longevity annuity

[skanney]

does anyone know a good place to get a longevity annuity? vanguard does not sell them but i am looking for a low cost version with cpi protection.

[JDCPAEsq]

Hartford Insurance Company is one choice.
John

[hicabob]

I have never heard of that twist on annuities before so I looked it up ...
here's metlife's (the purveyor) take on it ....

"
What's the Answer? MetLife has come up with a clever spin on the standard payout annuity that provides both protection against outliving your savings, and costs less money up front. Meet the advanced-life delayed annuity. I know. Its a mouthful. We'll stick to calling it a longevity annuity.

How Does it Work? Rather than paying out a significant portion such as 60% of your assets to the immediate payout annuity at the point of retirement and getting monthly payments immediately, you instead purchase the longevity annuity with only, say, 10% of your money and it doesn't begin paying out for 5-20 years depending on your terms.

Why would the insurance companies provide this sort of annuity? Because the catch is that if you don't live to the age at which you can start receiving payments, you don't get anything out of it. This is why you get the security of a significant income in your later years for quite a small premium. "

Seems like an OK concept - It would be really nice if the expected "house odds" of various annuities were disclosed.

[skanney]

has anyone heard of this site? http://incomesolutions.com/HomePage.aspx seems they have institutional pricing for annuities which can save about 5% off retail, they show you how to compare quotes and it is run by some professor...that what it looks like...maybe they give the house odds?

[JDCPAEsq]

has anyone heard of this site? http://incomesolutions.com/HomePage.aspx seems they have institutional pricing for annuities which can save about 5% off retail, they show you how to compare quotes and it is run by some professor...that what it looks like...maybe they give the house odds?

Yes, it's available through the Vanguard Website.
John

[dpbsmith]

I can't seem to find the actual product description at MetLife's website... can you give us a link to it?

Without some kind of CPI adjustment, I'd be very leery of any such product. In general, I'd worry about any annuity that is making a lot of its total payout twenty or thirty years down the line, and I'd be doubly leery of one that doesn't make any payments for that length of time.

To put things in perspective, in 1975 I was making $13,500 a year and thought I was doing OK. If you'd said "this product will provide $13,500 a year when you retire" I'd have said that sounded fine. Thirty years later, it would have taken $50,000 to provide equivalent buying power.

But the problem isn't so much inflation, it's the unpredictability, the not knowing how much inflation there's going to be. I pay the insurance company to take on the risk for me. If the point is a comfortable income thirty years from now, there there needs to be protection against both inflation risk and longevity risk. If the insurance company is only providing one of them, it doesn't seem like very good insurance.

As for the "house odds," in a very rough-and-ready sort of way, some academic studies of "moneys' worth" numbers all seem to show that the insurance company is keeping something on the order of 2% to 10% of the premium, i.e. the true value of the payouts is maybe 90% or more of the premiums. That's credible, because when I did some very crude amateur actuarial calculations myself with a spreadsheet, the results I got were within striking distance of the actual annuity payments. That's not to say 10% is trivial, but it suggests that there's a lot of competition and one kind of annuity isn't going to be a hugely better "deal" than another kind. That's strikingly different from health insurance, where one of the things that's emerged in the wake of the new regulations is that some health insurance companies were paying out as little as 60% of the premiums.

[stlrick]

As far as I know, insurance companies have not yet figured out how to offer worthwhile advanced life delayed annuities (longevity insurance). The problem is inflation protection, which generally is not available for this product. The more typical immediate annuity, an SPIA, is most commonly purchased at an advanced age, 70-80, when the mortality credits result in a payout that cannot be matched by other strategies. SPIAs also can be found with inflation protection. On the other hand, longevity insurance is most interesting when it is purchased at a younger age, 50-60, so that a small investment can result in a large payout at age 85. Therefore, the inflation protection is even more important than it is with an SPIA, but it is not available. I am watching this product fairly carefully. When they figure out how to do it right, I'll be first in line.

Rick

[skanney]

i think insurance companies can hedge the inflation part with swaps or use TIPs in some fashion. perhaps they buy TIPs, repo the collateral out and take in low quality collateral to pick up some sort of credit spread. there may be alternatives in swaps that run it off balance sheet as well.

as a buyer of the product, i think the inflation protection can be done. my main concern is the credit risk. i want to run the math for something like metlife's product where you lose all your money at age 50 and only have an option to get payments at 80 or something. that should drive up the payments and allow a higher withdrawal rate on the rest of the portfolio. this is the math i want to check. but i could only see doing something like this in small quantities across a number of insurers.

[stlrick]

Skanney:

I don't know what you mean by "the inflation protection can be done." It is my understanding that insurance companies are not offering this product with investment options for the buyer to chose. It is a straight contract. You pay X dollars at age Y, and they will then pay you A dollars at age B. What they can offer and what they are offering are different issues. Inflation protection is not available. I would be happy to learn that I am incorrect.

Rick

[skanney]

rick,

if i am not mistaken, inflation protection is offered by american general, principal life and probably metlife as a minimum. i do not believe it is offered by integrity life or lincoln...

[hicabob]

I can't seem to find the actual product description at MetLife's website... can you give us a link to it?

Without some kind of CPI adjustment, I'd be very leery of any such product. In general, I'd worry about any annuity that is making a lot of its total payout twenty or thirty years down the line, and I'd be doubly leery of one that doesn't make any payments for that length of time.

To put things in perspective, in 1975 I was making $13,500 a year and thought I was doing OK. If you'd said "this product will provide $13,500 a year when you retire" I'd have said that sounded fine. Thirty years later, it would have taken $50,000 to provide equivalent buying power.

But the problem isn't so much inflation, it's the unpredictability, the not knowing how much inflation there's going to be. I pay the insurance company to take on the risk for me. If the point is a comfortable income thirty years from now, there there needs to be protection against both inflation risk and longevity risk. If the insurance company is only providing one of them, it doesn't seem like very good insurance.

As for the "house odds," in a very rough-and-ready sort of way, some academic studies of "moneys' worth" numbers all seem to show that the insurance company is keeping something on the order of 2% to 10% of the premium, i.e. the true value of the payouts is maybe 90% or more of the premiums. That's credible, because when I did some very crude amateur actuarial calculations myself with a spreadsheet, the results I got were within striking distance of the actual annuity payments. That's not to say 10% is trivial, but it suggests that there's a lot of competition and one kind of annuity isn't going to be a hugely better "deal" than another kind. That's strikingly different from health insurance, where one of the things that's emerged in the wake of the new regulations is that some health insurance companies were paying out as little as 60% of the premiums.

Here's the metlife link - quite the domain name eh?
http://www.longevityannuity.org/

[Watty]

Without some kind of CPI adjustment, I'd be very leery of any such product. In general, I'd worry about any annuity that is making a lot of its total payout twenty or thirty years down the line, and I'd be doubly leery of one that doesn't make any payments for that length of time.

That was a major problem that I saw too. The other was than the current interest rates are a major factor in determining how much an annuity pays. The current low interest rates make it a difficult time to buy any sort of annuity.

It hasn't been mentioned but delaying when you start social security would be an alternative that is in effect buying a longevity annunity. It is too bad that the latest age that you can delay it to is 70. If it was higher, like 80, and it kept increasing then many people could safely plan on mostly depleting most of their retirement funds by then and having a large enough social security check starting at 80 to safely cover allmost all needs.

[hicabob]

Without some kind of CPI adjustment, I'd be very leery of any such product. In general, I'd worry about any annuity that is making a lot of its total payout twenty or thirty years down the line, and I'd be doubly leery of one that doesn't make any payments for that length of time.

That was a major problem that I saw too. The other was than the current interest rates are a major factor in determining how much an annuity pays. The current low interest rates make it a difficult time to buy any sort of annuity.

It hasn't been mentioned but delaying when you start social security would be an alternative that is in effect buying a longevity annunity. It is too bad that the latest age that you can delay it to is 70. If it was higher, like 80, and it kept increasing then many people could safely plan on mostly depleting most of their retirement funds by then and having a large enough social security check starting at 80 to safely cover allmost all needs.

I like that ! Delaying SS to any arbitrary age would seem to be an excellent idea with no downside for the gov or the recipient. The only people hurt would be the Metlifes of the world.

[ndchamp]

rick,
if i am not mistaken, inflation protection is offered by american general, principal life and probably metlife as a minimum. i do not believe it is offered by integrity life or lincoln...

Yes, on Immediate Lifetime Annuities, but Longevity (deferred) Immediate Annuities.....not so much.

[skanney]

rick,
if i am not mistaken, inflation protection is offered by american general, principal life and probably metlife as a minimum. i do not believe it is offered by integrity life or lincoln...

Yes, on Immediate Lifetime Annuities, but Longevity (deferred) Immediate Annuities.....not so much.

you may be right...i am checking to see if i can find one that is inflation protected. but if you can replicate what i am looking for with a deferred annuity, who cares about the name? the problem i think will be finding a long enough term to lock in the interest rate during the accumulation period. i think you lock in the annuity payment as long as you stay with one carrier. but if you need to roll over the rate and they are suddenly unaggressive, you need to take the risk of a new annuity payment at the new insurance company. is there any solution to this?

[stlrick]

rick,
if i am not mistaken, inflation protection is offered by american general, principal life and probably metlife as a minimum. i do not believe it is offered by integrity life or lincoln...

Yes, on Immediate Lifetime Annuities, but Longevity (deferred) Immediate Annuities.....not so much.

you may be right...i am checking to see if i can find one that is inflation protected. but if you can replicate what i am looking for with a deferred annuity, who cares about the name? the problem i think will be finding a long enough term to lock in the interest rate during the accumulation period. i think you lock in the annuity payment as long as you stay with one carrier. but if you need to roll over the rate and they are suddenly unaggressive, you need to take the risk of a new annuity payment at the new insurance company. is there any solution to this?

What is the difference between a deferred annuity and a longevity annuity??? Mortality credits. The longevity annuity guarantees a payout that includes the fact that many who buy the insurance (and this should be thought of as insurance) will never collect. The deferred annuity includes no such component in the payout. Mortality credits are the whole point of SPIAs and longevity annuities. Without them, it is just a do-it-yourself SWR wrapped in fees.

[dpbsmith]

I can't seem to find the actual product description at MetLife's website... can you give us a link to it?

Here's the metlife link - quite the domain name eh?
http://www.longevityannuity.org/

That is not MetLife's website. I don't know what it is. And it has no product details but that ad-like short blurb, and an invitation to tell 'em your birth date and phone number and so forth. It does not actually tell us the name of the product.

Their MetLife link just goes straight to the general MetLife website for individuals. If I search on the phrase "advanced-life delayed annuity" I get "Your search did not match any document." longevityannuity.org, whomever or whatever it may be, does not want you to learn anything about his product from anybody but them. Despite the dot-org domain, I really have to wonder whether it's a nonprofit organization.

What I'd like to see detailed product description, fact sheet, brochure, something which lays out just what this product does, just what options are available for it--at what ages can you buy the annuity, for what ages can you choose the start date, is there a joint-and-survivor option, and, in particular, whether you can get any provisions for inflation at all. Is there a CPI-indexed option? If not, can you at least select a 3%-per-year compounded increasing option? It would be nice to see, at the very least, some representative premium illustrations.

[HueyLD]

How about this link https://www.metlife.com/assets/investme ... rantee.pdf ?

FLEXIBLE ACCESS Version

Guarantee how much income you’ll receive in the future
Plan more efficiently, knowing when you’ll receive income
Make the most of your assets (a case study)
Prepare for the unexpected

Maximum INCOME Version

Receive even more income with the Maximum Income Version
Keep track of your future income with quarterly statements
Provide a “safety net” of lifetime income with LIG

[ourbrooks]

You don't need to worry about inflation protection. Just make an estimate of what inflation will be; use 2.5% to 3.2%. If you deflate what you will receive by that amount, then it should give you a good idea of how much the annuity will give you to spend in real dollars.

But, what if inflation is much higher than predicted? Well, we're talking about a 20 year period here so the thing to be concerned about is the average inflation over 20 years, not a spike lasting a year or two. Suppose inflation turned out to be 4% as versus 2.5%. Well, your $1000 payment would be worth $460 in today's dollars as versus $610 for the lower rate. It's definitely a difference, but one most people could handle, especially since they can probably see it coming five years out. Alternately, you could plan on 4% inflation and be pleasantly surprised.

What if inflation averages 6% over the next 20 years? That's a serious, national disaster rate. Stocks might keep up but just barely; bonds certainly won't. It's one of many risks you probably can't insure against; the odds of your dying before the 20 years are up are vastly higher.

As it turns out, there are more serious problems with longevity insurance than inflation. According to this Metlife site https://www.metlife.com/assets/investments/products/annuities/Longevity-income-guarantee.pdf , the determination of the start of payout is based on just the single owner of the annuity, although the payments may be dual life. The member of a couple who is not the owner might not get anything if the owner is the one to die before age 85. The best bet is probably to split the purchase money so that each member of a couple purchases a separate annuity.

[skanney]

as i understand it, if i start a deferred annuity for 10 years and roll it over for another 10, when it comes up for the roll again i can annuitize the payments. this replicates a longevity annuity with a 20 year start period that becomes a life annuity at that time. the question is what do i do at the first roll over if the rate is not competitive? if i locked in an annuitized payment of say 5k just to make the numbers easy, if i change companies to get a better rate for the second term the annuity payment i lock in could be 3k. so i either sit for 10 years with very little income or i give up 2k/year on the annuity for the rest of my life. the question is how can i avoid this problem?

as for inflation, you need to look at more than just the history of the US from 1980 onwards. inflation became very unhinged in the 1920's & 1930's in the US. there have been periods of hyper-inflation in other countries and we are not immune. i am not going to lock up money for 30-40 years without inflation protection. even a slight increase in inflation for such a long period of time can wreak havoc on the lifestyle i am trying to protect...

[dpbsmith]

How about this link https://www.metlife.com/assets/investme ... rantee.pdf ?

Probably. I can't quite tell if that is a product in itself, or whether it is a rider that can only be obtained in conjunction with a variable annuity... or perhaps only in conjunction with a MetLife investment account of some kind. Has anyone encountered a fixed annuity that's described simply as a "guarantee?"

One thing that's clear: there is no reference to any sort of inflation adjustments, CPI-indexed or otherwise.

You don't need to worry about inflation protection. Just make an estimate of what inflation will be; use 2.5% to 3.2%.... But, what if inflation is much higher than predicted? Well, we're talking about a 20 year period here so the thing to be concerned about is the average inflation over 20 years, not a spike lasting a year or two.

In the CPI data, which start in 1913, the total cumulative inflation over 20-year periods has varied by more than a factor of four. That's a lot of uncertainty. I don't want to take that risk myself if I can avoid it. I am definitely willing to pay an insurance company to take it for me.

From 1/1913 through 11/2011 inflation has averaged 3.2%. So your suggested "2.5-3.2%" seems to be tilted in the optimistic direction. If you were to use 3.2% as a planning number, you would set aside $1.87 today in order to have $1.00 worth of buying power twenty years from now. But...

--if you set aside $1.87 in 1948, in 1968 the buying power would have been $1.30. Seriously overconservative that might be seen as an unnecessary sacrifice in retrospect.

--if you set aside $1.87 in 1967, in 1987 the buying power would have been $0.55. That, I would say, is a dangerous shortfall with real consequences to your plans.

--And just for completeness, if you set aside $1.87 in 1920, in 1940 the buying power would be, not $1, but $2.60!

No, inflation over a twenty-year period is not that stable or predictable. Historically, there's been more than a factor of four between the smallest and largest values.

In a sense, inflation uncertainty is comparable to longevity uncertainty. What do I mean by that? Very roughly, let's say life expectancy at age 65 is 20 years, so we wish to insure against the possibility of living longer than that. In reality, though, living an additional 40 years, to age 105, is pretty rare; the CDC life tables end at age 100. So, there's roughly a factor of two between the average case and the insured-against case. That's just about the same as the ratio between average 20-year total inflation and 1967-1987 total inflation.

[Mitchell777]

I like the concept very much, but I'm apprehensive about the financial strength of the insurance company 10 or 20 years out. If payments begin immediately at least I know the financial strength during some intial number of payment years. I admit a couple of prior experiences with insurance companies have made me a tad gun-shy

[nisiprius]

I like the concept very much, but I'm apprehensive about the financial strength of the insurance company 10 or 20 years out. If payments begin immediately at least I know the financial strength during some intial number of payment years. I admit a couple of prior experiences with insurance companies have made me a tad gun-shy

Sure, but how is that any different from the situation with life insurance?

[Johm221122]

I like the concept very much, but I'm apprehensive about the financial strength of the insurance company 10 or 20 years out. If payments begin immediately at least I know the financial strength during some intial number of payment years. I admit a couple of prior experiences with insurance companies have made me a tad gun-shy

It should be covered by state guaranty association
john

[hicabob]

I can't seem to find the actual product description at MetLife's website... can you give us a link to it?

Here's the metlife link - quite the domain name eh?
http://www.longevityannuity.org/

That is not MetLife's website. I don't know what it is. And it has no product details but that ad-like short blurb, and an invitation to tell 'em your birth date and phone number and so forth. It does not actually tell us the name of the product.

Their MetLife link just goes straight to the general MetLife website for individuals. If I search on the phrase "advanced-life delayed annuity" I get "Your search did not match any document." longevityannuity.org, whomever or whatever it may be, does not want you to learn anything about his product from anybody but them. Despite the dot-org domain, I really have to wonder whether it's a nonprofit organization.

What I'd like to see detailed product description, fact sheet, brochure, something which lays out just what this product does, just what options are available for it--at what ages can you buy the annuity, for what ages can you choose the start date, is there a joint-and-survivor option, and, in particular, whether you can get any provisions for inflation at all. Is there a CPI-indexed option? If not, can you at least select a 3%-per-year compounded increasing option? It would be nice to see, at the very least, some representative premium illustrations.

It is a funny little website indeed - a lot of .orgs are for profit - here's the registrant ... an insurance agent I suppose
Registrant Name:Joe Signorella
Registrant Street1:700 Commerce Drive
Registrant Street2:Suite 500
Registrant Street3:
Registrant City:Oak Brook
Registrant State/Province:Illinois
Registrant Postal Code:60523
Registrant Country:US
Registrant Phone:+1.6304959900
Registrant Phone Ext.:
Registrant FAX:+1.6304959900

[skanney]

i think the credit risk of the insurance company is a serious issue. i do not have much confidence in the state guaranty assn. aig got into trouble thinking they knew more than they did. some insurance companies have jumped into fixed income products with a "synthetic aaa." examples include putting many bbb credit instruments together with enough diversification that the rating agencies qualify it for aaa. sometimes they create mbs that are covered by other insurance companies that are rated aaa. but the insurance companies take on enormous mbs risk and when a problem does arise they lose their aaa pretty quick. there have historically been a number of "unnatural" games insurance companies have played to try to make a quick buck while maintaining their aaa rating, and you might note not very many of them have aaa ratings now. taking such long term risk with insurance companies are a serious problem. if they do cover longevity risk AND inflation risk, it might be worth trying some of this as part of a portfolio, but i was thinking in terms of a small committment to many insurance companies. i am not sure that strategy would pass muster witha rigorous credit analysis. that would be the main drawback of this approach...

[Mitchell777]

I like the concept very much, but I'm apprehensive about the financial strength of the insurance company 10 or 20 years out. If payments begin immediately at least I know the financial strength during some intial number of payment years. I admit a couple of prior experiences with insurance companies have made me a tad gun-shy

Sure, but how is that any different from the situation with life insurance?

I do not have life insurance, but if I did it would be term insurance. I would be paying premiums and know I am covered from day one. If I die tomorrow, someone gets paid. I do not need to go out 10 or 20 years for the coverage to begin to kick in. If I feel the insurance company is weakening in financial strength I can go elsewhere

[Johm221122]

i think the credit risk of the insurance company is a serious issue. i do not have much confidence in the state guaranty assn. aig got into trouble thinking they knew more than they did. some insurance companies have jumped into fixed income products with a "synthetic aaa." examples include putting many bbb credit instruments together with enough diversification that the rating agencies qualify it for aaa. sometimes they create mbs that are covered by other insurance companies that are rated aaa. but the insurance companies take on enormous mbs risk and when a problem does arise they lose their aaa pretty quick. there have historically been a number of "unnatural" games insurance companies have played to try to make a quick buck while maintaining their aaa rating, and you might note not very many of them have aaa ratings now. taking such long term risk with insurance companies are a serious problem. if they do cover longevity risk AND inflation risk, it might be worth trying some of this as part of a portfolio, but i was thinking in terms of a small committment to many insurance companies. i am not sure that strategy would pass muster witha rigorous credit analysis. that would be the main drawback of this approach...

Do you have life insurance? As was already mentioned it the same guarantee
John

[skanney]

no life insurance. if i am dead i don't need to support myself any longer...different risk...

[bsteiner]

This report (footnote 2 at page 8) says that the present value of an annuity is about 85% to 90% of the premium, or slightly less for inflation-protected annuities: Executive Office of the President, Council of Economic Advisers, “Supporting Retirement for American Families” (Feb. 2, 2012): http://www.whitehouse.gov/files/documen ... bility.pdf .

Of course, for any particular purchaser, how you would value the expected stream of payments would depend on, among other things, the interest rate you would use to discount the expected stream of payments to present value. Some people might use a low rate since the annuity payments are safe. Others might use a high rate since you can't easily sell the annuity.

In this regard, earlier this year, the IRS issued Revenue Rulings 2012-3 and 2012-4 and proposed regulations designed to facilitate the use of longevity annuities in retirement plans and IRAs:

Rev. Rul. 2012-3: http://www.irs.gov/pub/irs-drop/rr-12-03.pdf .

Rev. Rul. 2012-4: http://www.irs.gov/pub/irs-drop/rr-12-04.pdf .

Proposed regulations: https://www.federalregister.gov/article ... -contracts .

[skanney]

mathematically speaking, the exact irr can be calculated on any stream of cash flows if you know what they are. so if you know the up front cost and all of the individual montly or years payments, the irr will be an exact figure determined by a precise calculation. in a life annuity, the only uncertainty on the stream of payments is how long the recipient lives. i understand the irs publishes standard life expectancies to be used in this calculation, so i would expect the irr to be set based upon it. personally i would not place this sort of insturment in any ira other than a roth because of the minimum distributions rules would make it far too complicated for me...

[skanney]

How about this link https://www.metlife.com/assets/investme ... rantee.pdf ?

Probably. I can't quite tell if that is a product in itself, or whether it is a rider that can only be obtained in conjunction with a variable annuity... or perhaps only in conjunction with a MetLife investment account of some kind. Has anyone encountered a fixed annuity that's described simply as a "guarantee?"

One thing that's clear: there is no reference to any sort of inflation adjustments, CPI-indexed or otherwise.

You don't need to worry about inflation protection. Just make an estimate of what inflation will be; use 2.5% to 3.2%.... But, what if inflation is much higher than predicted? Well, we're talking about a 20 year period here so the thing to be concerned about is the average inflation over 20 years, not a spike lasting a year or two.

In the CPI data, which start in 1913, the total cumulative inflation over 20-year periods has varied by more than a factor of four. That's a lot of uncertainty. I don't want to take that risk myself if I can avoid it. I am definitely willing to pay an insurance company to take it for me.

From 1/1913 through 11/2011 inflation has averaged 3.2%. So your suggested "2.5-3.2%" seems to be tilted in the optimistic direction. If you were to use 3.2% as a planning number, you would set aside $1.87 today in order to have $1.00 worth of buying power twenty years from now. But...

--if you set aside $1.87 in 1948, in 1968 the buying power would have been $1.30. Seriously overconservative that might be seen as an unnecessary sacrifice in retrospect.

--if you set aside $1.87 in 1967, in 1987 the buying power would have been $0.55. That, I would say, is a dangerous shortfall with real consequences to your plans.

--And just for completeness, if you set aside $1.87 in 1920, in 1940 the buying power would be, not $1, but $2.60!

No, inflation over a twenty-year period is not that stable or predictable. Historically, there's been more than a factor of four between the smallest and largest values.

In a sense, inflation uncertainty is comparable to longevity uncertainty. What do I mean by that? Very roughly, let's say life expectancy at age 65 is 20 years, so we wish to insure against the possibility of living longer than that. In reality, though, living an additional 40 years, to age 105, is pretty rare; the CDC life tables end at age 100. So, there's roughly a factor of two between the average case and the insured-against case. That's just about the same as the ratio between average 20-year total inflation and 1967-1987 total inflation.

i forgot to address this. the increase in present value from inflation risk was a factor of 2:1. however, an increase in age for longevity risk translates into a much higher present value than inflation for the same 2:1 factor. for example, an 85 year old has an expected lifespan of about 6 more years. with roght 72k in payments, the pv at 4% is 375k or so. if his lifespan increases to 16 years, the pv goes above 800k. if you pushed it to 26 years, the pv would be 1.15m. so longevity risk is much more powerful than inflation risk short of some sort of hyper inflation...that could change things.

[bsteiner]

In response to skanney: under the proposed regulations, if you buy a longevity annuity in your traditional IRA for up to the lesser of $100,000 or 25% of the value of the IRA, it won't count in determining the required distributions: https://www.federalregister.gov/article ... -contracts . (Click on IRAs in the table of contents and it will bring you to the section dealing with IRAs.)

[skanney]

In response to skanney: under the proposed regulations, if you buy a longevity annuity in your traditional IRA for up to the lesser of $100,000 or 25% of the value of the IRA, it won't count in determining the required distributions: https://www.federalregister.gov/article ... -contracts . (Click on IRAs in the table of contents and it will bring you to the section dealing with IRAs.)

perhaps...but all income and principal paid out of this will be fully taxable in an ira. also the reason for using this product is to hedge against the risk of living longer than the tables project. if i convert the ira to a roth and pay the tax (particularly during a year when i am not working for example), then all income out of this will be nontaxable. if i outlive my life expectancy for 10 years, those 10 years of payments will be completely tax free versus completely taxable in a trad ira. i suspect the roth is significantly better for this and worthy of a conversion.

[bsteiner]

I wasn't recommending (or recommending that you not buy) a longevity annuity. I was just pointing out the new rulings and proposed regulations facilitating its use in retirement plans.

The Roth conversion is a separate issue. The Roth conversion is generally beneficial for many reasons, which I've discussed in other threads, and which I've explained in the CCH Journal of Retirement Planning where I wrote the column Retirement Plan Strategies for many years. However, (i) if you have a large IRA and aren't otherwise in a high bracket, you may want to spread the conversion out over a number of years, (ii) if you're in a high bracket now but expect to be in a lower bracket upon retirement, you may want to wait until retirement to convert (or to begin converting if you want to spread the conversion out over a number of years), and (iii) it doesn't make sense if you intend to leave your IRA to charity.

[skanney]

I wasn't recommending (or recommending that you not buy) a longevity annuity. I was just pointing out the new rulings and proposed regulations facilitating its use in retirement plans.

The Roth conversion is a separate issue. The Roth conversion is generally beneficial for many reasons, which I've discussed in other threads, and which I've explained in the CCH Journal of Retirement Planning where I wrote the column Retirement Plan Strategies for many years. However, (i) if you have a large IRA and aren't otherwise in a high bracket, you may want to spread the conversion out over a number of years, (ii) if you're in a high bracket now but expect to be in a lower bracket upon retirement, you may want to wait until retirement to convert (or to begin converting if you want to spread the conversion out over a number of years), and (iii) it doesn't make sense if you intend to leave your IRA to charity.

i agree here. but given the tax structure, i am just looking at it from a practical perspective. if you are in the top tax bracket and pay 100k to buy a 72k longevity annuity beginning age 85, you would have to drop down to the 10% tax bracket or below before it would make sense to buy the annuity in a traditional ira. the problem is with 72k in income, it would be difficult to get into the 10% tax bracket even if you inflation adjust the current tax brackets. and that's only if you live as long as the irs projects. since you are hedging against a long lifespan, it really makes no sense even there. i just find it difficult to imagine a realistic scenario where one would benefit with putting it in a trad ira vs converting to a roth and putting it there...

[Epsilon Delta]

i agree here. but given the tax structure, i am just looking at it from a practical perspective. if you are in the top tax bracket and pay 100k to buy a 72k longevity annuity beginning age 85, you would have to drop down to the 10% tax bracket or below before it would make sense to buy the annuity in a traditional ira.

I can't follow this logic and would appreciate it if you could expand.

There are a couple of ways to look at it that make the traditional IRA attractive.

First is if you are trying to put a floor under your (post tax) spending rather than maximizing expected spending. If your near the floor you are in a lower tax bracket, so putting the floor in a traditional IRA is attractive.

Second is if you are likely to have deductible medical expenses. In that case your tax rate could be zero. Speaking for myself and related to the first point I'm happy with a modest material life style, but will "splurge" on medical expenses if they have been shown to work. So for the case where I really need the money the traditional IRA is better.

[skanney]

i agree here. but given the tax structure, i am just looking at it from a practical perspective. if you are in the top tax bracket and pay 100k to buy a 72k longevity annuity beginning age 85, you would have to drop down to the 10% tax bracket or below before it would make sense to buy the annuity in a traditional ira.

I can't follow this logic and would appreciate it if you could expand.

There are a couple of ways to look at it that make the traditional IRA attractive.

First is if you are trying to put a floor under your (post tax) spending rather than maximizing expected spending. If your near the floor you are in a lower tax bracket, so putting the floor in a traditional IRA is attractive.

Second is if you are likely to have deductible medical expenses. In that case your tax rate could be zero. Speaking for myself and related to the first point I'm happy with a modest material life style, but will "splurge" on medical expenses if they have been shown to work. So for the case where I really need the money the traditional IRA is better.

as i understand it, at age 85 you are going to need to take money out of an ira at fairly aggressive rates according to the minimum distribution rules. take a 100k cost for an annuity that starts in 34 years and produces 72,028 in payments yearly. the cost of the taxes at say 35% rate is 35k if you convert to roth. then in 34 years you save say 15% taxes on the 72k for minimum 6 years. the npv at 4% is about -5k, which means there is a marginal advantage to the ira. but the tax rate for 72k in 34 years discounted at a 2% inflation rate is roughly half of the 72k, which is higher than the 34.5k for the 25% tax rate. in addition you will have social security and possibly some other income. at the 25% tax rate you will have a negative npv, meaning the roth is better. if you live 6 years longer than the irs expeceted lifespan of 6 years, you will be better in the roth even at a 10% tax rate.

[Epsilon Delta]

as i understand it, at age 85 you are going to need to take money out of an ira at fairly aggressive rates according to the minimum distribution rules. take a 100k cost for an annuity that starts in 34 years and produces 72,028 in payments yearly. the cost of the taxes at say 35% rate is 35k if you convert to roth. then in 34 years you save say 15% taxes on the 72k for minimum 6 years. the npv at 4% is about -5k, which means there is a marginal advantage to the ira. but the tax rate for 72k in 34 years discounted at a 2% inflation rate is roughly half of the 72k, which is higher than the 34.5k for the 25% tax rate. in addition you will have social security and possibly some other income. at the 25% tax rate you will have a negative npv, meaning the roth is better. if you live 6 years longer than the irs expeceted lifespan of 6 years, you will be better in the roth even at a 10% tax rate.

I'm not matching your calculations.
$72,028 * 15% = $10,840
1.04^34 = 3.79

So NPV is -$2,847 since the discounting continues in future years it takes over 16 years before the NPV reaches the -$35,000 of paying taxes now.

More importantly you are calculating the NPV of a contingent payment (The payment is contingent on you staying alive). You have to multiply by the probability of making the payment. It's possible that the probabilities all cancel, but its not clear to me that they do.

Lets do it another way.
a) You have $100,000 in a traditional IRA and buy the longevity annuity in the tIRA.
In 34 years you receive $72,028 pre tax and pay 15% tax. You get to spend $72,028 * 85% = $61,223.

b) You have $100,000 in a traditional IRA. You convert to a Roth, paying 35% taxes out of the withdrawal.
This leaves $65,000 in the Roth. You buy a longevity annuity in the Roth.
In 34 years it pays you $72,028 * 65/100 = $46,818 and of course you get to spend all of it.

But $46,818 is less than $61,223 so the traditional IRA is better.

[skanney]

"I'm not matching your calculations.
$72,028 * 15% = $10,840
1.04^34 = 3.79

So NPV is -$2,847 since the discounting continues in future years it takes over 16 years before the NPV reaches the -$35,000 of paying taxes now.

More importantly you are calculating the NPV of a contingent payment (The payment is contingent on you staying alive). You have to multiply by the probability of making the payment. It's possible that the probabilities all cancel, but its not clear to me that they do.

Lets do it another way.
a) You have $100,000 in a traditional IRA and buy the longevity annuity in the tIRA.
In 34 years you receive $72,028 pre tax and pay 15% tax. You get to spend $72,028 * 85% = $61,223.

b) You have $100,000 in a traditional IRA. You convert to a Roth, paying 35% taxes out of the withdrawal.
This leaves $65,000 in the Roth. You buy a longevity annuity in the Roth.
In 34 years it pays you $72,028 * 65/100 = $46,818 and of course you get to spend all of it.

But $46,818 is less than $61,223 so the traditional IRA is better."

there was a bug in my spreadsheet doing the npv calculation using =npv(r,a1:a50) for the rate and cash flows. that is supposed to sum up to your method of matching my calculations by addiding the individual npvs and it didn't. the problem you mentioned with my approach being the payments are contingent does not apply since you are calculating a breakeven lifespan when it makes sense to use trad vs roth ira. the method should be correct. you make a tax payment initially in order to save a series of tax payments later on and discount all of the above to present value. then you look for the break even lifespan. this is an incremental analysis of the decision to convert or not.

your method looks cleaner on the surface. it says just place the annuity in the ira dependent upon which has the lower tax rate - rate at conversion or future rate. the problem is you are buying different amounts of the annuity. to the extent the annuity is a more or less efficient investment than the irr of the tax component alone, you are implicitly comparaing apples to oranges because you have different amounts of PRETAX payments. i ran the irr of the cash flows using a 35% tax bracket for both early and later years. they showed no preference for roth vs trad using your method and mine as it should. even if you live longer than expected, they stayed the same. but if your initial forecast of the future tax rate turns out to be wrong, you will get different after tax payments in the trad ira than the roth, meaning the irr will change and you will have a preference. your method does not capture this effect because you do not look at each future payment. my method does. this is what i came up with...please show me if i am wrong...

[skanney]

i've had time to look over the bug and complete my model on roth conversions. the bug was if you have a number of years with zero cash flows in a npv(r,values) calculation and you just skip over the cells, the program assumes the years did not occur. you need to put in a zero for each cash flow for each year.

as for your method, it seems you have discovered an unusual general case of converting a trad ira to a roth. the formula for $1 of traditional ira compounded at rate r for n years with a tax rate tr is ((1+r)^n)*(1-tr). if you leave it in the ira, this is what you get. if you convert it to a roth and pay for the taxes inside of the ira account, then you get precisely the same formula. you have reduced your investment by the tax rate and are earning a tax free return going forward. so as long as the tax rate in the beginning and at the end are the same, there is absolutely no difference whatsoever if you convert to a roth or not. if the tax rate changes, you would always make the decision for the favorable tax rate. this approach is not particular to an annuity, but to every situation. this is your method.

but there are a couple of problems here. first, if you pay for the taxes with roth ira money, you may have penalties for early withdrawal. second, when you apparently have a choice between having the cash to pay for the taxes either outside of the roth or inside the roth, why would you leave the cash outside the roth? liquidity issues aside, you are foregoing the tax exempt status of the roth on the investment returns on those dollars. if instead you pay for the taxes outside the roth, you then are compounding the returns on those same amount of dollars used to pay those taxes inside the roth tax free. incrementally, if you left it in the trad ira you would have had to pay taxes on any additional returns upon withdrawal. so there is an incremental value to the roth conversion which you are missing.

so your analysis can be used only in cases where the taxes are paid with roth dollars, in which case you need to assume tax penalties. alternatively, if you pay for the taxes outside the roth, you then have that amount of money to invest in the roth, but you did not consider the returns for those dollars in your calculations.