#244 | AG 49-A & John Hancock Accumulation IUL 20
Slowly but surely, the AG 49-A landscape is beginning to show itself and a few things are becoming very clear. The illustration restrictions introduced by AG 49-A certainly have teeth. Multipliers and buy-up caps have reduced performance relative to AG 49 although, as we’ll see later, not nearly to the degree that was originally contemplated by the guideline. Equally as impactful are the new restrictions on illustrated income. AG 49-A reduces illustrated arbitrage from 1% to 0.5%, which in itself accounts for something like a 5-7% decrease in illustrated income, but it also specifies that illustrated loan credits cannot include any excess interest from multipliers or any other bonus.
Taken together, these restrictions will push down illustrated income for the most aggressive products in the market by 40-60%, depending on the product and the cell. Premium financing structures that rely on refinancing the bank loan with a participating policy loan will face similar headwinds in illustrated performance, potentially to the point where some programs just “won’t work” – not that they ever did in the first place, but that’s a different conversation. Every premium financing vendor will have to reevaluate their programs after AG 49-A is implemented.
Certain product changes also look like they may become prevalent as well. Take, for example, the issue of multipliers that appear later in the life of the contract. These features are very difficult to show under AG 49-A and, consequently, life insurers appear to be getting rid of them in favor of other designs. Prudential, for example, removed the embedded 20% multiplier from years 11-40 from Index Advantage UL 2018 in its new, AG 49-A compliant Index Advantage UL 2020. Curiously enough, Prudential doesn’t appear to have made an adjustment to the cap (still 8.5%) or other policy charges, except to add a new tier of asset-based charge (0.55%) if the policy is heavily overfunded. So where’d the savings from removing the multiplier go?
The only material change is that Target premiums have increased 10% across the board. In this particular case, Prudential took a consumer benefit under AG 49 and turned it into an agent benefit under AG 49-A. Strange choice, isn’t it? I know for a fact that other companies are planning on removing embedded multipliers and either increasing caps or converting the multiplier to a fixed interest bonus. That seems like the better call, but appealing to agents’ baser instincts has certainly worked for some companies in the past. But that’s the thing about AG 49-A – every feature has the potential to be handled differently by every company, despite the fact that regulation is theoretically applied equally to all companies.
This leads us to John Hancock and their new, AG 49-A compliant Accumulation IUL 20, which has everything in the same way that Bill Hader’s legendary Saturday Night Live character Stefon describes New York City nightclubs (clip here, not for the faint of heart) – multipliers, buy-up caps, non-S&P 500 indices and a dynamic persistency bonus. The fact that Accumulation IUL 20 (AIUL20) is stuffed to the gills with features makes it the perfect case study on how to apply AG 49-A to a variety of features and the strange things that arise from doing so. Be forewarned – this post is also not for the faint of heart. It’s complicated. It’s cumbersome. It might not make any sense. You’re going to have to read it a couple of times. Heck, I even had to rewrite it a couple of times and I’m still not sure I did it justice. Welcome to the strange and bizarre world of AG 49-A. You’d better get used to it.
At first blush, there’s only one substantive change between AIUL19 and AIUL20. AIUL19 offered a Base account with a 10.5% cap and a 0% floor that also doubled as one of several Benchmark Index Accounts (BIAs, which set the maximum illustrated rate for the non-BIA accounts in the product). Under AG 49-A, products can only have a single BIA for the purposes of setting the maximum illustrated rate that serves as a limiter for all available accounts. There is no BIA available to clients in AIUL20. Instead, John Hancock offers a Select account, which has a 10% cap and a 5% multiplier. The net result is that the Select account illustrates a slightly higher total rate (6.47%) than the Base account (6.39%). Hmm. Why would they do that? It’s not because of a structural advantage for the Select account. Hancock could have just increased the cap in the Base account to equalize the option costs between the two accounts and the illustrated rate for the Base would have likely been slightly better. So what’s the rationale for offering the Select account instead of the Base?
Blame AG 49-A. One of the substantive changes introduced by the new guideline is that products now have a single Benchmark Index Account that governs the maximum illustrated rate for all accounts in the product. Therefore, the BIA is itself now a part of competitive positioning for the product because maximizing the BIA rate provides aircover for all of the non-BIA accounts in the product. The higher the BIA rate, the greater the opportunity to find other means of maximizing illustrated performance up to the BIA rate at a lower option price. Hmm. Now, how might a carrier do that?
The first step, of course, is maximizing the hypothetical BIA rate so that it provides for plenty of wiggle room. That’s actuarially pretty easy to pull off. AG 49-A allows for aggregation of all of the indexed accounts for illustration actuary purposes, which means that the BIA rate can be set so high than the account barely passes (or even fails) without causing any illustration testing problems for the whole product. That’s presumably how the hypothetical BIA in Accumulation IUL 20 uses a 12% cap and has a towering 7.05% illustrated rate, but the Select account offers only a 10% cap with a 5% multiplier. Those two things are not the same. The hypothetical BIA cap is much richer – and that’s on purpose. If Hancock offered the hypothetical BIA for client allocations, the rate would likely be lower and close to the 10.5% available in the Base account for the old product with a commensurately lower illustrated rate.
The second step is getting more juice out of the option budget so that the maximum illustrated rate for the non-BIA accounts illustrates similarly to the hypothetical BIA but at a lower option cost to preserve profitability. Again, this is pretty simple and there are two primary ways of doing it. First, a company can offer a proprietary index with a significantly higher hypothetical lookback rate than the S&P 500 strategies but at the same option cost. I’ve written extensively about these indices, so I’m not going to refresh that topic now, but suffice a long story to say that these indices are engineered to deliver maximum lookback performance at low and stable option prices at the cost of complexity and questionable real-world applicability.
Many companies will go the route of a proprietary index because it’s easy, but John Hancock chose an alternative route – charge-funded multipliers. Yes, you read that correctly. One of the common misconceptions about AG 49-A is that it “outlaws” multiplier accounts. That’s not true. It was designed to restrict the illustrated benefits of multiplier accounts but, as I’ve written before, the final result was a convoluted mess with a host of unintended consequences. One of those unintended consequences was that it is possible – in fact, it’s inevitable – that multiplier accounts will still illustrate better than non-multiplier accounts as long as there’s a high enough BIA illustrated rate to accommodate them as there is in AIUL20. Take a look at a comparison of illustrated performance for the Base/Select, Core (2% charge) and Enhanced (5% charge) accounts for the old and new products.
|John Hancock||Base Capped / BIA||Select Capped||Core Capped||Enhanced Capped|
|AIUL19 Total Rate||6.39%||6.81%||7.87%|
|AIUL20 Total Rate||7.05%||6.47%||6.80%||6.80%|
You’ll notice a couple of key changes. First, the caps for the Select, Core and Enhanced were different in AIUL19 but are all the same in AIUL20 – they’re all 10%. Given that the asset-based charges for all 3 accounts are the same, the multipliers had to change. We can determine the effective illustrated rates for the products by multiplying the maximum illustrated rate by the multiplier and subtracting the asset charge. That produces the net rate that actually drives the illustrated performance of the product. Using that formula, the illustrated rate for the Core Capped account is the same for the old and new product and, equally as importantly, the net illustrated rate for the Core Capped account is significantly higher than in the Select account. AG 49-A was supposed to clamp down on the illustrated benefits of multipliers. But, instead, this product illustrates its Core Capped multiplier account just as well as in the old product and still better than the baseline, no-multiplier account option. Did anything even change?
Yes, but it appears that the change is only evident for accounts with larger charges and multipliers, which we can see in the Enhanced Capped account option. The net illustrated rate is the same for both the Core Capped and Enhanced Capped account. They’re both 6.8%. The product also offers Core High Capped and Enhanced High Capped accounts with higher caps and lower multipliers and the net rates for both of those accounts are also 6.8%. It appears that the actual illustrated performance limit for the product isn’t 7.05% – it’s 6.8%. That’s appears to be the true limit under AG 49-A for this product and it only comes into play when the illustrated rate goes high enough courtesy of the multiplier interest. Why 6.8%?
The answer, presumably, is the new restriction in AG 49-A that clearly states that the illustrated rate for the product cannot exceed 1.45 times the Net Investment Earned Rate or Hedge Budget. The 1.45 factor was part of the original AG 49 but only as a backstage limit for illustration actuary testing. But for AG 49-A, it moves to the front-stage as a limitation on the illustrated rate itself. If that’s the case, then a maximum illustrated rate of 6.8% implies a Hedge Budget of 4.7% (4.7% * 1.45 = 6.8%). This shows that the real guardrail for illustrations in AG 49-A is not the hypothetical historical lookback rate for the BIA but, instead, the 1.45 factor.
And that’s a real problem. The 1.45 factor is applied to either the Net Investment Earned Rate (NIER) or the Hedge Budget, neither of which is disclosed or declared anywhere in the illustration or the policy itself. It is a hidden, internal metric that is unique to each life insurer. It bears zero relationship to the expected performance of the policy. The 1.45 factor was a negotiated settlement to accommodate illustration actuary testing under the original AG 49. It was not ever intended to visibly impact policy performance except in extreme scenarios. And yet, under AG 49-A, it appears to be the primary restriction on illustrated performance.
How did Hancock ensure that none of the account exceeded the undisclosed limitation of 6.8%? By adjusting the illustrated rates for the accounts. Take a look again at the top-line illustrated rate for the Select, Core Capped and Enhanced Capped Accounts, which are 6.16%, 6.05% and 5.72%, respectively, despite the fact that all of the accounts have the same 10% cap. This is a bizarre effect of AG 49-A. Here we have 3 accounts with the same cap that have 3 different illustrated rates, which is exactly what the original AG 49 was trying to avoid. For the first time ever, we have non-disclosed, non-declared, internally-set rates that directly impact the illustrated rate. It is a bizarre outcome of AG 49-A that was surely not intentional. Surely. But I have no doubt that John Hancock followed the rule to the letter and that’s exactly where it led them.
Unbelievably enough, not even the 1.45 factor can fully explain what’s actually going on. We can get a feeling for the actual cost of providing the hedges by looking at real, current-market data to see if they line up with the 4.7% imputed Hedge Budget that produces a 6.8% illustrated rate. The formula for determining Net Hedge Cost is the current hedge price for the cap times the multiplier minus the fee.
|Account||Cap||Multiplier||Fee||Net Hedge Cost|
|Core High Capped||12.25%||130%||1.98%||4.91%|
|Enhanced High Capped||12.75%||180%||4.98%||4.73%|
This makes no sense. The Select account rings in at almost 5%, the two Core accounts at 4.9% and the two Enhanced accounts at about 4.75%. The figures for the Select and Core Capped accounts are meaningfully higher than the implied Hedge Budget of 4.7% that feeds the effective maximum illustrated rate of 6.8%. What gives? I’m not sure. Based on a Hedge Budget of 4.9%, the maximum illustrated rate for the Core Capped accounts should be 7.11%, which would have meant that the BIA limit of 7.05% would be in play. But it’s not. Why it doesn’t is irrelevant. What this highlights is that, under AG 49-A, you will literally never know why a product illustrates the way it does. You just won’t be able to tell. The problem isn’t John Hancock or John Hancock’s interpretation of AG 49-A. I’m sure they’re perfectly compliant. The problem is AG 49-A.
I would argue that AG 49 ushered an era of remarkable transparency and consistency across illustrations, both of which were stated goals of the guideline. All products with the same caps illustrated at the same maximum rate. Products with multipliers illustrated consistently with products without multipliers. AG 49-A, however, sacrifices consistency in order to curtail what was perceived as abusive illustrations – despite the fact that those illustrations were actually in perfect accord with both the letter and spirit of the original regulation. That’s what made creating AG 49-A so difficult. How do you separate something that is part and parcel to the whole? You end up creating a complicated guideline that gives up on the original goals which, in this case, were consistency and transparency. Under AG 49-A, it appears that products and accounts with the same cap will have different illustrated rates. It appears that otherwise identical features will have different illustrated performance. It appears that products will illustrate differently than their mechanics would indicate. How is this the optimal solution?
It’s not. I give AG 49-A a one-year shelf life. Other companies are going to follow the John Hancock playbook because it is consistent with the letter and arguably the spirit of AG 49-A. Some will use multipliers, but most will use proprietary indices to achieve the same results. Every company will figure out its own way to maximize their illustrated performance under the 1.45 factor, which is now the undisputed king of the guideline. There will be an arms race again, of course, and it will end the way all of the arms races in IUL have ended – with a new regulation. Only this time, the industry will get the regulation it deserves, not the one it wants.
The story on Accumulation IUL 20 needs one more piece to make it complete. When you run an AIUL19 illustration against an AIUL20 illustration, you’ll notice that the new product handedly outperforms the old one. Take a look at the graph below.
What gives? Take a look at the Base/Select account comparison. The illustrated rate difference between Base and Select is just 0.1%, but the illustrated CSV IRR difference is 0.4%. The net illustrated rate difference between the AIUL19 Core account and the AIUL20 Core account is effectively zero, but the new product allocated to the Core account has 0.3% better performance. Even with the new AG 49-A limitations on illustrated loan arbitrage, the Core Capped account in AIUL20 performs similarly to the AIUL19 Core Capped account. How is all of this possible?
The answer is that John Hancock trimmed about 25% of fixed charge fat out of the product – and make no mistake about it, those charges were fat. I wrote in my review of the outgoing AIUL19 that it was the most expensive product in the industry and implied, although I might not have said it directly, that AIUL19 must have been unbelievably profitable. The charges in AIUL20 are more in-line with the rest of the industry. That’s a good thing for the client, at least in the short-term. In the long-term, it’s hard to imagine that future management will not see it fit to adjust rates to ensure that similar profitability is reached. It’s also hard to imagine that maintaining rates that cost in the 4.9% range are going to be sustainable for any extended period of time in an environment where most new money yields are closer to 3%. But for the time being, AIUL20 is certainly a more efficient product than AIUL19. You can’t find fault in that. As always, John Hancock knows how to bend with the wind – and AIUL20 is no exception.