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. Why does the method suit working layers and not peak layers?
Because it measures what burned and not what can burn: applied to a layer never hit it returns zero, and zero is not a price while the exposure is not nil
The word burning carries its own warning, and it decides where the method belongs: a working layer sees losses every year, a peak layer has an empty history while its exposure is not empty. It is therefore priced on an exposure model. The answer on limit size invents a scale restriction. The one on settlement describes a real difficulty of long tail lines rather than of high layers. The one on disclosure moves the subject to a market practice, when the problem would exist even with perfect data.
Glossary entry · burning-cost2. Among the observed years, which calls for the most caution, and why?
The most recent: its losses are partly reserved, its development factor is the largest and least certain, and those reserves rise more often than they fall
Recent years are not finished, and projection to ultimate is the more uncertain the younger the year: the last observed carries the largest factor and the widest margin of error, in the direction of understatement. The three other answers name genuine weaknesses. Cumulative indexation on an old year is an important correction with a known method. The exceptional loss poses a representativeness problem the next question handles. And a varying base calls for the third correction, itself made by a named method.
Glossary entry · sinistralite-attritionnelle3. Which correction is most often forgotten, and what does its absence produce?
Adjusting the BASE: the denominator describes a portfolio that is no longer today's, and relating revalued losses to unrevalued premiums produces a ratio that means nothing
People readily correct the numerator, because it is made of losses examined one by one, and leave the denominator alone because it is a total. Yet the portfolio has grown, its rates have moved, its mix has changed, and the rate change measure used for this correction is one of the most argued figures of a renewal. The answer on losses outside the layer describes a construction error that does not arise, the method keeping only the part above the priority. The ones on recoveries and currency name real and particular adjustments, not the structural correction the method always requires.
Glossary entry · primes-cedees4. Once the three corrections are made, an expected cost is obtained. What separates that cost from a price, and why does the loading vary with the height of the layer?
A loading covering the cost of capital tied up, expenses, brokerage and a margin for volatility. It is the heavier the higher the layer, one touched once in ten years tying up capacity for nine years that bring in only a premium
The expected cost answers what the layer will pay on average; the price must additionally pay for tying up capacity nobody else can use meanwhile. That is what explains the tiering of rates along a program. The answer on a uniform margin contradicts precisely that tiering. The one on future inflation describes an adjustment that bears on the expected cost, upstream of the loading. The one on a reinstatement premium confuses the loading with the reinstatement premium, which is called after a loss and not included in advance.
Glossary entry · rate-on-line5. Two actuaries work on the same history and produce two different prices. Can one be said to be wrong?
Not necessarily: five years contain at best five realizations of chance, one exceptional loss makes the burning cost jump and its absence makes it collapse, so the period chosen is enough to separate two equally defensible prices
This is the limit no correction lifts, which is why a burning cost is always accompanied by a reading of exposure. The answer seeing only a division is right about the raw calculation and forgets that the choice of period precedes it. The one invoking standardized corrections credits indexation, development and base adjustment with an objectivity they lack, each resting on argued assumptions. The one attributing everything to the loading conjures away the technical part of the price, which is precisely what the burning cost illuminates.
Glossary entry · wording-de-traite