Welfare spending and mental health

With two kids in every class of 16 year olds on disability benefits, and spending rising by £50 bn, it's a problem that has to be addressed

Me and the children went to a playground the other day. My son (aged four) enjoyed the little plastic spiral slide.

There were some bigger kids there too. They were having a discussion about who did and didn’t have ADHD: “I’ve got ADHD, have you got ADHD?”

There seemed to be some competition going on as to who had the shortest attention span.

I was reminded of it the other day during Mel Stride’s statement on the Health and Disability Green Paper.

There’s a huge number of people - particularly young people - reporting mental health conditions, and this is leading to a larger and larger problem for the welfare state.

Not working

We can see the problem in different ways.

First, we can look at ONS data on unemployment and inactivity. This is not the same as data on benefits. This is based on the Labour Force Survey, asking people if they are working, and if not, why not.

The good news first. The numbers who are unemployed and looking for work roughly halved after 2010, from 6% to 3% of 16 to 64 year olds. Despite a blip in the pandemic, the share of people unemployed quickly returned to a low level. That’s the good news.

The bad news is that during and after the pandemic the share of people not working because they are sick has increased, even as unemployment has fallen. The net effect has been to wipe out about half of the gains made since 2010 - we went from 12% off work sick or unemployed to about 8.5%, but are back up to 10% now. That means one in ten working age people not working for these reasons.

What is driving the growth in the number not long term sick? One of the big factors over the long term is mental illness.

A 2022 report from the Institute of Fiscal Studies notes that:

6% of working-age individuals are now on disability benefits, up from 2% in 1992–93. This rise has been driven by claims for mental health and other psychiatric conditions, which are now the main disabling condition for 44% of claimants (27% back in 2002–03).

This IFS analysis predates the recent surge, but shows the upward trends have been there for a while. And as we will see below, for the flow of new benefit claims the share driven by mental health problems is much greater.

What types of benefits are we talking about?

There are two main types of benefits affected by the rise in claims:

1. Additional costs or disability benefits. These are intended to aid people with the costs they face from disabilities. The two main ones are Disability Living Allowance (DLA), which is gradually being replaced by the Personal Independence Payment (PIP). Those are the kinds of benefits the IFS chart above is looking at.

2. Out of work benefits which aim to substitute for employment income for people who are unable to work because they are sick or disabled. The main ones are Employment and Support Allowance, which is mainly being replaced by Universal Credit.

This piece looks at the trends for both - but as we’ll see, the data for out of work benefits has been scrambled by various factors.

But before we get there, let’s look at another way to look at the growing problem - the total cost.

The cost

The rising cost of disability benefits plus out of work benefits for sickness is really substantial, however you measure it.

By the end of the forecast period, we will be spending £50 billion more in real terms than we were during the financial crisis.

It’s gone from 5.1 to 7.8% of all government spending, or one and a bit percent of GDP extra. And if anything, this is slightly understating the increase, because of devolution of spending.¹

Additional costs / disability benefits (DLA/PIP)

Now let’s look at what’s driving this increase in spending in more detail. Disability benefits first. The announcement of the Green Paper pointed out that:

“the caseload and costs are now spiralling. There are now 2.6 million people of working age claiming PIP and DLA – with 33,000 new awards for PIP each month which is more than double the rate before the pandemic. This is expected to cost the taxpayer £28 billion a year by 2028/29 – a 110% increase in spending since 2019.

This is in part fuelled by the rise in people receiving PIP for mental health conditions such as mixed anxiety and depressive disorders, with monthly awards doubling from 2,200 to 5,300 a month since 2019.”

In the Commons, Stride pointed out that 7% of working-age people in England and Wales are now claiming PIP or disability living allowance, which is forecast to rise to 10% by 2028-29.

Alongside the Green Paper DWP published a bunch of previously unreleased data (which is available here).

Consistently, over the last decade, around 50% to 60% of those who claim PIP have also been in receipt of one of the out of work income replacement benefits (JSA, IB/ESA, Income Support or UC), so there is a big overlap. About 40% of PIP claimants are in work at the point where they claim, but only 16% of the stock of PIP claimants are in work. This means that there is a pretty large overlap between the claimants of disability benefits and those claiming the out of work benefits I will explore below. Many people are claiming both.

Here is the share of GDP we spend on DLA and PIP for children and adults, which has doubled as a share of GDP, and is forecast to grow.

Here’s the number of working age people and children claiming, which has doubled since 2002, and expected to keep rising.

You might look at the charts above and hope that the forecasts at the end are wrong, and that numbers and spending will level finally off.

Sadly, the history is that forecasts have consistently been too low, and too optimistic.

When I was at the Treasury we used to get a sub from officials at pretty much every fiscal event, saying that the PIP and ESA outlook had deteriorated and the Chancellor would have to find more money to fill the hole.

The biggest increases are concentrated among younger people. This chart shows how the numbers claiming have changed across the age range. Historically people have, (unsurprisingly) got sicker as they age, but 16 year olds are now as likely to be claiming as 50 year olds. This is a big change compared to the past.

We’ve gone from 2% of 16 year olds claiming in 2002 to 8% in 2023. That’s about two kids in every average classroom.

The number of children on disability living allowance has also grown sharply. It was 258,000 in 2002/3, is forecast to be 714,000 this year (24/25) and 948,000 by 2028/29.

What’s driving this? We can look at those who transfer from child DLA to adult PIP each year. Consistently over the last decade about 80% of transfer claims have been for mental rather than physical conditions. Claims for autism and ADHD are sharply up in absolute terms. Anxiety and Depression has gone from 4% of claims a decade ago to 7% now.

Out of work benefits

Moving on from disability benefits to out of work benefits, the trends become much harder to analyse.

That is partly because the system is transitioning from the old benefits (JSA, IS and ESA) to UC. One is going up as the other goes down, but different types of claimant (new, old, different severity) have been shifted across at different times, making it difficult to track trends over time.

The problem is compounded as UC isn’t set up to record people’s conditions in the same way as ESA was.  Unlike ESA no ‘primary’ condition is identified, so statistics are ‘overlapping’ - people often have more than one condition. Like for like comparisons of the caseload are tricky.

One common feature between ESA and UC is that when people say they are unable to seek work because they are ill, they go through a Work Capability Assessment (WCA).

DWP have been trying to analyse the WCA data, and published a report a couple of months ago. The data isn’t perfect: for around a fifth of claimants the data was missing on what medical condition they were claiming for.

Of the 1.3 million people who went through a Work Capability Assessment between January 2022 to November 2023, 69% reported a mental or behavioural disorder - that’s 84% of those who we have data for.

And if we home in on those who the Assessment concluded were unable to work², nine out of ten (91%) had a mental or behavioural disorder. This was much higher than other overlapping conditions - the second largest type were musculoskeletal problems (57%).

So we can see that the share of claims caused by mental or behavioural disorders is now high. But if we want to look at the trend over time, the complex changes to the different benefits and recording systems means we will need to look elsewhere.

Disability and employment

The lack of good trend data for out of work benefits takes us on to looking at data on disability and employment, which ONS have been collecting for some time.

For the working age people as a whole, the share who say they are disabled has been curving up over the decade since 2013. Many of the people who say they are disabled are in work. But since 2018 the numbers who are disabled and economically inactive has gone up too. Almost a quarter of the adult population now report a disability—up from 16% in 2013.

Here’s the same data again, but homing in just on those disabled and inactive plus those who are disabled and unemployed. The proportion of working age people in this situation has been rising:

This, together with the high level of overlap with disability benefits, mean it is very likely that the number and proportion of people on out of work benefits (ESA and UC for health) are also going up.

And what is cauising this increase? If we look all working age people, growing numbers are inactive because of bad nerves, anxiety and mental health problems. The picture is complicated by the fact that ONS started recording autism as a separate category in 2020, and some people likely shifted from other mental illness to that category. Some of the categories are clinically questionable. My colleague Rachel Maclean MP has been digging into the lack of a clear definition for “bad nerves”.

The increase since the start of 2019 is evenly spread across many categories. Mental health-driven inactivity is up, but so is everything else. Mental health seems less dominant as a driver of inactivity in this data than it does in the driving up numbers of benefit claims.

But if we home in on younger people, mental health is their primary reason for being long term sick.

An ONS breakdown found that in 2022, nearly two thirds (63%) of those aged 16 to 34 cited either mental illness or “depression, bad nerves or anxiety” as their reason for being long term sick. This was a much higher share than for those 35-49 (47%) or 50-64 (31%). If this is a cohort effect and continues, then mental health will be an increasingly important cause of inactivity over time, just as it has caused an increasing number of disability benefit claims.

What’s causing so many claims for mental health conditions?

Some of the rise in claims is likely due to “system learning” - the longer a particular system for claiming benefits has been in place the more that people learn how to ensure they score the greatest number of points in any assesment.

Groups like “Benefits and Work” provide increasingly detailed advice and forums explaining how to maximise your chances of a successful claim. They say that they have 110,000 people subscribed to their email newsletter, and the subscription fee of £19.95 a year “comes with a full money-back guarantee.”

But as well as system learning, I think there are two other factors at work.

a) Prevalence inflation and overinterpretation

In a recent interview DWP minister Mel Stride said that:

“While I’m grateful for today’s much more open approach to mental health, there is a danger that this has gone too far… There is a real risk now that we are labelling the normal ups and downs of human life as medical conditions which then actually serve to hold people back and, ultimately, drive up the benefit bill.”

Though it isn’t the whole story, I think there is a lot of truth in this, particularly for younger people.

To pick just once example, I am particularly concerned about the medicalisation of ADHD. There is a large body of evidence from around the world that children who are young for their school year are much more likely to be medicalised. For example, one study from Australia found that:

Among children aged 6–10 years, those born in June (the last month of the recommended school-year intake) were about twice as likely to have received ADHD medication than those born in the first intake month (the previous July)

Twice as likely to be medicated, ju]st because they are a couple of months younger!

In the case of ADHD we are mistaking fidgety or naughty or immature kids for having a mental illness and need of medication.

As well as excessive diagnosis, there may also be a real world increase in mental health problems. Ironically, one cause of this may be the ever-growing conversation about mental health, and exposure to a therapeutic mindset. There is a risk that the endless conversation about mental health actually worsens people’s mental health.

The Atlantic reported on one experiment in Australia which bears out the risk:

Researchers in Australia assigned more than 1,000 young teenagers to one of two classes: either a typical middle-school health class or one that taught a version of a mental-health treatment called dialectical behavior therapy, or DBT. After eight weeks, the researchers planned to measure whether the DBT teens’ mental health had improved.

Immediately after the intervention, the therapy group had worse relationships with their parents and increases in depression and anxiety. They were also less emotionally regulated and had less awareness of their emotions, and they reported a lower quality of life, compared with the control group.

That is not the only study like this. In one experiment one group of kids were given mindfulness lessons and another were not. Result? Mindfulness lessons:

resulted in worse scores on risk of depression and well-being in students at risk of mental health problems both at post intervention and 1-year follow-up, but differences were small and not clinically relevant. Higher dose and reach were associated with worse social–emotional–behavioural functioning at postintervention.

There is a review here of what is called the “prevalence inflation hypothesis” - and the role of “overinterpretation”.

I think there is something in this: if we had an endless conversation about gut health and constantly asked kids if their stomach felt OK, we would pretty soon have lots of kids reporting sore tummies.

I went back to visit my old university last year. The first thing that greets students, pinned to their noticeboard, is list of various forms of mental health support. This is well meaning, but may well not be helpful.

Therapy done properly can be a very good thing, though it can also lead to an unhelpful therapy culture (think Hollywood) in which people become detached from a sense of agency, or start to over think and medicalise normal emotional ups and downs.

B) Multiple real-world causes

Prevalance inflation is probably not the whole story though. I do think there is a mental health problem brewing up among younger people.

The main cause of really severe mental health problems among adults is some sort of adverse experience, often in childhood. When I worked with the street homeless, an astronomical proportion had been in care, or had abusive or junkie parents. Drugs and drink then compound these problems.

But these factors have been around for decades. And I don’t think there is now any debate left that something has changed in the last decade, and there is a teenage mental health crisis across the developed world.

As well as overinterpretation, there are multiple other factors at play, from increased family breakdown to increased drug use - as shown in this excellent piece by Post Liberal Pete.

But in contrast to Pete, I am persuaded that the simultaneous nature of the increase in mental health problems and self-harm all across the developed world do suggest that Jonathan Haidt, Jean Twenge and others are right about the combination of a) smartphones / social media and b) more coddling in childhood.

I don’t think the other explanations work. As Haidt says, we are being too protective of our kids in the real world and not enough online. The problems kids face as a result of this are not as great as the problems faced by people who were abused in childhood, but they are much greater in number: creating a shallower but mile-wide river of misery.

Conclusion

I think Mel Stride is doing the right thing in trying to get this uncomfortable conversation going. I see he has now announced a new support programme, which is all to the good.

But I am struck by the scale of the problem. We are going to be spending 50 billion extra on disability and sickness benefits, and a substantially larger share of public spending.

Apart from anything else, this is a huge opportunity cost. But more importantly, being stuck on these benefits is not good for claimants, often worsening their problems.

This is a problem with several factors driving it, so multiple answers are needed.

To tackle the causes of worsening mental health, we need action on everything from family breakdown to drugs to phones in schools. And we need to be wary of how we handle mental health because interventions and conversations that people hope will help may be making things worse.

Another part of the answer will also lie in changing the benefit system, preventing us spending more and more on people who don’t really need it, and whom it won’t necessarily help. And the sheer scale of the problem means that tinkering won’t work.

1. DWP notes on this state that: “Executive competence for Disability Living Allowance, Personal Independence Payment, Attendance Allowance, Severe Disablement Allowance and Industrial Injuries Disablement Benefit in Scotland was transferred to the Scottish Government on 1st April 2020. Figures up to and including 2019/20 present data for the whole of Great Britain. Figures for 2020/21 onwards present data for England and Wales only. Scotland accounted for around ten per cent of the caseload at the point of transfer.”

2. The combination of those recorded as having limited capability for work (LCW) or limited capability for work and work-related activity (LCWRA - the furthest away from the labour market). Source: Data tables: UC Work Capability Assessment, April 2019 to December 2023