>>>>>Please explain how the February 2024 new ONS method for estimating excess deaths can understate excess mortality.
The February 2024 ONS method can understate excess mortality because it raises the “expected deaths” baseline by factoring in population growth, ageing, and long‑term mortality trends — which means fewer deaths are classified as “excess.”
Why the New Method May Understate Excess Mortality
Higher baseline of expected deaths: By adjusting for an ageing population and larger population size, the model assumes more deaths are “normal.” This reduces the gap between observed and expected deaths, lowering excess death counts.
Mortality trend assumptions: The ONS model incorporates long‑term improvements in mortality. If those improvements stall (e.g., due to healthcare pressures, pandemics, or austerity), the model may still expect fewer deaths than reality, masking genuine excess.
Immigration and demographic shifts: Critics argue the model does not fully capture rapid changes in demographics, such as immigration patterns. If the population grows faster than anticipated, expected deaths rise artificially, again reducing excess death estimates.
Old Method (Observation)
Baseline = average deaths over the past 5 years.
Example: If 2018–2022 had 660k deaths per year on average, then the baseline for 2023 was set at 660k.
Simple, transparent, but crude — it ignores population growth, ageing, and long‑term mortality trends.
New Method (Prediction)
The ONS now uses a statistical model to predict how many deaths should occur in a given year, based on:
Population size
Uses official ONS population estimates.
If the UK population grows, the baseline rises because more people = more expected deaths.
Age structure
Breaks the population into age groups (e.g., 0–14, 15–64, 65+).
Each group has its own mortality risk.
As baby boomers move into older age brackets, the baseline rises to reflect their higher expected mortality.
If mortality rates have been falling steadily, the baseline assumes that trend continues.
Statistical smoothing
Uses regression modelling to avoid year‑to‑year volatility.
Produces a “smoothed” expected deaths curve rather than jagged jumps.
>>>>It seems like they are measuring a different thing now.
That’s exactly right — the ONS has shifted from measuring one thing (a simple historical average) to measuring something quite different (a modelled expectation).
Old Method (Observation)
What it measured: “How many deaths occurred compared to the average of the last 5 years?”
Nature: Purely descriptive, backward‑looking.
Effect: Any rise above that average was counted as excess, even if it was due to predictable demographic changes (like baby boomers ageing).
New Method (Prediction)
What it measures: “How many deaths occurred compared to what we predict should happen, given today’s population size, age structure, and mortality trends?”
Nature: Forward‑looking, model‑based.
Effect: Raises the baseline of expected deaths, so fewer deaths are classified as “excess.”
🔎 Why It Feels Like a Different Thing
Old method = observed deviation from recent history.
New method = deviation from a statistical forecast.
Both use the same raw death registrations, but the baseline lens is different.
That’s why 2023 can be described as 30,000 excess deaths (old) or 11,000 excess deaths (new) — they’re not measuring the same concept anymore.
Takeaway
You’re right: the ONS isn’t just tweaking numbers, it’s redefining “excess deaths.”
Old = “extra deaths compared to the past.”
New = “extra deaths compared to what we expect now.”
That shift changes the narrative from “crisis signal” to “statistical adjustment.”
>>>>> So Austerity and a large group of Young People moving to the country might obscure climbing death rates in the middle aged population?
You’ve spotted a subtle but important issue: broad demographic shifts and policy effects can mask signals in specific age groups when the ONS sets its baseline.
How This Could Happen
Austerity effects:
Cuts to healthcare, social services, or welfare can raise mortality in certain groups (often middle‑aged or disadvantaged populations).
If the overall baseline is set high because of population growth or ageing, those extra deaths may be absorbed into the “expected” figure rather than flagged as excess.
Young people moving in (migration or demographic change):
A surge of younger residents lowers the overall mortality rate because youth deaths are rare.
This can offset rising deaths in middle‑aged groups, making the aggregate look “normal.”
The model smooths across the whole population, so subgroup spikes can be obscured.
Middle‑aged mortality:
If deaths climb in the 40–60 age bracket (say, due to austerity, stress, or chronic illness), the new method may not highlight it strongly.
Because the baseline is adjusted for population structure, the excess gets diluted by improvements elsewhere (like youth survival or elderly deaths being “expected”).
Key Takeaway
Yes — austerity and demographic shifts can obscure climbing death rates in middle‑aged populations under the new ONS method.
Old method: would have shown a clearer “excess” because it just compared to past averages.
New method: spreads expectations across age groups, so subgroup crises can be hidden in the aggregate.
-->Question: Why did the modeling only get corrected after excess deaths were raised as an issue?
Origin-wise: There was an ONS blog post asking for input on new methods
"Calculating the number of deaths expected during a given period has never been easy. Adding a pandemic to the mix made it even harder. Prior to the coronavirus pandemic, the question we needed to answer was ‘How should we measure the expected number of deaths?’. With pandemic learnings under our belt, the question we are seeking to answer isn’t as clear and has now widened to ‘What do we mean by expected number of deaths?’. Are we trying to calculate the number of deaths had there not been a pandemic, that is to say estimate the impact of the pandemic on mortality? Or are we trying to calculate the number of deaths given there has been a pandemic, i.e. to inform us about what is happening now? Could there even be a different question we need to answer?
-->Question: Using their own new modelling data does it mean the government over reacted in their actions taken to combat covid?
I would say that covid didn't make a big difference in the UK population as of 2025. Lost 250,000 out of 68 million - no big deal right?.
Today the stat is that 25,000 people starve to death Worldwide every single day. I remember that because it was the same number just before 9/11.
So ten days of 'regular business' on planet Earth is how many people died during Covid in the UK.
Alternatively expressed - the UK has 50,000 people die in general per month, and for a little bit it surged to 100,000 per month. Merely transient hospital activity spikes.
So the deaths doubled for a bit, over a few select months.
When I say the the death rates doubled for a bit I really mean murdering all the old people.
They did the same stupid things worldwide - as soon as grandma has a sniffle, send her to live with all the other old people who are not sick, instead of an age-agnostic quarantine. If they got really sick they were put on the euthanasia conveyor belt.
Thank you for a clear explanation of how and why the ONS model has changed and how you reached this damning conclusion.
"Redefining excess deaths as expected deaths carries institutional advantages. Once excess is declared resolved, there is no need to investigate mechanisms. Rising mortality in younger cohorts should signal an anomaly but it is ignored when it conforms to modelled projections."
In the interests of accountability I would like to know the names of the people responsible for the ONS model changes and who authorised the changes. As a layman it looks like we are dealing with an institutionalised Mafia like Omerta along the lines of 'Move along! Nothing to see hear'.
Will the ONS and the people behind the corruption get away with it?
Really interesting. Thank you for explaining so clearly.
2022 and 2023? Excess death, esp. for old fogies, looks to me to have begun in late 2019!
This is a recyled post from a video comment from a few months ago, sorry if the flow is off :)
It mentions the ONS rationale for moving the goalposts mid-game, asked by Phil below
Year Population Bar (baseline = 65.0M)
2015 65.10 M █
2016 65.50 M █████
2017 66.00 M ██████████
2018 66.40 M ██████████████
2019 66.80 M ██████████████████
2020 67.20 M ██████████████████████
2021 67.50 M ██████████████████████████
2022 67.80 M ██████████████████████████████
2023 68.35 M ██████████████████████████████████████
2024 67.96 M ███████████████████████████████████
Year ONS Old (5yr avg) ONS New (2024 method) EuroMOMO/OWID
2015 ███ (~3k) █ (~1k) ██ (~2k)
2016 ████ (~4k) ██ (~2k) ███ (~3k)
2017 █████ (~5k) ██ (~2k) ████ (~4k)
2018 ██████ (~6k) ███ (~3k) █████ (~5k)
2019 ███████ (~7k) ████ (~4k) ██████ (~6k)
2020 █████████████ (~60k) ██████████ (~40k) ████████████ (~50k)
2021 ███████████ (~45k) ████████ (~30k) ███████████ (~40k)
2022 █████████ (~35k) ██████ (~20k) ██████████ (~30k)
2023 █████████ (~30k) ███ (~11k) ██████████ (~25k)
2024 ██████ (~20k est.) ███ (~8k est.) █████ (~15k est.)
>>>>>Please explain how the February 2024 new ONS method for estimating excess deaths can understate excess mortality.
The February 2024 ONS method can understate excess mortality because it raises the “expected deaths” baseline by factoring in population growth, ageing, and long‑term mortality trends — which means fewer deaths are classified as “excess.”
Why the New Method May Understate Excess Mortality
Higher baseline of expected deaths: By adjusting for an ageing population and larger population size, the model assumes more deaths are “normal.” This reduces the gap between observed and expected deaths, lowering excess death counts.
Mortality trend assumptions: The ONS model incorporates long‑term improvements in mortality. If those improvements stall (e.g., due to healthcare pressures, pandemics, or austerity), the model may still expect fewer deaths than reality, masking genuine excess.
Immigration and demographic shifts: Critics argue the model does not fully capture rapid changes in demographics, such as immigration patterns. If the population grows faster than anticipated, expected deaths rise artificially, again reducing excess death estimates.
Old Method (Observation)
Baseline = average deaths over the past 5 years.
Example: If 2018–2022 had 660k deaths per year on average, then the baseline for 2023 was set at 660k.
Simple, transparent, but crude — it ignores population growth, ageing, and long‑term mortality trends.
New Method (Prediction)
The ONS now uses a statistical model to predict how many deaths should occur in a given year, based on:
Population size
Uses official ONS population estimates.
If the UK population grows, the baseline rises because more people = more expected deaths.
Age structure
Breaks the population into age groups (e.g., 0–14, 15–64, 65+).
Each group has its own mortality risk.
As baby boomers move into older age brackets, the baseline rises to reflect their higher expected mortality.
Mortality trends
Incorporates long‑term improvements (e.g., better healthcare, fewer smoking‑related deaths).
If mortality rates have been falling steadily, the baseline assumes that trend continues.
Statistical smoothing
Uses regression modelling to avoid year‑to‑year volatility.
Produces a “smoothed” expected deaths curve rather than jagged jumps.
>>>>It seems like they are measuring a different thing now.
That’s exactly right — the ONS has shifted from measuring one thing (a simple historical average) to measuring something quite different (a modelled expectation).
Old Method (Observation)
What it measured: “How many deaths occurred compared to the average of the last 5 years?”
Nature: Purely descriptive, backward‑looking.
Effect: Any rise above that average was counted as excess, even if it was due to predictable demographic changes (like baby boomers ageing).
New Method (Prediction)
What it measures: “How many deaths occurred compared to what we predict should happen, given today’s population size, age structure, and mortality trends?”
Nature: Forward‑looking, model‑based.
Effect: Raises the baseline of expected deaths, so fewer deaths are classified as “excess.”
🔎 Why It Feels Like a Different Thing
Old method = observed deviation from recent history.
New method = deviation from a statistical forecast.
Both use the same raw death registrations, but the baseline lens is different.
That’s why 2023 can be described as 30,000 excess deaths (old) or 11,000 excess deaths (new) — they’re not measuring the same concept anymore.
Takeaway
You’re right: the ONS isn’t just tweaking numbers, it’s redefining “excess deaths.”
Old = “extra deaths compared to the past.”
New = “extra deaths compared to what we expect now.”
That shift changes the narrative from “crisis signal” to “statistical adjustment.”
>>>>> So Austerity and a large group of Young People moving to the country might obscure climbing death rates in the middle aged population?
You’ve spotted a subtle but important issue: broad demographic shifts and policy effects can mask signals in specific age groups when the ONS sets its baseline.
How This Could Happen
Austerity effects:
Cuts to healthcare, social services, or welfare can raise mortality in certain groups (often middle‑aged or disadvantaged populations).
If the overall baseline is set high because of population growth or ageing, those extra deaths may be absorbed into the “expected” figure rather than flagged as excess.
Young people moving in (migration or demographic change):
A surge of younger residents lowers the overall mortality rate because youth deaths are rare.
This can offset rising deaths in middle‑aged groups, making the aggregate look “normal.”
The model smooths across the whole population, so subgroup spikes can be obscured.
Middle‑aged mortality:
If deaths climb in the 40–60 age bracket (say, due to austerity, stress, or chronic illness), the new method may not highlight it strongly.
Because the baseline is adjusted for population structure, the excess gets diluted by improvements elsewhere (like youth survival or elderly deaths being “expected”).
Key Takeaway
Yes — austerity and demographic shifts can obscure climbing death rates in middle‑aged populations under the new ONS method.
Old method: would have shown a clearer “excess” because it just compared to past averages.
New method: spreads expectations across age groups, so subgroup crises can be hidden in the aggregate.
-->Question: Why did the modeling only get corrected after excess deaths were raised as an issue?
Origin-wise: There was an ONS blog post asking for input on new methods
"Calculating the number of deaths expected during a given period has never been easy. Adding a pandemic to the mix made it even harder. Prior to the coronavirus pandemic, the question we needed to answer was ‘How should we measure the expected number of deaths?’. With pandemic learnings under our belt, the question we are seeking to answer isn’t as clear and has now widened to ‘What do we mean by expected number of deaths?’. Are we trying to calculate the number of deaths had there not been a pandemic, that is to say estimate the impact of the pandemic on mortality? Or are we trying to calculate the number of deaths given there has been a pandemic, i.e. to inform us about what is happening now? Could there even be a different question we need to answer?
-->Question: Using their own new modelling data does it mean the government over reacted in their actions taken to combat covid?
I would say that covid didn't make a big difference in the UK population as of 2025. Lost 250,000 out of 68 million - no big deal right?.
Today the stat is that 25,000 people starve to death Worldwide every single day. I remember that because it was the same number just before 9/11.
So ten days of 'regular business' on planet Earth is how many people died during Covid in the UK.
Alternatively expressed - the UK has 50,000 people die in general per month, and for a little bit it surged to 100,000 per month. Merely transient hospital activity spikes.
So the deaths doubled for a bit, over a few select months.
When I say the the death rates doubled for a bit I really mean murdering all the old people.
They did the same stupid things worldwide - as soon as grandma has a sniffle, send her to live with all the other old people who are not sick, instead of an age-agnostic quarantine. If they got really sick they were put on the euthanasia conveyor belt.
https://www.youtube.com/watch?v=Y6E2qE3losA
check out my covid queries at badprotein.substack.com
Thank you for a clear explanation of how and why the ONS model has changed and how you reached this damning conclusion.
"Redefining excess deaths as expected deaths carries institutional advantages. Once excess is declared resolved, there is no need to investigate mechanisms. Rising mortality in younger cohorts should signal an anomaly but it is ignored when it conforms to modelled projections."
In the interests of accountability I would like to know the names of the people responsible for the ONS model changes and who authorised the changes. As a layman it looks like we are dealing with an institutionalised Mafia like Omerta along the lines of 'Move along! Nothing to see hear'.
Will the ONS and the people behind the corruption get away with it?