Have you ever stared at a chart long enough that the inflection point starts to feel personal? I have. Not because I enjoy doom-scrolling spreadsheets, but because labor data is usually boring until it is not. When the share of working-age people who report a disability jumps and then refuses to settle back, you stop talking about sampling noise and start talking about payrolls, insurance bills, and who is still able to show up on Monday.
That is the uncomfortable conversation sitting under the latest disability numbers on both sides of the Atlantic. In the United States, survey data on people aged 16 and over who report a disability climbed to an all-time high near 37 million. That is roughly seven million more than the level seen just before the early-2021 break. The old plateau did not come back. It snapped.
People argue about surveys. Fair enough. Surveys can be messy. Definitions drift. Respondents get tired. So the useful move is not to shout about questionnaires. It is to look at an administrative system that actually assesses bodies, codes the findings, and pays or refuses claims. The United Kingdom’s working-age disability program does exactly that. And when you put the two series next to each other, the calendar lines up in a way that is hard to unsee.
Why Two Different Systems Matter More Than One Headline
I keep coming back to a simple idea. One noisy series can be a glitch. Two independent systems, in two countries, breaking in the same window, is a signal. The American series is a household survey. It tells you how many people say they have a disability. It does not tell you which organ system failed, or whether a clinician signed off.
The British program is different. New claims are medically assessed. Decisions, often called clearances in the paperwork, get coded by body system and then by underlying cause. The positive award rate has hovered around two-fifths for a long stretch. That matters. If the acceptance rate had collapsed or exploded overnight, you could blame a sudden change in standards. It did not. More people showed up sick. More of those cases cleared.
You can look at total new claims, excess new claims versus a late-2010s trend, percent excess, or a simple z-score. You can slice by month or by year. You can split by age band. In my experience, the people who dismiss the whole file have rarely spent an evening clicking through those cuts. The ones who do tend to get quieter.
If you want to know whether a survey spike is real, stop arguing about the questionnaire and look at a system that diagnoses people.
That is the advantage. The US print tells you the population got sicker after February 2021, not during the first viral year. The UK file tells you where in the body the excess showed up, and how the timing sat against the mass-rollout calendar. Different clocks. Same year.
The 2020 Test Almost Nobody Wants To Run
Alternative stories fail a basic timing test. If the virus at its most aggressive strain were the main driver of a lasting disability wave, you would expect the break in 2020. Lockdowns, delayed care, fear, isolation, all of that was already in the water. Yet many body-system series stayed close to their old trend through that first year. Then 2021 arrived and the lines left the rails.
That does not prove a single cause. Correlation is not a court verdict. I will say that out loud because adults should. It does, however, knock down the lazy claim that “everything bad started when the pathogen arrived.” The data do not look like that. They look like a delayed break.
Perhaps the most interesting aspect is how consistent the calendar is across systems that should not move together if the only story were fraud or a sudden fashion for claiming. Hematology jumped early. Musculoskeletal claims sat still, then leapt later. Neurology printed extreme years. Breast-related clearances showed large excesses with high z-scores in later windows. Different latencies. Same inflection year.
Hematology Did Not Drift. It Broke.
Blood disorders are not the first thing most people picture when they hear “disability benefit.” That is part of why the series is so unsettling. New excess clearances in that group rose on the order of two hundred percent in 2021, then more than five hundred percent in 2022. Even after a pullback, 2023 still sat hundreds of percent above the old trend. Over three hundred percent above trend two years running is not a rounding error.
On a monthly view, the hematology line jumped early and hard, almost on top of the first-dose wave. That is a regime change in medically assessed claims. Something in the presentation of illness changed at a scale that, in a normal decade, would have triggered a public-health review. It did not, at least not in public.
I am not a clinician. I have said that before and I will say it again. Doctors were asked to explain the pattern. Mostly silence. Charts do not replace a differential diagnosis. They do tell you when an administrative system started seeing a different world.
Musculoskeletal Claims Tell A Later Story
Joints and soft tissue are a different clock. Monthly clearances sat near a familiar ten thousand through 2020 and into early 2021. Then, around early autumn 2021, the series jumped and stayed elevated above eighteen thousand a month. Inflammation, mobility limits, the kind of pain that takes people out of shifts and onto daily-living awards.
Notice what did not happen. The rise did not arrive with the first lockdowns. It did not arrive with the first viral wave. It arrived after mass rollout and into the booster period. You can dislike that sentence. You still have to explain the months.
In my view, this is the cut that should make labor economists sit up. Musculoskeletal disease is a work-hours story. It is absences, restricted duties, early exits, and higher private insurance costs. It is not an abstract health debate. It is a staffing problem with a medical code attached.
| Body system | Timing of break | What the series shows |
| Hematological | Early 2021 | Very large excess, early jump |
| Musculoskeletal | Late 2021 | Level shift that stayed high |
| Neurological | 2021 onward | Extreme yearly deviations |
| Cardiovascular | After 2020 | Clear rise versus prior trend |
| Oncology-related | 2022–2023 window | Large excess, high z-scores |
Fraud Stories Have To Survive The Split
Whenever benefits rise, somebody says the public learned how to game the form. Sometimes that is true. Systems get softer. Word spreads. Advisers get better at paperwork. I have seen that in other programs. It is not a cartoon.
But a fraud wave does not usually light up blood disorders first, then joints later, while neurological claims print twenty-plus-sigma years. An awareness campaign does not move breast-related clearances on that schedule. If the only change were “people discovered the benefit,” the body-system mix should look more like a general lift. It does not.
- Award rates stayed roughly stable near forty percent
- New claims, not just the stock of old claimants, drove the story
- Different systems peaked on different calendars
- Age-band cuts still show the post-2020 break
- Two countries, two measurement methods, one inflection year
Could there still be some behavioral overlay? Of course. People talk. Media coverage changes what people notice in their own bodies. Delayed hospital care from 2020 can show up later as worse disease. All of that can be true at once. The point is not that one factor explains one hundred percent of the line. The point is that “ignore the timing” is not analysis.
What The US Survey Adds That The UK File Cannot
The American series is blunt. About 37 million people 16 and over reporting a disability. Up about 23 percent from February 2021. A three-to-four sigma break from the pre-2020 plateau. No mean reversion. That last part is the part markets should care about. Temporary shocks fade. This one has not.
Labor force participation is not a morality play. If a larger slice of the 16-plus population is disabled in a lasting way, you get fewer hours, more absences, higher insurance loads, and more pressure on public disability insurance and medical programs. You can argue about the medical why. You cannot argue the arithmetic for long.
I’ve found that readers often want a single villain. Virus. Policy. Shots. Hospitals that shut elective care. Metabolic health that was already poor. All of those can sit in the same decade. A grown-up read holds more than one variable without pretending the calendar is random.
You can ignore a chart. You cannot ignore the payroll.
The Fiscal Shadow In Britain Is Already Visible
Claimant counts have roughly doubled since 2019. Psychiatric disorders are now the largest single category. That last fact gets used as a dodge. Mental health claims rose, therefore the whole file is “just anxiety.” Look again. Other systems moved too. Psychiatry can be the biggest bucket and still not explain hematology or the musculoskeletal step-up.
Spending is on a path that forces politicians to talk about sustainability and tighter points tests. That is the usual endgame. When the bill gets large, the debate shifts from “why are so many working-age bodies failing” to “how do we make the form harder.” I understand why treasuries do that. I also think it is a category error if the underlying morbidity is real.
The American version of that conversation is coming. A permanently larger disabled share means lower participation, higher absence, higher private premiums, and more pressure on disability insurance and Medicaid-type programs. Firms already complain they cannot staff shifts. Some of that is pay. Some of that is health. Pretending it is only one of those is how you get surprised by unit labor costs.
How To Read Excess, Z-Scores, And Age Bands Without Fooling Yourself
Excess versus a 2016–2019 trend is a tool, not a religion. Trends can be wrong. Demographics move. Coding practices change. If you only stare at percent excess, a small category can look like a moonshot. That is why absolute new claims still matter. A tiny base can produce a huge percentage and still be a rounding error in the budget.
Z-scores help you see how unusual a year is versus its own history. They do not tell you the cause. A twenty-sigma print is a flare. It says “look here.” It does not write the clinical note.
Age bands keep you honest. If only the oldest working-age group moved, you would tell a different story than if prime-age groups moved too. Play with the cuts. Monthly versus yearly. Absolute versus excess. One system versus another. The pattern that survives those cuts is the one worth arguing about.
- Start with the raw level so you know the size of the bucket
- Compare to a pre-2020 baseline, not to last month’s noise
- Check whether award rates changed or only volumes changed
- Split by body system before you invent a single story
- Ask whether 2020 or 2021 is the true break
Labor Markets Feel Health Shocks Late
Markets price earnings and rates. They are slower to price a thicker tail of people who cannot work full weeks. Disability is a lagging labor indicator in the worst way. By the time the stock of claimants is obvious, the flow has already changed hiring, overtime, and benefit costs.
Think about a warehouse, a hospital ward, a delivery fleet. One extra absence per crew does not look like a crisis on a slide. Ten extra absences a month does. Multiply that across a country and you get the soft constraint everyone calls “we just cannot find reliable people.” Sometimes that sentence is about wages. Sometimes it is about bodies.
Private insurers notice before speeches do. Claims severity, musculoskeletal caseloads, longer recovery times. Those show up in premiums and in which roles get automated first. Public programs notice when the caseload graph starts to look like a second staircase.
What This Is Not Claiming
Let me be blunt, because this topic attracts slogans. This is not a claim that every new disability is one product of one month in 2021. It is not a claim that surveys are perfect. It is not a claim that administrators never err. It is not medical advice. It is a description of timing, magnitude, and cross-country corroboration.
If a later study shows that delayed cancer screening explains a slice of the oncology excess, good. Put it on the table. If better coding explains a slice of neurology, put that on the table too. The adult move is to allocate the residual, not to pretend the residual is zero.
I’ve sat with people who lost months of work to pain they did not have in 2019. I’ve also sat with people who treat every chart as a conspiracy poster. Both groups can be wrong at the same time. The file in front of us is still the file.
Why Official Silence Is Part Of The Story
When an administrative system records a regime change in medically assessed illness, the default public-health response should be curiosity. What changed in presentation? Which codes? Which ages? Which regions? Which comorbidities? Instead, the conversation slides toward process: tighter tests, longer waits, more conditionality.
Process can be necessary. Fraud exists. Budgets are finite. But process is not etiology. If working-age health deteriorated, squeezing the form does not restore labor supply. It just hides the sick until they show up somewhere else: family income, private insurance, informal care, or early retirement that never gets labeled as disability.
That is why the two-country comparison is useful. You can wave away one survey. Waving away a survey and a medical-assessment system that break on the same calendar takes more work. Most official commentary has not done that work in public.
A Practical Checklist For Readers Who Handle Money Or People
If you run a team, underwrite risk, or just watch participation prints, you do not need a manifesto. You need a watchlist.
- Track new disability flows, not only the stock of existing claimants
- Separate mental-health volume from other body systems
- Watch musculoskeletal claims as a leading work-hours indicator
- Compare 2020 versus 2021 before you write the cause paragraph
- Assume higher absence and higher medical cost in planning, then update if the lines fade
None of that requires you to pick a tribe. It requires you to treat population health as a labor input. Weird how often that sentence still sounds radical.
The Compounding Problem Nobody Budgets For
A one-year spike can be absorbed. A multi-year level shift compounds. People who leave work at 35 or 45 do not only cost a benefit check. They lose skill accumulation. Households lose a second income. Tax bases thin. Care burdens move onto partners who then work fewer hours themselves. That is a second-round labor effect hiding inside a health statistic.
Psychiatric caseloads add another loop. Distress can follow illness. Illness can follow distress. If both rise together, you get a thicker left tail of people who are neither fully employed nor fully counted as unemployed in the clean textbook sense. They sit in the gray zone that makes “jobs reports” look tidy while managers still cannot fill rosters.
I do not enjoy that paragraph. I also do not think wishing it away changes the claimant count.
Simple way to hold the file in your head: 2020: many systems near old trend 2021: break across several body systems After: levels stay high, composition matters Policy talk: tighter tests, not root causes
Questions That Still Need Answers
Which diagnoses inside hematology drove the early jump? How much of the musculoskeletal step-up is inflammatory versus degenerative versus injury that never healed because care was delayed? How much of the oncology excess is catch-up screening versus true incidence? Do the same age bands move in private insurance data? Do hospital admissions line up with benefit codes, or do they diverge?
Those are ordinary questions. They should have ordinary working groups, ordinary data releases, ordinary arguments among people who actually treat patients. Instead we get culture-war noise and budget speeches. That gap is itself information.
If clinicians later show a clean non-intervention explanation that fits the 2021 calendar and the body-system mix, I will read it. Happily. The current public record is thin relative to the size of the break.
What Readers Should Take Away Without The Fog
The US survey says the 16-plus disabled population jumped after February 2021 and stayed high. The UK assessment system says new medically reviewed claims left trend in 2021, with hematology early, musculoskeletal later, and other systems joining on their own clocks. Award rates did not collapse. The stock of claimants grew. Fiscal talk in Britain already sounds like a squeeze. The American cost conversation is behind the data, not ahead of it.
Together, that is a corroborating signal. Not a courtroom closing argument. A signal. The kind you put on the wall if you care about labor supply over the next decade.
Pick a body system. Pick an age band. Watch 2020 hug the old line. Watch 2021 leave it. Then decide whether you still want to call the whole thing a vibe. I don’t. The charts are still there. The damage, if the levels hold, compounds slowly enough that people will call it normal until the budget line item gets a new name.
