Here's What Makes an Annuity 'Optimal' for Clients

This is not necessarily a surprise, since only an especially healthy individual would be expected to purchase a DIA, given the structure of income payments. However, these differences have important implications for pricing. For example, since individuals who purchase DIAs are notably healthier than those who purchase VAs with a living benefit, the payouts for DIAs (based on the pure mortality component) would be lower, ceteris paribus.

Ignoring some of the structural differences in annuities (e.g., around mortality rates) can result in an incomplete perspective on the relative efficiency of different annuities. To demonstrate this effect, I analyzed what happens when some of the key pricing assumptions for an immediate annuity are adjusted, in particular the assumed expense load and mortality rates.

More formally, I determined the payout rate (income amount dividend by initial premium) for an immediate annuity for a 65-year-old male, assuming Gompertz factors fit to mortality rates from the Society of Actuaries 2012 Individual Annuity Mortality table with improvement to 2022 and a 5.5% discount rate. (This is consistent with the yield on 10-Year High Quality Market Corporate Bonds as of Dec. 5, 2022, based on data from the St Louis Federal Reserve.)

Assuming a 10% expense load, the payout rate would be 7.15%, which is consistent with the median quote available from CANNEX.

The tables below include how the adjustments affect the payout rate for either a life-only immediate annuity or for an immediate annuity with a cash refund provision. Technically, for the model, I adjusted the modal coefficient in the Gompertz factors to capture the impact of a life expectancy change.

Adjusting mortality rates or expense load assumptions can significantly affect expected income levels for an immediate annuity (for better or worse).

For example, if we focus on life-only annuities (Panel A), attracting a cohort of annuitants that have a life expectancy that is two years shorter, on average (i.e., trending more like the general population, which has a life expectancy that is about five years shorter than the average annuitant) with a 5% assumed expense load (versus a base 10% assumed load) would result in an income level that is 10% higher than the base scenario.

Additionally, while including a cash refund provision to an immediate (life-only) annuity would generally be expected to reduce the benefit level (e.g., by 4.5% for the base scenario), the actual impact is going to depend on the expected attributes of the annuitants.

To the extent cash refund annuities can be issued at a lower cost (i.e., lower expense load) and attract a pool of less healthy annuitants, it would technically be possible to have higher payout rates for cash refund annuities versus life only annuities, something that has been observed in the past based on the quotes available from CANNEX.

Consider the Nuances

While preferences can change over time and be influenced through choice architecture (e.g., defaults in a defined contribution plan) it is naïve to assume the underlying mortality and expense components for different types of annuities are identical.

In reality, truly optimal strategies are going to be those that blend behavioral and economic considerations. The best strategies may require some type of “compromise” when viewed entirely from an economic lens, but ignoring market realities is not going to result in the best outcomes for retirees.