## #117 | The Death Benefit IUL Dilemma – Part 2

I’ve always made the case that Indexed UL and traditional UL are so closely related that viewing IUL as its own product category is somewhat misleading. The fundamental mechanics, portfolio yields, pricing regimes and risk profile of the products are virtually identical. The only functional difference between the two is that Universal Life pays the crediting rate to the policyholder whereas Indexed UL uses the crediting rate to buy call options, which are market-priced exposure to the movements of an external index. Although both products rely on the same foundational elements, they have entirely different illustration methodologies. In UL, the maximum illustrated rate is the crediting rate. In IUL, it’s AG49.

In isolation, AG49 makes a fair bit of sense as a methodology for determining the maximum illustrated rate of an Indexed UL product *used for accumulation.* AG49 looks back over 65 years of actual S&P 500 data and applies the currently available cap in the product to the index returns to produce over 20,000 annual returns calculated on a daily basis. The average of all of those 20k+ datapoints is the maximum AG49 illustrated rate. Overfunded accumulation IUL products can handle the variation from the average because they still do what they’re supposed to do, which is deliver tax efficient growth at a stable cost, generally regardless of the actual performance*.

For death benefit sales, though, using an Indexed UL product presents a host of thorny issues. The fact that the maximum illustrated rate on an IUL product is set at the *average* of the 20,000 datapoints brings up an obvious question – can the policy be successful even if the actual credited return varies from the average? Again, in accumulation IUL products that are not leveraged to the hilt with funky multipliers, premium financing or income with indexed policy loans, the answer is generally yes**. But what about in death benefit products? The whole point of using a DB IUL is to illustrate the lowest premium possible. The only way to get the lowest possible premium is to run the illustration at the maximum possible rate, which is AG49. If the producer solves for the lowest possible premium based on the average return over 20,000 datapoints from the last 65 years assuming that the cap never changes, then it’s fairly intuitive that half of those policies can reasonably be expected to fail.

* *

But when you have the Dynamic Illustration Tool, you don’t need to leave a question like this to intuition. Instead, I ran 500 scenarios through a generic DB IUL product to see what happens. For reference, I used a 45 year old Preferred male with a $1M DB. The product reflects a typical DB IUL design, with slightly higher upfront charges and lower mortality charges in the later years. The cap is 10% and the maximum AG49 rate is 6.09%. Using 6.09%, I got a premium solve of $7,450. Looking at a quick benchmark of DB IUL products, that premium is actually little bit higher than what you’d find in the market, primarily because I didn’t build the product with any shenanigans like multipliers or bonuses. The chart below shows 500 real world scenarios for the product ranked from the highest average crediting rate to the lowest. The black line in the middle shows where 6.09%, the maximum AG49 rate, falls on the chart. The average crediting rate for the entire dataset is 5.96%, which is lower than the AG49 maximum rate because the DIT stops calculating returns once the policy lapses. If I were to allow the calculation to continue, the average crediting rate for all of the policies would be within 1 basis point of 6.09%.

They say a picture is worth a thousand words and this one is certainly no exception. First things first, a lot of policies lapse – exactly half, actually. What this graph tells you is that solving for the minimum premium based on the maximum AG49 rate is solving for a 50% success rate *assuming no changes to any policy parameter*. The average age where a policy lapses is 92, with a minimum of age 86 and a maximum of age 94 (because the policy has zero charges after age 95), so it’s not like these lapses occur well after age 100 and no clients are around anyways. Instead, they happen right at the heart of mortality for a healthy 45 year old. So how much premium would you have to pay to get the success rate to a more acceptable level? To get to 80%, tack on an extra $1,000 to bring the total premium to $8,450. Another $1,000 will get you to a 99% success rate. But at that price, you can have your pick of Guaranteed UL products. Furthermore, the “success rate” I’m talking about here doesn’t mean that risk is eliminated. What it really means is that, holding all else constant, the product has the specified percentage of likelihood of remaining in force. For context, the success rate of a UL product with the minimum premium solve is 100% – as long as the crediting rate stays the same and the charges don’t change, it’s going to last. But with IUL, the mere presence of indexed credits throws the whole enterprise into doubt.

Second, the chart also shows that the average crediting rate is only part of the story. Policies that have average crediting rates higher than 6.09% can lapse – 32 of them, or 12.8% of all lapsed policies, to be exact. There were also 32 policies with average crediting rates lower than 6.09% that stayed in force. Putting those together, more than 25% of the policies did not lapse or remain in force based solely on their average crediting rate. Instead, the sequence of returns was the deciding factor. Not all average crediting rates are created equally when uneven policy charges are coming out each year. However, it appears that sequence of return risk cuts both ways, which makes intuitive sense. There’s not a systemic problem. But there is a problem for each policyholder. Exactly zero policyholders will send a “thank you” note because their policy stayed in force even when their average crediting rate is lower than expected. But plenty of policyholders will be irate because their performance was good, on average, but the policy still got into trouble. That’s the problem with sequence of return risk in DB IUL products. The 6.4% of all policyholders who have an issue just from sequence of returns will be more than a little annoyed.

Finally, this chart has another huge but less obvious problem. In Universal Life, a 4% crediting rate is a 4% crediting rate. In Indexed UL, a 4% option budget buys a 10% cap that illustrates at 6.09%. But option prices change more often and less predictably than option budgets do. A 4% option budget could buy a 10% cap a couple of years ago, but now it buys a 9% cap, a change mostly based on two things that don’t help out the option budget – index volatility and short-term interest rates. A UL product with a 4% crediting rate that never changes will perform at 4%. End of story. An IUL product with a 4% option budget that never changes could have caps that float from 7-11% depending on volatility and market rates, with maximum AG49 rates ranging from 4.5% to 6.5%. See the problem? Putting a “valuation” on an IUL product at any point in time is extremely difficult. Accumulation products can handle it. Death benefit products can’t. Take a look at what happens to the chart if we assume a 9% cap rather than a 10% cap.

With a 9% cap, 78% of policies lapse. Dropping the cap to 8% results in a 91% lapse rate. If 8% sounds unreasonably low to you, then here’s a gutcheck – pull up the option chain for SPY and you’ll see that an 8% cap costs right about 4% these days. And that’s not out of whack with long-term averages. The likelihood of a 4% budget going the other direction and buying an 11% cap is slim. That would require extremely low volatility and extremely low interest rates, exactly what we had from 2013 to 2015. As discussed at length in the Indexed UL in the Mirror series, that epoch was the perfect petri dish for Indexed UL to grow. And now it’s gone, but its echoes remain. Companies are still reluctant to lower caps even though option pricing has been moving against them. So take a 10% cap with a grain of salt – it’s probably going to be 9% or 8% if things keep going the way they are.

These three issues basically preclude the idea of using the maximum AG49 illustrated rate to solve for the minimum premium to keep the policy in force. Best case, 50% of those policies lapse. Worst case, the cap drops a little bit and suddenly 91% of them are at risk of lapse. Furthermore, the specific sequence of returns can put your client on one side of the fence or the other. It’s a very dangerous game that belies the fundamental reason why people are buying life insurance – to offload risk. Instead, by paying a premium based on the maximum AG49 illustrated rate, they’re taking a very different kind of risk.

I was tempted to write a follow up blog post that specifically analyzes particular IUL DB products being offered in the market, but I decided against it. I did enough of the analysis to see that no design eliminates or even materially reduces the risks outlined here. A few of them actually make the situation a bit worse. So what are we to do? The answer is not going to make me many friends, but it’s the only one I could come to. When you do an illustration with a minimum premium solve, run it at the option budget. If you don’t know the option budget, then run it at 4% or 3%. The “competitiveness” of the product will drop, but the likelihood that the product actually works will dramatically increase. If the product actually gets stellar performance, then the client can stop paying premiums earlier than originally anticipated. Plan for the worst, hope for the best. Or just buy a Guaranteed UL or Whole Life product. At the end of the day, stripping out DB IUL’s illustrated rate advantage reveals it to be what it actually is – a traditional Universal Life policy with a twist. Pay premiums accordingly.

*This is only true for policies and sales strategies that don’t create external leverage on the performance of the policy.

**Admittedly, this is 90%+ of the accumulation IUL market.