## #100 | JH AIUL18 – Part 2 – Bonus Structure Analysis

#### Executive Summary

PacLife PDX and John Hancock AIUL18 represent two different philosophies on how to deliver leverage to clients. PDX does it through the Performance Factor, part of which is funded by fixed policy charges. AIUL18 funds its constant 55% bonus through a 2% asset charge. Each has its merits. With real-world return scenarios, AIUL18 performs very similarly to products without indexed interest bonuses, but with more leverage and risk. In those same scenarios, PDX’s performance can deviate meaningfully from AIUL18 because the sequence of return itself impacts PDX more and differently than AIUL18. Sometimes that creates real problems for PDX and other times it actually works in PDX’s favor. In the final analysis, there’s not a right or wrong answer – but AIUL18 performs more consistently and intuitively than PDX and that’s a benefit unto itself.

*I want to preface this post by saying three things. First, there are a lot of ways to potentially look at the risk/return profiles of bonuses supported by asset charges and fixed charges and I suspect that other lenses might deliver different results. I looked at a few different ways to do the analysis and landed on this one because it delivered the most intuitive and consistent results. My results are less important than the questions they raise – if you disagree with where I land, then take it as inspiration to come to your own conclusions by tackling the same questions. Second, the analytics in this post are a bit more complex than usual, but I think this is what’s necessary to fully understand and interpret the tradeoffs that these products represent. If you’re not comfortable with this type of analysis, I would really encourage you to stick to products without funded bonuses. Finally, I apologize in advance for the sheer number of graphs in this post. In an effort to simplify the story, I’ve excluded Lincoln WealthAccumulate from the analysis. You can safely assume that it follows the same basic story as Hancock’s AIUL18 because the mechanics are very similar, although with less leverage. Lincoln will be back in the conversation for the final post in the series.*

As the complexity of life insurance policies increase, particularly Indexed UL products, the shortcomings of compliant illustrations with level illustrated rates become only more obvious. How is an advisor, much less a client, supposed to get an intuitive feel for the mechanics and tradeoffs embedded in something as obtuse as the choice between funding an indexed interest multiplier through asset charges or fixed charges when the illustration only shows a static rate? There’s just no way to do it – and yet, these are the questions being posed by the crop of modern Indexed UL products. And, of the purposes of this article, specifically the question posed by John Hancock Accumulation IUL 18 (AIUL18) and PacLife PDX. With level illustrated rates, these two products are spitting distance from each other in terms of illustrated performance, but their mechanics are fundamentally different. PDX funds its Performance Factor bonus with fixed charges. AIUL18 funds its indexed interest bonus with asset charges. These two methodologies are not the same. As much as they resemble each other in level rate scenarios, they should diverge in real world return scenarios. Answering the question of which one is “superior” is going to get no small bit of airtime as other carriers consider which one to append to their product. The table stakes are about to be raised.

Before I get into the numbers, let’s briefly talk through an intuitive approach to these two bonus structures and I would encourage you to not discount intuition here. Most of the criticism and support that I’ve heard for both structures are intuitive appeals. There’s a story to be told for both structures that exists, frankly, outside of the data that supports it. The knock on PDX and its fixed charge funded bonus is that down markets will crush the product because the fixed charge is constant no matter what the account value is. As a result, it’s mathematically possible that the fixed charge can force the policy to lapse after a string of zero returns. At first blush, then, the fact that AIUL18 funds its 55% multiplier through a 2% asset charge solves the problem because the asset charge shrinks as the account value shrinks. Unlike the fixed charge in PDX, the 2% asset charge can’t mathematically cause the policy to lapse by itself. Intuitively, therefore, IAUL18 is less risky than PDX.

But that’s only half of the equation. The other half is how the charges translate into the bonus. In AIUL18, the ratio is always the same – 2% charge for a 55% bonus. But in PDX, the charge is always the same amount but the bonus is based on the ratio of the charge to the account value. As a result, the bonus increases when the account value decreases. Put differently, the leverage in the product increases as the account value drops, which means that any positive credit will get a much bigger bonus than would otherwise happen in an asset-charge funded structure. Is more leverage a good thing? It’s kind of an interesting question. When things are bad, AIUL18 takes chips off the table and PDX goes all-in. And if you believe that the bet will eventually pay off, then you might actually prefer PDX. In any event, it’s not much of a stretch to see that these two methodologies might deviate over the short term in volatile return scenarios but deliver remarkably similar performance over the long run. It would be a mischaracterization to say that one is low risk and one is high risk – they’re both more leveraged and riskier than products without funded bonuses, regardless of whether the funding comes from asset charges or fixed charges.

The other consideration is how the leverage factor of the product changes in good times and in the long run. With AIUL18, the leverage is always the same no matter what. When the product has performed well and has lots of account value thanks to past leverage working in its favor, then it will continue to run the same bet with the same payoff structure. PDX operates differently. The fixed charges in the product are massive in early years, sometimes equating to 5% or more of the account value even in fully funded scenarios, but necessarily become smaller as a percentage of the account value as the account value grows. As a result, PDX starts off with a lot of leverage and naturally deleverages if the policy performs well over time. As you might have picked up on, this is the exact opposite of the mechanics outlined in the previous paragraph. When the winnings are good, PDX takes money off the table whereas AIUL18 always plays the same bet. Which one is better? AIUL18 is certainly consistent and that counts for something, but I think there’s a case to be made for PDX here. PDX naturally applies leverage when the client needs it most, during the accumulation phase, and naturally reduces leverage when the client is risk averse during the decumulation phase. There’s an inherent logic to the structure that is actually kind of elegant – excusing, for a moment**, all of the other issues with the product.**

I bring up the intuition behind the product because I think it’s a good guide to the analytics you’re about to see. The base setup for the analysis is that I put a basic, no-bonus IUL (modeled after PacLife PIA5), PacLife PDX and John Hancock Accumulation IUL 18 into the Dynamic Illustration Tool. The client modeled is a 45 year old Preferred male with $45,000 premiums for 7 years and a $1M level death benefit. The crediting rates use 500 scenarios of historical S&P 500 annual returns so that we can take a look at how these products compare with “real world” volatility. I sliced the analysis so that I only looked at three datapoints for account value performance – year 11, year 21 and year 41. The goal of the analysis is to look at 500 scenarios to see how these three products compare under each scenario. The graph below shows the 500 scenarios ranked from best to worst for year 11 account values.

The way to read the chart is that each line shows the account values generated for all of the 500 S&P return scenarios ranked from best to worst for the No Bonus product. These scenarios generally range from average returns in the S&P 500 of 2% to 12% with fairly high volatility, which is then illustrated in the product with its cap and floor. I chose to rank the scenarios by the performance in the No Bonus product because it’s the baseline design. Before I move on, I want to make sure I say it again – these are not time-series graphs. In the life insurance business, we spend so much time looking at graphs of how things change over time that we approach every graph expecting to see time-series data. This is not. The graph above shows 500 individual scenarios (illustrations, if you will) for returns driven by the S&P *500 at a single point in time*, which is policy year 11 in this case. You’re seeing 500 return scenarios for all 3 products simultaneously ranked from best to worst based on the No Bonus product’s illustrated account value.

As you can see, the account value in the No Bonus product ranges from a high of just over $500k to just under $300k, depending on the return scenario. The two products with bonuses range from $700k on the high end to under $200k in the case of PDX and just under $250k in the case of AIUL18. From this chart, you can get a feel for a couple of other phenomenon that are going to be obvious in the upcoming charts as well. First, the downside on PDX actually is greater than in AIUL18. Remember what I said about increasing leverage as the account value drops? Sometimes that pays off, sometimes it doesn’t, and in the tail end of the worst of the 500 scenarios, the bet just doesn’t pay off, at least not by year 11.

Below is the same exact chart using the same 500 S&P 500 return scenarios but showing account values at year 21.

By year 21, PDX has gained an edge – the leverage has paid off. This is the sweet spot for PDX and rightly so. Year 21 is at age 66, right about the time that distributions are usually illustrated to begin. By then, PDX has had enough time to ride out the bad streaks and has continued to double down on the bet and, based on historical S&P 500 returns, will come out on top. There’s still a bit of a tail in the worst case scenarios, but it looks like a fair tradeoff for consistently higher returns across the distribution for PDX.

By year 41, though, the edge has evened out. Why? Because the leverage in PDX has diminished as the account value grows (and therefore the Performance Factor shrinks), but AIUL18 has continued with the same 2% charge and 55% bonus as always. Historically, more leverage has paid off and so AIUL18 pulls back in line with PDX. Take a look at the chart below.

In a lot of ways, the stochastic result confirms the intuition. PDX naturally deleverages over time whereas AIUL18 keeps its foot on the gas. Worst-case markets are definitely more deadly in PDX than in AIUL18, but it’s not like AIUL18 is perfectly safe, either. Both products have risk and leverage. But the most important thing these graphs show is that PDX is not like AIUL18 and the No Bonus product. You can see in all 3 of these graphs that AIUL18’s returns are smooth and move in lock-step with the No Bonus product through the ranking of the 500 scenarios. The 100^{th} ranked scenario for the No Bonus product is the 100^{th} ranked scenario for AIUL18. What this tells you is that these products are mechanically very similar and produce the same pattern of results, even though AIUL18 obviously has a lot more leverage. Not so for PDX. The jagged lines in its results tell you that PDX is mechanically different from both the No Bonus product and AIUL18 and so its results aren’t ranked exactly the same as the other two. The 100^{th} ranked scenario for the No Bonus product and AIUL18 might be the 50^{th} or 150^{th} ranked scenario for PDX. This is a direct result of the fact that PDX funds its leverage through fixed charges. As I wrote in a previous post, *sequences* of returns matter to PDX far more than other products and you can clearly see that playing out in the graph above. Think about it. The sequence of return at any given rank is identical for all three products, and yet PDX’s results bounce around all over the place. In the end, PDX still performs well, it just doesn’t follow the same path as the No Bonus product and AIUL18 because it has different mechanics. This is an incredibly important distinction. Producers who approach the leverage in AIUL18 with the same intuition as a product with no bonus will not be surprised with their results. Producers who do the same for PDX will be surprised because the mechanics are different – and I don’t think many producers appreciate this distinction and how it will play out for their clients.

The analysis above presumes that the 500 scenarios pull from the S&P 500 historical returns as they are and assume that the caps for these products are always at their current levels (10% for AIUL18 and the No Bonus product, 10.25% for PDX). So what happens if we modify those assumptions? I ran a scenario that deducted 3% from all S&P 500 returns and, to spare you looking at too many graphs, take my word that the results are nearly identical to the graphs above. This is not surprising because IUL products live within their caps and floors but the S&P 500 has a much larger distribution. Taking 3% off of it will still force most of the returns for an IUL product to sit at the floor or cap. More interestingly, I ran a scenario that assumed a 7% cap for AIUL18 and the No Bonus product and a 7.25% cap for PDX. If this strikes you as an extremely remote possibility, I would caution you not to jump to that conclusion. Caps on Fixed Indexed Annuities are 5%. Rising interest rates and lagging general account yields, which is exactly what’s happening right now, could generate economics that lead to caps in the 8% range. But I did want to take a very conservative stance to see what would happen to these products. Below is the chart of the 500 S&P 500 scenarios ranked, again, by the account value in the No Bonus product in policy year 11.

There are several important observations to draw from this graph. First, the returns for getting leveraged exposure to the indexed credits through either an asset charge funded bonus or a fixed charge funded bonus diminish significantly. Why? Because this graph assumes that option prices increase and the possibility for returns in the option itself have diminished. Remember that the benefits of any leverage delivered through a bonus in an IUL product are contingent on the benefits of underlying option profit assumption. In a standard IUL illustration, option profits are assumed to be 50%. In the scenario above, it’s more like 10%. As a result, the benefits of more leverage in the product are muted.

Second, you can see that AIUL18 still tracks the No Bonus product but PDX starts to show some structural issues. Only about half of the PDX scenarios are better than the No Bonus product whereas AIUL18 keeps its edge through the 400^{th} scenario. Some of this is due to the inherent pricing of the bonuses – AIUL18’s bonus “breaks even” at about a 3.6% illustrated rate whereas PDX needs 4.1%. Given that the average return of the 500 scenarios is about 4.5%, you can see why PDX is barely staying above water. As I’ll talk about in the next post, I don’t think this a fair representation of how these two products will compare over time because the caps will adjust to mute the difference. Finally, the downside of PDX shows itself more clearly here than in when it has a higher cap, but IAUL18 still stays pretty stable. The comparison of tail risk is even more stark in the graph shown below for policy year 41.

I had to double check these numbers a few times to verify them because, frankly, I was pretty surprised at the results. But they check out. I was expecting to see that, if given a long enough timeframe, PDX’s risk would sort of work itself out. Instead, in the lower performing scenarios, PDX starts to show *more* risk. Look at the tail. There are scenarios in here where, in the 41^{st} year, the policy has literally just lapsed. There are no scenarios in either the No Bonus product or AIUL18 that lapse. Once I saw and verified it, I knew why the lapses in PDX were happening. It goes back to what I said about PDX automatically doubling down on the bet when the account value falls. This works fine if the other charges in the product are pretty modest, but at year 41 you have COI charges that are increasing as well and even more so as the Net Amount at Risk increases. In other words, PDX can ride out volatility in the short run because the other charges won’t kill it. But in the long run, a worst-case series of returns coupled with other charges in the product start to compound the problem and doubling down on the bet still won’t pull PDX out of the hole. It can just implode in the worst-case scenarios. This is the risk that people *feel *with PDX, even if they can’t see it. The analysis above lets you see it, although I’d be remiss if I didn’t note that we’re talking about pretty extreme scenarios – probably beyond anything we’ve seen historically. Remember that this analysis is randomly pulling its returns from historical S&P 500 annual returns and is not necessarily reflective of any *sequence* of historical returns. The tail end of this analysis, when PDX implodes, represents a tiny probability of happening and it would be a mistake to base a decision about the product on that risk alone.

But this analysis also lets you see something else that I was not expecting. For about half of the scenarios on the left side of the graph, when returns are strong, PDX outperforms AIUL18. This is remarkable because PDX starts off at something of a disadvantage. By year 41, the 2% charge in AIUL18 is bigger than the fixed charges in PDX, so the bonuses in AIUL18 are bigger. Furthermore, PDX’s hurdle rate for its bonus to work (as I said earlier) is 4.1% and the average return of all of the scenarios is just 4.5%. What’s going on? Two things. First, don’t forget about QX. If you’ve already forgotten, let me remind you – QX is an “unfunded” bonus that shows up in PDX in the later years and, in this case, is as big as 25%. I made the case in previous posts that QX is more important in PDX than the fixed charge funded portion of the Performance Factor that we’ve been talking about in this post. QX bonus is still applying leverage here and it does not have a hurdle rate, so that’s benefitting PDX. Second, I have a sneaking suspicion that PDX’s bonus mechanism actually feeds on volatility in a lower return scenario in a way that Hancock’s does not. Even though average returns across all of these scenarios are fairly low, annual returns are still mostly sitting at the cap and the floor. When you have a floor observation (0%), then both products post a 0% return. Their multipliers did no good. But the next year, when both products have a cap observation (7% and 7.25%, respectively), then PDX’s bonus will have increased from the year before, which allows it to gain back the ground it lost and then some (because of QX). This is the paradox of PDX. Volatility can actually work in PDX’s favor if returns are low, but it also can crush the product. PDX is leveraged in a way that AIUL18 and No Bonus IUL products are not. The sequence of returns matters. You’re not just betting on the average return, you’re betting on how that average return comes to pass. And that seems like a lot of bets to make.

That’s why, after all of this, I stand by the conclusion that I drew in the first post. The asset charge funded bonus design found in John Hancock Accumulation IUL18 and Lincoln WealthAccumulate IUL is not only simpler and easier to explain than the fixed charge funded bonus design in PDX, but it also holds up better across possible scenarios and delivers more consistent and intuitive returns. Where PDX can be choppy, AIUL18 is smooth – and that’s a real benefit. Which one will actually perform better over the long run? It’s hard to say. I think this analysis shows the relative strengths and weaknesses of both designs for the purposes of analyzing performance. But for me, simpler is always better. Smoother is always better. Intuitive is always better. And on those scores, this analysis shows why asset charge funded designs are probably a better fit for most clients.

Before I close this article, I want to pause for a second and reflect on why I’m doing this analysis. Admittedly, this is a lot of information. It’s complex. There are a lot of angles. The statistics here are not the usual ones we use for life insurance – they’re almost like another language. I promise you, my intention is not to write in technical terms for the sake of being technical. I don’t particularly enjoy writing highly technical articles. But I’m doing this analysis because it’s honestly what I believe is required to understand and communicate the value proposition of these policies. As an industry, we are increasingly stepping into the world of investments and asset management – which means that we’re going to have to start employing the language and tools of that trade for our products. If you’re not comfortable with that, then I would seriously encourage you to stick your knitting of selling life insurance for life insurance. I’ve not made a single mention of the death benefit of these policies because, frankly, it’s not relevant. These products are being marketed exclusively for their risk and return characteristics, like any other investment. The fact that life insurance comes along for the ride is an afterthought. We can debate all day long about whether or not this is a good thing for our industry (it’s not), but the fact remains – if we’re going to put on a new cape and mask, we’d better find some new superpowers.

In the final post, we’ll take a look at how to distill the insights in this post into simple rules of thumb for illustrating these products and, by extension, all IUL products. I’ll also dig into how I think these products will compare to each other over the long run in terms of their renewal caps and bonuses, which will be extremely important in setting client expectations and servicing these policies.