Executive Summary
In planning for retiring clients, it's crucial to get an understanding of what the client's goals are in the first place - so that recommendations can be made about how to financially secure those goals. In the context of setting a spending goal, a popular delineation is to separate retirement spending into "essential" versus "discretionary" expenses - not unlike "needs" versus "wants" for accumulators - with the idea of using guarantees to secure the essential expenses, and less certain growth assets with some risk to fund the discretionary expenses (since they're 'only' discretionary and not essential, by definition).
Yet in reality, even discretionary spending still constitutes an important part of a retiree's overall lifestyle - the loss of which could be very psychologically damaging. As a result, merely securing the essential expenses of retirement and leaving the rest at risk still, in the eyes of most retirees, would constitute a failure of the overall retirement goal. Instead, clients often choose to ensure that all their spending can be sustained - by continuing to work as long as necessary (as health allows) to secure all of their goals. Does that mean the distinction between essential versus discretionary retirement expenses isn't necessarily helpful after all?
The inspiration for today's blog post was a recent discussion I had with a retirement researcher with AARP regarding the "essentials vs discretionary" approach to retirement planning, where spending needs are separated into two categories, each of which may have their own funding and investment strategies. Yet in practice, the approach can be more problematic than its simple elegance suggests. To understand why, it's necessary to understand the basic "wants versus needs" framework upon which it is built.
Wants vs Needs
In most ways, the separation of essential versus discretionary retirement spending is not unlike the separation of wants versus needs for any individual's general levels of spending. As the theory goes, there are basic needs that everyone has - food, shelter, and clothing - but no matter how much we really think we need it, an iPad is not really a "need" but is simply a want.
In fact, even within the classic need categories of food, shelter, and clothing, there are still wants. For instance, while we need clothing, we don't need fancy designer clothes; similarly, while we need shelter, we don't need to have an apartment to ourselves when roommates will do, or a huge 5,000 square foot house when 1,500 square feet would still be more than enough to provide a roof over our heads.
This framework can be highly effective in helping people to establish appropriate spending behaviors. Where there is no clear delineation between needs and wants, it becomes difficult to control impulse purchases and make rational spending decisions. When there's no difference between needs and wants, the only constraint to spending is the amount of money available to spend - leading to the unfortunate outcome where spending on needs and wants-as-needs rises to fill all the available income until there's simply no money left at the end of every month (or potentially not even enough money to make it to the end of the month!).
Accordingly, by separating the wants and the needs, spending on needs is done because it has to be, and spending on wants occurs as a trade-off to saving, with the opportunity to apply constructive thought and a proactive decision about whether the next dollar should be spent or saved.
Essential vs Discretionary and Wants vs Needs
In the retirement context, the basic idea is that needs are essential and wants are discretionary. Accordingly, we can extend the needs versus wants framework into retirement, and then plan to fund it accordingly: guarantees can be tied to the client's essentials, ensuring there will always be food on the table, a roof overhead, and clothes on the back, while discretionary spending - the wants of retirement - can be funded with the excesses, if/when/as the portfolio provides. Thus, as with the working years, we secure the essential needs first, and then allow for greater discretionary spending on the wants as income allows.
As the strategy is typically applied, a prospective retiree's "essential" expenses are secured with a guaranteed stream of income like Social Security, pensions, or by purchasing annuities (or TIPS) to the extent necessary, while "discretionary" expenses are funded with less certain investments (e.g., a diversified portfolio) where the spending can be adjusted in light of the ongoing returns.
Yet in practice, it seems there is one major flaw in the approach: prospective retirees often have the flexibility to choose to treat discretionary expenses as being "essential" in the first place! Because in the end, if you've only guaranteed a part of the goal and fail to achieve the rest, isn't that still an overall failure of the goal?
"Essential" for Retirees vs Needs for Accumulators
In the general case of spending, where people are ultimately are constrained by income and must make spending and saving decisions with the income they have, it is valuable to distinguish between needs and wants. In the case of retirees, however, the situation is somewhat different - for the simple reason that if retirees are not satisfied with the total level of spending available, the retiree can choose to spend less and save more before retirement, and/or can choose to work longer to accumulate more before retiring (at least to the extent health allows). Thus, in the case of the retiree, there is an interaction between "essential" needs in retirement and the decision to retire at all: if your accumulated savings can afford only the bare essentials and not the full amount of the "discretionary" desired lifestyle, the prospective retiree keeps working until the entire lifestyle they feel is "essential" is affordable!
As a result, in all but the situations where retirement is forced due to external circumstances, the distinction between essential and discretionary spending appears to be less relevant. If there are only sufficient assets to afford essentials but not the desired discretionary spending, the prospective retiree keeps working until they can afford their entire lifestyle, including the essentials and a desired amount of lifestyle discretionary spending. From there, if returns are even better than anticipated, lifestyle can always be upgraded further. But from the perspective of someone considering the retirement decision - who has a choice about whether to retire or not - the goal rarely is to secure essential spending and take the risk that the rest won't be affordable because it's just "discretionary" - instead, the typical goal is to afford the entire retirement lifestyle.
In fact, retirees are often willing to work longer specifically to ensure that their later retirement years aren't stripped of enjoyment (no more eating out, no more visiting the grandchildren) because there wasn't enough money to afford the desired "discretionary" spending. Or stated more simply, the psychological impact of losing one's discretionary expenses - those not required for food, shelter, and clothing, but still deeply tied to life enjoyment, personal satisfaction, social and family ties, etc. - can be so severe, that it's not really clear whether we can fairly call them "discretionary" at all, until/unless there are absolutely no other choices available and we have to.
In practice, I believe this is why an essentials vs discretionary discussion is often difficult to have with a retiree, especially a prospective retiree. If the answer is "you can afford the essentials in retirement, but not the discretionary expenses that also constitute lifestyle you are accustomed to" the typical answer is not "let's guarantee the essentials and hope returns carry us to the rest" but instead is "I'll keep working until I have reasonable certainly my entire lifestyle goal can be accomplished" - and if there's upside from there, all the better. In turn, this is why approaches like tying total spending to an amount sustainable under a safe withdrawal rate approach seems to be far more popular with planners and their clients than buying an annuity to secure "essentials" and investing for the discretionary expenses that remain.
So what do you think? Is the essentials versus discretionary distinction really meaningful for a retiree who can just work longer to ensure their entire lifestyle is affordable? Does the psychological impact of being unable to afford the discretionary expenses that constitute a lifestyle still make them somewhat essential? If the essential versus discretionary approach only meaningful for people who can't afford their lifestyles in the first place?
(Editor's Note: This article was included in Carnival of Personal Finance #364 - The Art of PF Blogging Edition - on One Cent At A Time and also Economy Freak.)
Michael, thank you for constantly challenging ideas that I hold dear. Personally, I am of a nature to want to “insure” that base level of income (my risk tolerance isn’t that high), so professionally I tend to like solutions (TIPS ladder, SPIAs) that mimic my own predilections. It’s good to have analysis that holds my tendencies/biases up to the (sometimes unfavorable) light.
At the end of the article, you write “this is why approaches like tying total spending to an amount sustainable under a safe withdrawal rate approach seems to be far more popular with planners and their clients than buying an annuity to secure “essentials” and investing for the discretionary expenses that remain.”
I can certainly see that what’s discussed in this article is ONE reason planners and clients might tend towards SWR (and not annuities). And I understand that it’s not possibly to always list all the reasons for everything. But I do want to direct attention to other, perhaps less savory reasons for the SWR preference: On the advisor side, in the fee-only AUM world, putting $ in annuities hurts business revenue. And while there is no shortage of bashing the commission world (not necessarily on this blog) for their conflict of interest, this AUM conflict of interest is ignored only at client’s peril, too. And as for the client side of things: many of our clients, I know, don’t give a rip about personal finance and investing and really really just want us to take care of it. If we are their sole or primary teacher of all things investing and personal finance, our bias against annuities and towards SWR could very easily just rub off on clients.
I do wonder how much of a planner’s bias towards SWR stems from financial self-interest (whether or not they’re aware of it), and how much of a client’s similar bias is simply by osmosis. And lastly, whether these considerations are stronger or weaker than the “essential vs. discretionary” analysis.
Gah…managed to post with no attribution. I definitely want to be notified of any of the comments.
Doesn’t the usual approach allow unrealistically optimistic assumptions on the extent and certainty of investment returns and retirement spending needs?
Jarrod,
The safe withdrawal rate research is based on the worst return sequences that have ever occurred in history. I’m not certain if that translates to ‘unreliastically optimistic assumptions’ – that data includes 15 year time periods where the real return on a 60/40 portfolio was NEGATIVE for 15 years. If we merely got 1% real returns on stocks between now and 2027, we’d drastically OUTPERFORM the research data.
-Michael
Sometimes when I read articles on this subject it isn’t clear that planners understand there is a middle ground – you can guarantee a spending stream and have assets to hand off to heirs. It is not an either or on SWR or annuity.
Today for example you can generate a $40,000, inflation adjusted income stream by buying zero coupon Treasuries out 25 years for $1,043,147.
Die in 6 years, say, and heirs get the remaining portfolio.
The whole exercise takes 20 minutes using the zero coupon quotes out of Barron’s.
Substitute corporate bonds and you can lessen the cost greatly. In fact, there should be software tied to bond inventory data bases that picks an optimal portfolio constraining individual holdings to 5% of total assets and investment grade and above.
On the historical data base used for SWR I share the concerns of others. Think where we would be today if stocks had remained unchanged from 3/2009 and bond yields were roughly the same.
Robert,
For the client who only has a million dollars and needs $40k/year, it IS an either-or. As your example notes, the client would have to consume the ENTIRE $1M account balance to produce $40k/year, leaving absolutely nothing left.
Also, just to clarify – if you’re buying zero-coupon Treasuries, isn’t the inflation risk still on the table? Or are you buying a laddered series of TIPS strips?
Regarding your final comment, “think where we would be today if stocks had remained unchanged from 3/2009 and bond yields were roughly the same” the answer is – the client would still be dramatically better off than a 1929 retiree for which a 4% withdrawal rate was sustainable.
– Michael
Michael,
I’m looking at it for an absolute certainty to age 90 assuming inflation of 3%. Not matter what happens to markets the client will get $40,000 adjusted for inflation. With the annuity you get the guarantee but nothing for heirs. With the SWR you get the assets to pass on but no guarantee.
With the zero coupon portfolio guaranteeing the payment adjusted for inflation there is no risk.
In terms of the numbers, I find that a year 2000 retiree who got market returns up through 2008 and then had zero returns for the remaining 3 years – in other words assuming markets didn’t recover after 2008 – would be in a shaky situation today. I find he would be drawing $52,350 on a portfolio valued at $552,759, giving an effective withdrawal rate close to 10%. Not even half way into retirement his portfolio would almost have been halved.
On the historical data all I can say is this. I spent way to many hours of my life listening to structured product salesmen explain how even if the mortgage default rate exceeded twice the highest that had ever ocurred historically the investment would still outperform the equivalent Treasury by 100s of basis points. Sad to say too many people bought into his argument. It is not easy to see the fat tail events.
Today with interest rates where they are we are in uncharted waters. I think we’ve got to be careful assuming that because we have a sample of overlapping periods inclusive of the 1930s that the SWR is perfectly safe, especially until we see that Bernanke et.al. can get us out of our present predicament.
Robert,
My point is simply that I don’t think a laddered bond portfolio ensuring cash flows with a 3% ASSUMPTION is particularly sure of anything.
There are plenty of time periods where inflation exceeds 3%, sometimes significantly.
Constructing a laddered bond portfolio with a flat 3% inflation assumption actually ‘fails’ to sustain lifetime income even more frequently than the 4% withdrawal rate approach. You’re radically exposed to even just “above average” inflation, much less a “fat tail” style inflation event.
You do realize that the safe withdrawal rate approach is built exclusively on tail events and has almost nothing to do with historical average returns at all?
As for the current status of a safe withdrawal rate retiree today, Bengen just recently ran these numbers in an article, and came up with results drastically different than what you just cited at http://www.fa-mag.com/component/content/article/10772.html?issue=190&magazineID=1&Itemid=73
I’d be very curious to know how your methodology is differing from his if you’re getting such different results.
– Michael
Michael,
I wrote a piece on my blog at http://rwinvesting.blogspot.com/2012/05/making-nest-egg-last-bengen.html about the Bengen article and recommended people read the article.
He points out that the 2000 retiree is in much better shape than his one failure case despite the 50% downturns we’ve experienced. My case is merely a counter factual. His 2000 retiree is in good shape because the market snapped back strongly after 3/2009. If it hadn’t, my numbers show that the retiree would have to consider one of the 3 possibilities Bengen lists: annuity, reverse mortgage, change investment approach. But I think there is an inflation assumption here going forward if you are setting up to avoid failure. Once the assumption is made zero coupon bonds can be used to guarantee future cash flows.
In using my admittedly simplistic approach I used actual inflation numbers. I took the diversified portfolio return reported by BlackRock at https://www2.blackrock.com/webcore/litService/search/getDocument.seam?contentId=23753&Source=SEARCH&Venue=PUB_IND, assumed 1/2000 retirement with $1.0 million, and $40,000/year on 1/1 adjusted for CPI-U.
My feeling is that we were closer to a “failure” on 3/2009 than can be appreciated by looking at annual numbers.
Robert
Robert,
Part of the issue, though, is that markets generally don’t stay pinned to the floor – which is part of what accounts for the resiliency of withdrawal rates in the first place. We only actually need a small slice of the portfolio in any particular year, so “merely” waiting 2-3 years for a rebound is not horribly destructive (although waiting 5-10 years can get pretty ugly).
But the underlying point is that I don’t view our tendency to rebound after crashes as pure randomness, as though in one future scenario the market might go the other way. It’s reflective of the fundamentals that underlie investing in the first place – stuff that declines in price is cheaper and has better return prospects, which subsequently leads to a sharper rise in price because of the decline that occurred. At some point investments just become so ridiculously profitable that we are ‘compelled’ to invest, as we see at the bottom of cycles like 1982. Granted, we can do a good deal of damage to ourselves in the intervening time period (as we saw in the 15-20 years that followed 1907, 1929, and 1966), but as the returns fail to manifest for many years, they steadily push higher the prospective returns from that point forward.
What this ultimately DOES rest upon as an underlying assumption – which I’ll grant is an assumption – is that regardless of the moderately severe harm we inflict upon our economies from time to time, and the 1-2 decade recoveries that they entail, ultimately our economic growth engine gets back and track and we begin to grow our economy again, which grows our earnings, and eventually renders the returns on markets as appealing again.
If we fundamentally break our growth engine not just for a decade or two at a time (as we have repeatedly throughout the history we’ve already tested), but for an indefinite period of time, I’ll certainly grant that all bets are off.
– Michael
Michael
A macro safe withdrawal rate – without segmenting fixed from discretionary – or not – isn’t really the question
True with an overall number clients will spend for the discretionary anyway.
But there is an exercise – which seems to work.
Give three scenarios to the 4% ‘safe’ withdrawal breaking down fixed, flexible, and variable into pessimistic, realistic, and optimistic.
Thus, the planner could show the client (without breaking down into guaranteed essential versus discretionary) his levels to choose from – and of course then go through the risk etc to meet these levels.
This approach also allows a fall back position.
schwartz
Michael Kitces has presented an interesting perspective on the distinction between Essential vs. Discretionary expenses for a retiree in his discussion entitled “The Problem with Essential vs. Discretionary Retirement Strategies”. (May 30, 2012) He likens these expenses to Needs vs. Wants for accumulators and discusses how these two categories of expenses are not separated from Total Lifestyle Spending for many retirees. This perspective lessens the importance of a separate categorization of essential and discretionary expenses.
Kitces makes an observation that discretionary expenses for some retirees are, in fact, lifestyle preference expenses making discretionary expenses closer to mandatory than optional. This is a valid observation for many retirees. However, there is another way of looking at essential vs. discretionary expenses that is more helpful to me as an aspiring retiree.
Modern Retirement Theory’s preference is to consider essential expenses as those necessary to maintain a base level of lifestyle that a client defines and as resources can support. These essential expenses must be funded with income from sources that are simultaneously stable, secure and sustainable. This is MRT’s preference to define, with the realization that others may define essential in a different way. As retirees have resources to support, they are encouraged to define essential as they see fit. Discretionary expenses are not defined as those expenses that a client can live without, but are such that clients can delay or accelerate as circumstances warrant. Therefore, discretionary expenses would not be defined as never expenses, but rather not now expenses.
This is a productive discussion to have, as it will indicate different approaches to funding retirement from which retirees may choose.
Michael goes on in this article to suggest that using funding sources to secure essential expenses and invest to fund discretionary expense is not the preferred approach by planners and their clients. Rather, he suggests “…this is why approaches tying total spending to an amount sustainable under a safe withdrawal rate approach seems to be far more popular with planners and their clients than buying an annuity to secure “essentials” and investing for the discretionary expenses that remain”. MRT has difficulty understanding the reliance on a safe withdrawal rate to fund total expenses and not relying on a safe withdrawal rate to fund discretionary expenses.
Respectfully, here is a possible flaw in Michael’s logic:
Assume TS = E+D, where TS = Total Spending, E = Essential, and D = Discretionary.
If a safe withdrawal rate is suitable to fund TS, then it must be suitable to fund D, a subset of TS. A safe withdrawal rate cannot be appropriate for total spending without being appropriate also for a component of total spending.
So, if investing for discretionary expenses and “hoping for the best” is not appropriate as Michael suggests, then it cannot be appropriate for total spending and “hoping for the best”.
Respectfully,
M. Ray Grubbs, Ph.D.
Modern Retirement Theory
Ray,
The strategies of “buy an annuity for E and invest for D” generally do *NOT* use a safe withdrawal rate amount for D (at least, not once I’ve ever seen). After all, if the safe withdrawal rate was used for D, you may as well use it for E as well. If you use a safe withdrawal rate (or annuity) for E, AND for D, then you really have changed the whole strategy – it’s no longer essential-plus-upside (for discretionary), it’s TS-plus-upside. Which I argue is a more appropriate framework from the perspective of the client goals, because in practice, if the portfolio for D usually has some material risk of failure or dramatic diminishment, it still fails the goal.
In essence, my point is that from the perspective of the client’s goal, the execution SHOULD be TS-plus-upside, not E-plus-upside-for-D (and then we can have a separate debate about the best way to secure TS, whether it’s via annuity or safe withdrawal rate, per my prior post last week).
– Michael
Michael,
I don’t look at essential -vs- discretionary as a strategy to the mutual exclusion of all others, including safe withdrawal rates, but more of a concept that is easy for both advisors and consumer to relate, understand and apply. That’s the beauty of its simple elegance.
Too many later stage retirees are now finding out that they do not have the resources to pay for their basic expenses. Regardless of the approach for drawing down income, there are no guarantees that a systematic approach by itself will work for millions of middle market retirees that have a real risk that they will run out of money before they run out of breath, even if they decide to work longer. There are too many variables and unknowns that impact a plan that may need to last 30 or more years.
When accumulating for retirement, you plan to the date of retirement. You can’t plan to a date once you retire because you don’t know how long you will live. Therefore, a primary focus for advisors helping middle market clients create an income plan should first be to ensure that at the very least there will be resources to meet essential needs.
Yes, a good plan will also give the client the best opportunity to live the lifestyle they desire as well.
Using a safe withdrawal rate, buckets or other accepted withdrawal strategies can be used to meet this goal and make adjustments to spending patterns over phases of retirement. When an annuity is used to fill the gap between essential needs and Social Security and Pension benefits, the rest of the managed assets will also last longer given that you get more income for a dollar annuitized than you can “safely” withdraw from managed assets. This is not an either/or decision.
When used properly, the essential -vs- discretionary “concept” allows advisors to incorporate a combination of income strategies that provides retirees the peace of mind that they never will be relegated to eating cat food while giving them the best opportunity to meet their overall lifestyle goal. Therefore, I would suggest that the essential-vs discretionary approach has a far less psychological damaging impact than a “safe” withdrawal rate only approach. Plus, by first filling a essential needs gap and then identifying discretionary expenses, pre-retirees will also be able to determine if their lifestyle expectations are realistic and can more easily determine what’s more important, retirement itself, or the way they want to retire.
Kevin