Loss development factors are critical to many different actuarial techniques used in the determination of ultimate losses. Unique factors, based on a company’s own historical triangles, can produce more accurate indications than industry factors if the triangles are sufficiently credible. Credibility of unique triangles depends on the number of evaluations in the triangle and the overall volume of losses. However, some companies lack credible historical data for the determination of unique factors. It is important to remember, that irrespective of HOW the factors are determined, they must be applied appropriately. Failure to apply loss development factors appropriately can produce unreasonable results with high variances from actual results. The following summarize four common mistakes in applying loss development factors.

Using the Wrong Type of Factor
Loss development factors are different for incurred losses, paid losses, and case reserves. Loss development factors also vary significantly by coverage. For example, loss development factors for workers compensation are very different from products liability. Additionally, there are differences even for the same coverage based on policy type. Occurrence medical professional liability factors are significantly different from claims-made medical professional liability factors. Even if unique factors are determined, the triangles should be constructed separately by coverage. Care should be taken to use factors that appropriately represent the risk, in terms of loss type, policy type, and policy retention. Furthermore, it is important to confirm that the factor being used reflects the appropriate maturity of the policy. Loss development factors are normally expressed in months from policy inception for twelve month policies.

Applying Incurred Factors Only to Case Reserves or Individual Claims
A common mistake is applying an incurred loss development factor to only the open case reserves or a few selected open claims. The reason almost always given is “my closed claims can’t develop, so why should I apply a factor to that?” Incurred loss development factors are calculated based on the total accident year experience and how this experience develops over different evaluations for several accident years. The factors, therefore, should be applied to total incurred losses. Incurred factors are normally lower than case reserve factors because they have already been inherently lowered by the lack of development on closed claims which serves to lower the overall incurred factor. Claims that are still open after many months often tend to be the more severe claims, with development much higher and often later than the average claim. If there are only a few open claims or if a range of possible outcomes is needed for a fixed small group of claims, a claims auditor may be a good resource.

Applying a Factor Too Early
Loss development factors prior to six months of development/maturity should not be used. Loss development factors between six months and twelve months should be used with caution. Prior to six months of development a policy is simply too immature and the loss development factors are too high and volatile. Between six months and twelve months the policy is still very immature and the ultimate losses obtained from development methods would be extremely volatile. Development methods are also very sensitive to the presence or absence of large claims at early stages of development. Actuarial methods that rely on longer term averages (projection methods and Bornhuetter-Ferguson methods) produce more reliable results until sufficient maturity is reached.

Limiting Development on Individual Claims
Companies with self-insured retentions often limit the development on specific large claims so that the claim is capped appropriately. This seems like a very logical approach and may be a reasonable adjustment in specific situations. However, it is important to understand if the development factor is a “limited” factor or an “unlimited” factor.

When historical loss development data is available to construct a loss development triangle with several years of history, the historical triangle usually contains data limited to the retention. For example, if a company has a self-insured retention of $500,000 per claim, the loss development triangle can be constructed with the claims limited to $500,000 per claim (on an incurred and paid basis for each respective triangle). Therefore, the selected loss development factors (really age-to-age factors) have already been adjusted lower versus unlimited development factors based on the capping of losses at each maturity. These factors are normally labeled as “limited”, meaning they already reflect the limit in place. In this case, no further adjustments should be made to large claims except in extreme circumstances. Again, the limited loss development factors are lower than the unlimited loss development factors and already take into account that some claims will be limited by the retention.

Many companies do not have unique historical loss development data. Industry benchmark loss development factors are normally used in this case. These industry factors are usually “unlimited” factors. Since the loss development factors assume unlimited development large loss adjustments are reasonable in this situation. Therefore, if a claim would develop in excess of the retention with the unlimited loss development factors, it would be appropriate, in most cases, to limit the claim to the retention.

Additional Information
For more information, please reference Chapter 2 of SIGMA’s An Actuarial Advantage. This chapter contains examples of loss triangles, data required to construct triangles, and examples of how loss development factors are selected.

We welcome your feedback by posting a comment, or contact Michelle at MB@SIGMAactuary.com.
Don't forget to register for our free resource portal by visiting www.SIGMAactuary.com/education
© 2011 SIGMA Actuarial Consulting Group, Inc.