What exactly does it mean when we say a product is lapse supported? Every product design has a persistency assumption built into the pricing. Generally if lapses are less than assumed, profitability improves because present value of the income stream increases more that the present value of the future benefits plus the acquisition cost. If instead profitability is reduced, the product is technically lapse supported. I say "technically" because whether this will become a problem for the company is a function of the sensitivity of the lapse assumption, and, of course, just how far off the assumption is. A more operational definition for the manager is that a product is lapse supported when there is a possible lower lapse rate that will create a problem for the company. If the profit margin built into the pricing is adequate, such that the lower lapse will still leave the product profitable, the issue of lapse support is unlikely to ever come up.

A tolerable degree of lapse support can be viewed as a positive, as it tends to reverse the more usual situation where the policyholders who persist are subsidizing those that lapse. Lapse support shifts the benefit pattern in favor of those that stay over those that lapse. The problem is that with any degree of lapse support the lapse rate becomes a moving target, and it moves in the opposite direction from your aim. Most companies have a stated profit target and interest rate assumption for product design, so with a lapse supported product the higher the lapse rate assumed, the lower the premium, and the lower the tendency to lapse. Lower the assumption and the rate goes up.

Notes

Discussion

A good way to size up whether a lapse support problem exists is to ask the actuary to give you the profitability test with a zero lapse rate, or, with the assumptions mandated by the life illustration rules for determining disqualifying lapse support. These allow actual lapse experience for the first five years, and zero lapse thereafter. If the product is still profitable under those conditions, you are unlikely to ever have a problem, and can just enjoy the higher profit if lapses are up to design.

How do reserves figure in here? The short answer is that they don't determine lapse support, they just determine when you are likely to find out about it. Profitability is measured at issue, when the reserve is zero. The function of the reserve is to take up the difference as the PV of the future premiums decreases and the PV of the future benefits increases. In other words, the reserve controls the incidence of profit, or apparent profit. If the reserve accumulates too slowly, you may be reporting a profit right up until the day it is discovered the reserve is inadequate.

The lapse assumption used for the reserve technically doesn't have to be the same as the one used in the product design and profit test, but it would sure look funny if they were very different. As a result, the inadequate reserve accumulation is the henchman of the lapse supported design. You can't have one without the other, because you wouldn't start with a product if you knew it was a loser.

The degree to which a lapse supported product is sensitive to the lapse rate is a function of when most of the benefits are paid. If there if a level premium and the bulk of the payments are late in the product cycle, then the lapse rate defines the number of people who pay premiums but do not last long enough to collect any benefits.

With most life insurance it is the cash values (which are benefits paid early) that eliminate problem lapse support. Some term insurance has some degree of lapse support, but on the longer terms and older ages there is enough cash value to keep profitability from being overly sensitive to lapsation.

When life products are specifically designed to work around the non forfeiture laws in order to eliminate or reduce cash values, (increasing the lapse support) low lapses are more likely to become a problem. The classic example is the term to 100 available in Canada, which is essentially level premium whole life with no values for those that lapse. The inherent lapse support became a huge problem for the industry. The U.S. has its own set of gimmicks to eliminate cash values, such as the 40 year level term that has low premiums the first 20 years and very high the second, and the rates are so low it would be surprising if it did not become a problem.

Most level premium A&H policies are lapse supported to some degree, since morbidity increases with age. Most do pay enough early benefits to reduce dangerous sensitivity to lower than expected lapse, with the notable exception of long term care policies, where there are no "early" claims. Pricing has already become a problem for the early entrants.

Both life and A&H can become dangerously lapse supported when they provide a refund of the premiums paid after some time. Generally the pricing on such products assume a lapse rate that will keep the premiums reasonable. The risk is that the refund feature will virtually eliminate lapse at the point where it becomes obvious to the policyholder that there is little or no cost to finishing out the period.

Lapse supported products have a bad reputation because they have resulted in disasters for so many companies. Why is that? As a practical matter, the forces are all against allowing a correct lapse assumption. The higher the lapse assumed, the lower the price of the product. The lower the price of the product the greater the perceived value is likely to be, and thus the lower the probable lapse rate.

Consider the interaction between marketing and actuarial in dealing with a product that is either inherently heavily lapse supported, like long term care, or has been designed to be lapse supported.

Since the goal is to have the cheapest premium that can be justified with the established profit goals, everyone becomes very pessimistic about persistency, and the lapse assumption is set as high as anyone can rationalize. While this contributes to lower lapse rates, it will be several years before the actual lapse statistics become credible, during which period the company is building a shortfall in the reserve. This comes due all at once.

Timing is everything on lapse support. A CEO once told me that the sale of a heavily lapse supported product would cause no problem because "the agent selling it has always had very poor persistency". That doesn't work. You not only have to have the projected lapse, you have to have it WHEN it is projected.

When you say agent persistency is poor, or the lapse on a product is high, almost inevitably you are talking about the first year or two. Most executives get their persistency information from the agency system, and generally those systems only give data for the first 13 or 25 months. After that it is generally assumed that the effects of how the sale was made wear off, and the product approaches its ultimate lapse rate, which depends more on the market than the agent.

Lapse support becomes relevant well after acquisition costs are recovered. Before that, you lose regardless of the ultimate lapse rate. After that there is usually a number of years where low lapsation increases profit, before the claims accelerate. It is clearly not fair to adjust your policies and procedures to discourage early lapse and encourage late lapse, although companies that cancel blocks of business or raise premiums disproportionately in later years are in effect doing exactly that.

Lapse supported products are very likely to create adverse results. The Canadian experience with term to 100 has been disastrous to the industry. Term to 100 is whole life at less than whole life premiums due to the elimination of non forfeiture values.

Term to 100 was designed assuming an ultimate lapse rate of around 6%, and the actual rate turned out to be 2%. The difference in Canada is that the companies guaranteed the premiums and will pay the price. In the U.S. on most of these devices there is a current rate and a significantly higher guaranteed rate, so the company can raise the premium. Therefore the policyholder bears the risk. Consequences to the company depend whether premiums can be raised enough, how soon premiums are increased, and the reaction to the premium increase.

Lapse supported products can be created by their structure, or created in normal structures by inadequate pricing assumptions. The fundamental problem with lapse support is that the higher the assumed lapse rate, the lower the premium, the better the perceived bargain, and the lower the resultant lapse rate.

Long term care and term to 100 products are pretty obvious, since the lapse support is created by pricing assumptions and the future is difficult to estimate. Since no one can prove conservative assumptions, the "everyone is doing it" test is likely to prevail. Perhaps less obvious are the universal life products where charges do not immediately affect premiums, and the class of "whole life" at half the cost, whether created by artificially low early cash values, or by the combination of non guaranteed annual renewable term with something that is supposed to increase, such as paid up additions or permanent dividends, to keep premiums and face level, like the product that caused the insolvency of Mid-Continent Life.

For a closer look at the Mid-Continent Life insolvency, seized by the insurance commissioner in 1997, put "Mid-Continent Life"+"Extra Life" in Google. The product was supposed to have level premiums and level face for life, thus making it function as a whole life product. The premium was less than half the cost of an equivalent whole life policy, something made possible (to calculate) by a very high lapse assumption coupled with high projected dividends on the small par whole life portion of the coverage. The reserves were based upon that lapse assumption and fell short when the lapse did not materialize.

Mid-Continent began aggressively selling the extra-life policies, a blend of traditional whole life and term insurance, outside the Southwest after its sale to Florida Progress. The majority of the policy was made up of term coverage that decreased in value every year. But the dividends were used to buy add-ons to the term insurance to guarantee a fixed level of protection -- a big selling point. Mid-Continent heavily promoted the policy as having "level" premiums, suggesting the annual cost of a policy would not change over the years.

Former Oklahoma Insurance Commissioner John Crawford seized Mid-Continent in 1997 after determining the company was insolvent because of a huge shortfall in reserves. Florida Progress fought the move, saying the company could raise premiums to cover future liabilities.
© St. Petersburg Times, published September 28, 2000

Note that the argument that the company could raise premiums did not forestall the seizure of the company, and that the commissioner sued the parent company, a utility, claiming it had control of the board and should be held responsible for the deficiency in the reserve. That suit was dropped by the next commissioner, and the issue never resolved.

Lapse supported products that perform as projected, i.e. have a lapse rate such that the expected number of premium payers is not around to collect benefits, still need to be reserved carefully so that too much of the lifetime profit is not reported in the early durations. While this is true of all products, ones with significant lapse support are the most difficult to gauge.

The essential nature of a level premium is that policyholders pay excess premiums early to support deficient premiums later. The reserve is posted to sop up the excess so it will be available later to support earnings when premiums are insufficient to pay benefits. Management needs to be sure that not only is the product lapsing at the rate designed, but that the reserve assumptions match the product design assumptions. There tends to be no complaint when reported earnings appear to be higher than expected on a given product or line of business, but it is pay me now or pay me later. Earnings emerging higher than expected warrant just as much concern as earnings that appear too low. You need to identify why. Failure to do that can create disclosure problems, at the very least.

Illustrations of non guaranteed values in lapse supported life products are prohibited by NAIC Model Regulation adopted 12/4/95 and effective in states on or after 1/1/97. Actuarial Standard of Practice 24 defines lapse support as one that is not self supporting given defined persistency assumptions in the first 5 years, and assuming 100% persistency thereafter. Questions remain as to whether the regulation has been effective.

The state laws adopting the model regulation do not prohibit the sale of lapse supported life products, they just prohibit showing the prospect any values that are not guaranteed. This can be an important consideration for management if the field force is accustomed to using illustrations or the proper presentation of the product requires an illustration. For example, it is hard to imagine how universal life can be explained without an illustration, since for any face amount there is a broad range of premiums that can be paid. The choice simply determines how long the policy can remain in force at the premium selected. A given premium at the guaranteed internal charges and interest credits might appear to maintain the policy for only a few years, not a sensible option, while at the current rates might appear to maintain it for life.

Companies in the high income markets, particularly those that market through brokers, would seem more likely than most to run into problems with lapse supported designs. Designing products that avoid cash values and have unrealistic lapse assumptions may appear to be the only way to get premiums low enough to be "competitive". The market, however, inherently has a very low lapse rate, that is, in the absence of replacements promoted by agents.

The company may be looking at lapse experience that is quite high for the market they are in, created largely by substantial replacement activity sponsored by agents. If the company relies on these statistics to issue products with artificially low premiums, the product may be replacement proof, and will reveal the inherent lapse rate of the large policy high income prospect, which is almost zero if there is no better value for the agent to switch him to.

Return of premium riders are popular with agents, as they make a product look much cheaper, and sometimes even "free" on a "net cost" basis. A net cost presentation, of course, compares the total premium outlay with the total dollars received, ignoring the time value of the dollars involved. Since payments under the rider go only to those who persist to the end of the specified period, it has the maximum possible lapse support, and profitability will be highly sensitive to the lapse rate.

For an excellent analysis see "Return of Premium Riders Under the Microscope", SOA Annual Meeting New Orleans, October 2001. Focusing on the disability insurance market where ROP is often a percentage (60% for example) of premiums less claims paid every 10 years, the panelists note that the demographics of ROP for the majority of purchases are the blue collar and gray collar markets. The presentation includes sensitivity testing of interest rate and lapse rate assumptions on profitability.

The discussion includes the CSV rider, which has the quite different effect regarding lapse support. CSV returns a percentage of premium, grading from 0% to 100% at maturity, when the base plan is surrendered, i.e., lapses. This has the early benefit effect of the traditional cash value, reducing sensitivity to lapse.

The ROP rider is clearly a tontine, modified by the deduction of any claims paid from the ultimate payout. Everyone pays in, but only those left at the end receive payment. The payouts are insufficient to make the ROP a rational investment for the policyholder. While the return is positive, usually north of 8%, for those staying the course, the risk of lapse makes that a bad investment. The ROP is best understood as a sales tool for the agent. For an considered view, and an consideration of the risks to the company, particularly on level term with ROP, see two articles in the Transamerica newsletter, the Messenger. here and here.

The relevance of the claim offset to the company varies by the nature of the product. Disability policies produce significant offset and enjoy claim deterrence as a result. Cancer policies produce low incidence high value claims, so the offset is less important to the pricing.

The purest tontine is the ROP available on term life policies, since the risk of reduction by claim is minor. One example of Zero Cost Life Insurance states that an extra premium of $14 a month will yield a "a tax-free check for $26,640 at the end of 30 years". The future value of $14 at 6% is $14,000 so they must expect just over half of the insureds to collect. At 9% the future value would cover the entire ROP. Put another way, the yield to the survivors is 9%, which in most periods is a good yield for a tax free return. On the other hand, it looks pretty bad if you factor in that each saver apparently has a 50% chance of losing their investment.

Like all heavily lapse supported products, ROP can be profitable, even highly profitable, if you can get the lapse rate right, or better yet, market with a design rate lower than you actually experience. If your perception of competition permits the higher premium dictated by the lower lapse rate, all of the excess goes straight to the bottom line.

While the strength of your niche will define competitive pressure and influence your lapse assumption, certain niches have additional pricing advantages. If the lapse rate is influenced by independent events rather than purely individual decision, the contrary affect of price on lapse may be significantly less, and therefore easier to forecast accurately. A prime example is work site marketing, or any market where premiums are collected by payroll deduction. Employee turnover and the cessation of premium deduction often results in a lapse, and for each industry, highly predictable. The normal rate is unlikely to be influenced by the relative cost of the policy.

There are some ROP offerings that do not take the lapse rate (and the resulting windfall) into consideration in pricing. For example, if the assumed long term new money rate is 7%, and the premium on the rider, invested at 7%, will give a future value equal to the return of premium, the lapse windfall flows exclusively to the company. To the same effect is the assumption of a very low lapse rate, like 2%, which would be pretty safe for about any product or market. Note that the Canadian disaster with term to 100 still had a 2% lapse rate.

Perhaps the definition of lapse support, at least with a negative connotation, needs further clarification. If a product is sensitive to lapse, and the lapse rate turns out to be significantly different from that priced into the product, then it is going to be unfortunate for somebody. Either the company takes a lower profit than projected or the policyholder paid a higher premium than necessary to produce a "reasonable" profit for the company. And of course, overpricing a product in a competitive market will be a negative for the company as well.

With some products the structure makes it is impossible to avoid lapse support, and ROP is one of them. The tendency of the lapse rate to move away from any reasonable projection explains the bad reputation, but unless design modifications that increase early benefits are possible (such as the CSV rider), all you can do is pay close attention to the lapse assumption at time of pricing, and then track it closely.

Any company that suffers from poor persistency in the early years should certainly consider emphasizing ROP. The standard nonforfeiture values are usually zero for at least the first 5 years, and many companies, particularly those in the lower income markets, lose more than half their policies in that period. On those policies the company received extra premium and paid no benefits. Poor early persistency will undercut the profitability of the base plan, so an ROP rider can turn a loser into a winner.

The pricing of ROP also favors companies with less than average persistency. From a purely theoretical standpoint, the worse the expected persistency, the less you would charge for ROP. But of course it doesn't work that way in practice. The rates are fairly uniform, and are kept high by the necessity to keep them adequate for the more persistent markets. For example, ROP is very popular with the level 20 and 30 year term sold for home mortgage protection, and the fairly good persistency historically has supported rates that add 80 to 85% to the base premium. There is no incentive for companies with poorer persistency to charge less than to going rate, so the result has been significant support to the profitability of the sale.

What about nonforfeiture values? Many states now require ROP riders to comply with the standard nonforfeiture law, so the profit to the company in a lapse will be reduced by the cash value. Even for sales in states with no requirements, market forces and standard industry practice has made the provision of cash values fairly universal. But "fairly" is not quite the word to use, since the ROP is still a loser for anyone lapsing before the bewitching hour, and hugely profitable for the company, when compared to the base product sold without ROP.

In most instances ROP nonforfeiture values are not as high as you would expect from a stand alone product. There are several reasons for this. Many ROP riders were filed as annuities, which they do resemble, and were thus exempt from the nonforfeiture requirements. Thus the cash values provided tended to be skimpy. Another reason, even in those states that require compliance, is to consider the base plan and the ROP rider together, and the base plan, particularly level term the way it is packaged today, often has a long expensive tail designed to reduce or eliminate reserves. An example is the "40 year" term with is really 20 year coverage, as the second half is prohibitively expensive (to keep the PV of future premiums more than the PV of future benefits). Taken together with a ROP rider, the tail also undercuts the rider CVs, leaving the cash values at the mercy of the market.

ROP cash values are drawing attention from the SOA, and some states are looking at such riders on a stand alone basis.

All limited pay policies have a degree of lapse support, but most are purely technical and will not cause problems for the company. For example, on a 7 pay 20 year term policy the real profit will decline after the 7th year (you have all the premium, and all you have left is cost) but the reserve accumulated during the premium paying period should cover the subsequent mortality. If, however, you add an ROP rider to the policy, you have back loaded the pay out and probably converted to a lapse support that needs to be carefully handled.

How do you handle reserves on a limited pay policy? To the business person it makes sense to accumulate the present value of the reserve during the premium paying period. To do otherwise, for instance, to accumulate the reserve pro rata over the life of the policy, is to seriously front load the profits. If all the profit emerges during the premium paying period, even small loses for the rest of the duration can become a problem, particularly on a closed block. But here we meet our old argument about ROI. The reserve counts, and accumulating the reserve early will significantly reduce ROI, even though there is no effect on the real profitability of the product. The actuary may then feel that accumulating the reserve over the life of the policy is a more accurate presentation for management, whom he feels may lack an in depth understanding of the process. Adding an ROP rider to this process will exacerbate not only the ROI problem, but also, if the reserve accumulation is deferred, create real problems for the company.