SOA: In A PBR World, Accountants Must Share Computers

When the actuaries tested how term blocks might do under a wide range of conditions, they found that PBR reserve amounts varied widely from the comparable statutory reserve amounts.

The gaps ranged from “0% for a relatively newer, less competitive, 30-year level premium term product to 82% for a more mature, competitive 20-year level premium term product,” the actuaries write.

Overall, the current statutory reserve levels for recently issued term policies are high enough that the proposed PBR requirements would result in no additional reserving requirements, the actuaries write.

When the actuaries looked at blocks of universal life policies with secondary guarantees, the found that 1 of the 6 UL blocks tested might need more reserves.

The actuaries found that 1 of the 4 blocks of participating whole life policies might need more reserves to meet PBR standards.

The actuaries also asked actuarial task force members who provided the case study modeling results about how well the modeling process really worked.

One reaction was that the practice of rounding up to the next higher underwriting criteria scores table appeared to be both conservative and subject to manipulation.

“Such a requirement may lead to underwriting criteria that just make it to the next lower UCS table,” the actuaries who wrote the study suggest. “A weighted average approach would be better at reflecting the incremental mortality improvements that should accompany the gradual tightening of underwriting standards.”

The actuaries also found strange problems cropping up when companies had to combine use of company date with industry-wide data when they had too little in-company experience to provide the “credible” data needed to handle some categories of business.

One insurer reported having better than underwriting results than the industry as a whole. It could use its own data for most categories of life insurance business, but it had too little experience to back its highest level of preferred business and ended up having to use industry experience for that category.

As a result, “valuation mortality rates for the most preferred class ended up being higher than [for] the next most preferred class,” the Milliman actuaries write. “This result is counter-intuitive and would be difficult to manage going forward.”

To meet PBR requirements, the actuarial task force members had to feed 1,000 scenarios into statistical modeling systems.

The sheer number of calculations involved with that process required use of a grid computing or distributed processing approach, and the actuaries who needed those computing resources had to compete for computing resources with the accountants who were preparing company financial statements, the Milliman actuaries write.

Most actuarial task force members “had trouble getting grid computing time for their research project since it took a back seat to financial reporting deadlines,” the Milliman actuaries write.