More on the OBR and the WCA
Since the last post, I've been able to find out more about whether the WCA is getting more generous - and my understanding has changed.
Blogging is different to other writing.
Most academic writing is painfully slow, crawling from an idea to the research to the writing over years, followed by endless revisions after feedback from others. Often writings are abandoned, never seeing the light of day.
Blogging, on the other hand, is about speed – a bit of digging into the data in response to current events, communicated without delay. It is an earlier part of the conversation, as others hopefully reply with new arguments or information. Which means that it’s more provisional and more likely to be wrong.
In this blog post, I want to go back to my recent post ‘The WCA is less generous than the OBR thinks’, revising it with things I’ve learnt in people’s responses to the post.
Why the combined series is not consistent
A quick recap: the problem with the OBR’s picture of the WCA is that it combined initial assessments from past years (for ESA) with both initial and repeat assessments for the most recent period (from UC).
I was right that this problem exists, but wrong to suggest that the combined initial + repeat assessment series provides a reliable view of the generosity of the WCA. This is because the balance between repeat and initial assessments changes over time.1 Because people are more likely to pass repeat than initial assessments (as I said in the original post), these fluctuations will bias the overall picture.
It’s tricky to tell how much of a difference this will make, as we don’t have data on UC WCAs split by initial assessments vs. reassessments. However, we can look at ESA using Stat-Xplore data.2
Share: Repeat assessments were 55-60% of all ESA assessments from mid-2016 to mid-2018, after which they rose to 85% on the eve of Covid, and further still to 92% in April 2020. They then fell during Covid so that repeat assessments were 50% of all ESA assessments in Feb/Mar 2021, and further fell to a low of 9% in March 2023. In the most recent Sep 2023 data they have risen back up to nearly half of assessments. So they’ve fluctuated a lot.
Absolute number: Interpreting the % share is tricky, because initial ESA assessments are generally going to become much less likely – new claimants are mostly going to go onto UC rather than ESA (unless they’re applying for contributory ESA). Nevertheless, it’s clear that the absolute number of ESA reassessments has gone down substantially – it was 35-40,000 per month in the few months before Covid, but then fell to around 1,000 per month in mid 2022, and the most recent rise only takes it to 5,000 reassessments per month, out of the 75,000 assessments done in Sep 2023. The OBR actually point this out in their 2023 report (I missed it in my original blog post as they don’t relate it to the chart I discussed, but elsewhere [§2.30] they note that “WCA reassessments were paused in response to Covid and have only restarted slowly in recent months”).
In other words: there have probably been relatively few WCA reassessments since Covid.
So has the WCA been getting more generous?
Hopefully the upcoming OBR Welfare Trends Report will have new figures on initial UC assessments. But in the light of the previous section, it’s defensible to argue that we can get a rough picture of WCA trends in the way the OBR do - that is, we should look at WCA trends for initial assessments only for ESA, and combine this with UC figures for initial + repeat assessments. (The figures are going to be biased slightly upward in the most recent period, but only slightly).
We can considerably improve on the OBR graph, by flagging this discontinuity, and also making clear the impact of the 2017 change (where new claimants to the ESA WRAG no longer get any more money than JSA claimants). The resulting trend looks like this:
This does make the recent rise look slightly sharper, compared to the previous version. It is still true that the recent generosity of the WCA (in terms of the % of claimants receiving extra payments) is not unprecedented - it’s similar, if not slightly lower, than Aug 2013 to Dec 2015. However, it’s about ten percentage points higher than 2011-13 or 2016-17 before the cut to the ESA WRAG.3 The absolute number of people is also higher at about 42k new approvals/month, compared to an average of 32k/month just before Covid, and 29k/month in 2015 (which was the highest previous year).4 But do remember: the recent figures will be biased slightly upwards because of the discontinuity.
One thing that I didn’t emphasise enough in the last post was that there has unquestionably been a large change - more people now are in the Support Group/LCWRA, which compensates for the restricted eligibility from the ESA WRAG/LCW cut. (And LCWRA was always worth more than LCW, with less conditionality). As I said, I think that there’s a straightforward connection between the WRAG cut and the rising number of people in the Support Group, but there’s a question about exactly how this happened, and what this means from the DWP.
From a bit of digging, my guess is that this shift must come from the DWP rather than the private sector assessor CHDA (the trading name of Maximus) - we know from Select Committee hearings that a sample of written assessment reports are continuously checked by DWP, so it would be difficult for CHDA centrally or individual assessors to lead this change. Anecdotally, it seems likely that this meant that the WCA softened a bit, particularly in terms of how much assessors trusted claimants’ self-descriptions of their functioning, vs. being sceptical and requiring medical evidence. But because of the WRAG cut, the end result is not as different as it first appears.
Finally, I guess it would be surprising if PIP applications were rising - which they unquestionably have been - but LCWRA applications were completely flat. We would expect a priori that the LCWRA rise will be smaller (because so many PIP applicants already have LCWRA, and because part of the shift is in awareness of PIP per se). But there’s clearly something going on, and that thing is likely to have some impact on incapacity benefits as well as disability benefits. As to what that ‘something’ is: well, more on that in coming blog posts.
Reflections
I take several lessons from this.
Firstly, when it comes to things that have already happens, the OBR are probably right! They’re pretty good, y’know. They also have large teams working on this, I’m assuming with lots of mutual checking; they have access to DWP and other data that are hidden to most people; and they chat to lots of experts (including people in DWP) to sense-check what they’re doing. My first lesson: if someone is disagreeing with the OBR, you should probably go with the OBR.
Secondly, that doesn’t mean that the OBR are perfect, as their influential claims are sometimes opaque. This is partly because they use private DWP data, so the rest of us literally can’t replicate what they do. But even when they use public data, they are pretty terrible at telling you the sources (‘Sources: DWP, OBR’ is not helpful…). And sometimes they don’t tell us things that we need to know if we’re going to interpret the data properly, like in this case. It’s fantastic that the OBR are transparent enough to write detailed reports setting out the thinking behind their forecasts, but I do think that it would be helpful to take this a level further so that their analyses are more transparent.
Connected to this, the DWP really do need to invest in creating better public data series. I’ve repeatedly blogged about the problems with the DWP’s ‘out-of-work benefits’ series, and the way the WCA data are presented is lousy too. This isn’t usually a deliberate decision; it’s often because DWP computer systems don’t work very well, or there hasn’t been an investment in creating a consistent series. But this needs to change under the next Government - the benefits system is important, and we simply need to have a basic understanding of what’s going on.
Third, partial transparency is a particular problem for OBR forecasts, which increasingly dominate the policymaking landscape. I would expect the upcoming Welfare Trends Report to forecast further rises in incapacity spending, partly because of a policy announcement that is pretty implausible.5 But there’s also a broader issue about the role and transparency of forecasts: what they take into account, and what they don’t. And this is something that I’ll come back to in a future post.
A further, but more minor, issues is that from 29th Sep 2017 some people are exempt from reassessments, if they meet the Severe Condition criteria. While data on this group aren’t regularly released, in mid 2018 there were only 25k people in this group, a tiny fraction of the ESA caseload – so unless things have changed this will make little difference to the statistics. See https://questions-statements.parliament.uk/written-questions/detail/2019-02-04/216179.
Note that this combines two different series in Stat-Xplore: the main ESA WCA series, and the ‘Incapacity Benefit reassessment’ series.
In the Aug 2013 to Dec 2015 period where the average % of people getting extra payments was 69%. Outside of this, the generosity averaged 52% from Sep 2011 to Jul 2013, and 55% in 2015 and Jan-Mar 2016. In the most recent period, this compares to an average of 64% (in the past 12 months, Oct 2022 to Sep 2023).
Figures are for the past 6mths (Apr-23 to Sep-23), ad the 6mths immediately prior to Covid.
In their Nov 2023 response to the WCA descriptors consultation (§82-4), the Government said that ‘We have committed to ensuring no one currently assessed as having LCWRA will face a WCA reassessment, save in some exceptional circumstances” (which includes those granted LCWRA on the basis of substantial risk, where the prognosis for recovery is short-term, or where the claimant reports a change in health condition). I would be astonished if reassessments were completely abolished in this way - this was an announcement for changes to take place from 2025 at the earliest, so the Government never planned to implement them before an election, and this only makes sense if the WCA is being abolished (which may well never happen; see my soon-to-be-released report on this). But it’s official policy, so presumably the OBR will have to take it into account, which will make lead to a slight increase in the forecast incapacity benefit caseload. To be clear - I’ve got no private access to the report and I’ve got absolutely no idea what it will show, this is just a guess based on how the OBR seem to work.
What effect will the raising retirement age have on this series? Its a small proportion of the working age population but you would expect older workers to have a higher burden of illness/disability.