Year loss table (YLT)
A catastrophe model output simulating thousands of fictitious years, each with its dated events and their losses, to price whatever depends on accumulation within the year.
The year loss table represents risk as a sequence of simulated years, often ten thousand to one hundred thousand, each containing the ordered list of events that occurred, with their date and the loss they cause the portfolio. It makes computable the clauses the event loss table ignores: an annual aggregate treaty, a paid reinstatement after the first loss, an hours clause grouping the damage from one episode, seasonal cover. It also keeps perils and regions consistent, since a single simulated year carries both the hurricane and the windstorm that hit it, which allows portfolios to be added without inventing a correlation. The aggregate exceedance curve is read by simply ranking the annual totals. Its cost is file size and sampling error on very high layers, touched by too few years for their price to be stable.
Hypothetical case: on a 50,000-year table, an annual stop loss attaching at 60 million euros of accumulated losses is hit in 1,150 simulated years. Its attachment probability is therefore 1,150 out of 50,000, or 2.3 percent a year, a figure no event loss table could give since it depends on adding up the year's losses.
Follow the research
One letter a week. First, the week's underwriting note, when one is published: a dated, argued decision, signed by Paul Bernardski. Then what the observatory published on cyber risk, political risk and insurance. No account, no tracking.
No spam, one-click unsubscribe, no data ever sold.