#151 | Lincoln WealthAccumulate IUL 2019 – Part 3
In traditional investment asset management, all fees are deducted as a percentage of the assets. Charging fees this way means that the performance of the asset is the same regardless of the sequence of returns. That’s why traditional risk measurement tools focus almost exclusively on return variance rather than return sequence. An asset that back-loads or front-loads returns could be equivalent in terms of risk and return for the purposes of determining, say, a Sharpe ratio. But the sequence of returns matters to the final result if fees are deducted in any way other than as a percentage of the assets. And, of course, virtually every fee in a traditional life insurance policy is calculated in some way other than as a percentage of the assets. As a result, all Universal Life policies have some level of sensitivity to the sequence of returns.
That’s even more true for Indexed UL products with some form of leveraged Index Credit Multiplier. Even though many products have ICM charges quoted as a percentage of assets, which means that the ICM itself doesn’t increase sequence of return risk, the fact that the ICM lowers the floor return and effectively increases the cap means that the overall product is more sensitive to sequence of returns because of the other policy charges interacting with a more volatile account value. It’s not so different than what happens in Variable UL.
By tying the benefits of the multiplier specifically to the sequence of index returns, the PPC in WA 2019 creates a whole new level of sensitivity to the particular sequence of returns that is virtually impossible to see in a traditional, level-rate illustration. That’s why the Dynamic Illustration Tool is so powerful. It allows us to actually see what WealthAccumulate 2019 and its PPC does under real-world S&P 500 return sequences.
For this analysis, I used only the Perform account in WA 2019 but I built two other alternative account options that Lincoln could have created but chose not to. The first is an all-ICM (“All ICM”) account where the 4% asset-based charge in the Perform account goes to purchase a 198% Index Credit Multiplier. The second is an all-PPC (“All PPC”) account where the 4% asset-based charge purchases only a PPC with a 235% Positive Performance Range Maximum. Finally, I also added an indexed account with no ICM (“No ICM”) and no asset-based charge to serve as the baseline product. For the 500 stochastic scenarios, I used a population of 17,700 20 year S&P 500 return sequences dating to 1950 with random start dates. Each 40 year observation period has the same 20 year S&P 500 scenario stacked sequentially. Not coincidentally, the average 20 year S&P 500 return of the 500 scenarios was 6.12%, exactly the same as the AG49 maximum illustrated rate. The methodology I use to select S&P 500 scenarios is nearly identical to the one used by the AG49 guideline. I should also note the methodology I use is a very simple and intuitive way to get a perspective on real-world expectations, but it’s certainly not the only way.
Let’s ease into the results with a very simple graph showing just the average index return and two versions of WealthAccumulate 2019 – the Perform account as it’s built and the No ICM version of the product. This is the baseline for the analysis.
Several things pop out on this graph. First, the yellow line of 40 year account value IRRs of the No ICM version of the product track closely to the black line, which is the geometric average of index credit returns over the same 40 year periods. This is the way all non-leveraged Indexed UL products look. In general, their performance follows the pattern of the index credits with a fairly consistent level of drag from policy charges, but it’s not exact because policy charge deductions do create variations based on the particular sequence of returns. That’s why the yellow line for the No ICM product jumps around a little bit while the index credits are smooth. Second, the red line representing the Perform account exhibits typical Leveraged IUL (any Indexed UL with a charge-funded ICM) behavior in that high crediting performance creates extremely high account value performance and low crediting performance creates much worse account value performance. And as is also typical with Leveraged IUL products, the Perform account in WA 2019 exhibits quite a bit more sequence of return risk than the No ICM version. As I wrote earlier, the ICM itself isn’t usually the culprit. Instead, it’s how the ICM interacts with the other policy charges to amplify policy volatility.
However, WA 2019 is not a normal Leveraged IUL because the PPC makes up a portion of the leverage structure of the Perform account. It’s very hard to just look at this graph and know what portion of its sequence of return risk is due to the ICM and what part is due to the PPC. For that, we need the other two clone products – a WA 2019 version with only an ICM and one with only a PPC. Take a look at what happens when we add those two clone products to the graph.
This graph puts the story in stark contrast – the Perform account in WA 2019 is truly a mixed blend of what would have happened had Lincoln chosen to do an All ICM version or an All PPC version. To the degree to which Perform outperforms in strong equity markets, the All PPC version does even better and the All ICM version does worse. The PPC clearly adds more leverage to the mix, which is why I said in the previous post that I anticipate that other life insurers will develop all-PPC leveraged products in order to take advantage of even better illustrated performance. The All PPC version of the product clocks an incredible average IRR over 40 years of 6.78%, which is 0.4% better than Perform as-built, 0.9% better than an All ICM version and 2.3% better than the No ICM version. If you think a 6.8% IRR sounds a little low, then save your curiosity for the latter part of this article.
But we still didn’t answer the question about sequence of return risk. For that, we need another graph to tell the story. Starting with the first index credit return scenario, I subtracted the difference between that scenario and the next scenario to get the average change in index return from one scenario to the next. All else being equal, that would be the expected change in each product’s IRR as well. The degree to which the change from one scenario to the next for each product is positive or negative, larger or smaller than the change in the average index credit is a clear indicator of how the particular sequence of return for that scenario impacted policy IRR. Take a look.
This graph clearly shows that the PPC is hugely sensitive to sequence of returns relative to a typical Indexed UL or even an All ICM design. Looking at the outliers over 40 years, policy IRR can easily deviate from the expectation by 1% or more just because of the sequence of returns. Contrast that extreme sensitivity with the No ICM version of the product, which varies by maybe 0.3% at the most due to sequence of returns. But I was surprised about a couple of things. First, I was surprised by how closely the All ICM and Perform accounts look in terms of variance due to sequence of returns. I think that’s an argument that the design Lincoln chose, which mixes ICM and PPC at about a 60/40 ratio, is an optimal one. It has the potential for more upside without dramatically exacerbating sequence of return risk beyond what a normal ICM structure would have done.
The second thing that surprised me was that the deviances don’t usually work in favor of the PPC. Borrowing the logic that the S&P 500 usually has long grinds upwards followed by steep downturns, I was expecting that the deviances of the All PPC product and the Perform account would skew towards the positive, but that’s only partially true. For the All PPC account, only 46% of the deviances are positive – but when they are positive, they tend to be slightly larger (0.75% vs. -0.69%). For the Perform account, however, 52% of deviances were positive but tended to be slightly smaller than negatives so that the average is almost perfectly zero. The All ICM account was almost identical. Again, Lincoln appears to have picked the optimal blend of the two bonus types. There’s something that works well in the real world when you combine an ICM and a PPC. Average returns for the combination are better than with just an ICM, but the sequence of return risk isn’t meaningfully higher.
To go back to the original analogy, it appears that the best bet is not $100 to Bet A or $100 to Bet B but, instead, to put $50 on each. That’s what Lincoln’s Perform account does and it’s hard to find fault in the structure based on this analysis and in the context of other multipliers in the Indexed UL market. In the crazy world of Indexed UL multiplier designs, you can’t support ICMs and throw dirt on the PPC. And you definitely can’t say that the way Lincoln mixed the two for their account options is a bad thing. Sorry. If you don’t like charge-funded ICMs (and I don’t) and you think they’re fundamentally inappropriate for Indexed UL products (and I do), then feel free to disparage it. But in context, you can’t say the structure is any worse and even might arguably be better than competitor offerings.
Which is not to say that the way this thing illustrates is an accurate representation of the reality of how it will perform. Far from it. The analysis so far has focused exclusively on how it performs in real-world scenarios, not how its real -world performance lines up with what the client sees on a level rate illustration. The breakdown occurs when you start to compare the real-world scenario returns with what practically every client will see when the agent shows them an illustration run at the AG49 maximum rate. Take a look at the table below, which compares the illustrated performance of the 4 designs with the real-world average performance.
|No ICM||All ICM||Perform||All PPC|
In the case of the No ICM product, there is a relatively small difference between the illustrated AV performance at the AG49 maximum illustrated rate and the average of the real-world stochastic scenarios. The gap only slightly increases with the inclusion of even a large charge-funded ICM because both charges and credits are calibrated to the account value. But things really get interesting once a PPC enters the mix. Perform blows the doors off of the All ICM option on the illustration, posting an incredible 0.9% increase in the AV IRR at the AG49 maximum rate, but the real-world stochastics increase only a little more than half as much. Switching the remainder of the Perform account to an All PPC design yields another 0.9% increase in illustrated performance at the AG49 maximum rate but only a 0.4% bump in the average real-world stochastic result. One thing is for sure – the PPC dramatically increases illustrated performance. The fact that the illustration operates off of a level rate and a level rate only is nothing but a boon for illustrated performance for the PPC.
This is the core problem with the PPC. All Indexed UL products illustrate slightly better under a level rate than they would have performed with variable rates. This is commonly known and an oft-cited problem with life insurance illustrations. I think one can reasonably argue that charge-funded ICMs don’t necessarily exacerbate that disconnect. But PPC-type multipliers? That appears to be a different story. Here, we have a product feature that creates a massive gap between illustrated performance in the level rate scenario and the average return in real-world scenarios. How likely is it that a PPC-type structure will underperform the illustration at the AG49 maximum rate? Take a look at the graph below, which compares the stochastic results with the illustrated performance at the AG49 maximum rates for each of the different products.
In all of the designs, the real-world scenarios are below the AG49 maximum illustrated rate most of the time. For the All ICM and No ICM designs, that happens 58% of the time. For the Perform and All PPC designs, the real-world scenarios are below the AG49 maximum illustrated rate about 61% and 63% of the time, respectively. The real differentiator, however, is the incredible gap between the top and bottom of the distribution for each design. The highest performing scenario for No ICM is 6.7% and the lowest is 1.9%. The same figures for the All PPC design are 14.1% and -1.5%, an absolutely massive swing relative to, well, anything. That’s what leverage does to the return distribution and the PPC only amplifies it to soaring heights. That’s also why the average returns are so low relative to performance at the AG49 rate. It’s not so much that the All PPC design underperforms more often, it’s just that when it does underperform, it massively underperforms. To give you another perspective, the biggest gap between the illustration at the maximum AG49 rate and the lowest real-world scenario for the No ICM design is just 2.6%. For the All PPC design, the same gap is a whopping 9.6%. Even though the downside risk for the two designs is actually not that different, the gap between the illustrated return shown to the client and the average return, let alone the low return scenarios, is absolutely massive. The problem with the PPC (and ICMs) in general is one of client expectations, not so much absolute downside risk.
It is tempting to conclude this article by saying that the PPC is a solid multiplier design if client expectations are managed properly, but that would be a shallow-end look at what’s going on. The PPC, particularly the way that Lincoln built it into their account options, is not much more dangerous than a normal ICM and arguably generates even better real-world results. But when you contrast how it works with how it illustrates, the PPC poses a huge challenge. Producers almost certainly will unwittingly over-illustrate WA 2019 due to the PPC. Expectations will not be managed properly. But more than anything else, the PPC is yet another way for life insurers to lever up the illustrated (but unproven) expectation of huge systematic option profits. The entirety of this analysis hinges on the idea that today’s option budget and option prices would have been the same back to 1950. That is obviously not the case and real option prices would have absolutely wreaked havoc on designs like this one. So before you rush off to sell the PPC (or disparage it), remember that this entire discussion boils down to the same question that underpins every discussion about Indexed UL illustrations – what are reasonable long-term option profit expectations? If the number is high, the PPC is magical. If the number is low, the PPC implodes. You’ve seen this movie before.
If the PPC is so effective at increasing leverage, then how effective is Lincoln’s Return of Premium feature at providing a viable exit option? The answer, as it turns out, is far less straightforward than you might think.
Want more great content like this? Join the hundreds of other people who have subscribed to The Life Product Review to get access to nearly 150 archived articles and weekly new content. Just click here