#70 | AG49 – Part 4 – What AG49 Got Right
The big win in AG49 is that it eliminated the gaming that carriers were doing to get the highest illustrated rate possible. This is generally universally seen as a positive outcome, even by ACLI & Co. Allowing back-tested returns without any sort of restriction is an invitation for gaming the system. All you have to do is run a bunch of indexes, crediting strategies, durations and currently affordable participation (caps, participation rates, spreads) through a model to figure out which one kicks out the highest illustrated rate over the given lookback period. Remember, because IUL backtesting doesn’t have to show rules and only uses spot pricing, the way to find the highest illustrated rate is to find the biggest gap between current and historical spot pricing. The bigger the gap, the higher the illustrated rate.
So where might you find significant gaps in current vs historical pricing? As it turns out, there are plenty of places. For example, there are volatility regimes in indexes as the composition of the index changes. The classic example of this is the Hang Seng, which was mostly comprised of private enterprises pre-1996, but is now heavily concentrated in Chinese state-owned enterprises. Current spot pricing for options does not look like historical spot pricing because the volatility regime is different. Even if the volatility regime is still sort of the same, there are all sorts of weird effects across when volatility is particularly low, as it is today. For example, in a low volatility environment, the difference in price between a 6% cap and a 7% cap might be 0.5% and the difference between a 13% cap and a 14% cap might be 0.2%. This means that carriers can “stretch” their budget a lot further. But just a slight uptick in volatility would result in the price difference between a 6% cap and a 7% cap staying at about 0.5% (because the cap is well within the volatility range in either case) but the price difference between a 13% cap and a 14% cap might double to 0.4%. In other words, today’s low volatility really rewards high option budgets if you’re doing IUL-type backtesting – but that won’t be the case as volatility increases.
But the granddaddy of them all is interest rates. Look at a chart of equity returns over time, and you’ll see a similar pattern somewhat regardless of how long of a period you choose. To use an analogy, you could successfully walk across an equity river that is 4 feet deep on average. Look at a chart of the 10 year Treasury, though, and you get something totally different. There are clear regimes of rising and falling, high and low rates. If you were to walk across an interest rate river that is, on average, 2 feet deep then you would drown in the middle. This is hugely important when you consider stretching today’s spot pricing back over 50+ years for IUL-type backtesting for two reasons. First, interest rates inform the carrier’s option budget, so historical option budgets would be markedly different from today’s. But the second is far more important for this conversation – even if you hold budgets constant, interest rates are a key factor in option pricing, especially when volatility is low. To use another analogy, in option pricing, interest rates are the music and volatility is the static. When there’s a lot of static, all you can hear is static and the music doesn’t matter. But when there’s not a lot of static, the music comes through loud and clear. Historically, option prices are constantly oscillating between music and static. Today, when rates and volatility are low, option prices are exceedingly cheap because the low rate music is coming through loud and clear. That is, of course, not extendable to history. But if you’re doing IUL-type backtesting and you want to gin up a high return, a surefire way is to incorporate options that price well in today’s low vol, low rate conditions. Voila, you’ll have yourself a winner. And if you want to go even a step further, do what some companies have done and just incorporate a fixed income rate into the index – which an obvious and egregious way to game the method.
Of course, you can take this logic to ridiculous extremes. I’ll give you a really simple example that we actually used during the AG49 discussions as a way to illustrate the problem to regulators. Imagine that we have an IUL with a 5% earned rate and option budget. We could buy a 12% cap on the S&P 500 and have a 7.5% pre-AG49 illustrated rate, or we could buy 200% participation in the 10 year Treasury, which at the time was sitting at about 2.5% (so 200% participation in a 2.5% rate is 5%). Now, if we were to do a standard, pre-AG49 historical lookback over 25 years on that what a 200% participation in the 10 year Treasury would have yielded, that strategy would have produced something well north of 11% returns, the highest illustrated rate in the market. That, of course, makes absolutely not a shred of logical sense and has absolutely nothing to do with how the product will actually perform, but it would have been consistent with how insurers were gaming the system with weird index and crediting mechanisms that generated ridiculously high illustrated rates. It also highlights, again, the difference between rules-based backtesting and pricing-based backtesting. In a rules-based backtest, the result of would have been 5% because the rule of the index is that the earned rate of 5% purchases whatever multiple of the 10 year Treasury costs 5%. AG49 disallows any illustrated benefit from gaming the IUL lookback mechanism by capping all rates in the product at the Benchmark Index Account rate. AG49 also happens to disallow any indexes with fixed income rates from historical lookback testing, so I guess our little example hit home with the regulators.
AG49 also slowed down the scourge of leveraged loans in IUL. Prior to AG49, it was not uncommon to see an illustration with an 7.5% illustrated growth rate and with loans illustrated at a 5% cost. For people who are mercifully unfamiliar with this kind of chicanery, what I’m saying is that the illustration showed that the life insurer is paying the policyholder 3% every year for taking money out of their policy. Yes, I’m serious and no, that’s not a joke and yes, that’s a practice that some IUL companies were defending as reasonable. Why? Well, remember, if you believe that buying equity call spreads produces 50% annual profits (as demonstrated by the returns from IUL backtesting), then allowing policyholders to borrow at a 5% cost and reinvesting that “asset” into equity call spreads and then paying it back to the policyholder with the 50% profit is quite reasonable. It’s basically a standard-practice wash loan provision where the 50% profits are paid back to the policyholder courtesy of the 7.5% illustrated rate. Practically thought, what it meant was that I got to show regulators an illustration from a certain company headquartered in St. Paul, Minnesota, where the premium was $100,000 for 10 years but the income over the life of the policy was $120 million, courtesy of a roughly 4.5% positive illustrated loan spread. There were lots of jaws on the floor. Thankfully, AG49 requires that the spread between the illustrated rate and the loan rate not be greater than 1%. Why 1%? Well it was going to be zero until certain companies said that they would fight it tooth-and-nail and regulators compromised at 1%. I wish there was more to the story than that. In retrospect, it does seem like a rather cavalier way to determine something that will drive the buying decision of hundreds of thousands of consumers. Ah, I’m sorry, nevermind. I forgot that illustrations are only used to show how a product works and not for illustrating performance that might inform a buying decision.
At this point, you’re probably looking back to my first post in this series and wondering when I’m going to talk about the twist that’s changing the IUL market. So far, to use my apple analogy, I’ve been debating different flavors of apples. When am I going to get to the whole apples tasting like glass bit? Well, what’s interesting is that the biggest impact of AG49 is what it doesn’t say, not what it does say. In particular, there were two issues that everyone knew were left unresolved. First, should AG49 be applied retroactively to inforce policies, so that a carrier didn’t have the predicament of having the same policy prior to the implementation date illustrating at significantly higher rates than policies sold after the implementation date? Eventually, after a bit of handwringing but far less than I thought there would be, AG49 was applied retroactively. The second issue is the one that is now changing the face of the IUL market by making a juicy looking apple taste like glass. I’ll tackle that in a series of upcoming posts.