|
|
|
Brainstorms
Responding to the Challenge by Stephen W. Philbrick
In her February “From the President” column, Mary Frances Miller discussed the press accounts of loss reserve increases. Those reports did not paint the actuarial profession in a favorable light. Her column closed with a challenge to the actuarial profession to improve our science and the communication of our results. I’d like to add a third goal to the list: alignment of incentives with goals. This cannot be achieved by the actuarial profession alone; it requires top management involvement, but actuaries can contribute.
All companies set goals and create incentives, often financial. If the incentives are misaligned with the goals, the incentives are likely to prevail. For example, if a call center claims its main goal is customer satisfaction, but pays bonuses solely on how many calls are handled in an hour, employees will likely try to maximize their income. This will make it difficult for the call center to achieve its goal.
For the property/casualty industry, a major goal is accurate loss reserves. (A related goal is clear communication of the inherent difficulties of the process.) I’m confident that people are working on improving both. However, we must recognize that there are considerable pressures working to prevent us from achieving the goal of accurate reserves. Meeting earnings projections often entails its own set of powerful incentives, which may be weighted toward short-term results. This often exerts a powerful downward pressure on reserves, but the direction is not always the same. Few companies would admit to managing earnings themselves, but few would deny that some other companies do. This can lead to upward pressure on reserves in some cases. The IRS also exerts pressure, usually downward. The very nature of the reserves—amounts set aside for events that are not yet settled and may not even yet be reported—means the numbers are inherently “soft.” In a changing environment, there is a large range of possible reserve amounts consistent with plausible assumptions and an even wider range consistent with possible assumptions. However an actuary does not opine on a range but on a single number. Thus, an actuary is almost never in the position of stating that a desired reserve level is flat wrong—at best, it is characterized as less likely than some other alternative.
In short, the selection of a reserve number involves considerable pressure that may not be counter-balanced by sufficient incentives to book the best possible number.
I know of two ideas for increasing the incentives to get the number right. Thomas Stanford and Bob Conger were batting around what might happen if companies could get a full deduction for reserves booked in the current and first prior accident year, but took a “haircut” on reserves booked beyond that time. For example, assume that companies can only take, as a tax deduction, 90 percent of additional reserves booked two years after the accident year, 80 percent for additional reserves booked three years late, etc. Both of them emphasize that this should be considered a brainstorm, not a specific proposal. They can find flaws in it. Indeed, everyone I’ve talked to considers it a bad idea. One person warned of the law of unintended consequences. However, the concept of brainstorms is first concentrating on the positive aspects, adding and refining and, only later, addressing negative issues. If enough time has been spent on the positive aspects, there may be sufficient incentive to address, solve, or work around the downsides.
Rather than focus on the specific attributes of this proposal, think of it as an attempt to align incentives with goals. Stanford and Conger have moved a step in the right direction, if only by creating disincentives for understating reserves. Can this be coupled with some positive incentives?
Rob Painter took a different approach. What if some portion of premium were set aside in a fund, held until the reserves were sufficiently mature so we could “grade” the original reserves. The premium would be returned in such a way that those who were most accurate got the larger shares and those who missed the mark got nothing back. I was concerned whether enough money could be set aside to provide a sufficient incentive, so my contribution to the idea is to couple it with a dedicated fund, for say, terrorism. Set aside a large fund for terrorism coverage and should we be fortunate enough to go a number of years without an event large enough to exhaust the fund, begin paying back the early contributions to those companies with the most accurate reserves. There are obvious issues. The metric for meeting an accuracy goal is more complicated than being within X percent of the original reserve. It should vary by line of business and arguably by size of company (but not necessarily!). Again, if we start by focusing on the positive aspects, and only later address the negatives, we might be able to create an incentive system that will help the actuary and the company work together to book accurate reserves.
If you are tempted to write an e-mail to point out a major flaw in either idea, please don’t. I can list several. If you can provide an enhancement to either idea, let me know. There will be plenty of time to address the problems after we get the basic ideas in better shape.
