Why Don't Retirees Buy Annuities? They Get Something Most Economists Don't

I think you missed part of running_man's plan. Did you notice the \$180,000?

I'll convert your numbers to annual and do a little rounding because it's easier for me to see what's happening.

You say the choice is:

A. Take SS today.
Starting now, \$30,400 from SS + \$40,000 from 4% SWR = \$70,400 total, indefinitely.

B. Defer SS for four years. Your plan is has two income levels.
First four years, \$0 from SS + \$40,000 from 4% SWR = \$40,000
After that, \$40,200 from SS + \$40,000 from 4% SWR = \$80,200

Given the choice between A and B, you'd take A. I suppose most of us would.

But, rm has a different plan. Again using your numbers:

A. Take SS today.
Starting now, \$30,400 from SS + \$40,000 from 4% SWR = \$70,400 total, indefinitely.

C. Defer SS for four years
Move \$160,000 from my \$1 million portfolio into short assets (CD's, short TIPS), leaving \$840,000 invested in stocks etc.
First four years, \$0 from SS + \$40,000 from ST + \$33,600 from 4% SWR = \$73,600 total.
After that, \$40,200 from SS + \$0 from ST + \$33,600 from 4% SWR = \$73,800 total

Notice that Plan C allows me to spend more money in the first four years, and spend more money in the remaining years then Plan A. I prefer Plan C over both Plan A and Plan B. (Note that if I'm a more conservative 3% SWR person, the advantage of C is even larger.)

In fact, that's what we did. When I retired, we had a laddered portfolio of CDs and I-Bonds that was designed to fill the gap between retirement date and SS start date. We've mostly spent that portfolio down.

Modified Plan C. As the SSA has been mentioning for years, benefits are reduced by 25% starting in 2033 (perhaps earlier) due to the trust fund running out and general economic malaise. Then, payments become \$30,150 from SS + \$33,600 from the reduced kitty for a total of \$63,750.

Modified Plan C. As the SSA has been mentioning for years, benefits are reduced by 25% starting in 2033 (perhaps earlier) due to the trust fund running out and general economic malaise. Then, payments become \$30,150 from SS + \$33,600 from the reduced kitty for a total of \$63,750.
Yes, there are lot of variables to consider in the deferral issue. The potentially unequal treatment of people who started earlier vs. those who started later is certainly high on many lists.

I was just trying to point out that people who defer usually do not have a two tier spending plan. Instead, they plan to make extra withdrawals in the early years to fill the income gap. For them "deferring SS" is not the same as "spending less in the early years".

Getting to your numbers, note the total for plan A after 2033:
(75% of \$30,400) + \$40,000 = \$22,800 + \$40,000 = \$62,800

It turns out that Plan C is still better in this particular case. But, I haven't given any thought to whether that is typical.

Obviously it all depends on how one emotionally views it. The SSA itself says that whether you take at 62 or 66 or 70, it's actuarially identical. Therefore there is no *financial* reason to prefer one age to another......

For someone who is single, I would agree that it doesn't matter since the benefits for any one person are designed to be actuarially equivalent. Ditto for a couple where each person has worked and their respective SS benefits are about the same.

But for a couple where the benefits are unequal, especially where the person with the lower benefit is female, there is a world of difference because of joint mortality and the fact that the lower earning spouse gets the higher earning spouse's benefit when the higher earning spouse dies. And if the lower earning spouse is female and also much younger, then the preference to wait to 70 is even greater.

But for a couple where the benefits are unequal, especially where the person with the lower benefit is female, there is a world of difference because of joint mortality and the fact that the lower earning spouse gets the higher earning spouse's benefit when the higher earning spouse dies. And if the lower earning spouse is female and also much younger, then the preference to wait to 70 is even greater.

If they are highly dependent on the SS benefit, arguably yes. Well, arguably perhaps.

The surviving spouse gets 50% of the deceased spouse's benefit, not 100%. So the max amount she'd get from SS would be \$1675 (70) vs. \$1266 (66), or \$409/mo more.
Note that she'd still get the full 4% SWR of the (assumed) \$1M portfolio, so her total income would be \$5008 (70) vs. \$4599 (66). That's only 8.9% more, so the benefit of deferring is even smaller than before.

I would submit that there's no significant difference in lifestyle for a (newly) single person between \$4600 and \$5000. And, again, in order to get that \$409/mo for rest of her life, they had to give up \$2533/mo for 4 years.
Using just back-of-envelope math: \$2533 * 48 = \$121,584. \$124,584 / \$409 = 297 months. It'll take her 25 years to get to break-even.

It's really difficult for some people to realize in their gut that actuarially identical indeed means ACTUARIALLY IDENTICAL. There's no magic way to wiggle around and get one outcome to be better than another.

The surviving spouse gets 50% of the deceased spouse's benefit, not 100%. So the max amount she'd get from SS would be \$1675 (70) vs. \$1266 (66), or \$409/mo more.
The surviving spouse gets 100% of the deceased spouse's benefit, not 50%. So, she'll get \$818/mo more if the now-deceased spouse has waited to 70 to start collecting SS. I think that amount of additional "I can count on this every month" money would make a significant difference in quality of life for many people.
When the surviving spouse starts collecting the full benefit of the deceased spouse, she must stop collecting any spousal benefit she got before or any benefits she was collecting based on her own work record.
It's really difficult for some people to realize in their gut that actuarially identical indeed means ACTUARIALLY IDENTICAL. There's no magic way to wiggle around and get one outcome to be better than another.
If a person's only COLA'd income stream is from SS, then "ACTUARIALLY IDENTICAL" may be a long way from IDENTICAL IN PRACTICE. As people tend to decrease their equities in later years and increase their percentage of fixed income holdings, deferring SS to get a higher monthly payout which will be COLA'd can make sense as a means to help preserve spending power in case these fixed income holdings don't keep pace with inflation. There's no cheaper inflation-protected annuity than that provided by delaying SS.

That said, I'm not entirely comfortable in depending on that government promise, and understand the appeal of taking the SS earlier to reduce draw from the portfolio in those early years, and keeping that "extra money" in more volatile assets that may also have a better chance to keep up with inflation (i.e. self-insuring against inflation rather than depending on an unchanging govt promise). I'm not sure what I'll do myself.

Last edited:
+1 . My FIRE spreadsheet includes small SPIAs after age 70, and already includes a couple of deferred annuities bought a few months ago. Three of my grandparents passed after the age of 90.

Personally I am thinking of buying a small SPIA if/when I reach age 80-85. It should be cheaper then than it would be at an earlier age. Inflation would not be as much of a concern at that age as it would be for a younger person. The point of the annuity would be to provide steady income that I could rely upon (along with SS and my tiny pension) for basic expenses, in case I turn out to be one of the few who survive until extreme old age.

Last edited:
It's really difficult for some people to realize in their gut that actuarially identical indeed means ACTUARIALLY IDENTICAL. There's no magic way to wiggle around and get one outcome to be better than another.

Oh, heck. Some individuals are dying of cancer at 62 and are definitely not actuarially identical. That's on average from SS's point of view, not any one individual. Given that it's on average, any little difference might throw your best option to either side of the average.

Anyway, I believe this SS stuff got started by a comment that if anyone was going to buy an SPIA they should be delaying SS until 70. I think that still stands. I assume rayvt won't be buying an SPIA any time soon.

Social Security is only actuarially neutral for an unusual set of circumstances -- that is, for an unmarried individual with an average life expectancy, when current real interest rates match the rates baked into the Social Security calculations.

For most people, if choosing between claiming as early as possible or waiting until 70, there is an answer that is preferable (that is, an answer that has a better than 50% chance of working out well).

Obviously it all depends on how one emotionally views it. The SSA itself says that whether you take at 62 or 66 or 70, it's actuarially identical. Therefore there is no *financial* reason to prefer one age to another.

I think the idea that this is actuarially identical is a false concept. This cannot be when interest rates are not figured in the payout decision nor is the sex of the recipient, nor are payouts adjusted for changes in life expectancy. However I do think under current market conditions it is an advantage to wait if one can afford to do so.

Actually you make a good point that I forgot RM. Even though the payouts are actuarially equivalent for a single, that is based on unisex mortality, so the decision would also vary based on the recipient's gender since females tend to outlive males.

If they are highly dependent on the SS benefit, arguably yes. Well, arguably perhaps.

The surviving spouse gets 50% of the deceased spouse's benefit, not 100%. So the max amount she'd get from SS would be \$1675 (70) vs. \$1266 (66), or \$409/mo more.
Note that she'd still get the full 4% SWR of the (assumed) \$1M portfolio, so her total income would be \$5008 (70) vs. \$4599 (66). That's only 8.9% more, so the benefit of deferring is even smaller than before.

I would submit that there's no significant difference in lifestyle for a (newly) single person between \$4600 and \$5000. And, again, in order to get that \$409/mo for rest of her life, they had to give up \$2533/mo for 4 years.
Using just back-of-envelope math: \$2533 * 48 = \$121,584. \$124,584 / \$409 = 297 months. It'll take her 25 years to get to break-even.

It's really difficult for some people to realize in their gut that actuarially identical indeed means ACTUARIALLY IDENTICAL. There's no magic way to wiggle around and get one outcome to be better than another.

Once again, let's write the numbers down:

Plan X: Both start today. Assume both are 66, husband has the high income, wife gets 50% at age 66.

Total while they are both living = \$30,400 + \$15,200 + \$40,000 = \$85,600
Total after the first death = \$30,400 + \$40,000 = \$70,400

Plan Y: Start wife's benefit today, defer husband's to age 70, move \$160,000 into short term assets to provide bridge income.

Total for first 4 years = \$15,200 + \$40,000 + \$33,600 = \$88,800
Total between 4 years and first death = \$15,200 + \$40,200 + \$33,600 = \$89,000
Total after first death = \$40,200 + \$33,600 = \$73,800

So, over three periods (first four years - then till first death - after first death) the totals are:

Plan X: \$85,600 - \$85,600 - \$70,400
Plan Y: \$88,800 - \$89,000 - \$73,800

Deferring provides more income in every period.

Last edited:
Once again, let's write the numbers down:

Plan X: Both start today. Assume both are 66, husband has the high income, wife gets 50% at age 66.

Total while they are both living = \$30,400 + \$15,200 + \$40,000 = \$85,600
Total after the first death = \$30,400 + \$40,000 = \$70,400

Plan Y: Start wife's benefit today, defer husband's to age 70, move \$160,000 into short term assets to provide bridge income.

Total for first 4 years = \$15,200 + \$40,000 + \$33,600 = \$88,800
Total between 4 years and first death = \$15,200 + \$40,200 + \$33,600 = \$89,000
Total after first death = \$40,200 + \$33,600 = \$73,800

So, over three periods (first four years, then till first death, after first death) the totals are:

Plan X: \$85,600 - \$85,600 - \$70,400
Plan Y: \$88,800 - \$89,000 - \$73,800

Deferring provides more income in every period.

Great for income, and I'm expecting to delay one or both of us to age 70, but you're out the \$160k in capital until the extra income makes it up. Could be good or bad depending on how long you expect to live and what kind of investment gains you assume.

Great for income, and I'm expecting to delay one or both of us to age 70, but you're out the \$160k in capital until the extra income makes it up. Could be good or bad depending on how long you expect to live and what kind of investment gains you assume.

Using the spending differential I've given, I don't see that the couple who defers ever makes up the difference. They would eventually if they had the same spending as the couple that started early, but I'm proposing that they spend more.

It seems to me that if investment returns are good and the \$1.0 million portfolio doubles to \$2.0 million, the \$840,000 portfolio will also double, but only to \$1.68 million. Similarly, if they're bad and the \$1.0 million shrinks to \$500,000, the \$840,000 will shrink to \$420,000.

OTOH, they are no more likely to spend their assets all the way to zero than the couple that starts earlier. And, if they should do that, deferring gives a higher safety net.

It gets really complicated really fast, doesn't it.

First thought: \$85.6K and \$88.8K are pretty close. And, of course, that's for a person who is getting the max SS benefit. I don't know how many people hit the cap, but the average benefit is more like \$1300/mo than \$3300. I don't know what the distribution is, though.

Ah-ha! Google to the rescue. The distribution is highly skewed. Almost everybody (46%) is between \$800 and \$1600. Only 9% are above \$1900. Only 2% are above \$2300. Above \$3100: 0.1%.

Retired worker beneficiaries distributed by benefit level

Aside from all that, I suspect that these 2 scenarios come out to be very close or perhaps even flip-flop when you take investment growth into account. The difference between \$89,000 and \$85,600 is \$3,200.

Assuming 8% total return...
Plan X: First year asset base is 96% of \$1M. Gain is \$76,800

Plan Y: First year asset base is 84% of \$1M. Gain is \$67,200.
That difference right there is \$9,600.
A naive comparison says that you'd be 3 times better off by *not* pulling 4 years of cash out.

Eh, whatever.... the numbers are so close that there is no practical difference. There's nothing you can do with \$89K that you can't do with \$86K.

I have purchased three annuities and the best way to buy them is NOT through a salesmen. Go to Vanguard or there is another company that caters to military that has four letters, I just can't remember it at this time, you will get a much better product for your money. Also, you don't have to be in the military or ever been it to apply for the annuity.
Kimo

Last edited:
I have purchased three annuities and the best way to buy them is NOT through a salesmen. Go to Vanguard or there is another company that caters to military that has four letters, I just can't remember it at this time, you will get a much better product for your money. Also, you don't have to be in the military or ever been it to apply for the annuity.
Kimo
USAA?

Yes, USAA is correct and thank you......it was late last night and the brain just wasn't working......

It gets really complicated really fast, doesn't it.
...
.
Yes, all the pros and cons of taking SS now or later can get pretty complicated.

I wasn't trying to cover everything, just make two simple points.

1. People who defer SS are not planning on a two-tier, low-high spending program. They have the same roughly level spending plans as people who start SS immediately.

2. People who defer SS are not planning to spend less, either sooner or later. Some of them plan to spend just as much as if they started SS immediately, others plan to spend slightly more.

I saw your post "But it's only 14% increase in your total income -- and to get it you have to take \$2533 LESS income for 4 years." and thought that you weren't getting #1, and probably not #2. I might have misread your meaning.

Yeah, I kinda got it and kinda discounted it.
When you are comparing different options, you have to compare like to like which means that you need to compute the NPV of the different alternatives.
When people say things like "you will get \$X more money if you defer", they are implicitly assuming that \$1 in 4 years (or 8 years in the case of deferring from 62 to 70) is the same as \$1 today.
You can change the shape of the distribution -- taking less today and more tomorrow, or more today and less tomorrow -- but these all have the same NPV.

Yeah, I kinda got it and kinda discounted it.
When you are comparing different options, you have to compare like to like which means that you need to compute the NPV of the different alternatives.
When people say things like "you will get \$X more money if you defer", they are implicitly assuming that \$1 in 4 years (or 8 years in the case of deferring from 62 to 70) is the same as \$1 today.
You can change the shape of the distribution -- taking less today and more tomorrow, or more today and less tomorrow -- but these all have the same NPV.
Good, at least we're talking about the same numbers.

Somehow the math in FireCalc or using the 4% SWR slightly favors deferring, even though the SS actuaries say the factors are "roughly" actuarially equivalent.

The issue is that calculating the PV of a life contingent annuity requires assuming a discount rate and a mortality table. The assumptions that produce the same PV for two SS choices are not the same assumptions that most people using FireCalc or 4% SWR are making when they calculate maximum "safe" spending.

I'm NOT suggesting that you or anyone else use any specific approach. I'm just saying that I think I can see how to reconcile this anomaly.

Almost the entire investing history of people investing today has been spent with falling interest rates and low inflation.

This makes it more or less impossible for us to really imagine the ravages of moderate to high inflation. Annuities are an extremely poor long term bet. I would not touch one with a stick, other than the inflation indexed one that we all will get some of.

Ha

Geez, I'm only 68 and wasn't a kid during the inflation of the 70's. I remember the stagflation discussions quite well.

For an older person who is considering a SPIA, their expected life span and current low inflation, make the prospect of future inflation less alarming than for a person who was 55 in the early-70's.

I do not have any annuities, but reserve judgment about them for now.

Last edited:
A typical payout rate on a SPIA for a 65 year-old male is about 6.2%. The Vanguard LT Investment Grade Fund (VWETX) currently has a yield of 4.8%. In order for the annuity to achieve a 4.8% IRR, the annuitant would have to live until age 97.

This is one of the things I hate in the annuity business- the confusion of payout rate vs IRR. Payout rate is really meaningless yet too many people fixate on it. The poster above has it right- IRR is what you're looking for to compare apples to apples.

The problem is, unless you know the exact day you're going to die and spend your last dollar that day, you can't compute your IRR, and you are always carrying your longevity risk. You can't spend principal peacefully if you risk outliving that principal....

The key thing to remember with annuities is that they are INSURANCE products and the insurance you buy with most contracts that are tied to your lifespan offloads that longevity risk to the carrier. You can spend principal prudently because the company holds the risk of you spending down to \$0.

True,the yield is not as good as you hope to get in your vanguard fund, but for most types of annuities, you gain a benefit in exchange for the lower yield.

Anyway, back to the point about IRR, if you're looking for a good safe annuity with a known, definite IRR based yield, there are fixed annuities with guarantees, or Secondary Market Annuities that are nothing more than discounted payment streams. they're sold priced based on their yield- the effective rate = IRR. Pretty transparent, and not a bad yield considering there is no volatility risk.

Secondary Market Annuities that are nothing more than discounted payment streams. they're sold priced based on their yield- the effective rate = IRR. Pretty transparent, and not a bad yield considering there is no volatility risk.
Could you give a link to a reputable broker of these things? I know they must exist, because my free UHF TV channels always feature ads offering to take those pesky annuities or structured settlements off people's hands.

BTW, I think annuities have a giant elephant in the room risk that rarely gets deeply considered. A person does not have to be crazy to expect a humdinger inflation at some point near or far in our future.

Ha

The problem is, unless you know the exact day you're going to die and spend your last dollar that day, you can't compute your IRR, and you are always carrying your longevity risk. You can't spend principal peacefully if you risk outliving that principal....
Not necessarily true.

We have a joint/life guaranteed term SPIA and it was very easy to calculate the base IRR. Simply plug in the premium, the total annual payments, and the number of term/years (cells) on a spreadsheet.

We know the minimum return, since our guaranteed term (either we get paid, or our estate does) establishes a minimum period of payments.

And if either/or live beyond the minimum term (calculated at 28 years for us)? The calculated return actual rises.

We can't outlive our SPIA; however if our SPIA "outlives" us, there is residual value that goes to our estate.

As far as IRR vs. payment? You are correct; IRR must be computed for any SPIA product in order to make correct comparisions. Also, folks forget that an SPIA is an income product, not an investment product. It's not there to make you money but rather an instrument in your retirement toolbox to provide you with distribution of investment proceeds over a period of time. That's why our current SPIA was purchased with only 10% of our then current joint portfolio value at retirement (early 2007). The remaining 90% of our portfolio stays invested (relative to our AA) and does the heavy lifting for the future.

Last edited:
BTW, I think annuities have a giant elephant in the room risk that rarely gets deeply considered. A person does not have to be crazy to expect a humdinger inflation at some point near or far in our future.

Ha

+1. But that risk exists with long bonds (& funds) as well. Locking in to a ~4-5% yield, even assuming holding safe indiv long bonds to maturity, risks a huge loss of purchasing power over time if inflation goes back to double-digits (as too many of us can recall ).