While I have already quoted the link upthread, I recommend Pfau's approach (https://retirementresearcher.com/income-annuity-101/) as reasonably non-mathematical and, for those sufficiently interested, fairly easy to reproduce in spreadsheet form. Since a version of this material appears in his book, 'Safety First retirement planning' I guess it sort of counts as an actuarial textbook!Thanks again to all who have posted, it has been very educational, and I am glad I did this prequel. The more so, since I decided to draft the paper first, before launching the “Should You Hedge Your SPIA with a TIPS Ladder?” thread. Which will be forthcoming.
*Technically, the insurance company looks at the entire mortality vector that averages out to life expectancy, rather than simply applying the PMT function with life expectancy as the period (as i do for purposes of getting a rough estimate). If you need page references to an actuarial textbook, please post to that effect.
Actuarily, it has to be a wash. And if interest rates go down while you wait, or if you can’t invest the premium at a high enough rate while you wait … or maybe, if you left your premium in BND through 2022, waiting for 2023 to annuitize—you lose the wager.
Milevsky and Young's seminal paper ("Annuitization and asset allocation", 2007) looked at delaying the purchase of the annuity using a stock/bond portfolio to support payments during the delay period and found the optimum age for annuitization was about 80 (females) and 85 (males). These results were well reported at the time and I've seen them quoted on occasion on this board. However, I note that these are for retirees with low relative risk aversion (i.e., willing to hold a high proportion of stocks) and that the optimum age decreases for those with high risk aversion (for those most risk averse, immediate annuitisation is optimum). I also note that there was a probability of between 35% (females) and 39% (males) that a purchase of an annuity at the optimum age would provide less income than one purchased immediately at 65. Work that I did with historical data (https://papers.ssrn.com/sol3/papers.cfm ... id=4289339) suggests that the historical retirements where this was the case were amongst the worst ones (in other words, delaying the annuity purchase often made a bad retirement worse).
While self-delay (i.e., using some portion of the remaining amount in the portfolio to annuitize at some time after retirement) is one approach, another is to use deferred income annuities. In order to avoid derailing this thread, I've posted the results of some modelling of DIAs at viewtopic.php?p=7897923#p7897923 that might be of interest.
cheers
StillGoing
Statistics: Posted by StillGoing — Wed Jun 05, 2024 4:00 am — Replies 212 — Views 16966