Common Misconceptions about Safe Floor Retirement Planning

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
bobcat2
Posts: 5222
Joined: Tue Feb 20, 2007 3:27 pm
Location: just barely Outside the Beltway

Common Misconceptions about Safe Floor Retirement Planning

Post by bobcat2 » Sat Oct 03, 2015 12:22 pm

Suppose your retirement date is 2045. Your retirement strategy is floor and upside. Let's say your goal is to have $1MM in cash in 2045 to purchase a SPIA for guaranteed income. What is needed to achieve this goal?

You only need two assets: an index stock fund and 2045 TIPS. Start off in 2015 investing in the stock index fund. As time goes by, and the stocks have appreciated, start moving from the stock fund to the 2045 TIPS. Over time, move more and more into 2045 TIPS until you have reached your goal. In 2045, the TIPS matures into $1MM cash, you purchase a SPIA which is the floor for your retirement spending. Leave the remaining in the stock fund, which is the upside.
Late yesterday I became aware of a recent thread about investing in TIPS funds vs TIPS bonds where my name is involved, although I have not participated in the thread. I checked out the thread to see if I was accused of crimes I have not committed. I did not find any such accusations. :D
Link to thread -viewtopic.php?f=1&t=174265

However, I came across the above quote in the thread, which I believe deserves its own thread. I’m not including the name of the poster, who is usually quite astute, but instead only the statement, which is a viewpoint that is widely held but wrongheaded.

The above quote is referring to a basic tenet of life-cycle finance (LCF), the economics based approach to financial planning. In that approach one has two financial retirement planning goals. The lower goal is a very safe retirement income floor. The other income goal is a higher aspirational level of retirement income that we try to reach relatively safely, but typically more risk is taken to attempt to reach this higher goal.

At Bogleheads the safe floor goal is often described as the liability matched portfolio (LMP). However, what the life-cycle approach is actually doing is somewhat different in setting both floor and aspirational income goals and using an overall liability driven investment (LDI) strategy, not simply forming a sub-portfolio. The distinction is important, not simply academic grammatical nit-picking. The LDI strategy is a much safer approach to obtaining targeted floor retirement income than the LMP approach.

Let me illustrate the difference between how the LDI strategy of life-cycle finance is implemented for meeting the retirement floor income goal as opposed to how the poster uses the LMP portfolio approach. A 63 year old male wants to retire in two years at age 65 and generate from his portfolio $64,000/year in safe retirement floor income in the form of a real life annuity.

Here is his strategy using the LMP approach. Real life annuities for an age 65 male are currently yielding 6.4% payouts. Ergo, I need $1,000,000 real in safe assets in two years using as safe assets TIPS maturing in 2 yrs, or near equivalents such as T-bills, MM funds, etc. Today my portfolio consists of $940,000 in such safe assets. Interest on the assets and my future contributions to the portfolio will get me to $1,000,000 in two years. I am in good shape. (Note that this example is similar to the example in the quote with both using the $1 million goal.)

Actually the LMP people don’t think of it quite that way, they instead think of it as in the above quote and set as the goal a level of portfolio wealth at retirement, which in this case is $1,000,000 real. Either way this approach is muddled finance.

Here is the LDI life-cycle approach for the same investor. He estimates accurately that the duration of the life annuity is 15 years and that currently a life annuity delayed two years producing the income floor of $64,000 costs $940,000 - the value of his safe assets today. He also knows that when the life annuity is purchased in two years it will be more costly because it will no longer be delayed. However, interest on the assets and future portfolio contributions should cover the additional cost of the delay going away.

Because the duration of the life annuity is 15 years, the safe LDI strategy is to invest in real income assets where the duration of the assets match the duration of the annuity (the liability). This means that the invested assets are not intended to be worth $1,000,000 in two years, but rather the invested assets will be duration matched with the duration of a life annuity that will pay out $64,000/year in real income, regardless of the price of that annuity in two years. Therefore, the safe assets are a portfolio of TIPS, bonds or funds, with a duration of 15 years, not 2 years or less. The duration of these TIPS holdings will be held constant over the next 2 years until the annuity is purchased from these holdings.

The above is an extremely important distinction because the price of real annuity income moves with LT real interest changes. In this example where the life annuity has duration of 15 years, a 1% move in LT real interest rates produces a 15% change in the price of the life annuity. If LT real interest rates move up 1.5% in two years the price of the annuity drops by 22.5%, but the value of the TIPS portfolio with duration of 15 years also falls by 22.5%. If, OTOH, real LT interest rates fall by 1.5%, the price of the annuity rises by 22.5%, but the value of the TIPS portfolio with duration of 15 years also rises by 22.5%. By using the LDI strategy the $64,000/year in floor retirement income is safely preserved, regardless of changes in interest rates in either direction.

If instead the investor uses the LMP approach and interest rates rise 1.5%, the value of his safe short duration assets falls slightly, but the price of the annuity will be much lower. The investor will be able to purchase significantly more annuitized income than $64,000/year. If, on the other hand, the investor uses the LMP approach and interest rates fall 1.5%, his wealth at retirement will be slightly higher, but the price of the annuity will be much higher. In this case his ‘safe’ assets will reduce his annuity income from $64,000/year to roughly $51,000 to $52,000 per year - a very significant shortfall from his floor income goal. The LMP approach was risky, not safe, and the risk materialized.

Such large swings in LT interest rates can happen over relatively short periods in the real world. In 2007 a 65 year old male could receive a 6.4% payout from a real life annuity. Two years later in 2009 LT real interest rates had fallen significantly and the payout had fallen to about 5.0%.

When using the LMP approach to attempt to obtain a safe retirement income floor, an investor makes two critical mistakes. One mistake is trying to fit a retirement income goal into a wealth target at retirement. The other is not matching the duration of the pre-retirement assets with the duration of the income goal – the duration of the annuity. The range of possible income per year outcomes in the LMP approach is so wide that the approach fails, in any meaningful sense, to provide a safe targeted retirement income floor. Instead what is produced is guaranteed retirement income where the level of that annual income is essentially, whatever.

If instead of purchasing an annuity at retirement the investor instead builds a TIPS ladder out of wealth at retirement the same analysis holds. For the targeted retirement income stream to be safely realized, the duration of the assets pre-retirement must be matched to the duration of the retirement TIPS ladder.

I am sympathetic to those who have not realized these problems when investing to reach a safe retirement income goal. For a long time I was unaware of these problems, which are extremely important, but subtle and non-intuitive.

BobK
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.

RetiredinKaty
Posts: 112
Joined: Mon Sep 09, 2013 6:52 pm

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by RetiredinKaty » Sat Oct 03, 2015 8:18 pm

Thank you for the analysis.

The Wizard
Posts: 12159
Joined: Tue Mar 23, 2010 1:45 pm
Location: Reading, MA

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by The Wizard » Sat Oct 03, 2015 9:19 pm

I have attempted to read BobK's lengthy OP, but am unclear on the point he is trying to make.
It has something to do with annuitizing a portion of your assets to meet your basic expenses or using a TIPS ladder to do likewise.

For the benefit of readers with even less one-pass comprehension than me, it would be good if someone could post a Cliff's Notes synopsis of whatever the salient points were...
Attempted new signature...

User avatar
in_reality
Posts: 4529
Joined: Fri Jul 12, 2013 6:13 am

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by in_reality » Sat Oct 03, 2015 9:57 pm

The Wizard wrote:
For the benefit of readers with even less one-pass comprehension than me, it would be good if someone could post a Cliff's Notes synopsis of whatever the salient points were...
My take is that if you want a SPIA for guaranteed income, you have to be careful in your planning to have enough money to buy it.

If you hold short term bonds up to your purchase, you won't lose much if rates spike, but the amount you have will not be enough to buy the SPIA you planned because it's cost jumped with the rates.
Last edited by in_reality on Sat Oct 03, 2015 10:44 pm, edited 2 times in total.

User avatar
ObliviousInvestor
Posts: 3359
Joined: Tue Mar 17, 2009 9:32 am
Contact:

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by ObliviousInvestor » Sat Oct 03, 2015 10:15 pm

in_reality wrote:
The Wizard wrote:
For the benefit of readers with even less one-pass comprehension than me, it would be good if someone could post a Cliff's Notes synopsis of whatever the salient points were...
My take is that if you want a SPIA for guaranteed income, you have to be careful in your planning to have enough money to buy it.

If you hold short term bonds up to your purchase, you won't lose much if rates spike, but the amount you have will not be enough to but the SPIA you planned because it's cost jumped with the rates.
Yep. That's my understanding.

Prior to the annuity purchase, you want to:
A) match the duration of your bond holdings to the duration of the planned annuity purchase (so that price changes in the annuity will be approximately offset by value changes in the bonds), rather than
B) holding bonds set to mature as of the planned annuity purchase date.

With strategy B, an investor might think that they're safe because they have very little interest rate risk. But they actually have quite a bit of interest rate risk -- not in their portfolio but in the cost of the annuity that they need to purchase.
Mike Piper, author/blogger

User avatar
bobcat2
Posts: 5222
Joined: Tue Feb 20, 2007 3:27 pm
Location: just barely Outside the Beltway

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by bobcat2 » Sat Oct 03, 2015 10:28 pm

Inflation has nothing to do with the following – both dollars and interest rates are in real terms.

If you want a safe income floor in retirement of a certain amount per year, say $20,000/yr, thru an annuity or a TIPS ladder, the invested assets saved pre-retirement for the annuitized retirement income, must be duration matched to the duration of the annuity or the TIPS ladder.

The targeted retirement income level will not be safe, if your target is an amount of wealth at retirement. That’s because the level of income in retirement the annuity or TIPS ladder produce is sensitive to changes in interest rates.

For example, suppose I target $500,000 in wealth at retirement to purchase an annuity. Assume I will reach at retirement several years from now the $500,000 target. I don’t know how much income a $500,000 annuity will generate at retirement because that payout depends on the level of interest rates at that time, which is unknown to me today. If LT interest rates are high then, the annuity payout could be 6.8%. If LT interest rates are low then, the annuity payout could be 4.2%. That makes a huge difference as to how much guaranteed income my $500,000 annuity will produce for my retirement years.

If I want to lock in my $20,000/yr income thru an annuity or a TIPS ladder, I must match the duration of the annuity or the ladder with pre-retirement assets that have the same duration. In that way changes in interest rates will not knock me off track from achieving my $20,000/yr income goal. So if the annuity has a duration of 15 years, then the TIPS being saved to purchase the annuity must keep a duration of 15 years. You will not have invested safely for the annuity if your TIPS saved for the annuity all mature at your retirement date. Having the TIPS mature at retirement will lock in a level of wealth at retirement. But what you want to do is lock in a $20,000/yr income stream, and that requires matching the duration of the assets saved for the annuity purchase to the duration of the annuity.

BobK

This stuff is tricky. In the past I had trouble understanding it. Now I have trouble explaining it. :D
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.

User avatar
bobcat2
Posts: 5222
Joined: Tue Feb 20, 2007 3:27 pm
Location: just barely Outside the Beltway

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by bobcat2 » Sat Oct 03, 2015 11:01 pm

in_reality wrote:
The Wizard wrote:
For the benefit of readers with even less one-pass comprehension than me, it would be good if someone could post a Cliff's Notes synopsis of whatever the salient points were...
My take is that if you want a SPIA for guaranteed income, you have to be careful in your planning to have enough money to buy it.

If you hold short term bonds up to your purchase, you won't lose much if rates spike, but the amount you have will not be enough to buy the SPIA you planned because it's cost jumped with the rates.
This stuff is tricky and the above has the right feel, but the specifics are wrong. If rates spike ST bonds will decline slightly in value, but the price of the annuity will drop sharply and the annuity payout will spike. You will be able to buy a lot more annuitized income with your slightly smaller in value portfolio of bonds.

OTOH if interest rates fall sharply the ST bond portfolio will rise slightly in value, but the price of the annuity will rise sharply in price as the payout shrinks. You will be able to buy a lot less annuitized income with your slightly larger in value portfolio of bonds.

When rates rise you will be able to purchase more annuitized income than your target, but when interest rates fall you will be able to purchase less annuitized income than your target. In both cases the mismatch in duration between the ST bonds and the annuity cause the change in the amount of annuitized income available.

Duration matching of the bond portfolio to the annuity alleviates this mismatch.

BobK
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.

User avatar
in_reality
Posts: 4529
Joined: Fri Jul 12, 2013 6:13 am

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by in_reality » Sat Oct 03, 2015 11:22 pm

bobcat2 wrote: When rates rise you will be able to purchase more annuitized income than your target, but when interest rates fall you will be able to purchase less annuitized income than your target. In both cases the mismatch in duration between the ST bonds and the annuity cause the change in the amount of annuitized income available.

Duration matching of the bond portfolio to the annuity alleviates this mismatch.
Got it! Thanks for your help!! It's much appreciated!!!

itstoomuch
Posts: 5343
Joined: Mon Dec 15, 2014 12:17 pm
Location: midValley OR

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by itstoomuch » Sun Oct 04, 2015 12:59 am

I'm toast
:sharebeer
Rev012718; 4 Incm stream buckets: SS+pension; dfr'd GLWB VA & FI anntys, by time & $$ laddered; Discretionary; Rentals. LTCi. Own, not asset. Tax TBT%. Early SS. FundRatio (FR) >1.1 67/70yo

User avatar
grayfox
Posts: 4932
Joined: Sat Sep 15, 2007 4:30 am

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by grayfox » Sun Oct 04, 2015 1:48 am

bobcat2 wrote:
However, I came across the above quote in the thread, which I believe deserves its own thread. I’m not including the name of the poster, who is usually quite astute, but instead only the statement, which is a viewpoint that is widely held but wrongheaded.

...

BobK
Good topic and good explanation. I will also not name the poster of the original quote. But it is likely that he and others do not fully understand Life-cycle Investing and have a few misconceptions. Some of this subject has subtle details and implications which require further thinking.

One thing that I am thinking is that, instead of TIPS now and SPIA at retirement, the best option might be to purchase a Deferred Income Annuity (DIA). That way you are buying a known future income stream and don't have to worry about the complexities of duration matching.

User avatar
in_reality
Posts: 4529
Joined: Fri Jul 12, 2013 6:13 am

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by in_reality » Sun Oct 04, 2015 2:18 am

grayfox wrote: But it is likely that he and others do not fully understand Life-cycle Investing and have a few misconceptions. Some of this subject has subtle details and implications which require further thinking.
Well clearly I did not not understand it, and so am grateful for the post.

I was of the mind to shorten duration so as to not be at risk of loss from rising rates.

It never occurred to me that SPIA pricing moved as bonds so that matching the SPIA's duration prior to purchase would cover me from rate movements.

For funds I hold now that I envision going into a SPIA, I wonder if I should start matching duration some years (15-17) before retirement. It's a much longer duration than I currently hold.

Even funds with a 20 year maturity only seem to have a 10 year duration. If I am really to match duration, I'll have to think about how to best do it...

Suggestions?
Last edited by in_reality on Sun Oct 04, 2015 2:19 am, edited 1 time in total.

User avatar
White Coat Investor
Posts: 13379
Joined: Fri Mar 02, 2007 9:11 pm
Location: Greatest Snow On Earth

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by White Coat Investor » Sun Oct 04, 2015 2:19 am

ObliviousInvestor wrote:
in_reality wrote:
The Wizard wrote:
For the benefit of readers with even less one-pass comprehension than me, it would be good if someone could post a Cliff's Notes synopsis of whatever the salient points were...
My take is that if you want a SPIA for guaranteed income, you have to be careful in your planning to have enough money to buy it.

If you hold short term bonds up to your purchase, you won't lose much if rates spike, but the amount you have will not be enough to but the SPIA you planned because it's cost jumped with the rates.
Yep. That's my understanding.

Prior to the annuity purchase, you want to:
A) match the duration of your bond holdings to the duration of the planned annuity purchase (so that price changes in the annuity will be approximately offset by value changes in the bonds), rather than
B) holding bonds set to mature as of the planned annuity purchase date.

With strategy B, an investor might think that they're safe because they have very little interest rate risk. But they actually have quite a bit of interest rate risk -- not in their portfolio but in the cost of the annuity that they need to purchase.
No wonder your books sell so well.
1) Invest you must 2) Time is your friend 3) Impulse is your enemy | 4) Basic arithmetic works 5) Stick to simplicity 6) Stay the course

sadie wess
Posts: 262
Joined: Wed Jan 19, 2011 9:42 am

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by sadie wess » Sun Oct 04, 2015 6:11 am

Very helpful posting indeed.

Thank you.

BTW, what is the book that Emergdoc is talking about?

User avatar
VictoriaF
Posts: 18525
Joined: Tue Feb 27, 2007 7:27 am
Location: Black Swan Lake

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by VictoriaF » Sun Oct 04, 2015 6:35 am

Let's assume that in 2015 a 63-year old man wants to perform a liability driven investment (LDI) strategy of Bob's example in the OP. Does it mean that the man would buy 2 individual 15-year TIPS, one maturing in 2030 and the other one in 2031, and sell both of them prematurely in 2017 to buy an annuity? The rise or fall in the individual TIPS selling price would reflect the intervening changes in the interest rates. However, it goes against a common advice to buy individual TIPS and hold them to maturity. Please clarify this point.

Victoria
WINNER of the 2015 Boglehead Contest. | Every joke has a bit of a joke. ... The rest is the truth. (Marat F)

Dandy
Posts: 5287
Joined: Sun Apr 25, 2010 7:42 pm

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by Dandy » Sun Oct 04, 2015 8:41 am

Interesting point about the interest rate risk in the planned future annuity and to offset that risk by matching the duration of the fixed income pot intended to fund that future annuity purchase to the time frame of that annuity purchase. People are often so confused and uninformed about annuities that they often don't consider how changing interest rates affect annuities. Almost no chance of them coming to understand the above on their own.

User avatar
bobcat2
Posts: 5222
Joined: Tue Feb 20, 2007 3:27 pm
Location: just barely Outside the Beltway

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by bobcat2 » Sun Oct 04, 2015 9:09 am

Dandy wrote:Interesting point about the interest rate risk in the planned future annuity and to offset that risk by matching the duration of the fixed income pot intended to fund that future annuity purchase to the time frame of that annuity purchase. People are often so confused and uninformed about annuities that they often don't consider how changing interest rates affect annuities. Almost no chance of them coming to understand the above on their own.
How very true! :(
People who claim personal finance is easy do a disservice IMO. It's not easy. Investing safely for a targeted safe level of retirement income is one area of personal finance that isn't easy at all.

BobK
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.

User avatar
bobcat2
Posts: 5222
Joined: Tue Feb 20, 2007 3:27 pm
Location: just barely Outside the Beltway

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by bobcat2 » Sun Oct 04, 2015 9:48 am

grayfox wrote:Good topic and good explanation. I will also not name the poster of the original quote. But it is likely that he and others do not fully understand Life-cycle Investing and have a few misconceptions. Some of this subject has subtle details and implications which require further thinking.
No kidding! :D

IMO here's a good place to start.
Link to "Applying life-cycle economics"
https://www.nestpensions.org.uk/schemew ... cs,PDF.pdf

A key quote from early in the article.
The effectiveness of the post retirement design is not independent of the accumulation period design. In this respect, there should be a continuity concept and strategy from saving for retirement through dissaving in retirement.
BobK
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.

User avatar
ObliviousInvestor
Posts: 3359
Joined: Tue Mar 17, 2009 9:32 am
Contact:

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by ObliviousInvestor » Sun Oct 04, 2015 11:54 am

sadie wess wrote:BTW, what is the book that Emergdoc is talking about?
I have a series of short books that attempt to explain various financial topics clearly and concisely, much as I was trying to do above.

I do not have any books on the topic of this thread, however. For that, we have BobK/bobcat2 to keep us informed. :)

And on that note, thank you, Bob, for this thread. This isn't a topic I've heard anybody bring up before.
Mike Piper, author/blogger

User avatar
bobcat2
Posts: 5222
Joined: Tue Feb 20, 2007 3:27 pm
Location: just barely Outside the Beltway

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by bobcat2 » Sun Oct 04, 2015 1:42 pm

VictoriaF wrote:Let's assume that in 2015 a 63-year old man wants to perform a liability driven investment (LDI) strategy of Bob's example in the OP. Does it mean that the man would buy 2 individual 15-year TIPS, one maturing in 2030 and the other one in 2031, and sell both of them prematurely in 2017 to buy an annuity? The rise or fall in the individual TIPS selling price would reflect the intervening changes in the interest rates. However, it goes against a common advice to buy individual TIPS and hold them to maturity. Please clarify this point.

Brief Answer –

I think the easiest way to do this would be to purchase a long duration TIPS fund or ETF and a short duration TIPS fund or ETF. Invest in the two funds so that the weighted average duration of the funds is equal to the duration of the life annuity. If no TIPS fund has duration as long as the annuity’s duration, invest in the longest duration TIPS fund you can find.


Longer Answer & Additional Thoughts –
I’ve assumed the duration of the annuity is 15 years. That’s probably not a bad generic estimate. But a more precise estimate of a particular life annuity’s duration depends upon at least
- the gender of the beneficiary
- the age the payouts commence
- the prevailing current level of LT interest rates

Absent access to a financial engineer knowledgeable in this area, a precise estimate will be difficult to come by. Individual investors are stuck with the generic estimate, assuming the annuity begins in their 60s.

Investing safely for floor retirement income requires more than thinking in terms of a liability matched portfolio (LMP). You don’t want to be thinking in portfolio terms. You want to be thinking in income terms. Meeting a targeted level of retirement income with high probability requires investing in safe assets during pre-retirement. For those assets to be safe they need to be duration matched to the duration of the financial products generating the targeted retirement income stream – whether that product is an annuity, or a TIPS ladder, or a weighted average of TIPS funds replicating the ladder with the same duration as the ladder.

Safe pre-retirement assets for a life annuity or a TIPS ladder to be purchased at retirement will not look safe from a portfolio view point. These fixed income assets will have long duration and their portfolio values will be bouncing all over the place as interest rates change. But in terms of retirement income units they will be very safe. ST fixed income assets will look safe from a portfolio POV. In terms of targeted safe retirement income they are risky.

Thinking in terms of portfolio values pre-retirement can be extremely misleading. An increase in portfolio wealth from high investment returns resulting from falling interest rates looks good from a portfolio POV. But lower interest rates for the investor seeking safe retirement income means the investor’s situation has gotten worse, not better.

BobK
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.

itstoomuch
Posts: 5343
Joined: Mon Dec 15, 2014 12:17 pm
Location: midValley OR

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by itstoomuch » Sun Oct 04, 2015 2:28 pm

in_reality wrote:
grayfox wrote: But it is likely that he and others do not fully understand Life-cycle Investing and have a few misconceptions. Some of this subject has subtle details and implications which require further thinking.
Well clearly I did not not understand it, and so am grateful for the post.

I was of the mind to shorten duration so as to not be at risk of loss from rising rates.

It never occurred to me that SPIA pricing moved as bonds so that matching the SPIA's duration prior to purchase would cover me from rate movements.

For funds I hold now that I envision going into a SPIA, I wonder if I should start matching duration some years (15-17) before retirement. It's a much longer duration than I currently hold.

Even funds with a 20 year maturity only seem to have a 10 year duration. If I am really to match duration, I'll have to think about how to best do it...

Suggestions?
The solutions we have come look at are Either options or the prepackaged deferred GLWB annuities coupled with Indexes and trading accounts. YMMV.
Disclaimers: 65/68. On SS but not pulling from personal retirement assets. No debt. Have LTCi.
Rev012718; 4 Incm stream buckets: SS+pension; dfr'd GLWB VA & FI anntys, by time & $$ laddered; Discretionary; Rentals. LTCi. Own, not asset. Tax TBT%. Early SS. FundRatio (FR) >1.1 67/70yo

User avatar
VictoriaF
Posts: 18525
Joined: Tue Feb 27, 2007 7:27 am
Location: Black Swan Lake

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by VictoriaF » Sun Oct 04, 2015 3:13 pm

bobcat2 wrote:Investing safely for floor retirement income requires more than thinking in terms of a liability matched portfolio (LMP). You don’t want to be thinking in portfolio terms. You want to be thinking in income terms.
...
Safe pre-retirement assets for a life annuity or a TIPS ladder to be purchased at retirement will not look safe from a portfolio view point.

BobK
Thank you, Bob. It's an interesting distinction between LDI and LMP, and particularly that different things look "safe" and "unsafe" in them.
ObliviousInvestor wrote:I have a series of short books that attempt to explain various financial topics clearly and concisely, much as I was trying to do above.

I do not have any books on the topic of this thread, however. For that, we have BobK/bobcat2 to keep us informed. :)

And on that note, thank you, Bob, for this thread. This isn't a topic I've heard anybody bring up before.
I think Bob could turn some of his posts into short books, too. Right?

Victoria
WINNER of the 2015 Boglehead Contest. | Every joke has a bit of a joke. ... The rest is the truth. (Marat F)

User avatar
ObliviousInvestor
Posts: 3359
Joined: Tue Mar 17, 2009 9:32 am
Contact:

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by ObliviousInvestor » Sun Oct 04, 2015 3:28 pm

VictoriaF wrote:I think Bob could turn some of his posts into short books, too. Right?
I can't tell if you're asking me or asking Bob. If you're asking me, I would say: yes, absolutely, if he was so inclined.
Mike Piper, author/blogger

User avatar
bertilak
Posts: 6053
Joined: Tue Aug 02, 2011 5:23 pm
Location: East of the Pecos, West of the Mississippi

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by bertilak » Sun Oct 04, 2015 3:32 pm

There is one concept I can't get past. What is the duration of an SPIA?

How about an example showing dates of events and the appropriate duration(s)?
May neither drought nor rain nor blizzard disturb the joy juice in your gizzard. -- Squire Omar Barker, the Cowboy Poet

User avatar
VictoriaF
Posts: 18525
Joined: Tue Feb 27, 2007 7:27 am
Location: Black Swan Lake

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by VictoriaF » Sun Oct 04, 2015 3:35 pm

ObliviousInvestor wrote:
VictoriaF wrote:I think Bob could turn some of his posts into short books, too. Right?
I can't tell if you're asking me or asking Bob. If you're asking me, I would say: yes, absolutely, if he was so inclined.

I was asking both of you. You have the experience to assess the suitability of Bob's information for this media, and Bob knows his inclinations. Thank you for a quick response!

Victoria
WINNER of the 2015 Boglehead Contest. | Every joke has a bit of a joke. ... The rest is the truth. (Marat F)

jalbert
Posts: 3585
Joined: Fri Apr 10, 2015 12:29 am

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by jalbert » Sun Oct 04, 2015 6:28 pm

There is still an important detail missing to get this right.  The liabilities you are trying to cover are the future cash flows, not the price of the annuity.   If, in 2 years, you will purchase an annuity with a 15 year duration, then the liability-matching portfolio has a 17-year duration at that time, not 15 years.  At 15 years, the bond portfolio won't keep up with the annuity price if rates fall.

The duration you need at any point is:

(life expectancy minus age)

And both life expectancy and age change yearly. 

-jalbert

heyyou
Posts: 3095
Joined: Tue Feb 20, 2007 4:58 pm

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by heyyou » Sun Oct 04, 2015 6:50 pm

BobK
This stuff is tricky. In the past I had trouble understanding it. Now I have trouble explaining it.
bertilak
How about an example showing dates of events and the appropriate duration(s)?
Please.

SGM
Posts: 2707
Joined: Wed Mar 23, 2011 4:46 am

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by SGM » Sun Oct 04, 2015 7:04 pm

I am also confused about an SPIA with a duration of 15 years. I expect to purchase an SPIA that will payout over a lifetime that may last much longer than 15 years. Is there another meaning of duration?

User avatar
alec
Posts: 2920
Joined: Fri Mar 02, 2007 2:15 pm

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by alec » Sun Oct 04, 2015 7:32 pm

SGM wrote:I am also confused about an SPIA with a duration of 15 years. I expect to purchase an SPIA that will payout over a lifetime that may last much longer than 15 years. Is there another meaning of duration?
See the annuity duration puzzle.

http://papers.ssrn.com/sol3/papers.cfm? ... id=2021579
"It is difficult to get a man to understand something, when his salary depends upon his not understanding it!" - Upton Sinclair

jalbert
Posts: 3585
Joined: Fri Apr 10, 2015 12:29 am

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by jalbert » Sun Oct 04, 2015 7:36 pm

When you buy an annuity, you are using principal and future interest to fund your liabilities until you reach the age of your life expectancy on the day of purchase. If you live longer, subsequent cash flows are provided from unused assets in your actuarial pool that would have covered liabilities of annuitants who did not live to their expected age. You get these liabilities covered if you live past your life expectancy in exchange for your heirs giving up your unused credits if you die before reaching your expected age of death.

So the liabilities you need to cover are from the present until your life expectency age, hence the liabilities have an effective duration of (life expectancy minus age).

One often overlooked service an insurance co. is providing when you buy an annuity is managing the annuitant pool bond portfolio to match the liabilities of the pool.

Delaying SS benefits emulates buying an inflation-adjusted annuity using the funds that would have been spent on the annuity to cover liabilities until the delayed SS benefit starts. Delaying SS usually wins when interest rates are low. This also changes the calculus of the liabilities you need to match.

-jalbert

User avatar
bobcat2
Posts: 5222
Joined: Tue Feb 20, 2007 3:27 pm
Location: just barely Outside the Beltway

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by bobcat2 » Sun Oct 04, 2015 10:05 pm

I came up with 15 years duration for a generic life annuity in an informal why which I suspect is fairly accurate. In one of Robert Merton's talks that you can access on the internet he briefly discusses annuity duration while talking about the DC retirement plan at DFA that is his brainchild. He mentions that a life annuity at age 65 typically has a duration of about 15 years. He talked as if this a good default estimate until you know more about the specific situation. I figured if it was a good enough rule of thumb for Robert, it was a good enough rule of thumb for me. :D

BobK
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.

jalbert
Posts: 3585
Joined: Fri Apr 10, 2015 12:29 am

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by jalbert » Sun Oct 04, 2015 10:30 pm

But we do know more about the specific situation. The person is planning the annuity purchase 2 years in the future, and is 63. They will be 65 when they buy the annuity in 2 years. What would be the duration of a deferred income annuity purchased at age 63 to start paying benefits at age 65? That is what is being emulated.

The analogy would be that you are today age 63, and at age 65, you plan to buy a zero coupon bond to fund a liability that occurs 15 years after that. The fact that in 2 years you need to buy a 15 year bond to fund the liability that occurs 15 years in the future, doesn't mean you need a 15 year bond today.

-jalbert

jalbert
Posts: 3585
Joined: Fri Apr 10, 2015 12:29 am

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by jalbert » Sun Oct 04, 2015 11:06 pm

in_reality wrote:
The Wizard wrote:
For the benefit of readers with even less one-pass comprehension than me, it would be good if someone could post a Cliff's Notes synopsis of whatever the salient points were...
My take is that if you want a SPIA for guaranteed income, you have to be careful in your planning to have enough money to buy it.

If you hold short term bonds up to your purchase, you won't lose much if rates spike, but the amount you have will not be enough to buy the SPIA you planned because it's cost jumped with the rates.
Actually, the problem with the short duration portfolio is if rates fall. The appreciation of the shorter portfolio won't keep up with the larger increase in cost of the annuity. If rates spike, you come out ahead with the shorter duration bonds because the annuity price falls further than the bond portfolio value.

Taxes on the bond interest and any capital gain at time of annuity purchase are another issue.

-jalbert

User avatar
#Cruncher
Posts: 2652
Joined: Fri May 14, 2010 2:33 am
Location: New York City
Contact:

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by #Cruncher » Mon Oct 05, 2015 12:32 pm

bertilak in [url=https://www.bogleheads.org/forum/viewtopic.php?p=2645342#p2645342]this post[/url] wrote:There is one concept I can't get past. What is the duration of an SPIA? How about an example showing dates of events and the appropriate duration(s)?
Using this longevity estimator with the SSA 2011 Period Life Table I calculated the present value of a $1,000 per month life annuity for a 65 year old male at 1.0% and 0.9% discount rates. It shows the present value increasing 1.03%. This corresponds to a modified duration of 10.3%.

Code: Select all

PV @ 1.0%     $190,679
PV @ 0.9%     $192,648
% Change         1.03%
% Change X 10   10.3%
Assume one is a 63 year old male planning to buy a CPI-indexed life annuity in two years. One could hedge against falling real interest rates by buying today a TIPS that will also have about a 10.3% modified duration in two years. The 2.5% Jan 2029 TIPS comes pretty close. Using the yield from the WSJ TIPS Quotes 10/2/2015 but with a settlement date two years from now, the modified duration is 10.0%.

Code: Select all

10.0% =MDURATION(DATE(2017, 10, 5), DATE(2029, 1, 15), 2.5%, 0.748%, 2, 1) / 100
Excel MDURATION function.

User avatar
bobcat2
Posts: 5222
Joined: Tue Feb 20, 2007 3:27 pm
Location: just barely Outside the Beltway

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by bobcat2 » Mon Oct 05, 2015 3:02 pm

#Cruncher wrote:Using this longevity estimator with the SSA 2011 Period Life Table I calculated the present value of a $1,000 per month life annuity for a 65 year old male at 1.0% and 0.9% discount rates. It shows the present value increasing 1.03%. This corresponds to a modified duration of 10.3%.

Code: Select all

PV @ 1.0%     $190,679
PV @ 0.9%     $192,648
% Change         1.03%
% Change X 10   10.3%
Assume one is a 63 year old male planning to buy a life annuity in two years. One could hedge against falling interest rates by buying today a bond that will also have about a 10.3% modified duration in two years. The 2.5% Jan 2029 TIPS comes pretty close. Using the yield from the WSJ TIPS Quotes 10/2/2015 but with a settlement date two years from now, the modified duration is 10.0%.

Code: Select all

10.0% =MDURATION(DATE(2017, 10, 5), DATE(2029, 1, 15), 2.5%, 0.748%, 2, 1) / 100
Excel MDURATION function.
#Cruncher's 10.3 estimated duration calculation for a 65 year old male looks reasonable to me. :sharebeer
This means I'm saying goodbye to the 15 year duration rule of thumb I espoused yesterday. :wink:

Bobk
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.

Gattamelata
Posts: 189
Joined: Mon May 05, 2014 7:41 pm

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by Gattamelata » Mon Oct 05, 2015 4:17 pm

bobcat2 wrote: This stuff is tricky. In the past I had trouble understanding it. Now I have trouble explaining it. :D
Thanks, Bob, your efforts are sincerely appreciated. I had never considered duration matching before I saw you mention it.

One way of characterizing the value occurred to me as I was reading your OP: Similar to the way that bogleheads and other indexers attempt to capture the market return by indexing, duration matching is designed to capture the interest rate at retirement (or at annuity inception) by matching the duration of their funds earmarked for annuity purchase to the anticipated duration of their annuity. In both cases it's attempting to eliminate the impacts of variable market conditions by tethering ones own financial plan to whatever conditions prevail.

Does that seem roughly accurate to you? It helped me to understand what you were going for, at any rate. (Har! "Any rate.")

Rodc
Posts: 13601
Joined: Tue Jun 26, 2007 9:46 am

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by Rodc » Mon Oct 05, 2015 8:07 pm

Very interesting.

Thanks.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.

User avatar
alec
Posts: 2920
Joined: Fri Mar 02, 2007 2:15 pm

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by alec » Mon Oct 05, 2015 8:42 pm

bobcat2 wrote:
#Cruncher wrote:Using this longevity estimator with the SSA 2011 Period Life Table I calculated the present value of a $1,000 per month life annuity for a 65 year old male at 1.0% and 0.9% discount rates. It shows the present value increasing 1.03%. This corresponds to a modified duration of 10.3%.

Code: Select all

PV @ 1.0%     $190,679
PV @ 0.9%     $192,648
% Change         1.03%
% Change X 10   10.3%
Assume one is a 63 year old male planning to buy a life annuity in two years. One could hedge against falling interest rates by buying today a bond that will also have about a 10.3% modified duration in two years. The 2.5% Jan 2029 TIPS comes pretty close. Using the yield from the WSJ TIPS Quotes 10/2/2015 but with a settlement date two years from now, the modified duration is 10.0%.

Code: Select all

10.0% =MDURATION(DATE(2017, 10, 5), DATE(2029, 1, 15), 2.5%, 0.748%, 2, 1) / 100
Excel MDURATION function.
#Cruncher's 10.3 estimated duration calculation for a 65 year old male looks reasonable to me. :sharebeer
This means I'm saying goodbye to the 15 year duration rule of thumb I espoused yesterday. :wink:

Bobk
Maybe the difference comes from insurance companies assuming some adverse selection in that only healthy people who expect to live a long time buy life annuities.
"It is difficult to get a man to understand something, when his salary depends upon his not understanding it!" - Upton Sinclair

Snowjob
Posts: 1482
Joined: Sun Jun 28, 2009 10:53 pm

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by Snowjob » Tue Oct 06, 2015 12:05 pm

ObliviousInvestor wrote:
in_reality wrote:
The Wizard wrote:
For the benefit of readers with even less one-pass comprehension than me, it would be good if someone could post a Cliff's Notes synopsis of whatever the salient points were...
My take is that if you want a SPIA for guaranteed income, you have to be careful in your planning to have enough money to buy it.

If you hold short term bonds up to your purchase, you won't lose much if rates spike, but the amount you have will not be enough to but the SPIA you planned because it's cost jumped with the rates.
Yep. That's my understanding.

Prior to the annuity purchase, you want to:
A) match the duration of your bond holdings to the duration of the planned annuity purchase (so that price changes in the annuity will be approximately offset by value changes in the bonds), rather than
B) holding bonds set to mature as of the planned annuity purchase date.

With strategy B, an investor might think that they're safe because they have very little interest rate risk. But they actually have quite a bit of interest rate risk -- not in their portfolio but in the cost of the annuity that they need to purchase.

Maybe some of this affect can explain the questions raised in this thread :D --

viewtopic.php?f=10&t=175148&newpost=2647068

Author poses the question -- if people are targeting certain amount of wealth and asset prices appreciate, why are we not seeing increased spending as less savings will be needed to reach that target?

User avatar
Kevin M
Posts: 10002
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by Kevin M » Tue Oct 06, 2015 2:05 pm

Great post!
bobcat2 wrote: But in terms of retirement income units they will be very safe. ST fixed income assets will look safe from a portfolio POV. In terms of targeted safe retirement income they are risky.
Is this statement within the context of a pre-retirement portfolio targeted for purchasing an LMP at some point in the future, in which case I understand it and I think it's impossible to refute?

However, I believe the common wisdom is that short-term fixed-income is pretty safe in terms of real income, as a substitute for a more traditional LMP, since nominal rates tend to track inflation pretty well. Even the likes of Fama and French have stated this.

So with ST FI, your nominal rate will tend to move with inflation, thus your real income will be fairly safe, although not as safe as with a TIPS ladder or SPIA. So although there is risk with respect to purchasing power for a TIPS ladder or SPIA, isn't there less risk with respect to longer-term safe income?

To put it another way, even though the portfolio value of ST FI does not move with the price of the TIPS ladder or SPIA, the longer-term real income from the ST FI portfolio should be about as safe (but not quite) as that from the ladder or SPIA, but without the large potential changes in portfolio value.

This does seem a bit paradoxical, so looking forward to your insights, bobcat2!

Thanks,

Kevin
Wiki ||.......|| Suggested format for Asking Portfolio Questions (edit original post)

User avatar
bobcat2
Posts: 5222
Joined: Tue Feb 20, 2007 3:27 pm
Location: just barely Outside the Beltway

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by bobcat2 » Tue Oct 06, 2015 3:36 pm

Kevin M wrote:
bobcat2 wrote:But in terms of retirement income units they will be very safe. ST fixed income assets will look safe from a portfolio POV. In terms of targeted safe retirement income they are risky.
Is this statement within the context of a pre-retirement portfolio targeted for purchasing an LMP at some point in the future, in which case I understand it and I think it's impossible to refute?

However, I believe the common wisdom is that short-term fixed-income is pretty safe in terms of real income, as a substitute for a more traditional LMP, since nominal rates tend to track inflation pretty well. Even the likes of Fama and French have stated this.

So with ST FI, your nominal rate will tend to move with inflation, thus your real income will be fairly safe, although not as safe as with a TIPS ladder or SPIA. So although there is risk with respect to purchasing power for a TIPS ladder or SPIA, isn't there less risk with respect to longer-term safe income?

To put it another way, even though the portfolio value of ST FI does not move with the price of the TIPS ladder or SPIA, the longer-term real income from the ST FI portfolio should be about as safe (but not quite) as that from the ladder or SPIA, but without the large potential changes in portfolio value.
Hi Kevin,
It's possible I misunderstand portions of your post, but assuming I don't, I am in disagreement with parts of it.
Kevin M wrote:[When] this statement [is] within the context of a pre-retirement portfolio targeted for purchasing an LMP at some time in the future, ... I understand it ...
This is not for the purpose of purchasing a liability matched portfolio (LMP) at some point in the future. It is for the purpose of securing a targeted level of income at some time in the future. That distinction makes all the difference. A major point in all this is thinking in terms of a targeted level of portfolio assets (LMP) at some future date is itself a serious error. You don't care what the portfolio asset level is, you care about hitting the income target. The level of safe assets is irrelevant.
I believe the common wisdom is that short-term fixed-income is pretty safe in terms of real income, as a substitute for a more traditional LMP, since nominal rates tend to track inflation pretty well.
In retirement the liability is consumption, which is always real not nominal. Therefore income, the financial asset that supports the consumption, needs to be real, at least for the safe floor, IMO. We are not trying to track safe floor income "pretty well". We are instead attempting to meet our targeted safe income floor as best we can.

BobK
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.

User avatar
Artsdoctor
Posts: 3448
Joined: Thu Jun 28, 2012 3:09 pm
Location: Los Angeles, CA

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by Artsdoctor » Tue Oct 06, 2015 4:40 pm

"This is not for the purpose of purchasing a liability matched portfolio (LMP) at some point in the future. It is for the purpose of securing a targeted level of income at some time in the future. That distinction makes all the difference. A major point in all this is thinking in terms of a targeted level of portfolio assets (LMP) at some future date is itself a serious error. You don't care what the portfolio asset level is, you care about hitting the income target. The level of safe assets is irrelevant."

Break it down. "It is for the purpose of (i.e, 'the goal is') securing a targeted level of income at some time in the future." Yes, the only question is how much this is going to cost you. Some of us have individual TIPS maturing each year we need them to spend money (essential living expenses). Forgeting about the lapses of some years because TIPS don't mature in every year, we may have 15-20 years where TIPS are maturing and we're spending the money generated from maturing TIPS and regular coupons to fund our financial needs. However, my TIPS ladder going from 2025 to 2044 and paying out $50,000 per year may have cost me a completely different sum than someone else because we've purchased it at different times and at different prices.

If you're following the traditional investing methods in hopes of one day purchasing the entire "secured level of income" in one fell swoop (say, at age 65 when you retire), that's very risky because you have no way of knowing what that amount is going to be. what that annuity is going to yield, or how many TIPS you'll be able to purchase. In another words, the amount you have at the time you're wanting to "purchase" your income stream will decide the income stream, and this figure may be completely different than what you want or need.

"A major point ... is thinking in terms of a targeted level of portfolio assets (LMP) at some future date is itself a serious error." True, because you have no way of knowing what that targeted level is going to buy you. "You care about hitting the income target." Yes, because the question is whether or not you can afford to buy that income stream. This is presumably why you should be purchasing your income stream while in your working years (perhaps 10 years prior to retirement) because you have to ability to adjust your purchases according to what price that income stream costs you. If you wait until you have little to no new money coming in, you're going to have to purchase what the market tells you you can purchase. The longer you wait to buy that income stream, the more risk you're assuming because the less control you have.

Bobcat, Is this what you're trying to say?

User avatar
Kevin M
Posts: 10002
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by Kevin M » Tue Oct 06, 2015 7:17 pm

#Cruncher wrote:Assume one is a 63 year old male planning to buy a CPI-indexed life annuity in two years. One could hedge against falling real interest rates by buying today a TIPS that will also have about a 10.3% modified duration in two years.
Why? The duration risk is for the two years starting now, so why wouldn't you want a security with the same current duration as the current duration of the SPIA?
Using the yield from the WSJ TIPS Quotes 10/2/2015 but with a settlement date two years from now, the modified duration is 10.0%.
How does shortening the maturity by two years but using the current yield for the original YTM produce a meaningful result? We don't know what the yield will be two years from now, so we can't know what the duration will be two years from now.

:confused

So back to the original point, I'd think you'd just want to buy a TIPS that has a correctly calculated current duration equal to the current duration of the SPIA. But that doesn't really accomplish the objective either, since the duration of both will change over the next two years.

Based on some of his other posts, I'd think that bobcat2's solution would be to use long-term and short-term TIPS funds, and rebalance 4-6 times per year to keep the duration matched to the duration of the SPIA. Using funds for this objective makes more sense to me than using funds to build an LMP (oops, I meant "to secure a targeted level of income") now.

Kevin
Wiki ||.......|| Suggested format for Asking Portfolio Questions (edit original post)

itstoomuch
Posts: 5343
Joined: Mon Dec 15, 2014 12:17 pm
Location: midValley OR

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by itstoomuch » Tue Oct 06, 2015 9:12 pm

Time for the wrench;
The anunity company also has very similar risks. Consequently, in the attempt to minimize the long-term risks, it will try to price the SPIA higher than the current data would suggest.

So what about price competition among the insurers? Let's assume annuity company has a good portion, most or even all of annuities. Does this say that the annuity company under priced. Or other annuity cos over priced? Or just don't want or need the risk in their holdings, at this time. ?
IMO, SPIA and TIPS are overly pushed to many who don't understand the risks.
YMMV. :annoyed :beer
Last edited by itstoomuch on Wed Oct 07, 2015 1:17 am, edited 1 time in total.
Rev012718; 4 Incm stream buckets: SS+pension; dfr'd GLWB VA & FI anntys, by time & $$ laddered; Discretionary; Rentals. LTCi. Own, not asset. Tax TBT%. Early SS. FundRatio (FR) >1.1 67/70yo

User avatar
bobcat2
Posts: 5222
Joined: Tue Feb 20, 2007 3:27 pm
Location: just barely Outside the Beltway

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by bobcat2 » Tue Oct 06, 2015 9:42 pm

Hi Kevin and #cruncher,

I believe the duration is for the time the annuity is generating an income stream. It is not a hybrid combining delay years with no income with the years of annuitized income.

What you want to do is essentially mark to market. If you want the annuity to start safely in two years (by purchasing an immediate life annuity at age 65) price today a two year delayed annuity for a 63 year old. That should be the level of of duration matched assets you are currently holding in order to be fully funded. (The duration is the duration of the immediate life annuity at age 65.) Now price a one year delayed annuity for a 64 year old. That will be more expensive because there is only one year of delay. The difference in price of the two delayed annuities is the amount interest on the duration matched assets and future contributions to those safe assets need to be over the next year in order to be fully funded next year.

In one year check to see if you are fully funded at age 64. You will be fully funded next year if the amount of duration matched assets you are holding then is equal to the price of a one year delayed annuity at age 64. Now repeat the previous process. This time for the difference in price of a one year delayed annuity at age 64 to the price of an immediate annuity at age 65. The difference in price of the one year delayed annuity and the immediate annuity at age 65 is the amount interest on the duration matched assets and future contributions to those safe assets need to be over that year in order to be fully funded at age 65 to purchase the immediate life annuity safely.

One thing not accounted for in the above is the change in life expectancy at age 65 over time. That will be minor, however, if the LT averages hold - probably about six weeks over the two years. That would lengthen the duration of the annuity by a small amount. You could redo the duration calculation next year and estimate it for the second. I doubt this is worth it, unless a terrible pandemic or war hits the US in the next two years. And in either of those cases you will have bigger things to worry about than price changes in life annuities. :wink:

Kevin wrote.
So back to the original point, I'd think you'd just want to buy a TIPS that has a correctly calculated current duration equal to the current duration of the SPIA. But that doesn't really accomplish the objective either, since the duration of both will change over the next two years.

Based on some of his other posts, I'd think that bobcat2's solution would be to use long-term and short-term TIPS funds, and rebalance 4-6 times per year to keep the duration matched to the duration of the SPIA. Using funds for this objective makes more sense to me than using funds to build an LMP (oops, I meant "to secure a targeted level of income") now.
That seems about right to me. I wouldn't expect the rebalancing would have to be done quite that frequently unless interest rates are changing a lot , but you might be right.
Yes LMD - liability matched duration. :D

BobK
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.

User avatar
bobcat2
Posts: 5222
Joined: Tue Feb 20, 2007 3:27 pm
Location: just barely Outside the Beltway

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by bobcat2 » Tue Oct 06, 2015 10:14 pm

Gattamelata wrote:
bobcat2 wrote: This stuff is tricky. In the past I had trouble understanding it. Now I have trouble explaining it. :D
Thanks, Bob, your efforts are sincerely appreciated. I had never considered duration matching before I saw you mention it.

One way of characterizing the value occurred to me as I was reading your OP: Similar to the way that bogleheads and other indexers attempt to capture the market return by indexing, duration matching is designed to capture the interest rate at retirement (or at annuity inception) by matching the duration of their funds earmarked for annuity purchase to the anticipated duration of their annuity. In both cases it's attempting to eliminate the impacts of variable market conditions by tethering ones own financial plan to whatever conditions prevail.

Does that seem roughly accurate to you? It helped me to understand what you were going for, at any rate. (Har! "Any rate.")
Yes.
It's essentially immunizing the targeted annuitized income stream against changes in interest rates. You are hedging the risk to your targeted level of floor retirement income. A duration matching strategy practically locks in the target income at the cost of giving up the opportunity for a higher income stream. This seems like a quite reasonable trade-off in the context of securing safe floor retirement income.

BobK
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.

User avatar
bobcat2
Posts: 5222
Joined: Tue Feb 20, 2007 3:27 pm
Location: just barely Outside the Beltway

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by bobcat2 » Tue Oct 06, 2015 10:23 pm

alec wrote: Maybe the difference (between Merton's duration estimate and #Crunchers estimate) comes from insurance companies assuming some adverse selection in that only healthy people who expect to live a long time buy life annuities.

Hi Alec,
That would account for some of the difference, but since we are talking about people at age 65, I doubt it's much more than one year of the difference in the duration estimates. Maybe I didn't remember correctly what Merton said. Maybe he didn't say anything about average annuity duration and instead my memory is beginning to fail me. :shock: Maybe he had an age other than 65 in mind. Maybe he was considering annuities with a joint second to die property. Hard to tell.

BobK
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.

thx1138
Posts: 774
Joined: Fri Jul 12, 2013 2:14 pm

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by thx1138 » Tue Oct 06, 2015 11:52 pm

bobcat2 wrote:
Kevin M wrote:
Kevin M wrote:[When] this statement [is] within the context of a pre-retirement portfolio targeted for purchasing an LMP at some time in the future, ... I understand it ...
This is not for the purpose of purchasing a liability matched portfolio (LMP) at some point in the future. It is for the purpose of securing a targeted level of income at some time in the future. That distinction makes all the difference. A major point in all this is thinking in terms of a targeted level of portfolio assets (LMP) at some future date is itself a serious error. You don't care what the portfolio asset level is, you care about hitting the income target. The level of safe assets is irrelevant.
I believe the common wisdom is that short-term fixed-income is pretty safe in terms of real income, as a substitute for a more traditional LMP, since nominal rates tend to track inflation pretty well.
In retirement the liability is consumption, which is always real not nominal. Therefore income, the financial asset that supports the consumption, needs to be real, at least for the safe floor, IMO. We are not trying to track safe floor income "pretty well". We are instead attempting to meet our targeted safe income floor as best we can.

BobK
First of all, wonderful thread Bob! Provides a lot of insight on a topic that often defies intuition. I hadn't ever considered this issue until you brought it up and described it so well.

I'd like to try and provide a perspective that might encapsulte both Kevin and Bob's perspectives - though I suspect the more likely outcome is I just confuse the issue more :( I'll try anyway...

The scenario Bob describes is to purchase a known real income stream at some point in the future by paying a single (unknown) lump sum for a SPIA at that future date. The key point is that neither the future real nor nominal cost of the SPIA is known at present. It is not a fixed predictable amount as the cost of that SPIA depends not just on inflation but also on long term real interest rates. The gotcha is that for relatively short periods of time (e.g. wanting to purchase the SPIA five years from now) the biggest threat is not inflation but rather changes in long term real interest rates which can change quickly and dramatically affect SPIA costs.

Therefore if we want to view our future SPIA purchase as a future "liability" we've got a problem as we can't predict what the nominal or real cost of that liability actually is in the future. The classic LMP approach to financial planning is to just account for inflation and provide the best safe return to support that future liability at the lowest present cost. The typical example for a future real liability would be a TIPS zero coupon bond. Bob points out that would not be the right thing to do for an SPIA because the future real liability is not know in the case of an SPIA. Doing the "typical" LMP thing will give us a known real portfolio value at the time of SPIA purchase but that's no good as the SPIA cost isn't fixed. So Bob provides the solution, instead of investing in the "typical" LMP asset - a TIPS maturing at the date of the liability - instead invest in an asset that tracks the cost of the SPIA and not just inflation.

In a sense Bob is still advocating a LMP approach, but just a more complicated one that accounts for the variable cost of that future liability - it is still "liability matching" but in this case is matching a liability whose future value varies. This is actually not very different a choice than what is often already done with many LMP securities. Some need to match future real costs and so in those cases things like TIPS are used to construct the LMP. In other cases the future liability is actually a nominal cost and so a TIPS is not the right instrument and instead nominal treasuries of the proper maturity are used instead. Here the same thing is being done except the duration of the security is being extended so that the "LMP" used to purchase the future SPIA tracks the expected cost of the SPIA. It is still "liability matching" but it is matching a liability with a more complicated cost than just simple nominal or real values. A similar example would be creating a LMP for a future liability in a foreign currency, you better make sure the instrument you use to construct the LMP tracks the liability in the proper currency or else you won't actually have "matched" the liability.

The added wrinkle is that Bob is talking about the future "liability" being the purchase of a safe income stream via an SPIA. There is actually no need for it to be a SPIA necessarily, you could at that time instead purchase a TIPS ladder assuming the ladder could extend far enough to cover your "worst (best) case" lifetime. The concept would be the same though, the future cost of that TIPS ladder would vary as long term interest rates changed and so you'd better park your money in something that will match that cost of the ladder, which would again be something of similar duration to the ladder.

Essentially the concept of a LMP can be applied to any future liability as long as safe securities exist that properly track the liability. As shown above common examples for "typical" liabilities would be nominal treasuries, TIPS or foreign denominated bonds. These are easily understood as they track future liabilities that are in some sense "fixed". Purchasing a future TIPS ladder or SPIA is a bit more complicated because their future cost varies, but as Bob points out securities do exist that are safe and track that cost. Use TIPS or TIPS funds that have a duration that matches the effective duration of the price of the SPIA (or TIPS ladder).

As others point out this is in someway exactly what various kinds of deferred SPIA products provide as an existing product that could be purchased in the present - the SPIA provider does the duration matching for you and adds some cost along the way. But it is still important to understand what Bob is saying here or else you might not be able to properly understand why the prices of deferred SPIAs are what they are...

Another thing to realize is that a "income stream" is really just a ladder of LMPs with different liability dates and amounts. So a TIPS ladder is just a cascade of LMPs that form a guaranteed real income stream for a finite period of time. The SPIA adds the benefit of longevity insurance so that the income stream is indefinite. We get a mixing of the concepts of LMP and guaranteed life income streams in Bob's example because he is discussing saving for a future liability (so an LMP) used to purchase a guaranteed life income stream (aka SPIA). And he points out the perils of using conventional LMP securities for such a scheme since the SPIA cost is not fixed as far as a nominal or real cost. If you are going to think of a LMP to purchase a SPIA in the future you must realize the "LMP" you construct is going to be matching a liability whose cost varies along with long term interest rates.

None of this is rocket science in the end, saving for future "liabilities" with variable costs is exactly what the finance departments of major corporations do day in and day out. And in fact the futures market is essentially based on meeting the needs of creating LMP for highly variable liabilities. So Bob's example while more complicated than "typical" LMP construction is really not that complicated.

To come back to Kevin's point I think he is pointing out that if I want a security that tracks inflation then short term nominals have done a pretty good job of that. So if you are trying to create a LMP for a future real liability then short term nominals can meet that need. However, they aren't really a very good choice for most LMP needs because they are short term - you are forgoing a better return by taking on the "duration risk" of longer term securities. Of course for a LMP there is no "duration risk" at all and it is better to think of the premium you are getting as an illiquidity premium, you know you don't need the money until a future date and so you don't care about the gyrations of the portfolio at dates in between. The short term nominal (or short term TIPS) are the right instruments for a known liability at an unknown time. Using them for liabilities that are known to be in the distant future would be leaving money on the table - making your LMP cost more than it should.

As Bob points out short term nominals are a bad choice for his example for another reason. Because the short term nominals only track inflation and in fact for a SPIA inflation is not the big risk, long term rates are the big risk, they are the wrong instrument to match the future cost of the SPIA.

Whew... Maybe the was too long and rambling to be helpful to anyone.

User avatar
#Cruncher
Posts: 2652
Joined: Fri May 14, 2010 2:33 am
Location: New York City
Contact:

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by #Cruncher » Wed Oct 07, 2015 9:49 am

Kevin M in [url=https://www.bogleheads.org/forum/viewtopic.php?p=2647583#p2647583]this post[/url] wrote:The duration risk is for the two years starting now, so why wouldn't you want a security with the same current duration as the current duration of the SPIA?
Because it wouldn't protect us as much. Assume we want to buy the SPIA in two years (2017) and we want to protect ourselves against its price being higher at that time. We believe that its price will rise 1.03% for a 0.1% decrease in interest rates. (See my previous post.) This is equivalent to a modified duration of 10.3%. Which would provide better protection, a bond with a duration of 10.3% today or one we estimate will have that duration in two years?

Consider two $1,000 TIPS each with a 1% coupon selling at par. One matures in 11 years (2026) and has a duration of 10.4% today. The other matures in 13 years (2028) and we estimate it will have a duration of 10.4% in two years. Assume yields decrease from 1.0% today to 0.9% in two years. We predict the SPIA will then cost 1.03% more. Which bond will better protect us against this?

Code: Select all

Matures                       2026       2028
Duration today                10.4       12.1
Est duration in 2 years        8.6       10.4
Value today               1,000.00   1,000.00
Value in 2 years          1,008.61   1,010.43
Gain                          8.61      10.43
The bond maturing in 2026 will gain $8.61 in value (0.86%) due to the lower yield. This partially offsets the 1.03% higher SPIA cost. But the bond maturing in 2028 will gain $10.43 in value (1.04%). This comes much closer to offfseting the higher cost of the SPIA.

This is more apparent if we assume that we want to buy the SPIA in ten years, not two.

Code: Select all

Matures                       2026       2036
Duration today                10.4       18.9
Est duration in 10 years       1.0       10.4
Value today               1,000.00   1,000.00
Value in 10 years         1,000.99   1,010.43
Gain                          0.99      10.43
In this case, the bond with the duration of 10.4% today provides hardly any protection. What we need is a bond maturing in 2036 that will have a estimated duration of 10.4% in ten years.
Kevin M in same post wrote:How does shortening the maturity by two years but using the current yield for the original YTM produce a meaningful result? We don't know what the yield will be two years from now, so we can't know what the duration will be two years from now.
You're right, Kevin, we don't know what the yield will be in two years. But using today's yield provides a good estimate. This is because modified duration doesn't depend much on yield. For example, here it is for the 2.5% coupon Jan 2029 TIPS using the same formula from my previous post except with three other yields besides 0.748%:

Code: Select all

        Modified
Yield   Duration
------  --------
0.000%    10.05%
0.748%     9.95%
1.500%     9.85%
2.500%     9.72%
An increase in yield from 0% to 2.5% causes only a 0.33% point change in modified duration.

User avatar
bobcat2
Posts: 5222
Joined: Tue Feb 20, 2007 3:27 pm
Location: just barely Outside the Beltway

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by bobcat2 » Wed Oct 07, 2015 11:17 am

Hi thx1138,

I agree with nearly everything you wrote. It naturally follows that I will only write about a portion with which I disagree. :D

I disagree with calling this a liability matched portfolio approach. It isn’t. It is instead a particular liability driven investment (LDI) strategy similar to, but not the same as, the liability matched portfolio approach.
Links to LDI definition and more
https://en.wikipedia.org/wiki/Liability ... t_strategy
http://www.investopedia.com/terms/l/ldi.asp
https://institutional.vanguard.com/iam/ ... main=false
http://www.lyxor.com/fileadmin/_fileup/ ... cs/all.pdf

One type of LDI strategy is the liability matched portfolio approach. Another is the liability matched duration approach. This is the latter approach.

The liability matching portfolio approach is the appropriate LDI strategy if the future liability is an asset level at a particular point in time. Examples – I want to buy a boat in three years and I am very confident in what the price will be. I want to pay off my remaining home mortgage in ten years.

The liability matched duration approach is the appropriate LDI strategy if the future liability is a stream of future payments, such as an income stream over time.

It doesn’t make good sense to talk about portfolio matching when no portfolios (asset values) are being matched, but instead only durations are being matched.

When you ask people about the DB pension plan they have or their Social Security, they reasonably reply with the level of retirement income they will be receiving. If you instead ask people about their DC retirement plan, they will respond by telling you the current level of assets in the plan. The connection between that level of assets and reliable retirement income is tenuous at best.

Have you ever asked anyone about their DB retirement plan or their Social Security retirement plan and received a reply in terms of the present value of their DB plan benefit or their Social Security plan benefit. That would seem nutty. But when we get that same reply when inquiring about a 401k retirement plan or 403b retirement plan, we consider that reply reasonable. It isn’t. Levels of saved assets tell us little about reliable retirement income streams from those assets, but those income streams are what’s important.

For retirement planning we need to nudge people into thinking much more about retirement income, and away from portfolio levels now or expected at retirement. Talking about liability matched portfolios (LMPs) is one of the impediments in getting people to think in terms of income. That’s in addition to its being incorrect, or at a minimum a very clumsy approach, in the context of retirement income. I believe the main reason your post seemed rambling to you is that you were continually trying to cram duration matching into a portfolio matching framework. That’s very awkward, as this paragraph from your post illustrates.
Another thing to realize is that a "income stream" is really just a ladder of LMPs with different liability dates and amounts. So a TIPS ladder is just a cascade of LMPs that form a guaranteed real income stream for a finite period of time. The SPIA adds the benefit of longevity insurance so that the income stream is indefinite. We get a mixing of the concepts of LMP and guaranteed life income streams in Bob's example because he is discussing saving for a future liability (so an LMP) used to purchase a guaranteed life income stream (aka SPIA). And he points out the perils of using conventional LMP securities for such a scheme since the SPIA cost is not fixed as far as a nominal or real cost. If you are going to think of a LMP to purchase a SPIA in the future you must realize the "LMP" you construct is going to be matching a liability whose cost varies along with long term interest rates.
None of this is rocket science in the end, saving for future "liabilities" with variable costs is exactly what the finance departments of major corporations do day in and day out.
Actually rocket science underlies all these ideas. It’s just that you don’t have to be a rocket scientist to apply them. :wink:

BobK
In finance risk is defined as uncertainty that is consequential (nontrivial). | The two main methods of dealing with financial risk are the matching of assets to goals & diversifying.

User avatar
Leif
Posts: 2471
Joined: Wed Sep 19, 2007 4:15 pm

Re: Common Misconceptions about Safe Floor Retirement Planning

Post by Leif » Wed Oct 07, 2015 11:37 am

Interesting. However, it only matters if you think a SPIA makes sense to establish a floor on income.

My SPIA (income floor) is SS@70.
Last edited by Leif on Wed Oct 07, 2015 3:55 pm, edited 2 times in total.

Post Reply