Committee on Reserves Announces 1997 Call for Papers
June 12, 1996
Committee on Reserves Announces 1997 Call for Papers
The Casualty Actuarial Society is pleased to extend a call for papers on the topic of measuring the performance of reserving methods. Several authors of accepted papers will be invited to present their work at the 1997 Casualty Loss Reserve Seminar. In addition, the papers will be published in conjunction with the Seminar, and a prize will be awarded to the best paper submitted. Both CAS members and non-members are invited to submit proposals for papers.
A variety of actuarial projection techniques have been developed, ranging from simple chain-ladder techniques to complex statistical models. Techniques vary as to the required data elements, the stability of the projection, and their ability to handle various 'turning points' in the data such as those caused by changes in inflation rates or re-engineering of a claim process. In developing a reserve indication, actuaries typically apply several of these projection techniques to the available data, testing alternate assumptions and injecting a dose of actuarial judgment along the way. Most typically the techniques, assumptions, and judgments employed in the review are also influenced by prior reserve reviews.
Broadly speaking a reserving method can be said to encompass all of aspects of the process used by the actuary to select an appropriate reserve. This includes the data elements used, the segmentation of the experience, the projection techniques employed, the diagnostic tests applied, and the reconciliation of competing indications into a single best estimate. A reserving method also includes the timing and frequency of reviews, and the procedures for setting reserve levels in the interim.
While reserving methods have been described in the literature, very little quantitative analysis has been published as to the performance of reserving methods. It seems logical that the profession would seek to assess performance, as a diagnostic tool to identify areas for improvement. Questions that might be asked relating to the performance of reserving methods include:
- What is the comparative performance of various projection techniques? For example, do the sophisticated computer models actually outperform the simpler traditional methods? Does the incurred method always improve over the paid method? Are there any generally accepted projection techniques that consistently underperform others?
- How do the inherent characteristics of the line of business, an abrupt change in external economic conditions, or variations in internal company circumstance act as drivers of reserve performance? Can these drivers help to tell us when a particular projection technique will underperform or outperform others? Which techniques are particularly suited to volatile underwriting environments?
- What is the value of various data elements, in terms of their improvement in performance? For example, do claim count, exposure, or price level index data improve performance sufficiently to warrant their capture? Does segmentation of the data, for example by type of payment, improve performance? How does the volume of data influence performance?
- Do multiple projection techniques used in combination outperform individual techniques? How does the way the techniques are combined affect performance?
- What is the value of actuarial judgment in terms of performance improvement? Does the actuary outperform the raw mechanical projections or an artificial intelligence approach?
- Has reserving performance improved, deteriorated, or remained constant over the last twenty years? Have advancements in technology and actuarial science improved performance? Is reserving performance since the imposition of opinion requirements better than it was prior to them?
- How should the performance of reserving methods be measured? Should it focus solely on the variance of overall errors, or should special consideration be given to bias or to the potential for underestimation vs. overestimation? What are appropriate tests of performance? Is hindsight a fair test? Should the measurement of performance vary from one business context to another?
- What tests should an actuary perform to identify the strengths and weaknesses of a reserving method? How should one actuary appraise the reserve work of another actuary? How should differences in indicated reserves produced by different actuaries, using different methods, be reconciled?
- Does improved performance, and the associated increase in the accuracy of estimates translate into a competitive advantage for the insurance company? If so, how can the value of that advantage be measured?
- Papers may be based on analysis of historical data or hypothetical data, or alternatively on survey or test results. For example, a paper could be based on a performance test conducted among competing actuaries, or against competing reserving methods. The Committee on Reserves encourages thinking 'outside the box' on this important topic, and is particularly interested in novel and innovative approaches to the questions presented above.
Timely submission of the discussion papers is critical to the
success of the call. The procedures and timetable enumerated below
- By September 3, 1996, authors should submit proposals for
their papers including the title, a short description of the topic
to be addressed, and the approach that will be taken. Proposals
should be submitted to:
Casualty Actuarial Society
1100 N. Glebe Road, Suite 600
Arlington, VA 22201
Attn: Committee on Reserves
- By October 15, 1996, the Committee will make a decision on
all proposals. The number of accepted proposals may be limited.
The Committee will contact authors regarding their proposals,
and assign a Committee member to work with each author and monitor
the paper's progress. By mutual agreement the author and the assigned
Committee member will establish a schedule for the production
of interim drafts.
- By March 14, 1997, a first draft of the complete paper will
be submitted to the assigned Committee member for review.
- By April 15, 1997, all comments on the first draft will have
been forwarded to the author.
- By June 2, 1997, the completed paper and a 200-word abstract
must be received by the Committee on Reserves for their review.
You will have received further instructions from the Committee
for the name and address of the committee member to receive the
completed paper. Each paper will be screened by the Committee
to assure that it meets the minimum standards established by the
Committee (relating either to relevance of material or quality
of exposition). The Committee may require further rewriting of
the paper to bring it to an acceptable standard.
- By July 15, 1997, all authors will have been notified as to
the results of this screening process. Accepted papers will be
printed, bound, and distributed for the September 1997 Casualty
Loss Reserve Seminar.
- Authors may be invited to present their papers at the CLRS;
authors who are invited to present will be contacted by the Joint
Program Committee for the CLRS in advance of the meeting regarding
the nature of the presentation.
- A prize in the amount of $1,000 will be awarded to the best
paper submitted in response to the call, and will be presented
at the CLRS. The best paper will be selected by a review committee
that will judge the papers anonymously. The criteria for evaluation
Originality of ideas
Clarity of presentation
Contribution to the literature on loss reserving
Thoroughness of analysis
- Papers must be no more than 10,000 words, and should be prepared in accordance with the Guides for Preparation of Papers and Procedures for Submission of Papers in the 1996 CAS Yearbook (pages 255-258).
It is hoped that the authors will also submit their papers for publication in the Proceedings in accordance with the procedures in the Yearbook. However, acceptance of a discussion paper for this call does not guarantee its acceptance for publication in the Proceedings.
The Committee on Reserves looks forward to receiving proposals to the call, and is happy to respond to inquiries from interested parties.