TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
ResearchMed
Posts: 12070
Joined: Fri Dec 26, 2008 11:25 pm

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by ResearchMed »

afan wrote: Wed Jun 16, 2021 9:23 am
<snip>

Second question.
The discussions I have heard suggest that the annuity performs best when one wants to spend all their money in their lifetime. However, I have heard it claimed that even one with bequest motives can benefit by partially annuitizing, living on the annuity payments and protecting the rest of the assets for heirs.I don't see how this would work but does anyone know?

RM: In some cases, this is intended more to allow someone to give gifts to heirs while the donor is still living. This allows some of the money be transferred earlier, perhaps when it could be more useful to the recipients. And it also allows the donor to enjoy watching how it helps the recipients in various ways. Taylor Larimore, here on BH, has written about how they did this, and specifically because they had annuitized enough to live on.

Third.
From the big picture POV mortality credits work to the advantage of those who live a long time at the expense of those who die young. I don't see how everyone can be better off under such a system. Without attempting to work through an example, it seems that the annuity transfers money to the longer lived people at the expense of the shorter lived people, minus the costs of the insurance company and tax consequences. How can it be that an annuity would be appropriate for everyone without money dropping from the skies to make those with short lifespans whole?

RM: Life annuities are *insurance* - insuring one will not "outlive their money", not matter how long one lives. Insurance is pooling of risks. The fact that some people end up receiving more and some less is part of the "deal", just as it is with insuring one's home or auto. Everyone "benefits" from having the security of knowing that coverage is there, IF needed. But not everyone "needs it". With life annuities, this means one did *not* live long enough that one may well have outlived one's money.

One real difference between life annuities and auto/home insurance is that with auto/home insurance one typically hopes one will NOT "need to collect" on the insurance. With life annuities, one would typically hope that one WOULD "collect"... with many extra years of income. :happy

RM
This signature is a placebo. You are in the control group.
User avatar
CWRadio
Posts: 482
Joined: Sat Feb 02, 2008 10:04 am
Location: In the Michigan ionosphere

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by CWRadio »

Little off subject. Two questions.
What age do you think the average Boglehead buys a SPIA?
What is the best age to buy a SPIA?
Thanks Paul
afan
Posts: 6029
Joined: Sun Jul 25, 2010 4:01 pm

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by afan »

In that narrow set of circumstances, I could see the value. But annuities are claimed to have far broader appeal, as in the claims that it is an economic puzzle why annuities are not used by everyone.

I am still working through a good academic paper on this, but so far the author's have assumed that the potential annuitant seeks to maximize expected consumption for each period. That is definitely not my goal and I assume there are many people who are not trying to increase the amount of money they spend.

What about whose goal, instead of maximizing consumption, is to maximize networth? To the extent that the cost of the annuity reduces networth and this catches up only if the person lives a long time, it would appear to be counter to that goal. Even if it resulted in a higher networth at say, 95 years old, one would have to discount this by the likelihood of living that long. If the odds on living to 80 are high and to 95 are low, then, in addition to time value of money, one should discount the age-95 networth to reflect the fact that it may not happen.

It seems that an annuity can be actuarially fair and still a poor choice.

Back to taxes. Even if one sells only enough assets to fill the current tax bracket and uses the proceeds to buy more annuities each year, the taxes still have to be paid. For those would would not otherwise liquidate their taxable holdings during life, this looks like a waste of money.
We don't know how to beat the market on a risk-adjusted basis, and we don't know anyone that does know either | --Swedroe | We assume that markets are efficient, that prices are right | --Fama
afan
Posts: 6029
Joined: Sun Jul 25, 2010 4:01 pm

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by afan »

ResearchMed wrote: Wed Jun 16, 2021 1:32 pm
afan wrote: Wed Jun 16, 2021 9:23 am
<snip>

Second question.
The discussions I have heard suggest that the annuity performs best when one wants to spend all their money in their lifetime. However, I have heard it claimed that even one with bequest motives can benefit by partially annuitizing, living on the annuity payments and protecting the rest of the assets for heirs.I don't see how this would work but does anyone know?

RM: In some cases, this is intended more to allow someone to give gifts to heirs while the donor is still living. This allows some of the money be transferred earlier, perhaps when it could be more useful to the recipients. And it also allows the donor to enjoy watching how it helps the recipients in various ways. Taylor Larimore, here on BH, has written about how they did this, and specifically because they had annuitized enough to live on.

Third.
From the big picture POV mortality credits work to the advantage of those who live a long time at the expense of those who die young. I don't see how everyone can be better off under such a system. Without attempting to work through an example, it seems that the annuity transfers money to the longer lived people at the expense of the shorter lived people, minus the costs of the insurance company and tax consequences. How can it be that an annuity would be appropriate for everyone without money dropping from the skies to make those with short lifespans whole?

RM: Life annuities are *insurance* - insuring one will not "outlive their money", not matter how long one lives. Insurance is pooling of risks. The fact that some people end up receiving more and some less is part of the "deal", just as it is with insuring one's home or auto. Everyone "benefits" from having the security of knowing that coverage is there, IF needed. But not everyone "needs it". With life annuities, this means one did *not* live long enough that one may well have outlived one's money.

One real difference between life annuities and auto/home insurance is that with auto/home insurance one typically hopes one will NOT "need to collect" on the insurance. With life annuities, one would typically hope that one WOULD "collect"... with many extra years of income. :happy

RM
Right, I get the longevity insurance motivation. I am not completely convinced, since it does not insure against inflation, which could drastically reduce the value of fixed payments decades in the future.

But not everyone needs longevity insurance. This is worth paying for only if one has a realistic concern of running out. For many BHs, that is not a worry. But if annuities are such good deals that those who do not need longevity insurance should buy the anyway, then there has to be more to it than that.

I am glad things worked out as they did for Taylor. I do suspect the story might have been very different if his annuity purchases had been followed by decades of high inflation. Those payments would now have a real value only a small fraction of where they started.

One can make gifts without having an annuity. The analysis is the same- how much can I afford to give away taking into account the risks of investment losses and inflation? I would think having annuities would be a secondary consideration at best.

And, of course, for it to work, one has to live a long time.
We don't know how to beat the market on a risk-adjusted basis, and we don't know anyone that does know either | --Swedroe | We assume that markets are efficient, that prices are right | --Fama
NiceUnparticularMan
Posts: 2526
Joined: Sat Mar 11, 2017 7:51 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by NiceUnparticularMan »

afan wrote: Wed Jun 16, 2021 1:48 pm
ResearchMed wrote: Wed Jun 16, 2021 1:32 pm
afan wrote: Wed Jun 16, 2021 9:23 am
<snip>

Second question.
The discussions I have heard suggest that the annuity performs best when one wants to spend all their money in their lifetime. However, I have heard it claimed that even one with bequest motives can benefit by partially annuitizing, living on the annuity payments and protecting the rest of the assets for heirs.I don't see how this would work but does anyone know?

RM: In some cases, this is intended more to allow someone to give gifts to heirs while the donor is still living. This allows some of the money be transferred earlier, perhaps when it could be more useful to the recipients. And it also allows the donor to enjoy watching how it helps the recipients in various ways. Taylor Larimore, here on BH, has written about how they did this, and specifically because they had annuitized enough to live on.

Third.
From the big picture POV mortality credits work to the advantage of those who live a long time at the expense of those who die young. I don't see how everyone can be better off under such a system. Without attempting to work through an example, it seems that the annuity transfers money to the longer lived people at the expense of the shorter lived people, minus the costs of the insurance company and tax consequences. How can it be that an annuity would be appropriate for everyone without money dropping from the skies to make those with short lifespans whole?

RM: Life annuities are *insurance* - insuring one will not "outlive their money", not matter how long one lives. Insurance is pooling of risks. The fact that some people end up receiving more and some less is part of the "deal", just as it is with insuring one's home or auto. Everyone "benefits" from having the security of knowing that coverage is there, IF needed. But not everyone "needs it". With life annuities, this means one did *not* live long enough that one may well have outlived one's money.

One real difference between life annuities and auto/home insurance is that with auto/home insurance one typically hopes one will NOT "need to collect" on the insurance. With life annuities, one would typically hope that one WOULD "collect"... with many extra years of income. :happy

RM
Right, I get the longevity insurance motivation. I am not completely convinced, since it does not insure against inflation, which could drastically reduce the value of fixed payments decades in the future.

But not everyone needs longevity insurance. This is worth paying for only if one has a realistic concern of running out. For many BHs, that is not a worry. But if annuities are such good deals that those who do not need longevity insurance should buy the anyway, then there has to be more to it than that.

I am glad things worked out as they did for Taylor. I do suspect the story might have been very different if his annuity purchases had been followed by decades of high inflation. Those payments would now have a real value only a small fraction of where they started.

One can make gifts without having an annuity. The analysis is the same- how much can I afford to give away taking into account the risks of investment losses and inflation? I would think having annuities would be a secondary consideration at best.

And, of course, for it to work, one has to live a long time.
So a lot of conversation here and elsewhere about retirement planning is about what withdrawal rate is "safe". And of course it is true once you are wealthy enough, that doesn't necessarily mean much anymore. But for many people, they feel like a relatively low withdrawal rate is necessary to stay safe, even though ideally they would be spending more, including on others.

And a lot of the reason why "safe" withdrawal rates are so low is self-insuring against longevity risk. Not all of it, but enough that some use of annuities can raise what your overall withdrawal rate can look like and still be safe. Or alternatively you can give away portions of your portfolio and still be "safe".

There is no real magic to any of that, and again if you are wealthy enough none of this might be much of a concern. But most people are not so wealthy this isn't a reasonable concern, and that is why various experts think use of annuities should be more widespread.
Wrench
Posts: 330
Joined: Sun Apr 28, 2019 10:21 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by Wrench »

afan wrote: Wed Jun 16, 2021 1:48 pm
ResearchMed wrote: Wed Jun 16, 2021 1:32 pm
afan wrote: Wed Jun 16, 2021 9:23 am
<snip>

Second question.
The discussions I have heard suggest that the annuity performs best when one wants to spend all their money in their lifetime. However, I have heard it claimed that even one with bequest motives can benefit by partially annuitizing, living on the annuity payments and protecting the rest of the assets for heirs.I don't see how this would work but does anyone know?

RM: In some cases, this is intended more to allow someone to give gifts to heirs while the donor is still living. This allows some of the money be transferred earlier, perhaps when it could be more useful to the recipients. And it also allows the donor to enjoy watching how it helps the recipients in various ways. Taylor Larimore, here on BH, has written about how they did this, and specifically because they had annuitized enough to live on.

Third.
From the big picture POV mortality credits work to the advantage of those who live a long time at the expense of those who die young. I don't see how everyone can be better off under such a system. Without attempting to work through an example, it seems that the annuity transfers money to the longer lived people at the expense of the shorter lived people, minus the costs of the insurance company and tax consequences. How can it be that an annuity would be appropriate for everyone without money dropping from the skies to make those with short lifespans whole?

RM: Life annuities are *insurance* - insuring one will not "outlive their money", not matter how long one lives. Insurance is pooling of risks. The fact that some people end up receiving more and some less is part of the "deal", just as it is with insuring one's home or auto. Everyone "benefits" from having the security of knowing that coverage is there, IF needed. But not everyone "needs it". With life annuities, this means one did *not* live long enough that one may well have outlived one's money.

One real difference between life annuities and auto/home insurance is that with auto/home insurance one typically hopes one will NOT "need to collect" on the insurance. With life annuities, one would typically hope that one WOULD "collect"... with many extra years of income. :happy

RM
Right, I get the longevity insurance motivation. I am not completely convinced, since it does not insure against inflation, which could drastically reduce the value of fixed payments decades in the future.

But not everyone needs longevity insurance. This is worth paying for only if one has a realistic concern of running out. For many BHs, that is not a worry. But if annuities are such good deals that those who do not need longevity insurance should buy the anyway, then there has to be more to it than that.

I am glad things worked out as they did for Taylor. I do suspect the story might have been very different if his annuity purchases had been followed by decades of high inflation. Those payments would now have a real value only a small fraction of where they started.

One can make gifts without having an annuity. The analysis is the same- how much can I afford to give away taking into account the risks of investment losses and inflation? I would think having annuities would be a secondary consideration at best.

And, of course, for it to work, one has to live a long time.
How do you or anyone else KNOW how long you are going to live? Unless you have a terminal disease (and even then...) nobody knows for sure. My FIL, born in 1930, always said he hoped to live to see the turn of the (21st) century. Today, at age 91, he is still going strong. Look at those who had AIDS in the early-mid 90s. Very few expected to live for more than a few years after diagnosis. 25 years later, many are still going strong with every expectation they will live a normal life span. Or breast cancer survivors today. Or compare heart disease survival in the mid-20th century to today. You just never know. And that's why longevity insurance is valuable to many. If you want to roll the dice, more power to you. I prefer to not take that chance. Oh, and you can at least partially protect against inflation, either by buying an annuity with a fixed increase every year (1, 2 or 3%), or by buying a ladder of annuities over several years (or decades), or by some combination of the two.

Wrench
NiceUnparticularMan
Posts: 2526
Joined: Sat Mar 11, 2017 7:51 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by NiceUnparticularMan »

Wrench wrote: Wed Jun 16, 2021 4:27 pm
afan wrote: Wed Jun 16, 2021 1:48 pm
ResearchMed wrote: Wed Jun 16, 2021 1:32 pm
afan wrote: Wed Jun 16, 2021 9:23 am
<snip>

Second question.
The discussions I have heard suggest that the annuity performs best when one wants to spend all their money in their lifetime. However, I have heard it claimed that even one with bequest motives can benefit by partially annuitizing, living on the annuity payments and protecting the rest of the assets for heirs.I don't see how this would work but does anyone know?

RM: In some cases, this is intended more to allow someone to give gifts to heirs while the donor is still living. This allows some of the money be transferred earlier, perhaps when it could be more useful to the recipients. And it also allows the donor to enjoy watching how it helps the recipients in various ways. Taylor Larimore, here on BH, has written about how they did this, and specifically because they had annuitized enough to live on.

Third.
From the big picture POV mortality credits work to the advantage of those who live a long time at the expense of those who die young. I don't see how everyone can be better off under such a system. Without attempting to work through an example, it seems that the annuity transfers money to the longer lived people at the expense of the shorter lived people, minus the costs of the insurance company and tax consequences. How can it be that an annuity would be appropriate for everyone without money dropping from the skies to make those with short lifespans whole?

RM: Life annuities are *insurance* - insuring one will not "outlive their money", not matter how long one lives. Insurance is pooling of risks. The fact that some people end up receiving more and some less is part of the "deal", just as it is with insuring one's home or auto. Everyone "benefits" from having the security of knowing that coverage is there, IF needed. But not everyone "needs it". With life annuities, this means one did *not* live long enough that one may well have outlived one's money.

One real difference between life annuities and auto/home insurance is that with auto/home insurance one typically hopes one will NOT "need to collect" on the insurance. With life annuities, one would typically hope that one WOULD "collect"... with many extra years of income. :happy

RM
Right, I get the longevity insurance motivation. I am not completely convinced, since it does not insure against inflation, which could drastically reduce the value of fixed payments decades in the future.

But not everyone needs longevity insurance. This is worth paying for only if one has a realistic concern of running out. For many BHs, that is not a worry. But if annuities are such good deals that those who do not need longevity insurance should buy the anyway, then there has to be more to it than that.

I am glad things worked out as they did for Taylor. I do suspect the story might have been very different if his annuity purchases had been followed by decades of high inflation. Those payments would now have a real value only a small fraction of where they started.

One can make gifts without having an annuity. The analysis is the same- how much can I afford to give away taking into account the risks of investment losses and inflation? I would think having annuities would be a secondary consideration at best.

And, of course, for it to work, one has to live a long time.
How do you or anyone else KNOW how long you are going to live? Unless you have a terminal disease (and even then...) nobody knows for sure. My FIL, born in 1930, always said he hoped to live to see the turn of the (21st) century. Today, at age 91, he is still going strong. Look at those who had AIDS in the early-mid 90s. Very few expected to live for more than a few years after diagnosis. 25 years later, many are still going strong with every expectation they will live a normal life span. Or breast cancer survivors today. Or compare heart disease survival in the mid-20th century to today. You just never know. And that's why longevity insurance is valuable to many. If you want to roll the dice, more power to you. I prefer to not take that chance. Oh, and you can at least partially protect against inflation, either by buying an annuity with a fixed increase every year (1, 2 or 3%), or by buying a ladder of annuities over several years (or decades), or by some combination of the two.

Wrench
And obviously it is quite nice that Social Security is adjusted for inflation. Of course that is only a partial backstop, but if you, say, delay Social Security to maximize the payments, and combine that with a TIPS ladder to get you to that point, and then plan to keep buying fixed annuities as necessary to maintain whatever more above that you want to provide for, it could get you pretty far.

Of course then you need some source of funding for the TIPS ladder, and some other source of funding that can be used to keep buying more fixed annuities. So we are backing to needing some significant wealth, although again it could still all be preferable to just a really low "safe" withdrawal rate.
Wrench
Posts: 330
Joined: Sun Apr 28, 2019 10:21 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by Wrench »

NiceUnparticularMan wrote: Wed Jun 16, 2021 4:43 pm [<SNIP>
And obviously it is quite nice that Social Security is adjusted for inflation. Of course that is only a partial backstop, but if you, say, delay Social Security to maximize the payments, and combine that with a TIPS ladder to get you to that point, and then plan to keep buying fixed annuities as necessary to maintain whatever more above that you want to provide for, it could get you pretty far.

Of course then you need some source of funding for the TIPS ladder, and some other source of funding that can be used to keep buying more fixed annuities. So we are backing to needing some significant wealth, although again it could still all be preferable to just a really low "safe" withdrawal rate.
What, are you my financial advisor or did you read my plan?! :happy Delaying social security until 70 coupled with a TIPS/iBonds ladder coupled with an inflation adjusted annuity is EXACTLY my plan, and what I already have in place. Those three together cover ~110% of my expected retirement expenses. The balance of my portfolio will be used for long term care (if needed) and/or a bequest and/or lumpy one time "wants". I have executed this plan because I wanted to be exempt from worrying about withdrawal rates, stock market returns, ABW approaches, declining mental capacity to handle my portfolio, inflation, etc. Will I end up with a smaller portfolio at the end? Probably. Will I sleep better between now and when I leave this earth? Yup.

Wrench
afan
Posts: 6029
Joined: Sun Jul 25, 2010 4:01 pm

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by afan »

Wrench wrote: Wed Jun 16, 2021 4:27 pm How do you or anyone else KNOW how long you are going to live? You just never know. And that's why longevity insurance is valuable to many. If you want to roll the dice, more power to you. I prefer to not take that chance. Oh, and you can at least partially protect against inflation, either by buying an annuity with a fixed increase every year (1, 2 or 3%), or by buying a ladder of annuities over several years (or decades), or by some combination of the two.

Wrench
Agree that, absent severe current disease, no one knows how long they will live.

But there are plenty of people who can plan for a sub 2% withdrawal rate supporting themselves in perpetuity. It would not matter how long they would live if their networth were increasing throughout retirement.

Some people can live quite well on SS and a pension. Others on SS and income, without touching principal, from their taxable accounts. Others can SPEND under 2%, by saving some of their after tax RMDs and spending the rest. Some can use a combination of these approaches, with no realistic worry of running out.

If annuities were ALWAYS better than simply holding the funds, then truly wealthy families with >$100 million should be buying massive annuities. I am not sure what they should be doing with that cash flow, which gets back to my question.

Part of my question is whether there is a role for annuities when one does not anticipate spending the cash flow from the product. I don't see how, but if there is no role then that seems to be a big gap in the argument that annuitizing is for everyone.

Economically, when you start receiving monthly distributions from an annuity, you are simply spending down the principal you put in. Once you have received back all the principal and the expected return on that principal, the annuity starts paying you from mortality credits. But you have to live long enough for getting annuity payments to result in more money than simply spending down your principal with no insurance company fees. If you have enough money that you do not expect ever to spend down all the principal, it is hard to see how the annuity helps. It is easy to see how it hurts if you die before you get at least your money back.

I suppose one could say that by saving cash flow from the annuity could save, rather than spend, SS or RMDs. Since money is fungible, it does not matter whether your money comes from SS, RMDs, taxable accounts or an annuity. That means buying an annuity to preserve capital is a weak argument.
We don't know how to beat the market on a risk-adjusted basis, and we don't know anyone that does know either | --Swedroe | We assume that markets are efficient, that prices are right | --Fama
afan
Posts: 6029
Joined: Sun Jul 25, 2010 4:01 pm

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by afan »

As I said, I don't know the answers.
The article I am reading starts with the assumption that one plans to spend all the money coming back from the annuity. The entire analysis is based on this point. It allows that one might spend only the annuity money, or that one might spend all the annuity money and more on top of that from a side account. But it explicitly excludes the situation where one need not spend the annuity money. Again, one could spend the annuity money and save the SS money, but it is hard to see how this helps.
We don't know how to beat the market on a risk-adjusted basis, and we don't know anyone that does know either | --Swedroe | We assume that markets are efficient, that prices are right | --Fama
afan
Posts: 6029
Joined: Sun Jul 25, 2010 4:01 pm

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by afan »

For those who are interested, the article to which I am referring is

"Journal of Economic Dynamics and Control
Volume 70, September 2016, Pages 18-35
Journal of Economic Dynamics and Control
The annuity puzzle remains a puzzle☆
KimPeijnenburga TheoNijmanbBas J.M.Werkerb"

It contains lots of references to other articles on the topic.
We don't know how to beat the market on a risk-adjusted basis, and we don't know anyone that does know either | --Swedroe | We assume that markets are efficient, that prices are right | --Fama
Topic Author
Northern Flicker
Posts: 8043
Joined: Fri Apr 10, 2015 12:29 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by Northern Flicker »

Wrench wrote: What, are you my financial advisor or did you read my plan?! :happy Delaying social security until 70 coupled with a TIPS/iBonds ladder coupled with an inflation adjusted annuity is EXACTLY my plan, and what I already have in place.
Where do you plan to buy an inflation-adjusted annuity? I'm not aware that they are still offered in the US.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
User avatar
Stinky
Posts: 8207
Joined: Mon Jun 12, 2017 11:38 am
Location: Sweet Home Alabama

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by Stinky »

CWRadio wrote: Wed Jun 16, 2021 1:39 pm Little off subject. Two questions.
What age do you think the average Boglehead buys a SPIA?
What is the best age to buy a SPIA?
Thanks Paul
From reading Forum posts, it seems like those Bogleheads who purchase a SPIA might be looking at age 70 or 75 for their purchase. Especially now, when interest rates are so low, monthly payouts for lower ages don't look too attractive.

Additionally, some Bogleheads have a traditional defined benefit pension plan in which they need to choose whether to take the monthly income or the lump sum. Many of those plans will require a decision by age 65. Taking the monthly income is the rough equivalent of "buying a SPIA".
It's a GREAT day to be alive! - Travis Tritt
NiceUnparticularMan
Posts: 2526
Joined: Sat Mar 11, 2017 7:51 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by NiceUnparticularMan »

Wrench wrote: Wed Jun 16, 2021 6:28 pm
NiceUnparticularMan wrote: Wed Jun 16, 2021 4:43 pm [<SNIP>
And obviously it is quite nice that Social Security is adjusted for inflation. Of course that is only a partial backstop, but if you, say, delay Social Security to maximize the payments, and combine that with a TIPS ladder to get you to that point, and then plan to keep buying fixed annuities as necessary to maintain whatever more above that you want to provide for, it could get you pretty far.

Of course then you need some source of funding for the TIPS ladder, and some other source of funding that can be used to keep buying more fixed annuities. So we are backing to needing some significant wealth, although again it could still all be preferable to just a really low "safe" withdrawal rate.
What, are you my financial advisor or did you read my plan?! :happy Delaying social security until 70 coupled with a TIPS/iBonds ladder coupled with an inflation adjusted annuity is EXACTLY my plan, and what I already have in place. Those three together cover ~110% of my expected retirement expenses. The balance of my portfolio will be used for long term care (if needed) and/or a bequest and/or lumpy one time "wants". I have executed this plan because I wanted to be exempt from worrying about withdrawal rates, stock market returns, ABW approaches, declining mental capacity to handle my portfolio, inflation, etc. Will I end up with a smaller portfolio at the end? Probably. Will I sleep better between now and when I leave this earth? Yup.

Wrench
Yep, as long as you have the funds necessary for all that, it is a solid plan.
NiceUnparticularMan
Posts: 2526
Joined: Sat Mar 11, 2017 7:51 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by NiceUnparticularMan »

afan wrote: Wed Jun 16, 2021 9:13 pm
Wrench wrote: Wed Jun 16, 2021 4:27 pm How do you or anyone else KNOW how long you are going to live? You just never know. And that's why longevity insurance is valuable to many. If you want to roll the dice, more power to you. I prefer to not take that chance. Oh, and you can at least partially protect against inflation, either by buying an annuity with a fixed increase every year (1, 2 or 3%), or by buying a ladder of annuities over several years (or decades), or by some combination of the two.

Wrench
Agree that, absent severe current disease, no one knows how long they will live.

But there are plenty of people who can plan for a sub 2% withdrawal rate supporting themselves in perpetuity. It would not matter how long they would live if their networth were increasing throughout retirement.
Well, when you say there are "plenty of people" with that much wealth, what do you mean by that?

I understand there are a variety of individuals here who are that wealthy, not least thanks to a very nice runup in asset prices in recent years. And I agree those people really do not need to worry about longevity.

But if you look at U.S. retirees as a whole, that amount of wealth is fairly rare. And the literature on annuities is not limited to relatively wealthy Bogleheads.
If annuities were ALWAYS better than simply holding the funds, then truly wealthy families with >$100 million should be buying massive annuities. I am not sure what they should be doing with that cash flow, which gets back to my question.
I don't think anyone is making the claim that annuities are always better regardless of wealth and personal spending plans. I think the claim is more that many more households would likely benefit from using annuities than are actually using annuities, but that is in reference to the many households who are not as wealthy as you are hypothesizing.
Part of my question is whether there is a role for annuities when one does not anticipate spending the cash flow from the product. I don't see how, but if there is no role then that seems to be a big gap in the argument that annuitizing is for everyone.
The argument I have seen is about providing funding for personal consumption. I don't believe I have ever seen anyone offer the sort of argument you are rebutting here.

I hope all this clarifies some things. I think you are right that very wealthy households don't necessarily need annuities. I also think you are right that the argument for annuities only applies to households planning to use the proceeds for personal consumption. And to my knowledge, I don't think anyone is actually suggesting otherwise.
TN_Boy
Posts: 2308
Joined: Sat Jan 17, 2009 12:51 pm

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by TN_Boy »

afan wrote: Wed Jun 16, 2021 9:24 pm For those who are interested, the article to which I am referring is

"Journal of Economic Dynamics and Control
Volume 70, September 2016, Pages 18-35
Journal of Economic Dynamics and Control
The annuity puzzle remains a puzzle☆
KimPeijnenburga TheoNijmanbBas J.M.Werkerb"

It contains lots of references to other articles on the topic.
I'm trying to wade through this a bit (lots of math which I am not trying to decipher in detail). Here is part of the conclusion:
This paper analyzes whether optimal annuity demand is affected by incomplete annuity markets and background risk. If no variable annuities are available and borrowing constraints are imposed, it can potentially be optimal to annuitize only a part of your wealth. However, we find that (almost) full annuitization remains optimal, irrespective of whether nominal or real annuities are available if agents save adequately out of their annuity income. In case of nominal annuities, the agent will save considerably out of the annuity income during retirement to gain equity exposure and hedge against background and inflation risk. If an individual receives a real annuity income, the agent saves a smaller amount as a buffer against (real) background risk. In all cases (close to) full annuitization at age 65 remains optimal. As a side result, we find that access to variable annuities is less welfare enhancing than previously found in the literature. The argument is similar: the buffer saved can be used to get sufficient equity exposure. These results are robust for realistic parameters of a bequest motive and default risk of the annuity seller.
If you look at the part I put in italics, they seem to be saying that the approach is to annuitize almost everything at retirement and then save what you don't spend to build up a buffer (hopefully you won't need extra money soon after you annuitize ....!). Background risk is the risk that you cannot borrow against the annuity income stream if you have a sudden large spending need; incomplete annuity markets refers to the fact that inflation adjusted annuities may not be available, ie. the income stream decreases over time.

I mention this because either I am misreading the conclusion and part of the text (quite possible as I skim it) or I'm not understanding your comment in the previous post:

"The article I am reading starts with the assumption that one plans to spend all the money coming back from the annuity."

Which is not quite how I read the article. But, again I might be misunderstanding.
NiceUnparticularMan
Posts: 2526
Joined: Sat Mar 11, 2017 7:51 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by NiceUnparticularMan »

TN_Boy wrote: Thu Jun 17, 2021 9:02 am
afan wrote: Wed Jun 16, 2021 9:24 pm For those who are interested, the article to which I am referring is

"Journal of Economic Dynamics and Control
Volume 70, September 2016, Pages 18-35
Journal of Economic Dynamics and Control
The annuity puzzle remains a puzzle☆
KimPeijnenburga TheoNijmanbBas J.M.Werkerb"

It contains lots of references to other articles on the topic.
I'm trying to wade through this a bit (lots of math which I am not trying to decipher in detail). Here is part of the conclusion:
This paper analyzes whether optimal annuity demand is affected by incomplete annuity markets and background risk. If no variable annuities are available and borrowing constraints are imposed, it can potentially be optimal to annuitize only a part of your wealth. However, we find that (almost) full annuitization remains optimal, irrespective of whether nominal or real annuities are available if agents save adequately out of their annuity income. In case of nominal annuities, the agent will save considerably out of the annuity income during retirement to gain equity exposure and hedge against background and inflation risk. If an individual receives a real annuity income, the agent saves a smaller amount as a buffer against (real) background risk. In all cases (close to) full annuitization at age 65 remains optimal. As a side result, we find that access to variable annuities is less welfare enhancing than previously found in the literature. The argument is similar: the buffer saved can be used to get sufficient equity exposure. These results are robust for realistic parameters of a bequest motive and default risk of the annuity seller.
If you look at the part I put in italics, they seem to be saying that the approach is to annuitize almost everything at retirement and then save what you don't spend to build up a buffer (hopefully you won't need extra money soon after you annuitize ....!). Background risk is the risk that you cannot borrow against the annuity income stream if you have a sudden large spending need; incomplete annuity markets refers to the fact that inflation adjusted annuities may not be available, ie. the income stream decreases over time.

I mention this because either I am misreading the conclusion and part of the text (quite possible as I skim it) or I'm not understanding your comment in the previous post:

"The article I am reading starts with the assumption that one plans to spend all the money coming back from the annuity."

Which is not quite how I read the article. But, again I might be misunderstanding.
So that article does start with a model which assumes away any sort of bequest motive, and the only utility comes from personal consumption. As an aside, it also assumes you can only buy an annuity once, at retirement, although they argue some of their proposed findings are robust to relaxations of that assumption.

It is true they do not assume you will IMMEDIATELY consume all annuity proceeds, however. Their model does allow for saving proceeds. But, they still assume the eventual purpose of those savings is to fund personal consumption, just at a later time.

They then relax the bequest assumption, and as one would expect, the higher the bequest motive, the less one should annuitize. But they are pointing out you can use the savings to fund a contingent bequest motive, so that does imply a higher annuitization even with a bequest motive than one would otherwise conclude, and they argue for many that savings plan is sufficient to satisfy a reasonable bequest motive.

Edit: By the way, I find it interesting their model suggests more annuitization in the face of a bequest motive if you can only use nominal annuities and not real annuities. This might seem counterintuitive, but it makes sense when you realize the payments on nominal annuities will be higher at first, and thus there is likely more spinoff of savings.

I also note it is interesting what happens to the model when you add default risk. This is potentially highly relevant to wealthy households where a lot of annuitization would mean exceeding available state guarantees.

Edit #2: Sorry, but one more note. Section 3.6 discusses the levels of wealth they are assuming, and they cap out at what I think would strike some Bogleheads as a relatively low level--$335,000 at age 65, which includes Social Security and any defined benefit plans along with liquid financial wealth.
Last edited by NiceUnparticularMan on Thu Jun 17, 2021 9:50 am, edited 1 time in total.
hirlaw
Posts: 405
Joined: Tue Sep 29, 2009 10:20 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by hirlaw »

Stinky wrote: Thu Jun 17, 2021 5:05 am
CWRadio wrote: Wed Jun 16, 2021 1:39 pm Little off subject. Two questions.
What age do you think the average Boglehead buys a SPIA?
What is the best age to buy a SPIA?
Thanks Paul
From reading Forum posts, it seems like those Bogleheads who purchase a SPIA might be looking at age 70 or 75 for their purchase. Especially now, when interest rates are so low, monthly payouts for lower ages don't look too attractive.

Additionally, some Bogleheads have a traditional defined benefit pension plan in which they need to choose whether to take the monthly income or the lump sum. Many of those plans will require a decision by age 65. Taking the monthly income is the rough equivalent of "buying a SPIA".
I agree with this, except I would add that if you are in poor health when you hit your 70's, I would would avoid purchasing a SPIA.
TN_Boy
Posts: 2308
Joined: Sat Jan 17, 2009 12:51 pm

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by TN_Boy »

NiceUnparticularMan wrote: Thu Jun 17, 2021 9:27 am
TN_Boy wrote: Thu Jun 17, 2021 9:02 am
afan wrote: Wed Jun 16, 2021 9:24 pm For those who are interested, the article to which I am referring is

"Journal of Economic Dynamics and Control
Volume 70, September 2016, Pages 18-35
Journal of Economic Dynamics and Control
The annuity puzzle remains a puzzle☆
KimPeijnenburga TheoNijmanbBas J.M.Werkerb"

It contains lots of references to other articles on the topic.
I'm trying to wade through this a bit (lots of math which I am not trying to decipher in detail). Here is part of the conclusion:
This paper analyzes whether optimal annuity demand is affected by incomplete annuity markets and background risk. If no variable annuities are available and borrowing constraints are imposed, it can potentially be optimal to annuitize only a part of your wealth. However, we find that (almost) full annuitization remains optimal, irrespective of whether nominal or real annuities are available if agents save adequately out of their annuity income. In case of nominal annuities, the agent will save considerably out of the annuity income during retirement to gain equity exposure and hedge against background and inflation risk. If an individual receives a real annuity income, the agent saves a smaller amount as a buffer against (real) background risk. In all cases (close to) full annuitization at age 65 remains optimal. As a side result, we find that access to variable annuities is less welfare enhancing than previously found in the literature. The argument is similar: the buffer saved can be used to get sufficient equity exposure. These results are robust for realistic parameters of a bequest motive and default risk of the annuity seller.
If you look at the part I put in italics, they seem to be saying that the approach is to annuitize almost everything at retirement and then save what you don't spend to build up a buffer (hopefully you won't need extra money soon after you annuitize ....!). Background risk is the risk that you cannot borrow against the annuity income stream if you have a sudden large spending need; incomplete annuity markets refers to the fact that inflation adjusted annuities may not be available, ie. the income stream decreases over time.

I mention this because either I am misreading the conclusion and part of the text (quite possible as I skim it) or I'm not understanding your comment in the previous post:

"The article I am reading starts with the assumption that one plans to spend all the money coming back from the annuity."

Which is not quite how I read the article. But, again I might be misunderstanding.
So that article does start with a model which assumes away any sort of bequest motive, and the only utility comes from personal consumption. As an aside, it also assumes you can only buy an annuity once, at retirement, although they argue some of their proposed findings are robust to relaxations of that assumption.

It is true they do not assume you will IMMEDIATELY consume all annuity proceeds, however. Their model does allow for saving proceeds. But, they still assume the eventual purpose of those savings is to fund personal consumption, just at a later time.

They then relax the bequest assumption, and as one would expect, the higher the bequest motive, the less one should annuitize. But they are pointing out you can use the savings to fund a contingent bequest motive, so that does imply a higher annuitization even with a bequest motive than one would otherwise conclude, and they argue for many that savings plan is sufficient to satisfy a reasonable bequest motive.

Edit: By the way, I find it interesting their model suggests more annuitization in the face of a bequest motive if you can only use nominal annuities and not real annuities. This might seem counterintuitive, but it makes sense when you realize the payments on nominal annuities will be higher at first, and thus there is likely more spinoff of savings.

I also note it is interesting what happens to the model when you add default risk. This is potentially highly relevant to wealthy households where a lot of annuitization would mean exceeding available state guarantees.

Edit #2: Sorry, but one more note. Section 3.6 discusses the levels of wealth they are assuming, and they cap out at what I think would strike some Bogleheads as a relatively low level--$335,000 at age 65, which includes Social Security and any defined benefit plans along with liquid financial wealth.
Okay, thanks for the summary there. I missed the part where they assume the savings are to handle the reduce in purchasing power of a nominal dollar annuity. I did see where they relaxed various constraints.

And I see your update, yes $335k at age 65 is perhaps not interesting to many folks on this board.

Many of these papers, even when they explain the reasoning, seem somehow to not view things the way I do :-). Obviously, the number crunchers could be right and I'm wrong (the answer to the puzzle: I'm dumb). But no matter what, the thought of writing a check for most of my savings while in my 60s is ..... a really awful thought. If I clearly have enough money to fund the retirement I want it strikes me as basically irrational to annuitize and thus create at least some short-term background risk and likely reduce legacies.

Interestingly, with respect to total wealth they say "Furthermore we see that when the total wealth level is lower, the optimal annuity demand is a bit lower. The reason is that an agent with a lower wealth level needs to keep a larger fraction liquid at the beginning of retirement to have the same absolute buffer against background risk."

Searching the doc I could find no mention of spouse or marriage except in one of the referenced papers. Which implies to me (without reading more carefully all the assumptions) that their calculations are for a single person, not a married couple. Am I missing something? I'd definitely want to see an analysis of a married couple, just to see if that affected any conclusions.
SGM
Posts: 3129
Joined: Wed Mar 23, 2011 4:46 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by SGM »

DW has a small percentage in TIAA traditional. Over a period of many years she transferred small amounts from TIAA traditional to a CREF world fund. She is very happy with the large payout she gets monthly after annuitization. She only worked 4 years for an institution that participated in TIAA-CREF. Early on after annuitization her payments went up a lot. I stopped following it several years ago. She now tells me that her CREF portion of the payments have gone up quite a bit in the last year. I think it may have gone up so much because she is one of the youngest of those in the CREF pool and older participants have passed on.
Wrench
Posts: 330
Joined: Sun Apr 28, 2019 10:21 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by Wrench »

Northern Flicker wrote: Wed Jun 16, 2021 11:00 pm
Wrench wrote: What, are you my financial advisor or did you read my plan?! :happy Delaying social security until 70 coupled with a TIPS/iBonds ladder coupled with an inflation adjusted annuity is EXACTLY my plan, and what I already have in place.
Where do you plan to buy an inflation-adjusted annuity? I'm not aware that they are still offered in the US.
Already own it. Bought it 2019 while they were still available - it actually is a deferred income annuity (DIA) for which inflation adjusted payments begin at age 70. You are correct,they are currently not being offered as far as I can determine.

Wrench
afan
Posts: 6029
Joined: Sun Jul 25, 2010 4:01 pm

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by afan »

NiceUnparticularMan wrote: Thu Jun 17, 2021 8:53 am

Well, when you say there are "plenty of people" with that much wealth, what do you mean by that?
Certainly hundreds of thousands of such people. Maybe more
I understand there are a variety of individuals here who are that wealthy, not least thanks to a very nice runup in asset prices in recent years. And I agree those people really do not need to worry about longevity.

But if you look at U.S. retirees as a whole, that amount of wealth is fairly rare. And the literature on annuities is not limited to relatively wealthy Bogleheads.
I agree. But the claims about annuities posit a puzzle why they are not purchased by EVERYONE, or nearly everyone.
If annuities were ALWAYS better than simply holding the funds, then truly wealthy families with >$100 million should be buying massive annuities. I am not sure what they should be doing with that cash flow, which gets back to my question.

I don't think anyone is making the claim that annuities are always better regardless of wealth and personal spending plans. I think the claim is more that many more households would likely benefit from using annuities than are actually using annuities, but that is in reference to the many households who are not as wealthy as you are hypothesizing.
Part of my question is whether there is a role for annuities when one does not anticipate spending the cash flow from the product. I don't see how, but if there is no role then that seems to be a big gap in the argument that annuitizing is for everyone.
The argument I have seen is about providing funding for personal consumption. I don't believe I have ever seen anyone offer the sort of argument you are rebutting here.
Well, I am not rebutting, I am trying to understand the argument. I am not convinced that annuities are useful for everyone who would spend all the money, since I worry about inflation reducing the value of that nominal cash flow. But if annuities are ONLY worth considering when one would spend all the money, that seems like a more restrictive condition than I have heard. Economists might be confining their analysis to those people, but they do not seem to SAY "we only consider people who would spend all the money". In fact, at least some claim that spending from the annuity protects the remainder of the assets, potentially leading to higher bequests.
I hope all this clarifies some things. I think you are right that very wealthy households don't necessarily need annuities. I also think you are right that the argument for annuities only applies to households planning to use the proceeds for personal consumption. And to my knowledge, I don't think anyone is actually suggesting otherwise.
In the article I cited, they show in one of the equations that their proposed goal of maximizing an expression means they are trying to maximize consumption. But other than showing the equation I don't think they come out and say this. This particular article, and maybe the general thinking behind papers on the subject, is that higher consumption always has positive marginal utility. That is, that someone can always find something else to buy that they want. Thus, saving money is always a sacrifice, perhaps justified by fear of future consumption desires or the wish to leave a bequest, but absent one or the other reasons, one would always be happier with more consumption.

If this is what they mean, and it seems to be, then it excludes someone who would say "I have $X. I believe $X to be more than enough to support my desired lifestyle to the end of a plausible maximum age. Above that $X, which constitutes my saving for future personal consumption, I also have $Y. I do not see a realistic circumstance under which I would spend any of $Y. I will save it, along with whatever amount of $X I do not spend, and it will be left at my death. Or I may give some or all of it during life to the same people who would have inherited it at my death." For someone in these circumstances, would annuitizing some or all of $X lead to a higher utility? Assume the goal is not to maximize consumption but, above some level of consumption easily affordable, the goal is to maximize networth.

That is my question. So far I cannot tell whether there is an annuity puzzle because economists assume there are no people with those desires, or there are too few to bother with. Alternatively, maybe they are taking those people into account and saying that even they would be better off annuitizing. The answer may be buried in their assumptions about utility functions, I have not gotten far enough to have an opinion. Curious whether anyone else has chased this and if so, what they found.
We don't know how to beat the market on a risk-adjusted basis, and we don't know anyone that does know either | --Swedroe | We assume that markets are efficient, that prices are right | --Fama
afan
Posts: 6029
Joined: Sun Jul 25, 2010 4:01 pm

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by afan »

To round this out, "maximize networth" includes maximizing networth today and maximizing expected real networth in the future. For something like buying an annuity, which requires an expenditure now and receiving payments later, one would have to do a present value calculation. Discount the future cash flows BOTH by the discount rate for the future but ALSO by the ever decreasing probability of those flows materializing. Since the annuity payments stop at death, one must discount each future flow to recognize that it may not happen at all.

Using a minimum term that the annuity will pay out even after death does not solve this problem. Not unless the term is so long that it is beyond any plausible life expectancy. At that point, the annuity payment could not contribute any mortality credits to the pool and presumably would be priced to reflect the certainty that the insurance company could never expect to use any of the money to pay other longer lived annuitants.
Last edited by afan on Thu Jun 17, 2021 1:36 pm, edited 1 time in total.
We don't know how to beat the market on a risk-adjusted basis, and we don't know anyone that does know either | --Swedroe | We assume that markets are efficient, that prices are right | --Fama
Wrench
Posts: 330
Joined: Sun Apr 28, 2019 10:21 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by Wrench »

NiceUnparticularMan wrote: Thu Jun 17, 2021 8:53 am
afan wrote: Wed Jun 16, 2021 9:13 pm
Wrench wrote: Wed Jun 16, 2021 4:27 pm How do you or anyone else KNOW how long you are going to live? You just never know. And that's why longevity insurance is valuable to many. If you want to roll the dice, more power to you. I prefer to not take that chance. Oh, and you can at least partially protect against inflation, either by buying an annuity with a fixed increase every year (1, 2 or 3%), or by buying a ladder of annuities over several years (or decades), or by some combination of the two.

Wrench
Agree that, absent severe current disease, no one knows how long they will live.

But there are plenty of people who can plan for a sub 2% withdrawal rate supporting themselves in perpetuity. It would not matter how long they would live if their networth were increasing throughout retirement.
Well, when you say there are "plenty of people" with that much wealth, what do you mean by that?

I understand there are a variety of individuals here who are that wealthy, not least thanks to a very nice runup in asset prices in recent years. And I agree those people really do not need to worry about longevity.

But if you look at U.S. retirees as a whole, that amount of wealth is fairly rare. And the literature on annuities is not limited to relatively wealthy Bogleheads.
If annuities were ALWAYS better than simply holding the funds, then truly wealthy families with >$100 million should be buying massive annuities. I am not sure what they should be doing with that cash flow, which gets back to my question.
I don't think anyone is making the claim that annuities are always better regardless of wealth and personal spending plans. I think the claim is more that many more households would likely benefit from using annuities than are actually using annuities, but that is in reference to the many households who are not as wealthy as you are hypothesizing.
Part of my question is whether there is a role for annuities when one does not anticipate spending the cash flow from the product. I don't see how, but if there is no role then that seems to be a big gap in the argument that annuitizing is for everyone.
The argument I have seen is about providing funding for personal consumption. I don't believe I have ever seen anyone offer the sort of argument you are rebutting here.

I hope all this clarifies some things. I think you are right that very wealthy households don't necessarily need annuities. I also think you are right that the argument for annuities only applies to households planning to use the proceeds for personal consumption. And to my knowledge, I don't think anyone is actually suggesting otherwise.
Clearly, if you are uber-wealthy, you do not need an annuity, or really any kind of insurance. (Aside - do you think Bill Gates or Elon Musk have homeowners insurance? Seems to me they don't need it, but I wonder if they have it anyway?) If you have little or no retirement savings, you can't afford an annuity. Somewhere in between no savings and uber wealthy, it may makes sense for some to buy one at some point. But where that wealth break point is, and whether it does make sense will depend on each individual, their personal situation, their risk tolerance, their health, single or married, etc.

Wrench
afan
Posts: 6029
Joined: Sun Jul 25, 2010 4:01 pm

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by afan »

I will never be at that wealth level, more is the pity. But I suspect those billionaires might have homeowners insurance not because it would be a hardship to replace a house but because the policy would provide risk mitigation advice from an expert source.

If they have enough expensive art and collectibles they may have their own team of conservators. If they are not that deep into it then the insurance company would consult about how to maintain and protect what they have. Could be a good way to locate the appropriate experts.
We don't know how to beat the market on a risk-adjusted basis, and we don't know anyone that does know either | --Swedroe | We assume that markets are efficient, that prices are right | --Fama
User avatar
dodecahedron
Posts: 5699
Joined: Tue Nov 12, 2013 12:28 pm

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by dodecahedron »

Wrench wrote: Thu Jun 17, 2021 1:31 pm Clearly, if you are uber-wealthy, you do not need an annuity, or really any kind of insurance. (Aside - do you think Bill Gates or Elon Musk have homeowners insurance? Seems to me they don't need it, but I wonder if they have it anyway?)
I would imagine they would have, at the very least, auto, homeowners' and umbrella liability coverage, since they would be very attractive targets for expensive and/or nuisance lawsuits.

Coverage for casualty property losses might also make sense, partly for tax reasons. Under current law, most unreimbursed casualty losses are no longer deductible (the only exception is losses due to officially declared federal disasters), but insurance reimbursement for casualty losses are generally tax-exempt.

There are other types of insurance high wealth individuals may not actually *need* but may find attractive for tax reasons, e.g., certain types of life insurance policies to manage gift and estate tax issues.

But yes, it is hard to imagine that standard SPIA or DIA policies would make sense for very high net worth folks. However, certain kinds of specialized charitable annuity trust products may be useful for reducing income and estate taxes and passing along bequests to subsequent generations in a tax-efficient manner.
NiceUnparticularMan
Posts: 2526
Joined: Sat Mar 11, 2017 7:51 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by NiceUnparticularMan »

afan wrote: Thu Jun 17, 2021 1:24 pm
NiceUnparticularMan wrote: Thu Jun 17, 2021 8:53 am

Well, when you say there are "plenty of people" with that much wealth, what do you mean by that?
Certainly hundreds of thousands of such people. Maybe more.
So I believe there were something like 46 million people receiving Social Security in 2020. Accordingly, you are talking about a very small portion of that population.
I understand there are a variety of individuals here who are that wealthy, not least thanks to a very nice runup in asset prices in recent years. And I agree those people really do not need to worry about longevity.

But if you look at U.S. retirees as a whole, that amount of wealth is fairly rare. And the literature on annuities is not limited to relatively wealthy Bogleheads.
I agree. But the claims about annuities posit a puzzle why they are not purchased by EVERYONE, or nearly everyone.
How high enough is "nearly" everyone? You have identified a group which may make up less than 1% of the retiree population.
If annuities were ALWAYS better than simply holding the funds, then truly wealthy families with >$100 million should be buying massive annuities. I am not sure what they should be doing with that cash flow, which gets back to my question.

I don't think anyone is making the claim that annuities are always better regardless of wealth and personal spending plans. I think the claim is more that many more households would likely benefit from using annuities than are actually using annuities, but that is in reference to the many households who are not as wealthy as you are hypothesizing.
Part of my question is whether there is a role for annuities when one does not anticipate spending the cash flow from the product. I don't see how, but if there is no role then that seems to be a big gap in the argument that annuitizing is for everyone.
The argument I have seen is about providing funding for personal consumption. I don't believe I have ever seen anyone offer the sort of argument you are rebutting here.
Well, I am not rebutting, I am trying to understand the argument. I am not convinced that annuities are useful for everyone who would spend all the money, since I worry about inflation reducing the value of that nominal cash flow. But if annuities are ONLY worth considering when one would spend all the money, that seems like a more restrictive condition than I have heard. Economists might be confining their analysis to those people, but they do not seem to SAY "we only consider people who would spend all the money". In fact, at least some claim that spending from the annuity protects the remainder of the assets, potentially leading to higher bequests.
You are leaping between extremes. The paper you cited neither makes the argument that full annuitization makes sense for all households regardless of bequest motive or wealth, nor that annuities are only worth considering if there is no bequest motive at all. Rather, it explains how full or partial annuitization might make sense for households up to a certain wealth even with some bequest motive.

Generally, I think you would benefit from being much more cautious about your use of words like "always", "everyone", or "only" when describing the results of stylized models like the one in the paper you cite.
If this is what they mean, and it seems to be, then it excludes someone who would say "I have $X. I believe $X to be more than enough to support my desired lifestyle to the end of a plausible maximum age. Above that $X, which constitutes my saving for future personal consumption, I also have $Y. I do not see a realistic circumstance under which I would spend any of $Y. I will save it, along with whatever amount of $X I do not spend, and it will be left at my death. Or I may give some or all of it during life to the same people who would have inherited it at my death." For someone in these circumstances, would annuitizing some or all of $X lead to a higher utility? Assume the goal is not to maximize consumption but, above some level of consumption easily affordable, the goal is to maximize networth.
So I think it is true economists would generally not see something like maximizing net worth as a plausible end in itself. They would want to know how that is supposed to achieve some goal.

But I think you are describing your $Y as a bequest. And they are allowing for some utility associated with such bequests.

Otherwise, I think the basic issue with your framing is that you have fixed the $X. Generally speaking, if you self-insure against longevity, you will need more money "to support my desired lifestyle to the end of a plausible maximum age" than if you use an annuity for that purpose.

So, here is a better framing. Suppose you need $X for an annuity that will serve this purpose. Suppose you will need an additional $Y to self-insure instead. And suppose you have $X + $Y available.

Are you better off:

(A) Annuitizing $X and investing $Y for bequests;

(B) Just investing $X + $Y in a self-insurance portfolio, and then bequeathing whatever happens when you die;

(C) Annuitizing all of $X + $Y, and using the excess proceeds to then invest for bequests.

Many have argued (A) is better than (B) given certain assumptions. Your paper is interesting in that it further argues (C) may also be better than (B) given certain assumptions.

However, if you also assume you have $Z, which is large in comparison to $Y . . . then ultimately this really stops mattering. That is because which of (A), (B), or (C) you choose starts getting swamped by $Z.

So you are assuming a large $Z, and I agree that means this doesn't matter. But the literature is generally about households which implicitly have little or no $Z, such that (A), (B), and (C) define their decision.
That is my question. So far I cannot tell whether there is an annuity puzzle because economists assume there are no people with those desires, or there are too few to bother with.
With this framework in mind, I think the answer is not so much about motives, as they are addressing households with little or no "$Z" such that the decision above matters. You are focused instead on the perhaps fraction of a percent of households for which there is a large $Z, and this decision more or less doesn't matter.
afan
Posts: 6029
Joined: Sun Jul 25, 2010 4:01 pm

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by afan »

In my example, $X included whatever amount the person felt was necessary to be sure they were OK for life. Thus, they had no need for a side fund to cover unexpected expenses. That money was included in $X. $Y was any amounts in excess of $X.
$Y did not have to be much larger than $X. It need not be as large as $X. It just needed to be truly in excess of what is needed. Therefore there was no Z. $Z was $Y and $X included the amounts in $Y in your nomenclature.
In the article, they assumed that even people who had bequest motives would consider annuities and that for most of the range of parameter values considered FULL annuitizing would be optimal.

I agree one has to be careful about over generalizing, but I read the conclusions of this paper to be very broad
If no variable annuities are available and borrowing constraints are imposed,
it can potentially be optimal to annuitize only a part of your wealth. However, we find that (almost) full annuitization remains optimal, irrespective of whether nominal or real annuities are available if agents save adequately out of their annuity income. In case of nominal annuities, the agent will save considerably out of the annuity income during retirement to gain equity exposure and hedge against background and inflation risk.
Here they assume that people establish annuity cash flows that exceed current spending needs, save the excess and use it to spend later or protect against inflation or other sources of higher than anticipated expense.

They analyze bequest motives in terms of a desire to leave some number of years of consumption to heirs, as opposed to "whatever is left after covering the lifetime expenses of the deceased". That may make the problem more tractable, but it may not capture the real goals of those doing estate planning.

They check whether high expected returns from investing in stocks outside an annuity would explain the low demand for annuities and reject that idea.
As a first robustness check we increase the equity premium to an expected stock return of 10% rather than 8%. Not surprisingly this implies a reduction in annuity demand, but the numerical effect is small. The optimal demand for real annuities reduces from 96% to 93%. For the nominal annuity case, full optimization always remains optimal.
In the latter sentence, I assume they meant. " For the nominal annuity case, full annuitization always remains optimal.

Although they ignore tax implications since they are trying to be general, they do discuss default risk in the context of the US. Of course, they find that high expected default rates and low expected recovery rates would lead to low use of annuities. However, they argue that the default rates at which this becomes a concern are far higher than realistic.
They treat all cash flows as if they are taxed at the same rates, whether from annuities or from a taxable portfolio. For a US investor who would pay a higher, and potentially much higher, tax rate on annuity distributions than they would on those from a taxable account, the analysis is clearly off base. If they had to reduce the size of the annuity by the taxes owed when assets were liquidated to buy it and then had to reduce the annuity payments by far more than the income from the taxable portfolio to account for the different tax treatments, I suspect their results and conclusions would have been quite different as well.

Their analysis does suggest that individuals with higher wealth and consumption, who might buy larger annuities, should be more concerned about default, since smaller portions would be covered by state guaranty associations.

They analyze the optimal demand for annuities under their assumptions and find
The pre-annuitized wealth is assumed to be an inflation-indexed (pension) income and the agent optimally chooses which part of his financial wealth to invest in a real annuity. So the optimal annuity levels are presented in two different formats: the numbers without brackets display the (1) optimal percentage in annuities as a fraction of total wealth and the numbers between brackets show the (2) optimal percentage in
annuities as a fraction of liquid financial wealth. First of all, we see that the optimal annuitization level as a fraction of total wealth is almost 100%. Naturally, when displayed as a percentage of liquid financial wealth it is lower, but still the annuity levels are much higher than found in reality.
In the end, they still conclude that more people "should" be buying annuities than are actually doing this. I suspect, at least for the US, the differences between their assumptions and real world tax considerations and bequest motives may be responsible.

Of course, this is just one paper. I do not yet know whether ignoring these tax and bequest issues are part of the standard economic approach. If they are, then it seems less of a puzzle. But they are so obvious that it would surprise me if all the economists who look at the question ignore them.
We don't know how to beat the market on a risk-adjusted basis, and we don't know anyone that does know either | --Swedroe | We assume that markets are efficient, that prices are right | --Fama
afan
Posts: 6029
Joined: Sun Jul 25, 2010 4:01 pm

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by afan »

On the number of people who do not have to worry about running out of money:

Last I saw, they estimated the cut point for top 1% of household wealth at about $11M. Now I am sure that some of those find it a struggle to make $11M last a life time. I am also sure that many with far less than $11M find no problem in living the life they want with no risk of going broke.

So call it 1% of the households, more or less. I would hope more than 1%, since I would hope those who cannot cover their lifestyles with $10M will tone it down before they go broke. Since this is household, not individual networth, estimate it at several million INDIVIDUALS who at least SHOULD NOT have to worry about running out of money.
We don't know how to beat the market on a risk-adjusted basis, and we don't know anyone that does know either | --Swedroe | We assume that markets are efficient, that prices are right | --Fama
NiceUnparticularMan
Posts: 2526
Joined: Sat Mar 11, 2017 7:51 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by NiceUnparticularMan »

afan wrote: Thu Jun 17, 2021 4:23 pm In my example, $X included whatever amount the person felt was necessary to be sure they were OK for life.
My point is $X is lower for an annuity than it would be for a self-insured portfolio. Until you account for that difference, you won't be able to make sense of any of this.
Therefore there was no Z. $Z was $Y and $X included the amounts in $Y in your nomenclature.
Right, but then you need a smaller amount than $X for the annuity amount.

I probably shouldn't have used the same variable names. So try this:

$A: annuity amount needed to provide for personal consumption for life.

$B: additional amount needed such that $A + $B is enough to safely provide personal consumption for life without an annuity.

$C: anything else you have.

Your $X = my $A + $B
Your $Y = my $C.

But you don't have terms specifically for $A or $B, because you haven't yet recognized the essential purpose of annuitization.

OK, then the rest of the discussion is as before. Is it better to put $A in the annuity and separately invest $B? Just self-insure with $A and $B? Or annuitize with $A and $B and then invest any extra proceeds?

But then you say, what about my $Y, your $C? And my point is that the bigger $Y/$C becomes, the less your answer to the above question matters.
In the article, they assumed that even people who had bequest motives would consider annuities and that for most of the range of parameter values considered FULL annuitizing would be optimal.
Yes, where the maximum wealth INCLUDING the value of Social Security and pensions at age 65 was $350,000.

You are imagining scenarios which are way, way outside of that range.
I agree one has to be careful about over generalizing, but I read the conclusions of this paper to be very broad
I don't know what to say to that. You really shouldn't do that.
Here they assume that people establish annuity cash flows that exceed current spending needs, save the excess and use it to spend later or protect against inflation or other sources of higher than anticipated expense.
Correct, they are modeling what happens if you annuitize $B and not just $A, and basically showing that because the excess savings are front-loaded, particularly with nominal annuities, it actually helps in terms of something like a bequest motive.
They analyze bequest motives in terms of a desire to leave some number of years of consumption to heirs, as opposed to "whatever is left after covering the lifetime expenses of the deceased". That may make the problem more tractable, but it may not capture the real goals of those doing estate planning.
I'd reverse that. My heirs getting whatever I leave when I die might be a simple solution, but it hardly counts as a description of a primary goal that is remotely plausible. Presumably my actual goal is to benefit my heirs in real ways.
They check whether high expected returns from investing in stocks outside an annuity would explain the low demand for annuities and reject that idea.
Correct, but within the boundaries of their other assumptions.

So, people with up to $350,000 to use at age 65. including the value of Social Security and pensions, aren't really very sensitive to exact stock returns when making these decisions. If you think about that scenario, I think you might intuit why.

Now $3.5 million? That could be a different situation.
Of course, they find that high expected default rates and low expected recovery rates would lead to low use of annuities. However, they argue that the default rates at which this becomes a concern are far higher than realistic.
Which again is true at the levels they are talking about, as they will be covered by state guarantees.
They treat all cash flows as if they are taxed at the same rates, whether from annuities or from a taxable portfolio.
Again at the levels they are talking about, that's not really a very powerful assumption.
In the end, they still conclude that more people "should" be buying annuities than are actually doing this.
Absolutely, because way, way more people fall into the wealth scenarios they are considering than the ones you are considering, which apply only to some very small fraction of U.S. retirees.
I suspect, at least for the US, the differences between their assumptions and real world tax considerations and bequest motives may be responsible.
Again, no, it is because you are only thinking about wealth scenarios which are way outside the boundaries of what they considered. Within the boundaries of what they considered, their simplifying assumptions and robustness tests make sense.

All you are doing is pointing out their argument doesn't really apply to the small percentage of U.S. households which have way higher wealth at retirement than their range.
NiceUnparticularMan
Posts: 2526
Joined: Sat Mar 11, 2017 7:51 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by NiceUnparticularMan »

afan wrote: Thu Jun 17, 2021 4:24 pm On the number of people who do not have to worry about running out of money:

Last I saw, they estimated the cut point for top 1% of household wealth at about $11M. Now I am sure that some of those find it a struggle to make $11M last a life time. I am also sure that many with far less than $11M find no problem in living the life they want with no risk of going broke.

So call it 1% of the households, more or less. I would hope more than 1%, since I would hope those who cannot cover their lifestyles with $10M will tone it down before they go broke. Since this is household, not individual networth, estimate it at several million INDIVIDUALS who at least SHOULD NOT have to worry about running out of money.
Again, this paper is looking at people (not households) with financial wealth (not all wealth) up to $350,000, INCLUDING the value of Social Security, at age 65. I don't know what percentile that takes you to, but I do know the median retiree household has way more in nonfinancial assets (usually a house and such) than financial assets. On the other hand, Social Security is pretty valuable on a lump sum basis, but then often people are sharing. So, my guess is that covers quite a lot of households.

If you want to talk about the potential role of annuities among households with far higher financial wealth, then that paper is not claiming to be easily applicable.
afan
Posts: 6029
Joined: Sun Jul 25, 2010 4:01 pm

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by afan »

Be careful about reading too much into the specific dollar values they use for their simulations. They could just as well have used 100x those figures, since they did not compare the behavior within their range to that of investors outside the range. Calculations were not driven at all by the dollar figures.

If they HAD compared the demand for annuities observed in the real world, as a function of wealth, then the dollar values would have mattered. The only place where any dollar value comes up is when they discuss default risk. But since they conclude that default risk is far too low to have any effect on real world behavior, it does not matter what would happen if this extraordinarily rare event were to occur.

Everywhere else the dollars values are cited is just illustrative for clarity. It is not tied to where someone stands on the wealth ladder whether one can live on the annual income it would produce, tax obligations (which are not discussed at all), or bankruptcy provisions.

In short, although it is true that they used low dollar values, the only place where that relates at all to the analysis is in the consequences of default and there they say the risk has to be far higher than it is to explain investor behavior.
We don't know how to beat the market on a risk-adjusted basis, and we don't know anyone that does know either | --Swedroe | We assume that markets are efficient, that prices are right | --Fama
NiceUnparticularMan
Posts: 2526
Joined: Sat Mar 11, 2017 7:51 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by NiceUnparticularMan »

afan wrote: Thu Jun 17, 2021 5:51 pm Be careful about reading too much into the specific dollar values they use for their simulations. They could just as well have used 100x those figures, since they did not compare the behavior within their range to that of investors outside the range. Calculations were not driven at all by the dollar figures.
That is incorrect. If you assume the same personal consumption goals in real terms, what they call FRAI, and then increase wealth beyond their parameters, it will change their model outputs.
The only place where any dollar value comes up is when they discuss default risk.
Again, that is incorrect. For example, they define what they call the "luxury parameter" in terms of multiples of FRAI, and then the lower the luxury parameter, the lower the optimal annuity demand given a particular bequest motive. This is modeling exactly the effect you are describing--if you have already provided for enough personal consumption for yourself and would value additional personal consumption (luxury spending) very little, you have less and less reason to annuitize any additional funds the farther you get from your baseline consumption.
But since they conclude that default risk is far too low to have any effect on real world behavior, it does not matter what would happen if this extraordinarily rare event were to occur.
Again, this is a place where dollar values matter because of how they model default risk. They assume a $100,000 state guarantee. How valuable that is depends on how much is annuitized. And they make this very explicit:
Second, we see that the fraction of wealth recovered is important for
the optimal annuity demand, which is similar to the findings of Babbel & Merril (2006). In this
paper we normalized all numbers in terms of the Full Real Annuity Income, but the wealth that
was used as a basis is $335,000, which is the median total wealth level at 65. Hence 33.5% of the
value of the annuity at 65 is guaranteed by the state.
If you use a much higher basis for wealth, the state-guaranteed value of the annuity will be lower, and they explain that is important to determining optimal annuity demand.

I really think you are reaching to find a problem where one doesn't exist. This is a constrained model which plausibly shows what it is intended to show: that annuity use should be much more common than observed even given factors like uncertainty about stock returns, default risk, and so on.

But they are very much not even trying to show that very wealthy retirees should fully annuitize, or use annuities at all for that matter. And that is not inconsistent, because there is a vast gulf between the percentage of households using annuities, and the percentage of non-very-wealthy households. So, the plausible logic of very wealthy households which you are expressing doesn't actually provide an explanation of why so many households without such wealth are also not annuitizing.
Topic Author
Northern Flicker
Posts: 8043
Joined: Fri Apr 10, 2015 12:29 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by Northern Flicker »

Stinky wrote: From reading Forum posts, it seems like those Bogleheads who purchase a SPIA might be looking at age 70 or 75 for their purchase. Especially now, when interest rates are so low, monthly payouts for lower ages don't look too attractive.
Unattractive conpared to what? Bond yields? Hypothetical annuity payout rates if and when interest rstes are higher?

The increase in payout rate that is due to being older is an illusion. It is just spreading the return of capital over fewer years from date of purchase to age of life expectancy. The main benefit of delaying is that you have more information about health outcomes and their effect on life expectancy when deciding on whether to annuitize.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
Wrench
Posts: 330
Joined: Sun Apr 28, 2019 10:21 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by Wrench »

Northern Flicker wrote: Fri Jun 18, 2021 3:11 am
Stinky wrote: From reading Forum posts, it seems like those Bogleheads who purchase a SPIA might be looking at age 70 or 75 for their purchase. Especially now, when interest rates are so low, monthly payouts for lower ages don't look too attractive.
Unattractive conpared to what? Bond yields? Hypothetical annuity payout rates if and when interest rstes are higher?

The increase in payout rate that is due to being older is an illusion. It is just spreading the return of capital over fewer years from date of purchase to age of life expectancy. The main benefit of delaying is that you have more information about health outcomes and their effect on life expectancy when deciding on whether to annuitize.
I would not quite say that the payout rate when you are older is an "illusion", because an SPIA pays out for a lifetime. And nobody knows how long they will live. It could pay out for 30 years whether you buy at 65, 70 or 75 because you could live to 95, 100 or 105. Of course, as you point out, you may have a better sense of your mortality at 70 or 75 than at 65. Or maybe not. A cancer diagnosis or heart attack or accident or whatever can happen anytime, even to one who is very healthy up to that point. One never knows.

It is also true the that the biggest factor in payout rate is age. BUT, a secondary factor that is not as impactful to payout rate IS interest rate. I saw that when I purchased my annuity. I quoted it multiple times over a period of ~2 months when rates where dropping. The payout rate went down as rates dropped, but only by a few tenths of a percent or so.

Wrench
NiceUnparticularMan
Posts: 2526
Joined: Sat Mar 11, 2017 7:51 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by NiceUnparticularMan »

Wrench wrote: Fri Jun 18, 2021 8:47 am
Northern Flicker wrote: Fri Jun 18, 2021 3:11 am
Stinky wrote: From reading Forum posts, it seems like those Bogleheads who purchase a SPIA might be looking at age 70 or 75 for their purchase. Especially now, when interest rates are so low, monthly payouts for lower ages don't look too attractive.
Unattractive conpared to what? Bond yields? Hypothetical annuity payout rates if and when interest rstes are higher?

The increase in payout rate that is due to being older is an illusion. It is just spreading the return of capital over fewer years from date of purchase to age of life expectancy. The main benefit of delaying is that you have more information about health outcomes and their effect on life expectancy when deciding on whether to annuitize.
I would not quite say that the payout rate when you are older is an "illusion", because an SPIA pays out for a lifetime. And nobody knows how long they will live. It could pay out for 30 years whether you buy at 65, 70 or 75 because you could live to 95, 100 or 105. Of course, as you point out, you may have a better sense of your mortality at 70 or 75 than at 65. Or maybe not. A cancer diagnosis or heart attack or accident or whatever can happen anytime, even to one who is very healthy up to that point. One never knows.

It is also true the that the biggest factor in payout rate is age. BUT, a secondary factor that is not as impactful to payout rate IS interest rate. I saw that when I purchased my annuity. I quoted it multiple times over a period of ~2 months when rates where dropping. The payout rate went down as rates dropped, but only by a few tenths of a percent or so.

Wrench
Yeah, a pretty good first-order model of a vanilla SPIA is they are just sticking a bond ladder inside a longevity-risk-pool wrapper. So, as a bond ladder that would make your desired nominal payments during your expected actuarial lifespan gets more expensive, so too will a SPIA that makes those payments.

This also more or less explains what happened to inflation-adjusted SPIAs. These days, a TIPS ladder has a negative real rate of return, meaning the price of a TIPS ladder for a given desired real payment has gotten really, really high. So marketing and selling such a thing, even with the longevity-risk-wrapper, at a reasonable profit is not so easy.

That being said--what ARE your other options? If you were counting on buying a DIY portfolio with a certain "safe" withdrawal rate, arguably you should be increasing the expected price (or decreasing the expected SWR) in light of these very same considerations.

In that sense, it is a bit unfair to say DIY portfolios are advantaged over SPIAs by high prices and low interest rates on available assets. That takes assuming your DIY portfolio is immune to the same issue, which is not exactly a safe assumption.
Topic Author
Northern Flicker
Posts: 8043
Joined: Fri Apr 10, 2015 12:29 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by Northern Flicker »

Yes, interest rates are low right now whether SPIA payout rates or bond yields. You can hold a fixed income portfolio with a shorter duration than the duration of the annuity to bet on interest rates rising, but it is a bet. It could go the other way.

SPIAs return principal by dividing it equally between the years to age of life expectancy. Those who live longer get principal returned to them from people who lived less long and did not receive it.

Insurance companies do not allocate a bond ladder for each annuitant but hold a bond portfolio with duration matched to the duration of their liabilities.

By delaying the purchase, you get more return of principal in each payment but get fewer of them. It is actuarially neutral except there is a wrinkle: life expectancy increases slightly every year. This might increase the cost of SPIAs over time. But it could go the other way if insurance companies model this change and price in the risk of life expectancy increasing faster than it is projected to increase.

But just getting age credits is actuarially neutral. If interest rates stay flat, delaying just means getting your principal returned to you at a faster rate. What it does do is alleviate much of the risk of tying up the capital and then needing to spend it at a faster rate. By delaying, the principal is tied up for fewer years and returned to you at a faster rate.

Delaying also has the benefit of having more information to make the annuitization decision. Without annuitization, years past the age of life expectancy have to be funded with our own principal, increasing the present value of retirement costs-- it is funding for the worst case outcome. A SPIA funds for the average case. This means that annuitization may be a good option if we find out retirement assets are running short for reasons that do not shorten life expectancy (such as high health care expenses).

But I think a SPIA can be a way to deal with low interest rates. By annuitizing a part of fixed income allocation, fewer years have to be covered with fixed income principal so the overall fixed income allocation including the SPIA can be a little lower, a benefit when yields are low. Such a strategy is likely to add the most value if the SPIA is bought early in the process to increase the horizon for the slightly larger equity holding.

For instance, instead of holding 50% stock and 50% bonds, one might hold 55% stock, 20% bonds, 25% SPIA. The SPIA level might be chosen to cover the gap between SS benefits and known non-discretionary expenses.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
NiceUnparticularMan
Posts: 2526
Joined: Sat Mar 11, 2017 7:51 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by NiceUnparticularMan »

Northern Flicker wrote: Fri Jun 18, 2021 2:12 pm Yes, interest rates are low right now whether SPIA payout rates or bond yields. You can hold a fixed income portfolio with a shorter duration than the duration of the annuity to bet on interest rates rising, but it is a bet. It could go the other way.

SPIAs return principal by dividing it equally between the years to age of life expectancy. Those who live longer get principal returned to them from people who lived less long and did not receive it.

Insurance companies do not allocate a bond ladder for each annuitant but hold a bond portfolio with duration matched to the duration of their liabilities.

By delaying the purchase, you get more return of principal in each payment but get fewer of them. It is actuarially neutral except there is a wrinkle: life expectancy increases slightly every year. This might increase the cost of SPIAs over time. But it could go the other way if insurance companies model this change and price in the risk of life expectancy increasing faster than it is projected to increase.

But just getting age credits is actuarially neutral. If interest rates stay flat, delaying just means getting your principal returned to you at a faster rate. What it does do is alleviate much of the risk of tying up the capital and then needing to spend it at a faster rate. By delaying, the principal is tied up for fewer years and returned to you at a faster rate.

Delaying also has the benefit of having more information to make the annuitization decision. Without annuitization, years past the age of life expectancy have to be funded with our own principal, increasing the present value of retirement costs-- it is funding for the worst case outcome. A SPIA funds for the average case. This means that annuitization may be a good option if we find out retirement assets are running short for reasons that do not shorten life expectancy (such as high health care expenses).

But I think a SPIA can be a way to deal with low interest rates. By annuitizing a part of fixed income allocation, fewer years have to be covered with fixed income principal so the overall fixed income allocation including the SPIA can be a little lower, a benefit when yields are low. Such a strategy is likely to add the most value if the SPIA is bought early in the process to increase the horizon for the slightly larger equity holding.

For instance, instead of holding 50% stock and 50% bonds, one might hold 55% stock, 20% bonds, 25% SPIA. The SPIA level might be chosen to cover the gap between SS benefits and known non-discretionary expenses.
Excellent post! And on that last point, I was actually kinda intrigued by the paper's examination of a strategy of buying a bigger nominal SPIA than you think necessary right at the beginning of retirement, and then using the expected extra proceeds, should they materialize, to invest in longer-term risky ways.

That originally struck me as a bit counterintuitive, but then it started occurring to me it actually sorta has the same logic as the "bond tent" concept.

So I am going to think about that one some more. Although I do think taxes are a concern as well. However, if you are using tax-deferred funds (like from a 401K or traditional IRA or cash-balance pension), I think that just becomes an IRA annuity, and the proceeds are even an exception to the 10% pre-59.5 rule.

You might want to watch your overall income levels anyway, but I do wonder if a nominal "SPIA tent" (maybe along with or instead of the tent portion of a bond tent strategy) could actually make sense for a lot of people, with that SPIA being bought earlier than normal around here.
Topic Author
Northern Flicker
Posts: 8043
Joined: Fri Apr 10, 2015 12:29 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by Northern Flicker »

NiceUnparticularMan wrote: I was actually kinda intrigued by the paper's examination of a strategy of buying a bigger nominal SPIA than you think necessary right at the beginning of retirement, and then using the expected extra proceeds, should they materialize, to invest in longer-term risky ways.

That originally struck me as a bit counterintuitive, but then it started occurring to me it actually sorta has the same logic as the "bond tent" concept.

So I am going to think about that one some more. Although I do think taxes are a concern as well. However, if you are using tax-deferred funds (like from a 401K or traditional IRA or cash-balance pension)...
While I would not rule out such a strategy, my first thought is that the benefit has to overcome the transaction cost of buying the SPIA (2-3% load hidden in the rate) and the higher admin cost also hidden in the rate (higher than a low cost bond fund ER).

As far as taxes are concerned, holding a SPIA in either tax-deferred or taxable space is efficient. The case for tax-deferred space is that it will be distributed by RMDs whether you use the IRS RMD algorithm or a SPIA or a combination of the two to distribute it. The SPIA makes the distribution more uniform.

In a taxable account, the interest is taxed but not the return of principal. Suppose as an example a 72-year-old has a 10-year life expectancy and buys a SPIA for $100K with a 12% payout rate. Then the annual payment is $12K. $100K of principal is returned over 10 years, so $10K of the annual payment is return of principal and $2K is taxable income. What is interesting is that if the annuitant lives longer than 10 years, the full principal will have been returned, but my understanding is that the payments received will still be $10K of return of principal and $2K of taxable income. The fact that the annuitant is getting the principal funded by other annuitants in the pool who did not live to their age of expectancy does not turn this into taxable income, at least that is my understanding. But there is a symmetry-- the estate of annuitants who die before the expected age also do not get to claim a loss of principal (also as far as I know).

Fixed nominal SPIAs in tax-deferred accounts have a little bit of inflation protection because tax brackets are indexed to inflation (to chained CPI). This becomes insignificantly small when the
SPIA is held in a taxable account.

My view is that it is preferable to hold a SPIA in tax-deferred space, but maybe not if it would consume most of tax-deferred space. Tax-deferred assets could be taxed much less if there were a future large deduction for medical or long-term care expenses.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
LivingTheDream
Posts: 100
Joined: Sun Oct 18, 2020 10:30 am
Location: Savannah, GA

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by LivingTheDream »

I admit that I didn't read all comments, so I apologize if I missed some important information.

However, several of the posts seem dumbfounded as to why people are averse to even considering buying annuities. I can't speak for others, but my reason for being biased is simple.

In 2019, the Life/Annuity Insurance Industry paid $61.2 billion in commissions, on $679.9b in premiums (9% sales load).
Source: https://www.iii.org/fact-statistic/fact ... -insurance

Even if an SPIA has a relatively low commission of "only" 3%, I personally am very unlikely to pay that for what is essentially a bond fund [VBTLX would cover an SPIA roughly from 80-95]?
Living The Dream
Topic Author
Northern Flicker
Posts: 8043
Joined: Fri Apr 10, 2015 12:29 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by Northern Flicker »

I've never seen SPIA's sold with 9% commissions. The lowest cost brokers have been 2% in the past.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
User avatar
Stinky
Posts: 8207
Joined: Mon Jun 12, 2017 11:38 am
Location: Sweet Home Alabama

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by Stinky »

LivingTheDream wrote: Sat Jun 19, 2021 12:06 am In 2019, the Life/Annuity Insurance Industry paid $61.2 billion in commissions, on $679.9b in premiums (9% sales load).
Source: https://www.iii.org/fact-statistic/fact ... -insurance
This is one of those statistics that is “correct, but irrelevant”.

The table shows total premiums and total commissions for all life insurance and annuity products.

The quoted “9% sales” load is -
—- Somewhere in the range of the commissions paid on many “bad” annuities (variable annuities, indexed annuities, etc)
—- Much higher than the commissions paid on “good” annuities (MYGAs, SPIAs)
—- Much lower than the commissions paid on first year life insurance premiums
—- Much higher than the commissions paid on renewal Life insurance premiums.

The quoted “sales load” is so aggregated that it is an essentially meaningless statistic.
It's a GREAT day to be alive! - Travis Tritt
NiceUnparticularMan
Posts: 2526
Joined: Sat Mar 11, 2017 7:51 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by NiceUnparticularMan »

Northern Flicker wrote: Fri Jun 18, 2021 7:41 pm
NiceUnparticularMan wrote: I was actually kinda intrigued by the paper's examination of a strategy of buying a bigger nominal SPIA than you think necessary right at the beginning of retirement, and then using the expected extra proceeds, should they materialize, to invest in longer-term risky ways.

That originally struck me as a bit counterintuitive, but then it started occurring to me it actually sorta has the same logic as the "bond tent" concept.

So I am going to think about that one some more. Although I do think taxes are a concern as well. However, if you are using tax-deferred funds (like from a 401K or traditional IRA or cash-balance pension)...
While I would not rule out such a strategy, my first thought is that the benefit has to overcome the transaction cost of buying the SPIA (2-3% load hidden in the rate) and the higher admin cost also hidden in the rate (higher than a low cost bond fund ER).
Absolutely. I think it is possible this could make sense anyway if, for example, you can do something like put less overall into low-risk/low-return assets, meaning your SPIA "tent" (or SPIA + bonds tent) would be less costly than a bonds-only tent, and the savings could be put into higher-risk/higher-return assets. But I haven't really tried to work out a model of that yet.
As far as taxes are concerned, holding a SPIA in either tax-deferred or taxable space is efficient. The case for tax-deferred space is that it will be distributed by RMDs whether you use the IRS RMD algorithm or a SPIA or a combination of the two to distribute it. The SPIA makes the distribution more uniform.


Just to be clear, I was referring to the issue of potentially having to sell stocks and realize capital gains in order to buy your SPIA tent. If you are using assets in a tax-deferred account, that is a non-issue. But if you would be selling assets in a taxable account, it might be.
In a taxable account, the interest is taxed but not the return of principal. Suppose as an example a 72-year-old has a 10-year life expectancy and buys a SPIA for $100K with a 12% payout rate. Then the annual payment is $12K. $100K of principal is returned over 10 years, so $10K of the annual payment is return of principal and $2K is taxable income. What is interesting is that if the annuitant lives longer than 10 years, the full principal will have been returned, but my understanding is that the payments received will still be $10K of return of principal and $2K of taxable income. The fact that the annuitant is getting the principal funded by other annuitants in the pool who did not live to their age of expectancy does not turn this into taxable income, at least that is my understanding. But there is a symmetry-- the estate of annuitants who die before the expected age also do not get to claim a loss of principal (also as far as I know).
I actually thought once you go past the expected lifespan, it all becomes taxable as ordinary income. But I am not a tax expert.
NiceUnparticularMan
Posts: 2526
Joined: Sat Mar 11, 2017 7:51 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by NiceUnparticularMan »

LivingTheDream wrote: Sat Jun 19, 2021 12:06 amEven if an SPIA has a relatively low commission of "only" 3%, I personally am very unlikely to pay that for what is essentially a bond fund [VBTLX would cover an SPIA roughly from 80-95]?
But it isn't a bond fund. It is not even a bond ladder, although it functions like one through the actuarial age. But it then continues as long as you live. No bond ladder is like that.

I of course wish the SPIA market was highly funded, hypercompetitive, and insurance company margins on these products were much lower.

But, if the longevity insurance aspects of these products has enough of a benefit for a given investor over trying to self-insure against longevity, then it can be worth that cost.
User avatar
Stinky
Posts: 8207
Joined: Mon Jun 12, 2017 11:38 am
Location: Sweet Home Alabama

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by Stinky »

NiceUnparticularMan wrote: Sat Jun 19, 2021 7:56 am I of course wish the SPIA market was highly funded, hypercompetitive, and insurance company margins on these products were much lower.
I will assure you, from personal knowledge, that the SPIA market is quite competitive, and that insurers do not make outsized profits on SPIAs.
It's a GREAT day to be alive! - Travis Tritt
Topic Author
Northern Flicker
Posts: 8043
Joined: Fri Apr 10, 2015 12:29 am

Re: TIAA White Paper on Fixed Annuities and SPIAs in Retirement Portfolios

Post by Northern Flicker »

I believe MetLife, USAA, and TIAA sell SPIAs directly without using brokers. Not sure if you have to be a USAA member or TIAA group retirement plan participant to buy from those respective insurers. Also not sure how their pricing compares to SPIAs sold, say by immediateannuities.com.

There are some annuity brokers who are unscrupulous, and when you ask for a SPIA quote, will quote for a variable annuity with equity exposure to try to make the product look cheaper for a given projected payout level and not indicate that they switched products in the quote. I had this happen when I once was trying to get a quote for informational purposes.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
Post Reply