Gaming the System in Annuity Illustrations

What You Need to Know

Using historical returns in annuity illustrations for index-linked strategies is ripe for potential misrepresentation.
It is incredibly important for consumers and advisors to understand the expected index performance and chosen crediting strategy.
We need to move to illustration approaches that incorporate forward-looking expectations.

To say that using historical returns in annuity illustrations for index-linked strategies is ripe for potential misrepresentation would be an understatement. 

The index-linked annuity industry, which is currently primarily for fixed indexed annuities, continues to evolve, offering crediting strategies that are anything but vanilla. Recently, though, a strategy caught my eye on the S&P 500 index that seems too good to be true — which it probably is.

Long story short, it is incredibly important for consumers and advisors to understand the expected performance of a given index and chosen crediting strategy, not purely historical. The longer the industry relies on pure historical illustrations, the more advisors and consumers run the risk of choosing and allocating to strategies based on unrealistic assumptions. 

An Index by Any Other Name

The index-linked annuity illustration space largely relies on historical returns. As much as this approach may be easy to explain and understand, this can obviously be problematic when an index is created such that when backtested it shows great historical performance, with questionable potential future benefits. Research and data have repeatedly demonstrated that the outperformance of many index strategies decays significantly after the index goes live.

In a recent Morningstar report that documents this effect, one of the report’s authors is quoted noting, “A typical index’s backtested performance looks great, but it usually fails to live up to those historical expectations once it goes live.”

Relying on historical performance, without the appropriate context, can also result in unrealistic expectations, even if you consider the performance of relatively well-known indices whose primary goal was not to be placed inside the insurance wrapper and have long-standing track records. 

For example, it recently came to my attention that a company was offering a 80% participation rate for the S&P 500 Futures Excess Return Index, while a participation rate up to around 50% is more common for strategies on the S&P 500 index (neither is assumed to include a cap).

To the untrained eye, these are seemingly very close index cousins, since both are based on the S&P 500, a representative sample of 500 leading U.S. companies; however, while the S&P 500 index actually involves buying the underlying stocks, the Excess Return index measures the performance of the nearest maturing quarterly E-mini S&P 500 futures contract trading on the Chicago Mercantile Exchange.

Not only are the names similar, but the performance has been relatively similar. If you look at calendar-year performance over the past 10  years, the S&P 500 index has slightly underperformed the Excess Return index, at 9.9% and 10.2%, respectively.

Among other things, this can be explained with near zero interest rates and efficiencies coming from “rolling” futures contracts. Additional performance differences are included in the exhibit below.

Products using the Excess Return index appear to have the potential to offer higher caps on an index with higher returns. For example, if I apply an 80% participation rate to the historical returns of the Excess Return index over the past decade, the average annual geometric return over that 10-year period would be 10.8% versus 6.7%, assuming a 50% participation rate for the S&P 500 index. Is it reasonable to expect this approximately 4.1% annual performance difference will persist on a forward-looking basis? Nope.

The Murky Past

Less clear, focusing just on performance, are the potential factors that could drive the differences in the performance of S&P 500 index and Excess Return index. One such factor is the impact of bond yields. The average yield on 1-year government bonds from 2013 to 2022 was about 1%. Fast forward to the end of 2023, and yields on 1-year government bonds have been closer to 5%. Does the sudden jump in yields matter? Definitely.