13.07.2015 Views

PROCEEDINGS May 15, 16, 17, 18, 2005 - Casualty Actuarial Society

PROCEEDINGS May 15, 16, 17, 18, 2005 - Casualty Actuarial Society

PROCEEDINGS May 15, 16, 17, 18, 2005 - Casualty Actuarial Society

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

52 RISKINESS LEVERAGE MODELSbackground for user input. The financial variables are the twonet underwriting results and the investment result, all of whichvary randomly. F9 will recalculate to a new set of results. Belowthe income variables are the starting and ending surplus, and calculatedmean and current (random) return. Interesting simulationresults such as allocation percentages are displayed to the rightof the surplus calculation.Starting with “basics,” Line A has a mean surplus of10,000,000 and a standard deviation of 1,000,000 and Line Bhas a mean surplus of 8,000,000 and a standard deviation of2,000,000. There is a correlation of about 25% between the lines(if the functions were normal rather than lognormal, it would beexactly 25%). Each line is written with a premium equal to themean loss plus 5%. We interpret this calculation as our estimateat time zero of the value at time 1 of the underwriting cash flows,including all investment returns on reserves and premiums.The investment income on the surplus is taken directly. Theinvestment is at a mean rate of 4% with a standard deviationof 10%. The total of the results, on which we will define ourleverage functions, is then added to the beginning surplus of9,000,000 to get the ending surplus. As a consequence of theinput values, the mean return on surplus is 14%. We would allbe happy to have such a company, provided it is not too risky.The simulation (“Sim basics”) shows the actual correlation ofthe lines and the coefficient of variation on the return, as well asthe distribution of total ending surplus and return. From the “Simbasics” sheet we can see that the probability of ruin is less thanone in a thousand, and the coefficient of variation on the returnis better than on the investment, which is good. We can also seefrom comparing the simulated means and standard deviations ofthe income variables to their known underlying values that thesimulation is running correctly.Management has decided that it wants to consider not justruin, but on-going risk measures. In particular, it wants to get

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!