New thoughts on the draw down phase

[walkinwood]

The August issue of the Journal of Financial Planning has 3 articles on the spend-down phase or as they put it "Portfolio Decumulation"

Journal Home Page

There is an interesting article on Modern Portfolio Decumulation: A New Strategy for Managing Retirement Income

He proposes:
- Annuities are a legit part of a retirement portfolio to hedge longevity risk.. The question is when (if at all) it makes sense to include it. He suggests a "check" to see if it is needed. I think we'll see a "Walmart" or "Dell" of the annuity business come in and eliminate the excessive fees.

- Portfolios behave differently when there is an regular outgoing cash flow than in the building phase when there is a regular incoming cash flow.

- Dynamic asset allocation to match the ability of your portfolio to deliver the cash flow you need. Not sure if I understand this concept. Will need to read it again.

I think we're going to see a lot of new ideas on the best way to draw down our retirement portfolios while handling market, inflation and longevity risks.

The only issue I have is finding the time to read this stuff while I'm still working!!

ww.

 

[Nords]

Holy crap, I can't believe that this thread is "more than 1329 days old", as vBulletin is [-]whining[/-] informing me. A very belated thanks to Walkinwood for digging up the article and posting it.

When I read something that piques my interest (or if I didn't make the time to understand it on the first read) then it goes into my "later" pile. During my seven-hour Mainland flights this month I finally got down to the bottom of that pile.

This article has a very neat approach to annuitizing longevity risk. FIRECalc (and even Monte Carlo) can't really assess the SWR for a 40-year-old ER who might live another 50-60 years. There just aren't that many data sequences in the history of investing, and even if they're tweaked for relevance they still might not have any application to the future. But this paper suggests that instead of trying to maximize a success rate which is really very speculative to forecast in the first place, it starts its analysis with whatever budget it's given and tries to generate enough portfolio cash flow to meet that budget.

Instead of the ER having to come to terms with a 80% or 95% or 99% FIRECalc success rate, along with all the arguments about past being prologue, this approach doesn't use portfolio failure as "game over". Instead every year it compares the amount of the remaining portfolio to the cost of a SPIA. Instead of caving in and buying an annuity at the beginning of ER, this holds off the annuity as long as possible... and longevity is only a risk if you live long enough to shrink the portfolio to that annuity cost.

For example, Raddr has been running a six-year thread on the hapless Y2K ER who relentlessly spends the traditional Trinity 4% SWR of a portfolio that's 75% S&P500. That guy was doing OK 1328 days ago. However in January 2011 it's lookin' pretty grim with a 10% withdrawal rate.

But instead of that Y2K ER spending to oblivion while hoping for a sustained market boom (or slashing his budget), he could start shopping annuities. Maybe it's too late for him now, but if he'd been doing this every year then he would've been able to fall off the tightrope in early 2009 and land safely in an annuity net. He could even have started pricing annuities in 2008 as the Fed was lowering interest rates, perhaps deciding to buy one or two small SPIAs (hopefully not from AIG) before yields got too low to support his budget.

The other side of the SPIA safety net is that ERs can confidently start with a 4-5% SWR (or Bob Clyatt's 4%/95% variable method) and see how things work out. They're pricing SPIAs every year against their portfolio so they'd have plenty of warning if things turned ugly. However if their portfolio exceeded expectations then they'd also have plenty of comfort from seeing how far they are ahead of the game, and they could adjust their spending or their AA to match their comfort level. It's even possible that they'll enjoy a lucky market sequence of 10-20 years of decent returns and won't ever have to worry about an SPIA. They could decide to purchase long-term care insurance with the excess of their portfolio. They could even set aside a reserve for the high medical expenses anticipated for the last year or two of late life.

Figure 9 of the linked article shows the summary. The failure line isn't at zero-- it's the declining curve showing the cost estimate of an SPIA. Yeah, yeah, we have no idea how much SPIAs will cost in 20 years, either, but that's not the point-- the point is that the portfolio is constantly checked against that cost, and an SPIA is only purchased if a portfolio is in danger of going below the line.

I think it's a very reassuring way to start an ER, and a very practical application of a low-cost SPIA.

I just wish I'd appreciated its significance at least two or three years ago.

Anyone aware of any updates or flaws since WW first posted this?

 

[pb4uski]

Here's what I'm thinking.

In the accumulation phase, I value averaged - had a target for the portfolio each month and would make monthly investments to bring the portfolio to that level - so in months where the portfolio was doing well I invested less and in months that the portfolio was doing poorly I invested more. It ended up being DCA on steroids because in the long run it was similar to DCA but had the nuance of buying more when price were lower and less when prices were higher.

I thinking of doing the inverse for withdrawals. Start retirement with a year or two of spending in cash, then draw more from the portfolio when it has a good month and less when it has a bad month and let the cash account be the shock absorber.

 

[Major Tom]

I like the idea of considering the failure rate of the portfolio as being the value required to purchase the desired annuity. However, I still have one problem with SPIA's and that is the fact that you're putting that large chunk of money in one basket. What if the company selling you the SPIA folds?

I know that if you buy a SPIA with someone like Berkshire Hathaway, the chances of failure are very slim, but it's a psychological thing for me. For some reason, I think I'd feel more comfortable with the money invested in stock and bond funds than with one company.

Also, doesn't the monthly payout from a SPIA always remain the same, with no annual COLA?

Can anyone more knowledgeable help me out with this one?

 

[chinaco]

..However, I still have one problem with SPIA's and that is the fact that you're putting that large chunk of money in one basket. What if the company selling you the SPIA folds?

...

One can create a SPIA ladder across time (over several years) to try to manage the risk of buying in at a low interest rate (get an average over time) and split the annuity purchases across several strong insurance companies (highest ratings) to mitigate some of the risk of company failure.

Insurance companies are heavily regulated and monitored for excessive financial risk and liquidity.

If one is going to buy a SPIA the choice to purchase a cola or go nominal depends on how one intends to manage inflation risk.

 

[BigNick]

I like the idea of considering the failure rate of the portfolio as being the value required to purchase the desired annuity. However, I still have one problem with SPIA's and that is the fact that you're putting that large chunk of money in one basket. What if the company selling you the SPIA folds?

I know that if you buy a SPIA with someone like Berkshire Hathaway, the chances of failure are very slim, but it's a psychological thing for me. For some reason, I think I'd feel more comfortable with the money invested in stock and bond funds than with one company.

I'm not aware of any law saying you can't have more than one SPIA. The only downside is a little more paperwork, and maybe a slightly reduced rate because of economies of scale (possible sliding scale of commissions).

Also, doesn't the monthly payout from a SPIA always remain the same, with no annual COLA?

Not necessarily. You can have any COLA you like, if you're prepared to spend more, or get less per month on day one. Non-COLA SPIAs are commonly used for comparison purposes as "typical" becase they're easier to compare (there are a zillion ways to "do COLA").

Back to the original point of the thread: another thing to consider is how much you want to, potentially, leave to your heirs. Putting your portfolio into an SPIA gives you, the retiree, more security, but your heirs - who, typically, will be choosing your nursing home - will know - if they find out - that the money is definitively gone, as far as they're concerned. But if your money is going to the cat's home / inheritance tax people / melt when you die, then putting it into an SPIA simply means that the residue will go into the pockets of the insurance company's shareholders, some of whom may well be right here on this forum.

 

[walkinwood]

This is definitely a "did I post that?" moment.

I just started re-reading the article, but wanted to contribute a thought regarding SPIAs. I think an overlooked risk of the SPIA is that retirement spending is not constant. A retiree has to have an emergency stash to accommodate an unexpected expense. This would be a relatively minor tweak to the method suggested in the article.

Back to reading the article.

 

[Major Tom]

Great answers chinaco and BigNick - thank you for taking the time and effort.

A SPIA ladder bought from several different companies - of course - hadn't thought of that. How simple.

The fact that I didn't realize you could purchase a SPIA with increasing payments over time just highlights how little research I have done on them. Thanks for setting me straight.

 

[Nords]

The fact that I didn't realize you could purchase a SPIA with increasing payments over time just highlights how little research I have done on them. Thanks for setting me straight.

This board needs a good quiver of online annuity-pricing calculators for comparison shopping.

But who am I kidding-- the last thing the insurance industry wants is an educated buyer.

 

[MasterBlaster]

This is definitely a "did I post that?" moment.

I just started re-reading the article, but wanted to contribute a thought regarding SPIAs. I think an overlooked risk of the SPIA is that retirement spending is not constant. A retiree has to have an emergency stash to accommodate an unexpected expense. This would be a relatively minor tweak to the method suggested in the article.

Back to reading the article.

The concept for a SPIA is to take some of your stash to buy a SPIA that would then provide (along with SS) a baseline standard of living.

The rest of the stash could then be used both to live better and for emergency situations.

If you can provide for the baseline standard of living then the withdrawal rate from the remaining pile can be more aggressive than a safe withdrwal rate. In that way one could have a higher standard of living without ever going broke.

 

[Htown Harry]

Here's what I'm thinking.

In the accumulation phase, I value averaged - had a target for the portfolio each month and would make monthly investments to bring the portfolio to that level - so in months where the portfolio was doing well I invested less and in months that the portfolio was doing poorly I invested more. It ended up being DCA on steroids because in the long run it was similar to DCA but had the nuance of buying more when price were lower and less when prices were higher.

Funny coincidence you should mention this. Last weekend I had less than 15 minutes to waste at the mall. I stopped in a bookstore and picked up Bernstein book and skimmed the TOC and a chapter or two.

Amazon.com: The Investor's Manifesto: Preparing for Prosperity, Armageddon, and Everything in Between (9780470505144): William J. Bernstein: Books

The only part I dwelled on was his description of value cost averaging. See page 152-153.

I had made a mental note to come back and read more on my next trip to the library or the bookstore.

It turns out Bernstein and a finance professor named Edleson co-wrote a book of the same name in 2006.

Amazon.com: Value Averaging: The Safe and Easy Strategy for Higher Investment Returns (Wiley Investment Classics) (9780470049778): Michael E. Edleson, William J. Bernstein: Books

The book review on Amazon says Edleson does the math that shows a potential 1% return advantage over DCA. After searching the board, I see it's been mentioned here from time-to-time.

Anyway, thanks for the reminder, pb4uski. I'm halfway through a rebalancing and account consolidation exercise, so I'll definitely check this out. I will be doubly interested if it applies the same math to draw-down strategies.

 

[pb4uski]

....I will be doubly interested if it applies the same math to draw-down strategies.

Same here. I wonder if any other people on this board who have been in withdrawal for some time have used this appraoach and how they fee it has worked for them. I would seem that it should work inversely, sell more when prices are relatively higher and sell less when prices are relatively lower.

 

[chemist]

But if your money is going to the cat's home / inheritance tax people / melt when you die, then putting it into an SPIA simply means that the residue will go into the pockets of the insurance company's shareholders, some of whom may well be right here on this forum.

Uhh, not really. If you die "early" and leave a "residual" it goes to continue paying the annuitants who die "late".

 

[haha]

Funny coincidence you should mention this. Last weekend I had less than 15 minutes to waste at the mall. I stopped in a bookstore and picked up Bernstein book and skimmed the TOC and a chapter or two.

Amazon.com: The Investor's Manifesto: Preparing for Prosperity, Armageddon, and Everything in Between (9780470505144): William J. Bernstein: Books

The only part I dwelled on was his description of value cost averaging. See page 152-153.

I had made a mental note to come back and read more on my next trip to the library or the bookstore.

It turns out Bernstein and a finance professor named Edleson co-wrote a book of the same name in 2006.

Amazon.com: Value Averaging: The Safe and Easy Strategy for Higher Investment Returns (Wiley Investment Classics) (9780470049778): Michael E. Edleson, William J. Bernstein: Books

The book review on Amazon says Edleson does the math that shows a potential 1% return advantage over DCA. After searching the board, I see it's been mentioned here from time-to-time.

Anyway, thanks for the reminder, pb4uski. I'm halfway through a rebalancing and account consolidation exercise, so I'll definitely check this out. I will be doubly interested if it applies the same math to draw-down strategies.

I read Edleson's earlier book describing this method. I still have it somewhere. It did not concern itself with withdrawals, only an investing method. It is a somewhat less automatic method than value cost averaging, in that you assume a benchmark equity return- it was really funny to reread my old edition where he assumes equity returns like 15% p.a. I suppose he may have modified the method.

Overall I feel that tweaking withdrawal strategies is mostly a delusional quest, although it can be a seductive one. It really boils down to AA and returns that are actually achieved in one's chosen classes, and the size and modifiability of the withdrawal demands that are made.

Ha

 

[chinaco]

... Putting your portfolio into an SPIA gives you, the retiree, more security, but your heirs - who, typically, will be choosing your nursing home - will know - if they find out - that the money is definitively gone, as far as they're concerned....

I know you were joking... Anyway... my 2 cents on the subject just in case someone reading this might take that seriously.

It is a Non Sequitur... Although a popular bit of illogical rhetoric.

Planning for LTC is a smart move (IMO)... some sort of plan. Weaving together that rationale would be a foolish move (it is no plan).


Bottom line: If one is put in the NH... it will happen anyway. Why... because caring for and managing the care is difficult and stressful. The heroics of family members are relatively rare when it is NH time (for real)... And I am not talking about getting mom to the grocery store! Even with money NH placement is a crap shoot...


Income management and LTC might be inter-related in terms of planning in some ways. But when I hear someone casually say something (as you joked and as some believe) that usually means there is no LTC plan. And I do not mean LTC insurance necessarily. One can self insure, but they would need to maintain a large reserve that is not invested in very risky assets.


Most people's LTC plan is based on little more than hoping they won't need it...

No free lunch there unless one is in poverty! For many... that is the plan.

 

[rescueme]

I know that if you buy a SPIA with someone like Berkshire Hathaway, the chances of failure are very slim, but it's a psychological thing for me. For some reason, I think I'd feel more comfortable with the money invested in stock and bond funds than with one company.

Also, doesn't the monthly payout from a SPIA always remain the same, with no annual COLA?

Can anyone more knowledgeable help me out with this one?

First of all, BH does not accept qualified funds (e.g. pre-tax, such as TIRA or 401(k) money) at this time. You would have to find an alternative such as VG or FIDO (who we have our SPIA from) that accepts direct rollover of qualified funds and is a "strong" company.

Yes, the monthly payout does remain the same with a "fixed" SPIA (e.g. not inflation index adjustments). They are not for everybody, especially for those that fear future inflation but for some (like DW/me) it fits quite nicely in our plan for ER income while we wait for SS - an inflation adjusted annuity.

For more info, I'll recommend the BH wiki on the subject of fixed SPIA's:

Immediate Fixed Annuity - SPIA - Bogleheads

 

[rescueme]

... the residue will go into the pockets of the insurance company's shareholders, some of whom may well be right here on this forum.

Not necessarily true. It all depends on the terms of your contract. If it is a "pure life" policy, you are correct.

However, if you take the option (as we have) for a lifetime term with a minimum amount of years of payout, any remaining payments go to your estate/beneficiary after your death.

Our SPIA was purchased at age 59, with a calculated minimum term of 28 years. As long as one of us is alive, payments continue (at 100%). If we both pass before age 87, remaining payments go to our estate.

BTW, if we beat the odds and one of us live to the age of 88 (or later), payments continue (at 100%) till we're both gone. It's a long shot on the odds, but it actually increases your IRR for the contract.

Any earnings of the provider/insurance company above the return of principal and our "guaranteed earnings" (which we know at contract execution) is of course retained by the company to cover their expenses and also used to pay others. That's why our contract did not specify any sales commission. It's included within the premium quote we received at the time of purchase and that's why we did not have a "salesman" to sell us the policy (but I'm sure the person we talked to from FIDO had the authority to sell insurance, or at least work under the direction of a person authorized to sell insurance through FIDO.)

 

[rescueme]

Also, doesn't the monthly payout from a SPIA always remain the same, with no annual COLA?

It's the same amount every month - the same as most traditional company (private sector) defined benefit plans were in the past.

DW will have two small pension incomes starting in two years. They don't go up (non-COLA) and they are for her lifetime only - no survivor benefit...

Just think of a standard non-inflation SPIA acting in the same way as the pension that most folks had in the old days - that is for non-government j*bs. They received the same "check" every month for the rest of their lives. One exception is if they worked for a union, and pension payments were increased as part of a union contract for those already retired. But that was the exception in the private sector - not the rule.

 

[cardude]

OK, this paper kind of makes sense to me. A better way to quantify an "end date" in the decumulation phase.

As to the mechanics of it, do I simply go to, say, the Berkshire annuity quote website and quote an annuity every year, and if the price of the annuity is less than our portfolio, then we are good to go? Is that the simple version?

But how does one account for inflation, since these annuity payments are fixed? I don't understand how to bake inflation into the annuity, in other words. For instance, I see that I can buy an SPIA annuity from Berkshire today that would take care of all our cash flow needs for life, at a cost of about 1/2 of what our total portfolio is. But those cash flows are going to be worth a lot less 40 years in the future, so to me that's not really covering my future cash flows. I guess the other 1/2 of our portfolio would be keeping up with inflation?

What am I missing here?

 

[jdw_fire]

OK, this paper kind of makes sense to me. A better way to quantify an "end date" in the decumulation phase.

As to the mechanics of it, do I simply go to, say, the Berkshire annuity quote website and quote an annuity every year, and if the price of the annuity is less than our portfolio, then we are good to go? Is that the simple version?

But how does one account for inflation, since these annuity payments are fixed? I don't understand how to bake inflation into the annuity, in other words. For instance, I see that I can buy an SPIA annuity from Berkshire today that would take care of all our cash flow needs for life, at a cost of about 1/2 of what our total portfolio is. But those cash flows are going to be worth a lot less 40 years in the future, so to me that's not really covering my future cash flows. I guess the other 1/2 of our portfolio would be keeping up with inflation?

What am I missing here?

one way would be to get a quote for a COLAed annuity instead of a fixed annuity. unfortunately since vanguard got rid of their COLAed annuity quoter i dont know where to get said quote from an online calculator.

 

[Ed_The_Gypsy]

Baldwin United

 

[Nords]

But how does one account for inflation, since these annuity payments are fixed? I don't understand how to bake inflation into the annuity, in other words. For instance, I see that I can buy an SPIA annuity from Berkshire today that would take care of all our cash flow needs for life, at a cost of about 1/2 of what our total portfolio is. But those cash flows are going to be worth a lot less 40 years in the future, so to me that's not really covering my future cash flows. I guess the other 1/2 of our portfolio would be keeping up with inflation?
What am I missing here?

Annuity-quote websites were everywhere in 2007. Gee, I wonder what happened.

One of the posters on Raddr's board suggested Immediate Annuities - Instant Annuity Quote Calculator. and thinks that buying a COLA annuity would start at about 60% of the fixed amount. But no link was provided and I don't know how that would change with interest rates & recent inflation history. I can remember old Boglehead threads about COLA annuities where some of the posters were adamant that a COLA cap of 10%/year would absolutely not be tolerable. Why, starting over again at 1981's interest rates would bankrupt you before the 1990s!

Keep in mind that you're kvetching about the size/color of your reserve parachute. The annuity option is supposed to be that 5% disaster among the 95% success rates, the last stop on the "game over" line. By the time you get to that point you'll be deciding that a Social Security COLA will be enough, or that you'll bag groceries to pay for those days when you'd rather eat steak than hamburger.

I think the reality is that a portfolio approaching the annuity line would inspire several years of cost-cutting & variable-withdrawal schemes, perhaps part-time work if the ER is physically/mentally capable, and one or two years of spread-out purchases of smaller fixed annuities in order to delay the inevitable. But it's hard to chart that on a graph for a research paper.

I don't think anyone would fly toward the brick wall at Mach 2 and jerk the ejection lever just before impact. Not even REWahoo...

 

[nun]

Being a TIAA-CREF investor I've had annuities pushed at me for a long time. It's the TIAA default and for many people it's a very good solution. But there are other ways to go to get income for life. As I've said before my solution was to buy a rental property. My money is locked up and my investment can go up or down in value, but it wins over the annuity as I still own the principal and it can pass to my heirs. The return isn't bad either. Over the course of my 15 year mortgage the property will have cost $150k and I get $12k a year in rent after costs are taken out, so thats an 8% return. That $12k and SS will cover my retirement expenses. My draw down phase will be from my ER date until 66 and after SS starts I go into another accumulation phase where I reinvest all dividends and gains. Now I'm worrying about RMDs.

 

[ronin]

This is a fairly detailed study on the SWR question and includes some assumptions regarding the use of annuities in addition to portfolio withdrawals. It's pretty long but a lot of it is charts.

SSRN-Revisiting Retirement Withdrawal Plans and Their Historical Rates of Return by Christopher O'Flinn, Felix Schirripa

"Abstract:
This paper examines the historical record of the so-called 4% rule, the popular guideline for sustainable real annual withdrawals in a self funded retirement.

Our findings indicate that a withdrawal plan following this rule (“4R”) carries an historical risk of failure for a long retirement that is much higher than generally acknowledged. For example, we find that 15% of the historical 35-year retirements failed when funded with equal parts of stocks and bonds. The “real” withdrawal plans that generated no historical failures were all less than 4%, sometimes far less, when retirements exceeded 25 years. The historical failure rates that we find for a 5R plan are higher than a 4R plan by a factor of at least three for all retirement periods.

The historical failures are not random. Rather they occur in clusters of years in which the majority of new retirement withdrawal plans fail. A key driver of these failures was a rapid, significant and lasting increase in the rate of inflation - this event increased withdrawals and contributed to a declining real rate of return that was ultimately unable to support the withdrawal plan."

 

[nun]

The historical failures are not random. Rather they occur in clusters of years in which the majority of new retirement withdrawal plans fail. A key driver of these failures was a rapid, significant and lasting increase in the rate of inflation - this event increased withdrawals and contributed to a declining real rate of return that was ultimately unable to support the withdrawal plan."

This just says to me that there needs to be feedback between inflation rate, expenses and withdrawal rate. So during spikes in inflation the retiree needs to economise rather than increase the withdrawal. Again this approach is the "Mcawber Rule" of never spending more that you have....or that your portfolio generates. So "don't eat your children", or spend principal........no rocket science here.