Executive Summary
The conventional view of sequence-of-return risk is that it’s something retirees face as they take ongoing distributions from a retirement portfolio, and requires caution in the early years of retirement.
Yet the reality is that sequence of return risk is equally relevant for accumulators in the years leading up to retirement as well. The only difference is the problematic sequence is reversed – for retirees, the dangerous sequence is to get bad returns at the beginning (of retirement), while for accumulators it’s getting bad returns at the end (of the accumulation phase) that causes the problem.
In fact, the volatility of an accumulator’s portfolio can also be viewed as its “retirement date risk” – the danger that a planned date of retirement (or financial independence) may turn out to be later than originally expected, due to a poorly timed sequence of bad market returns. And as it turns out, the greater the volatility of the portfolio and reliance on growth, the greater the retirement date risk that goes along with it.
Accordingly, accumulators in the final years leading up to retirement may wish to proactively manage risk and reduce the volatility of the portfolio, specifically as a means to reduce the retirement date risk they face. Of course, getting a portfolio more conservative at the end of the accumulation phase may also force an accumulator to work slightly longer, or save more, to make up for the lower expected returns in those last few years. Still, for those who don’t want the risk of a much later retirement date, a decreasing equity glidepath leading up to retirement may be a very appealing retirement date risk management strategy!
Sequence Of Returns Risk For An Accumulator
A fundamental challenge for retirees taking ongoing withdrawals from a portfolio is “sequence of returns” risk – the fact that even if markets average out to long-term returns eventually, if early returns are too low for too long, ongoing withdrawals can deplete the portfolio before the “good” returns finally arrive. Notably, this issue does not apply if the portfolio has no withdrawals, such that all the dollars which experience a decline can participate in the subsequent recovery. Sequence of return risk is limited to situations where there are ongoing cash flow distributions coming out of the portfolio.
Except as it turns out, sequence of return risk is not unique solely to retirees taking money out of a portfolio. It also applies to accumulators who are still making contributions into a portfolio a well.
For instance, imagine an investor whose goal is to accumulate $1,000,000 by saving $300/month every month starting at age 25. Assuming an 8% long-term rate of return on a balanced 60/40 portfolio, this systematic savings strategy would be able to accumulate a $1,000,000 portfolio balance by age 65, after 40 years of compound savings.
The caveat, however, is that the projected accumulation above doesn’t just assume the portfolio gets a long-term average return of 8%, but that it generates a return of exactly 8% every year. And just as any deviations from that return assumption can dramatically impact the sustainability of retirement distributions from a portfolio (sequence of return risk), so too can it adversely impact the projected accumulation of a retirement portfolio with ongoing contributions, too.
For instance, a 60/40 portfolio has a long-term standard deviation of about 12% per year, or about 3.8% per decade (the standard deviation declines by the square root of the number of years). Given that 95% of returns are expected to fall within plus-or-minus 2 standard deviations, this means that an average return of 8% would actually have a 95% chance in any decade of falling somewhere between 0.4% and 15.6% (per year compounded across the 10 year time window).
So what happens if, for our earlier accumulator, the first decade experiences a bad -2 standard deviation event, but the last decade (right before retirement) recovers with a +2 standard deviation bull market? Or conversely, what if the favorable bull market happens at the beginning, and the -2 standard deviation mediocre decade comes at the end. (Notably, these would not be “black swan” events, but merely a known-to-be-possible bad decade of returns.)
As the results reveal, the sequence of returns has a dramatic impact. In both of the above scenarios, the long-term returns are exactly the same. Both accumulators averaged out to the same long-term returns of 8%. The difference is simply that one had good returns early on (when few contributions had yet been made to the portfolio), and then had bad market returns later that impacted the bulk of the savings. The other had bad returns early on (with not much in savings to be impacted), and enjoyed the good returns at the end when most of the contributions had already been made.
For someone working towards retirement, the difference is significant. In the good sequence, the accumulator with the “plan” to retire at 65 with $1,000,000 actually reaches the goal by age 62. On the other hand, the accumulator with the bad returns is still 34% short of the $1,000,000 goal at age 65, and even if returns subsequently revert back to 8% that accumulator will still have to wait until age 70 to actually reach the original $1,000,000 goal!
There is a significant risk that their actual retirement date will be materially different than the goal that was set in the first place, as the age 65 retirement date goal ends out with an 8-year time window in which it may fall! (And of course, a severe bear market in just the final year or two before retirement could make the situation even worse!)
How Portfolio Volatility Causes Retirement Date Risk
Portfolio volatility is typically used to characterize the ways a portfolio’s value can go up and down from year to year. But in the context of an accumulator saving for retirement and facing sequence of return risk, volatility doesn’t just impact the value of the portfolio, per se, but the timing of retirement itself. Or viewed another way, portfolio volatility for an accumulator creates retirement date risk.
The reason that portfolio volatility causes retirement date risk is that the goal of retirement typically has two components: a target dollar amount (to cover retirement spending), and a target date by which that amount will be accumulated (so the accumulator can stop working!). This joint goal means that if the accumulator can’t achieve one part, the other must be adjusted.
In the context of an accumulator who is forced to retire early – e.g., due to an adverse health event – the retirement date becomes unchangeable, and thus spending must be adjusted (often downwards) to accommodate the retirement transition.
For those who can continue working, though – until they choose to retire – the path is typically to keep working until the retirement accumulation goal is achieved.
Yet this means that a volatile portfolio with an ill-timed bear market or poor results can end out forcing the accumulator to work longer, waiting until the portfolio recovers! And the greater the reliance on growth, the more sensitive the outcome is to the volatility of the portfolio (the greater the potential for a shortfall on the originally-intended retirement date), and thus the greater the retirement date risk!
How To Manage Retirement Date Risk
So how can a prospective retiree manage their exposure to retirement date risk?
The first key is to recognize that for the final decade leading up to retirement, it’s primarily the volatility of the portfolio and its returns that matter; the ongoing savings contributions are no longer the driver of the outcome. The reason, simply put, is that as a portfolio compounds, the relative impact of contributions themselves are less and less.
For instance, in the first year, adding $300/month (or $3,600/year) takes the account balance from $0 to $3,600! In the second year, if the portfolio earns its expected 8%, the growth is only $288 while new contributions are another $3,600 – a year’s worth of growth is less than one month’s contribution!
By the last decade, however, the tables turn. In the final years, return on the investment portfolio dominates its increases in value, and new contributions become trivialized. Over the very last year, as the portfolio grows from nearly $925,000 up to the final $1,000,000 target amount, growth contributes almost $75,000 to reach the finish line, dwarfing the $3,600 of new (final-year) contributions. In fact, while in the second year one year’s worth of growth is the same as just one month’s contribution, in the final year one month’s growth is nearly double one year’s worth of savings!
Accordingly, the reality is that managing retirement date risk really is all about managing the volatility of the portfolio itself, and not necessarily the contributions to it (except to the extent that a lower volatility portfolio with a lower growth rate may require larger contributions to meet the retirement goal in the first place).
A Decreasing Equity Glidepath Into Retirement (Using Lifecycle Funds?)
Since the driver of retirement date risk is the volatility of the portfolio, the easiest way to reduce retirement date risk in the final years is to reduce portfolio volatility itself in those final years. What this means in practice is that retirement date risk can be reduced by decreasing exposure to volatile assets (e.g., equities) in the final years leading up to that target retirement date.
In this context, the reality is that target date funds (or lifecycle funds), which typically take equity exposure off the table in the years leading up to retirement, arguably really do have it right when it comes to asset allocation for accumulators. Reducing equity exposure in the final years – as the portfolio gets largest and most sensitive to return volatility – is an excellent means to narrow down retirement date risk.
For instance, assume the earlier accumulator saving $300/month into a 60/40 portfolio begins to take equity exposure down by 3%/year in the final decade leading up to retirement. Each 3% equity reduction will reduce the expected return of the originally-60/40 portfolio by about 0.15% (from the original 8%), and the standard deviation by about 0.5% (from the original 12%). The impact of this on the trajectory of the portfolio – including getting a plus-or-minus 2 standard deviation event – is shown below.
As the results reveal, implementing this form of decreasing equity glidepath in the final accumulation years leading up to retirement significantly reduces retirement date risk, which is cut down from an 8-year time window to only 6 years! The "early" retirement scenario for good returns begins at age 63, but the bad return scenario only delays to age 69.
On the other hand, the reduced returns associated with getting more conservative in the final years do delay the retirement date slightly – as it moves from age 65 to age 66 instead. For most accumulators, this is probably an acceptable trade-off to manage retirement date risk. Though notably, the prospective retiree could also try to ramp up savings in the final decade to help make up the difference – an increase in the monthly savings rate from $300/month to $415/month in the final decade is sufficient to bring the retirement date back to the original age 65, even with the decreasing equity exposure in the final years. (Notably, the increasing savings requirement is material - a nearly 40% boost to required saving - precisely because the portfolio is so large that even just moderate shifts in returns have an outsized impact on the outcome.)
Ultimately, though, the fundamental point is simply this: for investors that have no cash flows coming out or going in to a portfolio, it’s feasible to just wait for long-term returns to manifest. However, for retirees taking distributions, or accumulators making contributions, the cash flows moving in/out of the portfolio introduce a sequence of return risk, which is only amplified for accumulators relying on very-long-term returns to compound their way to a targeted retirement date. And the greater the portfolio volatility, the greater the associated retirement date risk.
Which means in the end, portfolio volatility is not merely a risk to be taken and “waited out” through a series of adverse market returns – instead, proactively managing portfolio volatility, and the retirement date volatility it creates, can be a meaningful way to manage retirement date risk as well!
So what do you think? Is it helpful to think about portfolio volatility for accumulators as a form of “retirement date risk”? Would your clients consider their exposure to volatility in the final years before retirement differently if they considered the retirement date risks that result from aggressive portfolios?
It ultimately comes down to the trade-off the client is willing to make. Are they so determined to leave the workforce that the risk of a bad sequence is worth it to them? Are they happy in their job and want to minimize the volatility? I can think of 1 of my clients that would share each of those perspectives.
Regarding the use of target-date funds to reduce the glide path, I’d still prefer to use a basket of low-cost ETFs or DFA funds so I can maintain better control of the portfolio. Call me a control freak.
-Elliott Weir
Great diagrams! And great points. Recently demonstrated by the barrage of stories following the Great Recession from workers who were about to retire and found themselves suddenly looking at several more years of work, instead.
A good strategy to avoid such a foreseeable problem would be to target a retirement date by or before 55 so you have more flexibility, if the need were to arise, to work long enough to recover from a significant downturn. Also, it suggests that you might consider setting a nest egg target at 20% or so above what it would take to meet your post retirement cash flow needs in order to more easily survive a downturn that happens immediately after retirement.
Great topic. Relying on long term averages is so dangerous. We use similar analysis to show how active volatility management with options increases the likelihood if hitting your planning goal.
I know you are saying the contributions become less important over time, but looking at my husband’s retirement contributions I see he was contributing $0/year to his 401(k) in 1986, $9,500/year in 1996, $20,000/year in 2006 and $24,000/year in 2016. Do you have any sense of how much increasing contributions effect the outcome?
Thanks! I love your careful work and appreciate you making this work available.
Great article as always Michael. Just as accumulators might want to manage risk and reduce volatility of their portfolios leading up to retirement, I think it’s just as important to do with 529 plans as the beneficiary nears the start of college.
Is it naive for me to think that if I keep 5 years of near cash available to meet my spending needs as I enter retirement, that I don’t really need to be concerned about sequence of returns on my investment portfolio? If equity markets behave as they have historically there is a high likelihood that five years from now any downside correction will have been recovered, at which time (or prior in up markets) I can generate an additional cash cushion. I should admit that what traditionally would have been allocated to bonds has been moved to cash/short maturities given the likelihood of rate increases.
If people owned a high cash values life insurance policy or program-say $500,000 of accrued cash values, that would substitute for zero interest cash reserves or buffers, while earning near 5% tax deferred on cash values. At least that’s what I do.
T.E.,
Shifting bond allocations to shorter duration is a reasonable adjustment. We’ve shown in prior research that as long as you have a healthy equity exposure for long-term growth, the “fixed income” portion should really act more as a buffer to stocks, for which short-term bonds and cash are a better buffer than longer-term bonds (and especially corporates). See https://www.kitces.com/blog/accelerating-the-rising-equity-glidepath-with-treasury-bills-as-portfolio-ballast/
However, where you draw out the cash allocation is important. If you draw primarily from equities, you just end out with a more conservative portfolio with a significant cash drag – as 5 years of cash could be 15%-25% in total cash allocation – and just reduces sustainable spending in retirement. See https://www.kitces.com/blog/research-reveals-cash-reserve-strategies-dont-work-unless-youre-a-good-market-timer/
– Michael
It does seem that moving from one bucket to another is “robbing Peter to pay Paul” – you still end up with lower returns overall so is there reason to do anything – are you just fooling yourself thinking you’re any safer?
Excellent analysis Michael, I’ve been preaching adverse sequence of accumulation approaching retirement date for years now. Do any of the Fund Families offer an autopilot reducing equity glidepath product?
This would seem to represent an excellent marketing niche for the informed vendor.
Actually yes, most target date funds to this in the years leading up to the target [retirement] date.
We should be able to insure this risk with a tail risk annuity product.
Interesting analysis but the strategy you outlined doesn’t seem to help significantly. You are talking about moving the bad scenario from a hitting your magic number at 70 up one year to 69. The biggest benefit that financial advisers never seem to endorse forcefully enough is to save more! The amount you save is really the only aspect of retirement accumulation that you 100% control. You pick the investments as well but you have no idea what the future returns will be. If it looks like you will fall short of your goal, an appropriate response would be to allocate more money to savings to make it up.
Not completely true that accumulation is all that you can control. Like before retirement, you also have control over what you spend in retirement, especially if you manage your post retirement expenses in such a way that you have maximized for discretionary rather than fixed expenses like rent or loans where adjusting your cash flow needs is significantly more difficult.
Hi Michael,
Thank you for discussing a risk that is not addressed enough. I like to use the term “red zone” (so does Prudential) in reference to the 10 years before and 10 years after retirement.
To protect against this risk, I favor having enough stable income streams to cover essential living expenses, and then investing the remainder in a diversified balance of stocks, bonds and cash (this is a conversation for another day). Stable income streams, such as Social Security (don’t plan on getting all of it though), pensions, rental real estate, plain vanilla annuities, dividend and bond income, should provide enough cash flow to allow retirees to “wait out the storm.” Yes, that means they might need to postpone or minimize that round the world cruise, but it beats withdrawing a large amount from their stock portfolio in a down market.
Also, and this is unconventional among advisors because it hurts AUM, but I encourage my clients to purchase a residential rental home as part of their lifestyle living portfolio. By purchase I mean withdraw enough from their portfolio to make a significant down payment and borrow the rest. Time and time again that is where the smart money is. Why? Because real estate historically has been an excellent hedge against inflation, i.e. rents can be raised. Again, this may not be in the AUM advisor’s best interest, but I believe it is very much in our clients’ best interest.
Rob, you might benefit from an objective read of this Jason Zweig article: http://www.wsj.com/articles/SB10000872396390443493304578038811945287932.
Also, I might suggest you consider the possibility of vulnerabilities to some of your statements:
On buying bonds while also having a mortgage – http://www.wsj.com/articles/SB10001424127887323916304578402810211631842
On personal real estate relative to other assets –
http://www.economist.com/blogs/graphicdetail/2015/11/daily-chart-0
On the risks of rental real estate income –
http://www.nytimes.com/2013/03/30/your-money/investing-in-a-rental-home-isnt-as-safe-as-it-may-seem.html
On investing in indexed or variable annuities –
https://www.sec.gov/investor/alerts/secindexedannuities.pdf
https://www.sec.gov/news/studies/secnasdvip.htm
William,
It appears you didn’t read or understand what is meant by “plain vanilla” annuities judging by referring me to an indexed and variable annuity website. Plain vanilla means no bells or whistles annuities, e.g. no indexed or variable, but simply a contract between the insurance company and the buyer where they pay the policyholder a certain amount for a certain time period or as long as they live. They are also called income annuities, and help reduce longevity risk. They don’t get as much press because they are low-commission products. A good resource for those looking to steer clients in the right direction is: http://www.stantheannuityman.com.
Regarding real estate, I like the chart, especially since it drills down by city. But it doesn’t provide any data on rental income, it only shows housing appreciation which, on average, has exceeded inflation. But clearly in a number of cities prices have gone up a lot faster than inflation.
For information on rental costs and trends, check out: http://www.census.gov/housing/hvs/files/currenthvspress.pdf
This census report shows some major trends in place: Homeownership is trending down, renting is go up, and rental vacancies are down. This has cause average rental rates to rise over 20% since the Great Recession. These are good trends for someone considering buying investment rental property. Remember, the trend is your friend.
No, being a landlord is not for everyone, and not in every city either. It is a lot harder than clicking “buy” on a computer screen. I would encourage anyone considering it to read several books and talk to other landlords in the area. But the stories of someone calling you in the middle of the night are overblown, unless of course one owns a lot of properties, and then you can afford to hire someone to take those calls.
As with any investment, there are risks and rewards. But rental real estate offers several benefits:
1. Is it truly is a portfolio diversifier, which is a lot more difficult to achieve than it used to be since markets are increasing in correlation. When that rental check arrives month after month and the stock market is taking a beating, that is when you really appreciate it.
2. Having some leverage at today’s interest rates provides an opportunity for a higher return. It is easier to make this decision when looking at trends. Remember, the trend is you friend.
3. I like the idea of truly owning something, as opposed to a piece of paper. Yes, the US dollar is very strong now, but things change, and our $18+ trillion in debt, plus the overwhelming unfunded future obligations to SS and Medicare should concern anyone trying to protect their money from losing it’s purchasing power.
I can’t comment on the WSJ links since I subscribe to the WSJ paper version. Old habit. Yeah, gotta change that.
An “AUM” advisor, whatever that means, should not be influenced by their own compensation. I think I’m what you define as an “AUM” advisor and I’ve suggested clients purchase rental properties, or vacation homes to take advantage of the tax benefits, and the ability to long in a long term interest rate on a 30 year note in a rising rate environment. The times and circumstances make this advice prudent right now, but it doesn’t mean it will always be so. Be careful when over generalizing or taking passive shots at other business models. Real advisor’s do real work regardless of compensation.
Ken,
“taking passive shots at other business models”
My primary income is based on AUM. I do it because it makes sense to my clients, and for me too. Not sure how you can take what I wrote as some sort of shot at AUM.
The reality is some advisors will not recommend this because it would lower their income, since clients tend to use their investment money as a down payment for the real estate.
To further my thinking, I will add this: If an advisor says they are a fiduciary, then they should evaluate every investment alternative for the client, even if it results in lower compensation. Fiduciaries must put their client’s financial interest before their own. If the fiduciary advisor prudently evaluates rental real estate and determines that it is not in the best interest of the client, then they have done their job. And even though it might make financial sense, some clients do not want to deal with the inherent responsibilities that come with being a landlord (although there are companies that will do this for them).
I am glad you are an advisor that will consider rental real estate as a potential investment and income stream for clients. I hope more advisors do, because most of my generation and those that follow will not likely have a defined-benefit pension plan. Rental real estate can be a terrific replacement for that, or even superior to a DB pension.
Thanks for the enlightening article Michael. With just 5 years out till I retire, the asset allocation (currently 70/30) is where I struggle. Your article follows more traditional 60/40 scenario and you also mentioned the target date funds which; looking at Vanguard’s, show even more conservative allocation. On the other hand, a recent series of Articles that David Lavine wrote for the NY Times suggests leaving all your assets in stocks:
http://www.nytimes.com/2016/02/13/your-money/how-much-of-your-nest-egg-to-put-into-stocks-all-of-it.html
http://www.nytimes.com/2016/02/06/your-money/why-your-portfolio-needs-plenty-of-stocks-whatever-your-age.html
Naturally, the hope is over the next 5 years will see growth so that gains can be redirected in safer cash, bonds, etc., and lowering exposure to equities.
From what I gathered in your analysis, is that having more a more aggressive allocation would cause greater deviation and higher portfolio volatility risk.
Hi Michael,
Thanks for the excellent analysis.
In addition to using a target portfolio size number and a specific retirement date, I wonder if we also need to consider the scenarios with different annual withdraw rates in retirement phase (3%, 4%, 5%, 6%, etc. of the target portfolio size), and different annual saving rate (0.4%, 0.5%, 1%, 2%, etc. of the target portfolio size).
For example, is it true that for people with a 3% annual withdraw rate they are subjected to lower “retirement date risk” vs. people with a 6% annual withdraw rate?
Mike,
Great article and not unlike my own chart on SOR risk that I did last October on Seeking Alpha.
Your article was much more focused on just this topic, so thanks.
Dave
Except that inflation adjusted withdrawal sustainability in retirement requires 50 – 65% in equities. Closer to the higher end if you want to be a little more aggressive using the Guyton rules.
Another great article! Perhaps even start reducing stock allocation 15 years prior to retirement, if stock market valuation is high (implementing valuation-based tactical asset allocation – https://www.kitces.com/blog/valuation-based-tactical-asset-allocation-in-retirement-and-the-impact-of-market-valuation-on-declining-and-rising-equity-glidepaths/)
Why the keen focus on effectively decreasing equities so much right before retirement, when that decision more accurately would seem appropriately based upon how many down cycles a retiree can handle later in life? If a person is planning retirement at age 60, they have multiple bull market cycles ahead of them to compensate for bears. Bailing out of equities near retirement for them sacrifices much return.
What this article doesn’t address is the reality that the longer you wait to retire the less you need to accumulate for retirement. For example, assuming you only need a nest egg to get you to 85, there is a world of difference in what you need at 40 to make it that far than what you need if you retire at 75, 10 years rather than 45 years.