings.

In contrast, a smaller, per-family element of Child Tax Credit/Universal Credit (not shown) has not been uprated at all since its introduction in 2003, and is now being phased out. And Child Benefit (Figure 3) has fallen sharply in real terms. From 2009 to 2019 it was barely changed (rising only from £20 to £20.70 for the first child). This means it is far lower in real terms than in 2003.

Figure 4 combines all of the family benefits mentioned above, to show the effect on the overall income of an out of work couple with two children. This shows that while during the 2000s family benefits overall kept up with earnings and outstripped inflation, they are now back at a similar  level, relative to CPI, as they were in 2003. (And for families with all children born since 2017, not shown, family benefits have risen by exactly the same amount as CPI prices over this period.)

Finally, Figure 5 shows a contrasting trend for Pension Credit, the safety-net benefit for those above the state retirement age. This has outstripped both prices and earnings over the period shown.  Since 2008, Pension Credit has been uprated at least by earnings, as a statutory requirement.  However, in eight of the 14 uprating years since 2011, earnings have risen either by less than inflation or by less than 2.5%, causing Pension Credit to rise faster than earnings under a link with the State Retirement Pension.[6]  This creates a leapfrogging effect so that Pension Credit rises faster than either prices or earnings over the long term. In the 2000s, when this system was established, years where prices rose faster than earnings had been rare, so this effect was not anticipated.

Table 1 summarises how much safety-net benefits today differ from the level they would be if they had kept pace with earnings over the past 21 years. This shows how the minimum degree of protection people have if they are out of work, relative to what they might have earned if they are working, has changed as a result of different uprating policies. The working age households shown have up to £1000 a year less, and a pensioner has £1000 more, than if benefits had been pegged to earnings.

As described above, these variations have resulted not only from different benefits being subject to different uprating regimes, but also from the inconsistent ways in which these regimes have been applied. Many ad hoc or discretionary decisions have been taken, including suspension of any uprating for working-age benefits in 2016-2019. Pensioner benefits are more systematically linked to a formula, but this involves more discretion for Pension Credit than for the State Retirement Pension.

3. Creating a more adequate safety net

To ensure that everyone in the UK has enough regular income to provide for their basic needs requires substantial increases in entitlement for some groups, which cannot be achieved through small adjustments to the system. Given current public expenditure constraints, this is unlikely to be achieved quickly. However, if governments had a clear commitment to mending our broken safety net, and a vision of what they would like it to achieve, it would at least be possible to start moving in that direction.

For this to happen, the answers to three questions would be helpful.

First, what kind of living standard do we consider adequate for people on the lowest incomes?

Second, is it feasible to pursue such a goal for people from all backgrounds, rather than just for favoured groups?

And third, what kind of commitments could provide a reliable trajectory towards that goal, with uprating rules that steadily improve benefit levels for those for whom they are currently inadequate? This needs to be considered in relation to what is considered realistically affordable.

Identifying an “adequate” level

Debates about what comprises benefit adequacy are never conclusive.

One concept of what a safety net income should be doing is minimalist: it should be enough for “subsistence”, paying only for the bare essentials of life, notably food, shelter, warmth and clothing. There are two underlying problems with such a concept, applied in twenty-first century Britain rather than in a traditional subsistence economy.

The first is that within a category such as “food”, it is unclear what comprises a bare minimum requirement: consumption of the calories required to keep alive could be achieved very cheaply on a diet that bears no resemblance to how people actually eat. The other is that in practice, people with limited budgets do not spend them only on these most basic physical necessities, but also on things like transport, household goods and toiletries which are part of everyday living if not directly “subsistence” items. This means that a household with income just enough to afford, say, food, home energy and clothing would still fall short of meeting these needs, because they necessarily spent money elsewhere.

A contrasting concept of the minimum is enough income to lead a decent life, socially acceptable in terms of being able to buy material necessities and to participate in society. In recent years, this has been embodied in the Minimum Income Standard (MIS), drawn up through research into public consensus on what should be included as part of such a minimum.

It is widely accepted that having “enough” to live on, embodied in even the most austere benefit regime, should be more than just keeping alive. Yet there is certainly not agreement that the level described by MIS should be the income safety net. The consensus established by MIS explicitly does not address what benefit levels ought to be, a subjective question over which there are a variety of views. One view is that guaranteeing people a decent living standard when they are not working is morally wrong because it may enable people to choose to live satisfactorily at taxpayers’ expense. Even without this moral objection, a difficulty with guaranteeing such a level for those not working would make it necessary to ensure that working incomes are sufficiently higher than such a guaranteed level, in order to make work pay, which is likely to prove very costly in terms of in-work benefits.

In reality, a credible ambition for income adequacy will need to fall somewhere between these two paradigms of subsistence and decency, recognising that for the time being a guarantee of a MIS income is not practical, but also that living at an acceptable level is about more than just surviving. This has been recognised, for example, in the development of a “minimum income guarantee” in Scotland, which is intended to be “somewhere between the relative poverty line and the Minimum Income Standard”. The Joseph Rowntree Foundation and Trussell Trust’s “Essentials Guarantee” level is also intended to be a step along the road towards more adequate benefits, although not the final destination. They have recommended that the UK government establish an independent process, informed by expert knowledge and lived experience, to calculate annually a level covering essentials that the basic rate of Universal Credit is not allowed to fall below. An initial version of this level (£120 a week for a single person) is both indicative and minimalist. It is indicative in that it is not based on the deliberative process described, and minimalist in that it is judged to be the lowest level at which such a benchmark could be set.

The Essentials Guarantee is calculated principally based on existing expenditure levels in certain categories by low-income households. While this is a useful initial benchmark that it is hard to argue that anyone should have to live below, it would be problematic as a long-term goal for benefit levels. Any benchmark based on what low-income households actually spend, without reference to their needs, risks being circular: where benefits are set too low to meet needs, recipients will necessarily spend inadequate amounts.

A benchmark more directly linked to “need”, but still set somewhere between subsistence and the “decency” level represented by the Minimum Income Standard, could be a percentage of MIS. Rather than selecting a purely arbitrary percentage, it is possible to look at evidence of the relationship between falling short of MIS by various amounts and reporting hardship. One analysis suggested that, based on how many households report not being able to afford things that the majority of the public considers essential, the risk of hardship is particularly high for households below 75% of MIS. This is corroborated by qualitative research which shows that those between half and three quarters of MIS have distinctly worse experiences than those closer to the MIS level.

Note that referencing safety net benefits, initially, on a percentage of MIS would not necessarily entail directly indexing them to changes in MIS over time. This could be problematic due to unpredictable short-term changes in the MIS level. Having initially set benchmarks at a level informed by a MIS-referenced living standard, it may for example make more sense to base changes over time to an index of average earnings. This would ensure that the definition of what is considered an adequate safety net, for those unable to support themselves through earned income, progresses in line with improvements in living standards for those who can.

Translating a benchmark into a fair safety net across groups

A target level of the safety net of, say, 75% of MIS would currently be more than double the standard Universal Credit allowance for working age adults without children, about one and a half times that of families with children, and slightly below the safety net provided by Universal Credit for pensioners.

Clearly a policy to close the gap between working-age benefit levels and such a benchmark would either have to involve a very high increase in those benefits today or consistent, smaller increases over a long period of time. In either case, there would be important political decisions to be made about the relative treatment of different groups. The group with the strongest case for increases based on meeting the most basic of needs is working age adults without children, who fall furthest short of this level. Yet shortfalls for families with children are also high, and an emphasis on tackling child poverty has in the past caused this group to be prioritised.

In reality, the distinction between providing an adequate minimum income for working age adults with and without children is not as clear-cut as it first sounds. This is because child-related benefits come much closer to providing for the additional cost of children than adult benefits come to providing for adult needs, whether those of parents or adults without children.

For a couple the average additional cost of paying for a child, on average for the first two children at the full MIS level, was £81 a week in 2023, almost identical to the £82 minimum that an out of work family would receive in children’s benefits, including Child Benefit, on average for the first two children.[8] The fact that for such a family overall, benefits cover only about half of MIS can therefore entirely be attributed to the inadequacy of adult benefits, which are only a third of what is needed to cover an adult’s costs. Therefore in the case of couple parents, a focus on increasing the adult UC rates (which are the same for both parents and non-parents) would eventually close the gap between benefits and minimum needs.

For lone parents, this is not so clear-cut, since the additional cost of having children works out substantially higher for a single person than for a couple.[9] The cost of a child in this case is £130 a week to cover a full MIS living standard, or £98 to cover three quarters of MIS. Since this is still significantly above the £82 average received per child in benefits (or £87 for families with at least one child born before 2017), increasing the adult benefit rate alone would be less effective for a lone parent than for a couple, creating a case for a family element to be restored and increased for lone parents receiving UC.

It is worth noting that an increase in entitlements focused mainly on improving the adult or family element of UC would have implications for the relative adequacy of benefits for different sized families. Before the introduction of the two-child limit, the larger the family the greater the percentage of their needs were covered, because each additional child brought a higher proportion of their own costs than that of the adults in the family. The two-child limit has, in contrast, hugely disadvantaged larger families. Arguably, the abolition of the limit, but a focus on making adult or per-family benefits more generous relative to per-child benefits, would be a balanced approach in this respect. The devastating effect on large families of providing no additional means-tested income to cover the additional costs of the third and subsequent child would be removed. But the effect of over-weighting the coverage of needs to per-child rather than adult or family payments, which arguably distorts provision in favour of larger families, would also be modified.

Overall, then, a concentration of future improvements in safety-net benefits on the standard allowance for adults in UC, supplemented by additional help for lone parents, would over time help create a more balanced safety net, reducing the currently yawning gaps between the adequacy of benefits for working age adults without children, families with children and pensioners. Alongside these changes, a reassessment of whether extra benefits for groups such as disabled people and carers (which this paper has not addressed) will be essential.

Setting a course

This report has argued that safety net benefits have reached a dangerously low level, at which millions of households are being left with insufficient resources to afford the most basic necessities of life. The downward trend in the value of working-age benefits relative to need should be reversed as a matter of urgency.

Ensuring that such benefits are at least uprated in line with inflation will stop things getting worse, but will do nothing to improve the very low levels that benefits have reached. To achieve that will require systematic increases by amounts greater than inflation, with safety-net benefits progressing towards a long-term adequacy target.

If earnings and general incomes were growing healthily, an obvious solution would be simply to link benefits to earnings, ensuring that growing prosperity was evenly shared, including by lifting the incomes of those who have least. However, forecasts for real earnings growth over the next few years are depressingly gloomy. The Office for Budget Responsibility forecasts that up to the later part of this decade, earnings will be growing annually by 1% or less in real terms. An earnings link in this case would increase the value of benefits, but painfully slowly. At 1% real growth, working age benefits for people without children would be reaching 75% of MIS early in the twenty-second century.[9]

The alternative would be to make a bolder commitment to use part of the fruits of future growth to raise the incomes of the worst-off faster than average earnings. Designing a policy that gives greatest increases to benefits during periods of growth is likely to add to its political feasibility. For example, a formula could automatically increase the standard allowance by one percentage point above the growth in average earnings or one percentage point above the growth in prices, whichever is the higher. This would still produce only a gradual rate of improvement, but would guarantee continuous growth that would transform the adequacy of the safety net over a generation. It is worth recalling as a parallel that when pensions were reformed in the 2000s, they were also at a historic low relative to earnings; the Triple Lock, partly by accident,[10] has decisively reversed this situation.

As argued above, an additional boost could be given to lone parents with the reinstatement and gradual increasing of a family element (i.e. a higher allowance for the first child) in Universal Credit entitlements for this group.

None of this implies that an earnings link with the State Retirement Pension or with Pension Credit should be dropped. There are arguments for revising the Triple Lock, and by extension its effect on Pension Credit levels, to ensure that in future pension incomes keep up with but do not outstrip earnings growth over the long term. But to break the earnings link itself would be a step backwards, since it has given both working-age and retired people greater security that the minimum value of state retirement income will be retained relative to working-life earnings. This kind of security should not be removed from pensioners, but rather extended to people who need help during working life. As the relative consistency of pension upratings in recent years has demonstrated, what is really needed is greater certainty backed by clearer commitments about how governments will protect the incomes of those who have the least. The need has never been more urgent.