The book reports 66% and lands at 84%, and two programmes change category on the way
Restated at realistic scale, and developed to ultimate. Reported figures come from the loss run unchanged. Ultimate applies age-to-ultimate development factors by development band — Immature ×1.35, Developing ×1.12, Mature ×1.02 — as stated assumptions. These bordereaux carry reported and incurred amounts, not triangles, so this is a directional restatement, not a reserving opinion. The actuaries own the real number.
The book reports a loss ratio of . Develop the same business to ultimate and it is — a gap of 17.8 points and $9.81M of IBNR. That is not a deterioration; it is arithmetic. of the book's earned premium sits in cohorts too young to have settled, and claims arrive and grow for months after the period they belong to. Propane is the widest gap: reported, ultimate, with closure at — the largest fall in the book — while the book's own closure rate is .
So what? Every ratio in the other three briefs is a reported ratio. Propane's 144.4% combined in Brief 01 is built on a 111.4% reported loss ratio that develops to 145.7% — so the cancellation conversation is not about a programme at 119% on the year, it is about one materially past it. And Habitational, which reads a comfortable 73.8% today, lands at . Two programmes change category once the number is aged.
what the loss run says today
where the same business lands once it develops
claims that have happened and reached no bordereau yet
the largest fall in the book — a backlog behind the severity
Four numbers the rest of the brief explains. Reported and ultimate are the same business read at two different ages.
Loss ratio today, and where it lands
The solid band is what the loss run reports. The lighter band is the development still to come. Together they are the ultimate.
One shared scale to 150%. Development factors are applied by development band on each programme's own mix, so programmes with younger cohorts carry wider gaps.
Every programme grows. The question is which ones change category when they do. Propane goes from bad to far worse. Habitational crosses from comfortable to — a point under break-even on loss alone, before LAE, commission or expense are added. Golf has the narrowest gap in the book at and is still the best programme afterwards, which is what a genuinely well-priced book looks like at both ages.
June 2026 · ultimate loss, by how certain each part is
Cash already out the door, money set aside on claims that are known, and an estimate of claims nobody has seen yet. Only the first of the three is a fact.
Paid and case reserves come from the loss run. IBNR is the modelled development described at the top of this brief.
Less than half of the book's ultimate loss is cash that has actually been paid. A third is an estimate on claims that are known, and a fifth is an estimate of claims nobody has seen yet. Any decision taken on the reported figure alone is taken on 46 cents of every ultimate dollar.
Claim closure rate against each programme's own normal
Closure rate is claims closed as a share of claims reported. Below the line means the programme is closing fewer than it opens — the earliest sign that a severity event is turning into an operational one.
Change in percentage points against each programme's own trailing-twelve-month median. The book as a whole moved −0.2 pp, so this is not a claims-department capacity problem.
Propane is the only programme materially behind, at against a book that held at . That matters for the reserve read: a programme closing fewer claims than it opens accumulates open cases, and open cases are where reserve estimates live. The reserve ratio at and the closure rate are the same fact seen twice.
Reported loss ratio against modelled ultimate
The red line is what each month reported. The blue line is where that month lands after development. The band between them is what had not arrived yet.
Loss ratio only. LAE, commission and operating expense add roughly 33 points on top of every figure on this chart.
Read the calm months, not the spikes. Propane reports around 61%% in a quiet month and develops to about 77%% — which is a perfectly ordinary programme. The spikes develop from 152%% and 162%% to 202%% and 216%%. Development does not just add a fixed margin; it magnifies whatever is already there, because the bad months are the ones with open, growing claims on them.
Every KPI behind this brief, with its verdict
Moved and flat metrics together. Three of the twelve are here because they did nothing — and they are what stop this being read as a claims-operations failure.
| Metric | Actual | Against | Δ | Verdict | Reading |
|---|---|---|---|---|---|
| Loss Ratio · reported, book | 66.1% | 83.9% ultimate | −17.8 pp | outbreak | The number every other brief quotes. It is 46 cents in the ultimate dollar. |
| Loss Ratio · ultimate, book | 83.9% | 66.1% reported | +17.8 pp | outcome | Where the same business lands. Still under break-even on loss alone, but not by much. |
| IBNR · book | $9.81M | $0 in the loss run | +$9.81M | driver | Claims that have happened and have not reached a bordereau. Entirely modelled. |
| Immature earned premium | 68.2% | 68.2% | flat | structural | Two-thirds of the book is too young to read. This is normal, and it is why the gap exists. |
| Loss Ratio · Propane, ultimate | 145.7% | 111.4% reported | +34.3 pp | worst | The widest gap in the book, because 80% of its premium is Immature. |
| Loss Ratio · Habitational, ultimate | 96.4% | 73.8% reported | +22.5 pp | category change | Reads comfortable today, lands a point under break-even on loss alone. |
| Claim Closure Rate · Propane | 52.4% | 62.0% | −15.5% | amplifier | Closing fewer than it opens. Open cases are where reserve estimates live. |
| Claim Closure Rate · book | 61.7% | 61.9% | flat | rule-out | Flat. The backlog is one programme, not a claims-department capacity problem. |
| Reserve Ratio · Propane | 45.9% | 35.0% | +31.1% | precursor | Energy lines reserve near 35%. Movement here is the earliest reserve warning available. |
| Reopen Rate · Municipal Law Enf. | 7.6% | 5.1% | +49.0% | watch | Closed claims coming back — a different failure from Propane's, and worth its own look. |
| Late reporting share · book | 15.3% | 14.7% | +4.1% | rule-out | Reporting lag did not lengthen. The IBNR gap is cohort age, not a slowing feed. |
| Days to close · book | 50 | 50 | flat | rule-out | Unchanged. Claims handling speed is not the explanation for anything on this page. |
Ultimate figures are the modelled development described at the top of this brief and are directional. Reported figures, closure, reopen and lag rates come from the bordereaux unchanged.
The three rule-outs matter more than usual here. Reporting lag held at , days-to-close held at 50, and book closure held at . Nothing is arriving later or being handled slower. The gap between reported and ultimate is not a process failure — it is the ordinary consequence of reading a young book, which is exactly why it gets forgotten.
Every ratio in every other brief is 46 cents in the dollar
The book reports and develops to . Of the $46.18M ultimate loss, $21.33M is paid, $15.04M is reserved on known claims and has not reached a bordereau at all.
Why it matters: This is the single most useful thing this value pack does, and it applies to every other page in the set. A combined ratio built on a reported loss ratio is a combined ratio built on less than half of the eventual loss. It does not make the other briefs wrong — it makes them conservative, which is the direction nobody expects.
Two programmes change category once the number is aged
Habitational reads 73.8% today and lands at — a point under break-even on loss alone, before 33 points of LAE, commission and expense. Propane goes from to .
Why it matters: A watch-list built on reported loss ratios puts Habitational in the comfortable half of the book. Built on ultimate, it sits beside Commercial Trucking as a programme that does not make money. That is a different portfolio conversation, and it is available today rather than at the next reserve review.
Development magnifies, it does not add
Propane's calm months report about 61%% and develop to about 77%%. Its shock months report 152%% and 162%% and develop to 202%% and 216%%.
Why it matters: The instinct is to treat development as a fixed margin added to every month. It is not — it is proportional, so the months that already look bad are the ones that grow most in absolute terms. That is why a severity event should be reserved and escalated on the month it happens, not averaged into a quarter where its development gets diluted.
The backlog is one programme, and it is the same fact as the reserve
Propane's closure rate fell to while the book held at . Its reserve ratio moved to over a 35% base.
Why it matters: Closure rate and reserve ratio are not two signals, they are one. A programme that closes fewer claims than it opens accumulates open cases, and open cases carry case reserves. Reading them together is what separates 'a bad month happened' from 'a bad month happened and is still open', and only the second one keeps growing.
Nothing is arriving later or being handled slower
Late reporting share held at , days-to-close held at 50, book closure held at , and reopen rate held at 4.7% against 4.9%.
Why it matters: These four rule-outs are what make the rest of the page safe to act on. If claims were arriving later the IBNR estimate would be unstable; if they were being handled slower the reserve movement would be an operations story. Neither is true, so the gap between reported and ultimate is structural and predictable — which means it can be carried in every monthly read rather than discovered at the reserve review.
How have the reserves set in each of the last twelve months developed since?
The only way to test whether these development factors match this book's actual behaviour.
How many Propane claims opened in the shock months are still open?
Separates a backlog that is clearing from one that is still accumulating.
What did Habitational's older cohorts settle at against what they reported?
Checks whether the 22.5-point gap is real for this programme or an artefact of the factors.
Which programmes have the youngest cohort mix, and how has that changed?
A book growing fast carries a structurally wider reported-to-ultimate gap.
Which Municipal Law Enforcement claims reopened, and how long had they been closed?
A 7.6% reopen rate is a claims-handling question the severity story does not cover.
For the last four quarters, how did reported loss ratio compare with what each quarter eventually settled at?
Turns the development assumption into a measured factor for this specific book.