I found the article and it’s resulting to be interesting. That said, the authors do not mention any of the limitations of annuities. For instance, locking in an annuity at age 50 deferred to age 65 is likely to be unappealing to many given the irrevocable nature of annuities; most would prefer the flexibility of having capital that they can convert to income via a variety of methods later in life when they retire.This paper evaluates a new variant of the popular target date funds used in employer-based retirement savings plans. We call this new variant a “target retirement plan.” Instead of increasing the allocation to bond funds as retirement approaches, a target retirement fund gradually purchases deferred life annuities beginning at age 50. In the particular straw model target retirement fund examined in the paper, the defined contribution participant makes deferred life annuity purchases at ages 50, 52, 54, 56, 58, 60 and 62. We compare how a target retirement fund participant would fare compared with someone who stays with a traditional TDF until retirement and then buys an immediate life annuity. We examine 1,000 possible 30-year futures for stock returns, bond fund returns and Treasury interest rates. The main result from this paper is that buying a retirement annuity in advance (by accumulating deferred life annuities) is superior to sticking with a Target Date Fund until retirement and then buying an immediate annuity in most scenarios of future stock returns, interest rates and bond returns.
It’s also worth noting that many 401k plans do not even allow in-plan annuity purchases. As such, it wouldn’t be possible for an individual to purchase a deferred life annuity prior to age 59 1/2. That said, the target audience for this article is likely plan sponsors and investment companies building target funds, rather than individual investors.
Second, inflation is not explicitly mentioned, although I think it’s implicitly recognized in their results via the future stock returns, bond fund returns and Treasury interest rates that are used.
I anticipate that an individual could get similar results by beginning to build a non-rolling TIPS ladder 15-20 years before retirement. An individual could also use duration matched TIPS funds to accomplish this; the Dimensional Target Retirement Income funds provide an example glidepath that an individual could use. An individual could also convert these assets into annuities during retirement, if desired.
Last, it’s worth noting that the authors achieved better results using deferred annuities in their model (and using their assumptions), but the outperformance over traditional target date funds was not life-changing - it amounted to about 6-10 months of the individual’s salary before retirement. While potentially meaningful, I don’t think that this result would dramatically change one’s retirement plans.