70-some posts later, it's interesting how some of you think you'll end up poor... and a few of you think you'll end up penniless.
One advantage of annuitizing a portion of the portfolio is that you've hedged your longevity risk. Yeah, assuming the insurance company pays off. And assuming Social Security is still there. And assuming that a Greek butterfly flapping its wings doesn't cause hurricanes in Hawaii. And a bunch of other peripheral risk factors.
Another advantage of annuitizing a portion of the portfolio is that you can choose a more aggressive asset allocation with the rest of the portfolio. Before you had to hedge/diversify against loss of principal. Now if you happen to lose all of your remaining principal, well, at least next month you'll get an annuity check.
You still retain the flexibility of variable spending. I wish there was a better way to model variable spending in retirement, or at least a better way than ORP. Other than Bob Clyatt's 4%/95% rule, though, or Bud Hebeler's one-year negative-feedback loop, I don't know of any other method of deciding how much (less) to spend in a down year. But if a portion of the portfolio is annuitized to your bare-bones lifestyle then you can cut your remaining portfolio's spending to zero.
The reason I give a crap about this study is because of the nine million or so veterans of the U.S. military, approximately a million of them are drawing some amount of a military pension. Conventional wisdom among military retirees has been "You won't starve, but you can't live on your pension." Well, that conventional wisdom misses an important point: it's a COLA annuity that affords a bare-bones lifestyle. If you could save enough during the 20 years it took to earn it, then you can invest the savings more aggressively than someone who didn't have a COLA pension.
We've all agreed (as much as this board agrees on anything) that we should have enough of an ER portfolio to cover our spending gap between our expenses and our income. The percentage math works the same for a retiree with a spending gap of $40K/year and an ER portfolio of $1M as it does for a retiree with a spending gap of $250K/year and an ER portfolio of $6.25M.
The difference is what's below the spending gap. Almost every version of a retirement plan, no matter the SWR or the system, includes some form of annuity against longevity risk-- even if it's just Social Security. If your annuity happens to be too small for you to live on, then your spending gap could lead to failure. But if your annuity happens to be "just barely enough" for your lifestyle, then screwing up your spending gap won't doom you to Friskies.
If one ER with a spending gap of $40K has purchased a COLA annuity for $20K/year, then if he still has $1M left he can spend $60K/year. If an ER with a spending gap of $60K/year has no annuity, then he can still spend $60K/year but he has to figure out how to hedge longevity risk on his own. The first ER only needs a remaining portfolio of $1M to cover his $40K/year spending gap. The second ER still needs a portfolio of $1.5M to cover his $60K/year spending gap. The trick for the first ER is persuading an insurance company to sell him a $20K/year COLA annuity for $500K. Probably not achievable, and probably not financially trustworthy, not even for the Thrift Savings Plan.
But the solution to this multi-variable problem is somewhere among the parameters of a bare-bones budget, an annuity to cover that budget, an ER portfolio (with a suitable asset allocation) to cover the rest, and a plan to reduce spending if the market has a down year. Consider it the design criteria for FIRECalc v2013-- simulating not only variation in market returns but variation in portfolio spending.
A portfolio with an annuity will always survive a portfolio without one. But how much more aggressive does the non-annuity portion of the partially-annuitized portfolio have to be to outperform the portfolio without an annuity? Another design criteria for a new retirement calculator.
Perhaps the trick is to design two budgets, one for a bare-bones survival and the other for an ER standard of living that you're willing to keep working for. Then the intermediate step is being able to annuitize the bare-bones survival budget while you continue to accumulate a big-enough portfolio for the full ER. Once you're ER'd, the annuity hedges your longevity and frees you to have an 80/20 equity/bonds asset allocation. Or even a 92/8 equity/cash asset allocation.
The point is that you annuitize some of your portfolio to afford some minimum standard of living. If you end up poor-- or even penniless-- then it's your own darn fault for picking the wrong standard of living.
One variable in this spending model is medical care during the later stages of life. That bare-bones budget has to somehow afford the premiums for Medicare supplemental insurance and prescription drug coverage, or set aside enough money to pay for it. We can still utter platitudes like staying healthy and slim, but even a few skinny triathletes may need blood-pressure medication or heavy-duty arthritis anti-inflammatory/painkillers.
Once again, the military version of that solution is Tricare For Life.
I'll edit this for a blog post to go live in a week or two. If you subscribe to the blog or its Twitter feed or its Facebook page then you'll know when the post goes up. If you have a comment on the implications of this study for military retirees then you could post here or send me a PM.