Bonhuetter Ferguson Case Study

The Bonhuetter-Ferguson Method:
[18]This is a Bayesian technique, meaning it incorporates the independently derived prior estimates of the overall expected losses with the estimates generated in a similar matrix as the BCL method. To use the Bonhuetter –Fergusson method one must take the expected loss ratio and expected past/future split. Use these to calculate the expected position and compare with the actual position. This method is useful when the data are unstable. The Bonhuetter-Ferguson method assumes that the expected incremental losses are proportional to the ultimate for the accident year not to the emerged to date.
We will assume that to initial premium in 2000 is 500 and that this grosses up by 2.5%p.a. So by the year 9 the

In this method we will assume that the past inflation is 4% and the future inflation is 5%. Firstly you apply past inflation rate to the non cumulative triangle. Then you put it in cumulative form and apply the Basic Chain Ladder method as shown in sections 3.1.4.a to3.1.4.e. Once you have the lower triangle developed you express it in non cumulative form to get: 2 3 4 5 6 7 8 9 10
2001 5238
2002 5368 9099
2003 56531 24593 11693
2004 17515 31025 13497 6418
2005 34998 17279 30607 13315 6331
2006 130189 80282 39636 70208 30543

The basic chain ladder estimate is 1666246 derived by adding the total projected claims.
The Average cost per claim