Vol 7 No 1		January 2005

Katt & Company is a national fee-only life insurance advising firm. The June 2002 Forbes magazine, and a July 16, 2003 Wall Street Journal article, name Peter Katt as one of only four nationally recognized advisors. The Forbes article states that, "…advisers are well worth the money… These savants are working for no one but you…" For references please contact us.

Katt & Company News - Scott Witt has joined the firm. Scott is an actuary (FSA) with 10 years experience in life insurance pricing and related matters. Our clients will benefit from Scott's experience, talents and good sense. One of Scott's first assignments has been the development of a sophisticated mathematical model for assessing the appropriateness of variable life, which this Perspectives details.

The Fuzzy Math of Variable Life

Insurance agent boilerplate statement "The stock market has averaged 12.4% since 1928 - I am showing you an illustration that only uses a 10% average, which means the results are almost certain to be better than the illustrated values" is dangerous nonsense.

Variable life has various sub-accounts that policyowners can choose to invest their cash values. VL policy expenses are higher than universal and whole life expenses so investments are logically directed to equity funds for the potentially higher returns, otherwise it makes more sense to buy UL and WL. But the performance of stocks is very volatile with years of losses that are sometimes large. VL policies with level death benefit designs are intrinsically flawed because we can't know what the premiums will be as cash values surge and recede. This has created a huge problem for variable life buyers. Collectively the life insurance industry uses arithmetic averages (taking the year-to-year gains and losses and dividing by the number of years) to create premium certainty that is necessary to make sales. But there is no premium certainty. The only relevant measurement is a geometric average that defines actual performance of volatile equity investment results. A simple example will explain this. A $100,000 investment is made. The first year it grows 80% to $180,000. The next year is drops 50% to $90,000. The arithmetic average of these two years is a plus 15% (80% minus 50% = 30% divided by 2 = 15%) and this is the way it is understood by every insurance agent I have ever dealt with. But the actual result is a geometric average of a minus 5%. Therefore the insurance agent boilerplate statement "the stock market has averaged 12.4% since 1928 - I am showing you an illustration that only uses a 10% average, which means the results are almost certain to be better than the illustrated values" is dangerous nonsense. A significant additional problem associated with variable life is that when actual cash values have a big drop when the insureds are old, it exposes the policy to very high mortality costs. This compounds the problem and leads to policies terminating without the payment of huge premiums.

One of our most common engagements is assisting shell shocked VL owners, many of whom purchased their polices just prior to the busting of the 1990s stock bubble. But VL's inherent flaws exist even if we weren't living in the shadow of a major market correction. With Scott's assistance we have developed life insurance pricing models to test the probabilities illustrated premiums will properly fund a level death benefit VL policy. The test, using 1,000 or more runs, randomly selects investment yields based on a specified arithmetic mean and historical standard deviation (volatility). From the testing we can determine the number of times a policy is expected to fail and the average excess premiums that will be needed. For example, in a litigation case where we are serving as expert witness we found that the agent's promised (written in the application) single-premium funding of $650,000 for a VL policy had a 55 percent probability of failure and a present value average additional premium need of $360,000. Since the policy was sold in 1998, we chose testing parameters based on the information known at that time - an arithmetic average of 13.0 percent and a standard deviation of 20.3 percent, both covering 1928 - 1997. We extracted the subject VL policy's pricing from the illustration and tested 1,000 runs using the randomly selected patterns of annual stock results from the investment history described.

This testing, known as stochastic or Monte Carlo, is also essential in assessing the risk for increasing death benefit designed (superfunded) VL policies and their potential advantage compared with a participating whole life policy. A recent test produced remarkable results. Our client has a VL survivorship policy that is being superfunded. He is 72, his wife is 70 and both are in excellent health. The client also has a superfunded Northwestern Mutual whole life policy. He has $200,000 of additional funds for investing into the trust's life insurance policies. Using the methodology described above we tested the probabilities that investing the $200,000 into the VL policy would produce better results than investing into the WL policy. We also tested for the probabilities that the VL policy might fail. Using a 250 basis points advantage for VL equity results compared with Northwestern Mutual's dividend-interest-rate and a VL arithmetic mean of 10 percent as reasonable parameters we found there is about a 65 percent probability investing the $200,000 in the VL policy will produce better results. But we also found that there is a 7 percent probability that the VL policy will fail by the insureds' joint life expectancy and about a 10 percent probability of failure for the VL policy without the additional $200,000 premium. This has alerted us to the need for more careful management of the VL policy or a decision to replace it with another Northwestern Mutual WL policy. Because the client has the good judgment to spend most of the winter skiing in Utah, avoiding having to think about such things, a decision hasn't been made yet. But stay tuned and we will let you know what he wants to do.

You should endeavor to obtain life insurance policy inventories from your clients. All VL policies should be reviewed by fiduciary life insurance advisors.