Every answer and its explanation appears here once you have finished the path. Each one then links to the matching glossary entry, where the concept is set out in full with its worked example.
1. A finance director tells the board the program carries ninety million of capacity, after three events in the year. What must be corrected?
The nominal amount is right on day one and wrong since the first event: the only quantity that matters for the next one is RESIDUAL capacity, layer by layer
Risk departments track exhaustion layer by layer rather than the nominal amount, because a cedant announcing a capacity only has it if no layer has been touched. The answer on priorities describes a real correction of a different order, a priority being a per-event retention and not a consumption of capacity. The one on reinstatement premiums confuses a spend with a capacity. The one concluding nothing is wrong is the most dangerous because it is half true: reinstatements do restore capacity, and they are counted, paid for, and exhausted in turn.
Glossary entry · epuisement-de-tranche2. A program exhausts in two ways. Which, and why does sizing look at almost only one?
Vertically, one very costly event crossing the program bottom to top; horizontally, one low layer hit several times until its reinstatements are exhausted. Sizing looks at HEIGHT, hence the vertical mode
Horizontal exhaustion is the most frequent and least anticipated failure mode, precisely because sizing almost always compares the top of the program to a probable maximum loss. The answer on loss and event names a characterization distinction rather than two exhaustion modes. The one on expiry invents a mode. The one on frequency and severity comes close, since frequency is what the horizontal mode expresses, and explains it poorly: it is not ease of modeling that makes frequency neglected, it is that the question put to the program is one of height.
Glossary entry · epuisement-horizontal3. Midwest insurers exhausted their low layer by June 2023 and absorbed the summer hail net, high layers untouched. What does the peril's signature explain?
Hail, flood and wildfire REPEAT medium events instead of producing one extreme: more than twenty convective storms above a billion, some sixty billion of losses, with no major hurricane landfall
The exhaustion mode follows the peril's signature, and that is what makes 2023 demonstrative: sixty billion of losses without a single event able to reach a high layer. The answer to the vertical mode, raising the tops, would have changed nothing. The answer on an exclusion of convective perils invents a wording. The one on aggregate priorities describes a structure that exists and would on the contrary have answered frequency. The one calling it attrition mistakes the scale: events above a billion do reach a low layer, they simply do not climb.
Glossary entry · reassurance-catastrophe4. Three devices answer frequency, and only one really does. Which, and why do the other two fall short?
AGGREGATE cover, which adds events together instead of treating them one by one: buying reinstatements and lowering the low layer's priority push the threshold back without treating accumulation
The first two act event by event, which leaves frequency intact: more reinstatements cost dearly and exhaust in turn, a lower priority brings more events in and consumes the limit faster. Only a structure that adds up answers what adds up. The answer on raising the top is the most instructive because it is the response to the vertical mode, and 2023 shows it would have changed nothing about the horizontal one. Top and drop is an efficiency variant of the aggregate answer, and it carries its own counterpart.
Glossary entry · reconstitution-garantie5. A top and drop converts dormant capacity into frequency protection. What is its counterpart, and why is it complete?
A very large first event consumes it AT THE TOP and makes it unavailable AT THE BOTTOM: the same capacity cannot serve twice, and one trades protection against the extreme for protection against repetition
The mechanism is a reallocation, not an addition: a top layer the first event does not reach comes down to serve as a second line, which is exactly why it costs less than an extra reinstatement. The counterpart is therefore complete in the literal sense, you do not keep both uses. The answer on price inverts the advantage that makes this structure attractive. The one on the previous twelve months invents a condition. The one on the reinsurer describes an objection that does not arise, the drop being provided for in the contract and priced as such.
Glossary entry · point-attachement