#351 | Rising Rates & Policy Loans

blur chains chrome close up

Quick Take

One of the most obvious and profound impact of rising interest rates has been on variable policy loan interest rates that are pegged to the Moody’s Composite both in participating loans in Indexed UL and in non-direct recognition loans in Whole Life. The problem with both of those structures is that loans create their own set of incentives and tradeoffs in rising and falling rate environments. But Indexed UL has its own issue – the vast majority of Indexed UL illustrations show the policy being loaned to the maximum limit in order to create tax-free retirement income. If those illustrations are indications of reality, then future Indexed UL blocks will be backed almost entirely by policy loans. What kind of loan rate – whether variable, guaranteed or declared – may prove to be a major determinant of the long-term performance of the block and its policyholders. One solution is to carve off those loans into separate indexed loan accounts, each with its own declared rates. Sound familiar? It should – that’s what we call direct recognition loans in Whole Life.

Full Article

When I covered MassMutual’s newly released 10 Pay product in 2021, I made a point to highlight what I thought were the two most significant things about the new product. First, it was the first Whole Life product released that was priced using a 2% guaranteed rate, which meant that it had significantly higher premiums and illustrated performance than its predecessor and, at the time, higher than anything else on the market. That was a story unto itself, especially considering subsequent moves by Guardian and New York Life to match the 2% 10 Pay design.

But the second part of the story was equally as important. MassMutual has long set its minimum variable policy loan interest rates at 1% higher than the policy guaranteed interest rate. For 4% products, that meant a guaranteed minimum loan rate of 5%, which is what MassMutual had been illustrating and charging for a very long time thanks to a decade of ultra-low interest rates. But applying that same methodology to the new 2% 10 Pay product dropped minimum loan interest rate dropped to just 3%. Thanks to the ultra-low interest rate environment of 2021, the launch illustrated loan rate for the product was just 3.25%. To put it a different way, out of the box the illustrated loan arbitrage for the product was nearly 3% – far higher than anything in Whole Life or even in an Indexed UL product since AG 49. Here’s what I wrote about it at the time:

“Producers should increase the illustrated loan rate to 4.75% or 5.25% (options in the software) or run Fixed loans as the baseline. Either of these options will provide a more realistic and sustainable perspective on long-term performance. From what I understand, even MassMutual’s own rollout materials for this product don’t push the illustrated loan arbitrage to the limit and advocate for more conservative illustrations.”

Nonetheless, a lot of advisors pitched the 10 Pay with maximum illustrated income at the lowest possible loan rate. The timing couldn’t have been worse. As interest rates spiked in 2022, so did the adjustable loan rates for all life insurance policies that are directly pegged to the Moody’s Composite index*. That includes MassMutual. For Mass’s historical block of business with a 5% minimum guaranteed variable loan rate, the impact hasn’t been so big. The Moody’s Composite is sitting at just above 5%, which means that there is hardly a discernable or meaningful difference between the as-sold illustration in 2020 and the in-force illustration in 2023. Given the amount of chaos in the financial world, those policies and illustrations have been remarkably stable.

That is not the case for the 10 Pay Whole Life. I pulled a 2021 illustration for the product on a 45 year old Preferred Male and overlayed illustrated distributions onto it as pure policy loans**. Here’s how the results changed at various illustrated loan interest rates. I also included a baseline illustration with no variable loans to show the classic leveraged effect on illustrated performance that occurs with any sort of financing, whether through policy loans or external premium financing. When the loan rate is lower than the policy cash value IRR, then taking income increases illustrated IRR. When the loan rate is higher, then taking income decreases illustrated IRR.

Policy Loan RateIllustrated IncomeIllustrated IRR @ Age 90

An advisor who sold a MassMutual 10 Pay product in 2021 without making an adjustment to the illustrated loan rate at point of sale would be looking at a roughly 22% reduction income solely due to the increased policy loan interest rate. There’s nothing wrong with the product itself. Dividends have remained strong. The miss is entirely related to the policy loan interest rate – and could have been entirely avoided by illustrating with an assumption that is more realistic and in-line with the current dividend interest rate***.

The situation with this MassMutual 10 Pay product is a very clear, easily identified example of the variability and leverage introduced to illustrated performance when using policy loans with an assumption of long-term perpetual arbitrage, as is found in Whole Life with non-direct recognition loans and all Indexed UL policies. When things change, the illustrated performance for clients changes a lot. These sorts of loans apply leverage to illustrated (and real world) performance.

However, it is far from the only example, although it may be the most dramatic because of the difference between today’s loan rates and the originally illustrated loan rates just 18 months ago. The same phenomenon is playing itself out to varying degrees across the industry at other Whole Life companies selling policies with non-direct recognition loans such as New York Life and Security Mutual Life of New York, but also for Indexed UL, where it may take some bizarre and counterintuitive twists.

In general, Indexed UL loan rates are structured in one of three ways – fixed, variable or declared. Fixed loans have a guaranteed cost for the life of the contract. The most prominent and vocal proponent of this structure is Allianz, which guarantees a 5% loan cost for the life of the policy. Variable loans have rates that are pegged to the Moody’s Composite index. Declared rate loans are loans with a cost that is periodically declared by a life insurer. And, of course, these loan structures are not mutually exclusive. Some companies (such as John Hancock) offer two of these options in a single contract, the same way that MassMutual offers both non-direct recognition and direct recognition loans in its Whole Life policies.

In a rising rate environment, each one of these structures is going to react differently. The simplest to decipher is the variable structure. There is very little ambiguity in loan rates if the rate is pegged to the Moody’s Composite. The only question is how often the carrier resets the rate and how long the rate applies for, which are the same considerations as you’d find in Whole Life. Most companies set rates monthly. Some apply them monthly and some reset them annually at the prevailing rate for the month.

Regardless of how the loan interest rate is adjusted to follow the Composite, rates have been long enough now that all accounts with a variable loan rate will have seen a dramatic increase in the policy loan rate. At first blush, this is a bad thing. Illustrated incomes will go down and the cost of any interest on outstanding policy loans will go up. All else being equal, it makes policy loans look less attractive than they used to, although they may still be much more attractive than borrowing money from a bank.

But the second-order outcome is equally as important – any outstanding policy loan is essentially earning “new money” rates for the life insurer, which means that a portfolio with a large portion of variable policy loans will have a more rapidly increasing yield than a portfolio without them. Who’s paying those higher yields? Policyholders. In today’s environment, policyholders with outstanding loans at variable rates are essentially paying higher interest into the pot that benefits everyone else in the block, including the folks who aren’t taking policy loans. It’s a subsidy of non-loaned policyholders paid for by loaned policyholders. However, the inverse is true in a falling interest rate environment. There, loaned policyholders are being propped up by non-loaned policyholders. Variable loans cut both ways.

On the complete opposite end of the spectrum is the guaranteed fixed loan rate. As with variable loans, there’s an implicit subsidy between loaned and unloaned policyholders because the portfolio yield is going to be different than the guaranteed fixed loan cost. At certain period of times, it may be advantageous to unloaned policyholders and at other times it may be advantageous to loaned policyholders. The only thing that is clear is that at any given moment in time, one of those groups is on the right side of the trade and one is on the wrong side of the trade.

Companies that offer this feature usually tout it as a benefit because the loan cost is known and fixed. I think it’s worth noting that these sorts of fixed loan rates are fundamentally different than standard loans in Universal Life or direct recognition loans in Whole Life. With those sorts of loans, the carrier is charging a certain rate and crediting a certain rate. What’s important is the difference between the two rates which represents the net cost of the loan. There is no leverage. With guaranteed fixed rate loans, the gross cost of the loan is known but the net cost is not because the index credit is variable. These loans are still very risky.

Subsidization between loaned and unloaned values is a problem, in my view, because it creates a bizarre and counterproductive incentive structure in different interest rate environments. In a falling rate environment, policyholders are incentivized to take variable loans because the loan rate is below the portfolio yield, which creates more policy loans at exactly the wrong time and creates a quicker decline in the overall portfolio yield. The inverse is true for fixed loans. There, policyholders are disincentivized to take policy loans because the cost of the loan is higher than market interest rates, which means that the unloaned policyholders aren’t getting the benefit of lots of policyholders taking policy loans at above-market rates.

In a rising rate environment, the opposite effect occurs. Policyholders with variable loans will be disincentivized to take policy loans because the cost of the loan will be in excess of the portfolio yield, which means that the unloaned policyholders won’t get the benefit of policy loans coming into the block at new money rates. Policyholders with fixed loans will be incentivized to take policy loans, but those loans are at sub-market rates, which will drag down on the overall portfolio and eventually anchor the yield to the guaranteed loan cost.

Neither situation is a good one. This is a classic example of how narratives and illustrations get in the way of real-world considerations. That’s why Northwestern Mutual, for example, is so adamant about using direct recognition loans that eliminate subsidization between loaned and unloaned values. That’s also why Universal Life policies have traditionally relied on wash loan structures. But with Indexed UL, life insurers are allowed to illustrate perpetual arbitrage forever regardless of what kind of loan structure they use, which means the optimal solution is to avoid subsidization while still retaining the ability to illustrate loan arbitrage. 

Enter the declared rate loan option in Indexed UL, which allows the life insurer to essentially set the cost of the loan at a level similar to their portfolio yield. With properly set rates, there won’t be a subsidy between loaned policyholders and unloaned policyholders. There also won’t be a “new money” effect of outstanding policy loans, which means the block will move more slowly in reaction to a changing interest rate environment, as you’d expect for any portfolio yield. In my view, it’s the optimal approach for policy loans in Indexed UL and one that goes largely unnoticed because it’s not quite as sexy as the other options.

These different loan structures matter. Based on conversations with a few folks at IUL carriers and my own experience in seeing proposed and sold transactions, I’d guess that something like 90% of all accumulation-focused Indexed UL illustrations show a future policy loan that exceeds 60% of the future policy values. In most cases, the outstanding loan balance is more like 80-90% of the policy value. Taken at face value, that means that by far the biggest asset backing these blocks of business more than 20 years from now will be policy loans.

Portfolio yield is destiny. If future portfolio yields will be 60-90% determined by policy loan yields in Indexed UL, then that is by far the single biggest factor in the future performance of the block as a whole. Choose your structure wisely. And, to me, the choice is pretty clear. Variable Loans – including in Whole Life – push the block closer to New Money while creating subsidization between policyholders. Guaranteed fixed loan rates will drag the performance of the block down in a high-rate environment or force policyholders taking loans to subsidize unloaned policyholders in a falling rate environment. The optimal choices are direct recognition loans in Whole Life, standard loans in Universal Life or, if you really want to have index participation on loaned values, a declared rate loan in Indexed UL.

However, there’s a third way. So far, this article assumes that life insurers selling Indexed UL set Caps, Participation Rates and/or Spreads consistently across loaned and unloaned values. I’m making that assumption because it’s what the vast majority of life insurers actually do. However, as I wrote a couple of weeks ago, I expect that more and more life insurers will begin to use separate accounts with their own declared rates for loaned balances. One reason, as I wrote, may be to game AG 49-B and produce higher illustrated performance.

But another reason might be just to eliminate subsidization and incentive effects of certain loan structures. Lincoln just filed separate loan accounts without loan-specific bonuses, which presumably means they’re trying to provide a way to set different rates for those accounts than the unloaned accounts. It makes sense. For variable loan structures, for example, a loaned account would essentially offer new money rates that would today be much higher than the base unloaned values. Or a company declaring loan interest rates may decide that it doesn’t want to hamstring loan balances to the rest of the portfolio and wants the flexibility to set different rates.

Taken together, I think it’s reasonable to assume that in the future, more Indexed UL policies will offer specific loan accounts for the simple reason that there appear to be some upsides (including potentially illustrated performance) and no downsides. Or to put it differently, Indexed UL companies will start doing what Whole Life has been doing for years with the much-reviled, often-misunderstood direct recognition loan structure. There is nothing new under the sun.

*The Moody’s Composite is an odd index in that the only place I can find it quoted and used publicly is in reference to insurance applications. The more common indices are the Moody’s Aaa and Moody’s Baa. Taking an average of those two indices will get you within a few basis points of the Moody’s Composite. If you want to find a quote for the Moody’s Composite, you can find it on the NAIC website here.

**The math is pretty simple for Non-Direct Recognition loans because the loans are simply collateralizing the policy values. The only moving part is the margin of net cash value that remains in the policy as income is being drawn down. For simplicity, I assumed that all of these scenarios had to remain in-force for life and that the maximum net cash value to gross cash value would be 20:1.

***Policy loans are booked as assets. If the entire block was comprised of policy loans, then the yield supporting the Dividend Interest Rate would be the policy loan rate plus any earnings from outside businesses. This is how Direct Recognition Loans work.