Loss Reserve Estimates: A Statistical Approach for Determining "Reasonableness"

With the NAIC's adoption of the Accounting Practices and Procedures Manual, the statutory accounting practices for the P&C insurance industry have now been codified in a series of Statements of Statutory Accounting Principles (SSAP's). Within the SSAP's, various terms such as "Management's Best Estimate," "Ranges of Reserve Estimates" and "Best Estimate by Line" have been defined. In addition, the Actuarial Standard of Practice (ASOP) No. 36, adopted by the Actuarial Standards Board in March 2000, provides definitions for terms such as "Risk Margin," "Determination of Reasonable Provision" and "Range of Reasonable Reserve Estimates." While they are both well designed and a definite improvement, these new principles and standards of practice provide only broad guidance to the actuary on what is "reasonable." This broad guidance is based on the principle that "reasonable" assumptions and models lead to "reasonable" estimates. Unfortunately, this broad guidance can leave the low end of a range of "reasonable" reserves open to an interpretation which could lead to unintended consequences in practice. This paper will review some current actuarial practices and examine how they relate to the question of what is "reasonable" from a statistical perspective. Moreover, it will review and further develop some statistical concepts and principles that actuaries can add to their repertoire when developing ranges of liability estimates and then evaluating the "reasonableness" of management's best estimate of reserves within those ranges. It is hoped that the Actuarial Standards Board and others will consider adopting a more definitive definition of "reasonableness" in order to help avoid the unintended consequences of allowing the reserves to get "too low" in practice.