The paid chain ladder and the incurred chain ladder are two of the most frequently used methods that actuaries employ to develop indicated loss reserves for property and casualty companies. Their popularity stems from both their ease of use and simple familiarity. However, there are other reserve analysis methods, such as Bornhuetter-Ferguson, that can prove extremely valuable.
In an article entitled “A fresh look at actuarial reserving methods,” Milliman’s Susan Forray discusses the results from a test of 30 less-common analytic methods. Here is an excerpt:
The results of the analysis were surprising. In brief, we found that the methods exhibiting the greatest skill over time were not the most popular but rather those that best satisfied the following two criteria:
1. They relied, at least in part, on case reserves in their evaluations.
2. The paid-to-date data they used did not directly influence the indicated unpaid loss.
We identified different methods satisfying both criteria, each of which exhibited greater skill than the incurred chain ladder – and significantly greater skill than the paid chain ladder.
…The results of our study suggest that there are many more valuable methods for reserve analysis beyond the incurred- and paid-chain-ladder methods, and that the paid chain ladder, in particular, should not receive the weight if often does. Of course, this is a general observation, and a particular company’s circumstances always should be considered in selecting methods for any reserve analysis.
…The ongoing challenge is identifying which new methods to select among the handful indicated and how to weight them against the more common methods already in our actuarial toolbox. This is an area for possible future work.
Whichever methods actuaries ultimately decide to use when performing reserve analyses, this study strongly suggests we should all consider methods beyond the familiar chain-ladder approach.
This article was first published in the May/June issue of Contingencies.