#233 | Lincoln VULone and Prudential VUL Protector Review – Part 4
In some ways, the question of whether or not it’s reasonable to rely on separate account performance to bail out the guarantees in Guaranteed VUL is as simple as the head of product at Lincoln put it to me back in 2013 when describing how they justified their pricing on VULone – “well if equities do 7%, then the guarantee will be worthless and we won’t need to worry about it.” We’ve already shown that to be true in a previous article. If equities just clip along at 7%, then there’s no reason to have a guarantee in the product. You’d be better off buying a slightly cheaper version of a DB VUL without a secondary guarantee.
But equities don’t just do 7%. Equity returns are volatile, obviously. I’ve written in previous articles that volatility itself isn’t a problem in terms of lapsing overfunded VUL products which have sufficient account value to ride out the volatility. Not so for thinly funded death benefit VUL products like these. Here, the risk of account value lapse is real and tangible. When account value lapses, then the put option embedded in the Guaranteed VUL is in the money and chances are good that the life insurer is on the wrong end of the trade.
To get a sense of how volatility impacts these products, I ran them through my Dynamic Illustration Tool using a thousand stochastic equity return scenarios. This produces a mountain of data, but the best way to thin-slice it is to look at Death Benefit IRR at age 95. Because these are guaranteed products, they have a minimum DB IRR related to how the guarnatee is priced. If the scenario results in a DB IRR at the guaranteed minimum, then you know that the account value is less than the death benefit and has most likely gone to zero. If it’s above the guaranteed minimum, then you know that the policy hit corridor because cash value performance is pushing up the death benefit.
To provide a point of comparison, I also track the average credited interest while the policy remains in force. What matters is not the average equity return over 75 years – what matters is the average equity return prior to policy lapse. This allows us to get a feel for the drag from owning a Guaranteed VUL versus equities outright or, in other words, the true “cost” of the guaranteed death benefit put option supplied by the life insurer.
The key remaining question is what equity return assumption is reasonable for this type of analysis. As you can imagine, answering this question is a huge deal and there’s no “right” answer. Every company has different views of future equity performance that inform its willingness to take equity risk. Let’s start with a strict baseline historical model, which uses the historical S&P 500 Total Return distribution and randomly resequences the returns to create future strings of potential returns. The average annual S&P 500 Total Return since 1950 is around 12%, much higher than needed to make the guarantee in these products expire out of the money.
Here’s what the graph looks like under that scenario for the 45 year old Preferred Male for $1M DB using the guaranteed single premiums for each product. Each dot represents the result from an individual scenario within the 1,000 equity return scenarios modeled in the DIT. Lincoln is red, Prudential is blue. The gray line is the equity return for the scenario.
The first thing you might notice is that, as expected, the cash value growth of Prudential VUL Protector outperforms LincolnVULone across the board. VUL Protector really does deliver better cash value performance in level rate scenarios and in the real world. All of the observations falling neatly into lines to the far right hand side of the chart are scenarios where the guaranteed death benefit expires in the money and the DB IRR is equal to the guaranteed DB IRR. Under this baseline equity return assumption, the death benefit guarantee only has a 7.3% chance of being in the money on VULone and a 4.8% chance for VUL Protector. In other words, this is a solid bet for the life insurer. But it’s worth noting that under this assumption, VULone is the more profitable of the two products because it has a higher charge drag but the death benefit guarantee still expires out of the money. You can see the effect of the higher charges in the fact that VULone’s death benefit IRR lags VUL Protector’s by, on average, about 0.7%.
What happens if we switch to a different equity return assumption? Consider a future where equity returns aren’t quite so strong but are less volatile. Modeling a future like that requires a proportional change to all returns in the historical S&P 500 returns so that the best returns aren’t quite as good and the worst returns aren’t quite as bad. I’ve calibrated the adjustment so that the average of the return distribution is 7% rather than the actual historical average returns of nearly 12%. Here’s what that looks like:
In this assumption, the percentage of scenarios where the account value lapses and the death benefit guarantee expires in-the-money jumps to 43% for VULone and 31% for VUL Protector, a massive difference from the baseline historical scenario. In this assumption, again, VULone is the more profitable of the two because there are relatively few scenarios (11.7%) where the account value in VULone hits zero and VUL Protector hits corridor. But this assumption is actually pretty aggressive. A much more likely assumption is that volatility stays in-line with historical norms and that overall returns decrease, which is what basically all of the major asset managers are projecting. In this model, we’re assuming 7% average returns by subtracting a set value from all of the S&P 500 historical observations, which makes good years not quite as good and bad years a little bit worse. Here’s what that looks like.
Under this assumption, VULone has an account value lapse 70% of the time and VUL Protector isn’t far behind at 60%. Neither product would be even remotely profitable. Here you’re starting to see another effect come to light – sequence of return risk. The average long-term equity return assumption is still 7%, but just a quick look at the gray line graph will show that only about 25% of the scenarios actually have greater than 7% long-term returns. Instead, the average return looks to be about 5%.
Why is that? Because volatility is forcing account value lapses before the average return can recover. As I’ve written in other articles, this type of sequence of return volatility really isn’t a threat to overfunded VUL products because there’s enough account value to ensure that the product will survive and get the benefit of higher returns in the future. That’s not the case in these products, where the guaranteed premiums are very low and therefore consistent high returns are required to keep the account value in force. Again, to put it in option terms, high equity volatility is a risk unto itself. The higher the volatility, the more valuable the put option from the guaranteed death benefit. Based on this analysis, it’s pretty clear that the put option is extremely valuable due to volatility – even if equities do actually average 7%.
And I’m actually fairly certain that this is still an aggressive assumption. In Morningstar’s annual round-up of long-term equity return assumptions, the likely average is more like 5%, which actually is in accordance with the historical equity risk premium over risk-free assets. Here’s what that looks like:
Yikes. Now we’re looking at 89% and 83% account value lapses for VULone and VUL Protector, respectively, with both products likely producing nasty losses for the issuing companies. It’s not a pretty picture for the insurers, but it’s a fantastic bargain for consumers. The guaranteed death benefit put option that cost them very little in the way of slightly higher policy charges than a non-guaranteed VUL is in the money 80-90% of the time. Volatility around a 5% long-term return assumption is even more problematic than at 7% because there’s less of a chance that the product will revert to a high enough average to bail it out. Something like 20% of the scenarios don’t break 0% average return, which means that the product couldn’t survive even a short bad run of returns.
So far, we’ve only been looking at a 45 year old, which would be on the young end of a VULone or VUL Protector buyer despite the fact that this product is a screaming deal for someone that age. What happens if we look at, say, a 65 year old using the same single pay design? You get nearly identical results to the 45 year old. The observations above are applicable across the age range. They’re also broadly applicable to other funding patterns, although level premiums show a bit less lapse behavior than others because the rates are relatively high and level premiums smooth out some of the volatility of the account value.
But the rubber hits the road when you consider that under Principles Based Reserving (VM-20), stochastic analysis isn’t just an academic exercise – it’s the way the reserving works. Reserves are designed to reflect the expected flows arising from the liability, whereas capital reflects the unexpected flows. What’s a reasonable expectation for the flows of a liability with a long-term guarantee? Imagine that the life insurer runs 1,000 stochastic return simulations just like what I did for the previous analysis. Each one of these stochastic returns has a certain associated cash flow of gains and losses. Let’s say we take the 300 worst scenarios for the life insurer and average all of the results into a single number, what actuaries call Conditional Tail Expectation 70 (CTE 70). This level is what sets the reserve for the product. The worse the CTE 70 looks, the more reserves the carrier has to hold and the lower the profitability of the product.
Those familiar with Variable Annuity reserving will also note that CTE is the basis for both reserves and capital and the metric that life insurers often use to gauge their level of capitalization. PBR in life insurance essentially aligns Guaranteed VUL with what Variable Annuities have already been doing for nearly a decade. For any product with long-term guarantees, stochastic analysis is the only credible way to look at the true economics. It’s what can tell you what the guaranteed death benefit put option is really worth both to life insurers and consumers – and that’s why I’ve chosen to use it as the final piece of the analysis on Guaranteed VUL. It’s the key that unlocks the door.
The result is clear. Guaranteed VUL is an immense benefit to consumers and an immense problem for life insurers. We know from watching the Variable Annuity business that writing equity-based products with rich, interest-rate dependent guarantees isn’t sustainable. The same will go for Guaranteed VUL and for the same reasons. Companies watching the arms race between Lincoln and Prudential have been left scratching their heads, even if they too offer a Guaranteed VUL product. The smash success of the Guaranteed VUL category and Prudential’s rising prominence, in particular, isn’t a harbinger of the future. It’s the Battle of the Bulge – a final and fevered push when the outcome has already been determined.
Right now, there’s no such thing as a bad Guaranteed VUL. But there is such a thing as the best Guaranteed VUL and, undoubtedly, that’s Prudential VUL Protector. Sell the daylights out of it while you still can.
Prudential announced that as of 7/20, it will be reviewing any transaction with more than $1M of premium in any year in order to “ensure we are placing cases that meet our desired profitability and risk tolerance levels.” The previous review threshold was $5M. This might sound like an innocuous change because $1M is still pretty steep, but it’s not. A significant portion of flows into VUL Protector are single premium or 1035 exchanges, so a $1M threshold comes into play for quite a few cases. That’s especially true now that Lincoln has repriced its product and, invariably, Prudential is picking up a lot of what would have otherwise gone to Lincoln. Prudential is suddenly
But more importantly, why would these cases be cause for concern? What’s the profitability difference between a $100,000 single premium to guarantee for life and a $1M single premium to guarantee for life? On a product like this, it really shouldn’t matter unless your profitability was thin to begin with, as this series has signaled. My guess is that this is the first step in what will likely be other moves for Prudential to stem the flow into this product before they do a wholesale reprice. Pru might have been comfortable with a second place position behind Lincoln and being less concentrated on single pays, but now that Lincoln has stepped out, it’s all going to go to Prudential. That probably wasn’t part of the plan.