Executive Summary
As baby boomers continue into their retirement transition, two portfolio-based strategies are increasingly popular to generate retirement income: the systematic withdrawal strategy, and the bucket strategy. While the former is still the most common approach, the latter has become increasingly popular lately, viewed in part as a strategy to help work around difficult and volatile market environments. Yet while the two strategies approach portfolio construction very differently, the reality is that bucket strategies actually produce asset allocations almost exactly the same as systematic withdrawal strategies; their often-purported differences amount to little more than a mirage! Nonetheless, bucket strategies might actually still be a superior strategy, not because of the differences in portfolio construction, but due to the ways that the client psychologically connects with and understands the strategy!
The inspiration for today's blog post is a recent white paper by Principal Financial Group entitled "Comparing a Bucket Strategy and a Systematic Withdrawal Strategy" that compares and contrasts the two approaches. The FPA's 2011 Financial Adviser Retirement Income Planning Study recently found that approximately 3/4ths of advisors frequently or always use a systematic withdrawal approach, while 38% frequently or always use a "time-based segmentation" (i.e., bucket) approach (as the percentages add up to more than 100%, some planners apparently 'frequently use' both approaches).
For those who aren't familiar, the basic concept of a bucket strategy is relatively straightforward - the retiree divides assets into 3-4 "buckets" that will provide the required cash flows for varying time horizons. For instance, bucket #1 might provide for the first 5 years of retirement and be invested entirely in cash and short-term bonds. Bucket #2 might provide for the following 10 years, and be invested in intermediate and longer-term bonds and perhaps a modest allocation of conservative equities (e.g., a 40/60 allocation). Bucket #3, which won't be required for 15+ years, contains the remainder of the portfolio, which would be invested entirely in equities since the money won't be needed for a long time to come. On the other hand, systematic withdrawal strategies are essentially a safe withdrawal rate approach that simply generates ongoing cash flows by liquidating as needed from a regularly rebalanced diversified portfolio.
Strategies for Portfolio Construction
What's notable about the two approaches, though, is that while their methodology for portfolio construction is framed entirely different, the outcomes are remarkably similar. For instance, let's assume that the client's short-term bucket (years 1-5 in cash and short-term bonds) generates a real return of 0%, the intermediate bucket (years 5-15 in a conservative 40/60 portfolio) generates a real return of 3%, and the long-term bucket (years 15-30 in a full equity portfolio) generates a real return of 7%. If the client wishes to generate $40,000/year of real income, we can discount the future cash flows to derive the required investment portfolio:
| Stocks | Bonds | Total | |
| Short-Term | $0 | $200,000 | $200,000 |
| Intermediate-Term | $121,264 | $181,895 | $303,159 |
| Long-Term | $141,288 | $0 | $141,288 |
| Total | $262,552 | $381,895 | $644,447 |
| Allocation | 41% | 59% |
The end result? The bucket strategy leads the client to a conservative 41/59 portfolio, not unlike the portfolio any conservative client would select. If the client is less conservative, and adjusts the time horizons according - for instance, by making the second bucket only run for another 5 years (from years 5-10) - the buckets produce a moderate 50/50 portfolio! If the client is slightly more aggressive, and sets the first bucket to only run for 3 years, the portfolio is 60/40.
Which means in reality, the bucket strategy ends out producing substantively similar asset allocations as a systematic withdrawal strategy, ranging from about 40/60 for a 'conservative' client to 60/40 for a more moderate risk client. If you add in some additional dollars to the intermediate and especially long-term buckets to adjust for the danger that below-average returns occur (which can happen, even over long time horizons), you end out with an even more equity-tilted portfolio where most clients are about 50% to 70% in equities... remarkably close to the 'optimal' recommended allocations from the safe withdrawal rate research.
Of course, there are many different ways to assemble a bucket portfolio allocation, involving more buckets, or slightly different time horizons, or varying growth assumptions. Nonetheless, it seems that virtually all of them arrive at substantively similar results, with allocations down near 30/70 or 40/60 for conservative clients, and up to 60/40 or 70/30 for more moderate and risk tolerant clients.
Buckets Are The Same... But Different?
Notably, this isn't meant to suggest that since bucket strategies produce asset allocations that are similar to systematic withdrawal strategies, that they are irrelevant or inferior. In fact, as advocates of the strategy often point out, the bucket approach is arguably superior from the perspective of client psychology; it fits far better into our mental accounting heuristics, and makes the portfolio easier for clients to understand. Furthermore, clients may have an easier time staying the course through market volatility when they can clearly see where their cash flows will come from in the coming years, and that they truly have a decade or more to allow for any declines in the equity bucket to recover. And various buckets can be linked not just to time horizons, but also to separate goals (e.g., the bucket for the vacation house).
The bucket strategy also arguably articulates a clearer systematic strategy to generating cash flows each year than to simply say "we'll sell whatever we need to from a total return portfolio to generate your retirement spending every year."
While the details for many of these prospective psychological benefits are not well tested, much of the existing behavioral finance research supports these benefits conceptually, and many advisors using bucket strategies have anecdotally substantiated them. Which means the reality is perhaps that even if a bucket strategy merely produces the exact same asset allocation and portfolio construction, but does so in a manner that makes it easier for clients to stick with and implement the strategy, it is arguably a superior one.
Practical Challenges With Bucket Strategies
Unfortunately, though, as the white paper points out, implementing a bucket strategy as the advisor may actually be more difficult, even while it is easier for the client to understand. For instance, standard tools do not exist to calculate the allocations across the buckets (nor is there even a standard framework and set of assumptions to do so); portfolio reporting software generally only reports on investments in the aggregate or by account, but not necessarily by mentally constructed time-horizon bucket. Such software constraints make reporting and oversight of bucket strategies difficult, unless the advisor truly wishes to create one account for each time bucket (which may or may not be administratively feasible for the client, especially if there is a mixture of retirement and taxable accounts). And more monitoring may be required over time, as the advisor must adjust the amount of assets in each bucket from year to year to keep the client on track.
Nonetheless, the bottom line is that even if the asset allocation benefits of the bucket strategy might be a mirage, the psychological benefits may be real enough to merit more attention to the strategy, and more focus on ways that it could be implemented practically across a client base.
So what do you think? Do you use bucket strategies, systematic withdrawal strategies, or something else? Have you ever witnessed any of the 'psychological benefits' of bucket strategies for clients? If the strategies were comparable in portfolio construction but superior in helping clients to stay the course, would that be persuasive enough to make you shift how you build portfolios for clients in retirement?
Michael – I appreciate you blogging on this topic as I also saw the White Paper from Principal a few months ago. Our firm has been successfully using the bucket approach for many years as it is the perfect compliment to our Cash-Flow Based Financial Planning and our philosophy of Goals-Based Investing. I can attest to the impact it has made for clients – especially as we have experienced tremendous volatility – in maintaining their overall asset allocations vs. following the temptations to out-guess the market. I will also say that it DOES take additional work to execute and monitor this strategy. However, there is something powerful in clients knowing that the cash flows needed for supporting their lifestyle are held in assets less susceptible to short-term market drops.
Hi Michael,
i tend to think and use bucket strategies more with guaranteed income sources like annuities and other target maturity investments.
The difference between bucket and allocation is that there tend to be more guarantees (ie stability and certainty) in a bucket type plan. But I prefer the floor/upside concept much more (though I am a biased RIIA member:)
Chris
Haha! – it’s precisely a mirage, or perhaps better stated, a reframing of asset allocation in a way clients can better understand it in retirement.
We actually use the bucket concept for framing purposes with the client, but manage assets with an allocation / withdrawal strategy framework.
In other words, if we have a 60/40 allocation and a 5% withdrawal rate, the first 15% of the portfolio is the cash reserves bucket, equal to three years of cash withdrawals.
The second 25% of the portfolio, consisting of the remainder of the bond portfolio, is the second bucket…equivalent to the next 5 years of portfolio withdrawals. The framing that happens here is merely a way to express that the client has 8 years worth of withdrawals that are not exposed to the stock market in a 60/40 model.
The remainder is the equity (long-term) bucket.
This is an imperfect explanation of how our withdrawal strategy really works, but clients seem to understand it and take solace in knowing how many years of fixed income investments they have to burn through before touching their equities.
Joe – I agree with you, a systematic withdrawal rate means that there is an explicit rules-based system that tells you where to generate each years cash from. From a behavioral perspective, I would think sketching out buckets and showing how they fill one another back up (hopefully) each year would have a very similar effect to showing an actual bucket strategy.
This is very interesting information to read a few years down the road after additional research has been done and economic conditions have changed.
Micheal has recently cited data that shows the danger in keeping a cash-reserve bucket and how that approach could potentially cause a lower average per-share price when spending that cash, and then having to refill the reserve during a down market, than by just using systematic withdrawals ‘on the way down.’
Moreover, with interest rates at historic lows, I’m not sure 5 years in fixed income to ‘burn through’ is any safer than equities. Unless, of course, they are extremely low duration bonds, which poses another problem: additional drag on the performance of the portfolio that may lead to failure.
Michael, it is great to see an intelligent planner write about a subject that means so much to the client. We use the bucket strategy when the plan has to work. With forced spending and guaranteed money secure, this strategy greatly reduces the risk. I see the withdrawal method when there is room for error and possible overspending. Also, this would be used if a client could not determine their income need. It’s not a queston on which one is better, its a question of risk tolerence.
Also, a mirage makes us see something that is not there. If properly presented, there should be no comparison to a mirage. The benefit is there and so is the risk reduction.
Heath,
You state that the strategy greatly reduces risk, and therefore is not a mirage. But this is actually my point. There is little research to actually prove it IS reducing risk, beyond the mental accounting lines we draw in our head.
As I showed here, it’s the substantively SAME portfolio in the end. We can account for it differently, but how can the bucket strategy have less risk than the non-bucket strategy if they both have the same allocation and therefore hold the same investments?
It’s also worth noting that implementing the bucket strategy on an ongoing basis will result in a steadily declining equity exposure (as the long equity buckets slowly roll into the shorter-term buckets). Yet studies in the Journal of Financial Planning from Bengen in 1996 to Blanchett in 2007 have shown that steadily decreasing equity exposure REDUCES THE SUCCESS of the plan relative to a constant asset allocation. This would suggest that a bucket strategy could actually INCREASE the risk that the plan fails.
Ultimately, I don’t think this has quite been fully tested and could stand to be studied further. But the fundamental point is that I find little in the way of any actual research and evidence that substantively suggests the strategy really DOES reduce risk. In fact, the existing research suggests the opposite may be true.
– Michael
This piece by Moshe Milevsky on the topic is worth reading as well:
http://www.ifid.ca/pdf_newsletters/PFA_2006OCT_Buckets.pdf
I believe it’s important to focus on the difference between asset allocation and order of liquidation. We know that time greatly mitigates the risk of owning equities.
If you’re forced to sell in the short-term, you can get your rear-end handed to you.
If you can put together a plan to hold stocks for a very long period of time, i.e., a minimum of 15 years, you give yourself statistical assurance you’ll make money in stocks.
I don’t care anything about the fact that the allocation may be the same between the two methods. That’s not relevent; when evaluating these strategies, the focus should be on the order of liquidation between the two methods of distribution planning.
Michael,
I read that Principal study a couple of months ago, and from what I remember the study seems to have a methodological problem. The way I understand it, with bucketing you are supposed to use an asset/liability matching bond ladder for the shorter-term spending needs, but they seem to just be using constant-maturity bond funds instead.
I’m not suggesting a conclusion one way or the other, but just that I don’t think the Principal study is the last word on the matter.
Wade,
I’ll have to go back and take a look again.
But if you’re rolling equities (long bucket) into shorter buckets (bonds), then in practice won’t you keep refreshing the duration of the short bond bucket? I guess strictly speaking, you could implement it as a rolling bond ladder to preserve an asset/liability matching approach, but it would still end out being a bond ladder of constant maturity as you continue to roll it?
– Michael
Michael,
I’m thinking of it more as it should be a rolling bond ladder. Each year, you add a new 5 year bond, for instance. The duration of your bond ladder may stay constant, but you wouldn’t ever experience capital losses this way.
I think they are using a volatile bond fund (I’m piecing this together based on their description on the last page) that can experience capital losses and still has some small sequence of returns risk. So it’s not really the way I would want to look at this.
I will give them credit for using Monte Carlo simulations though, as a usual side effect of bucketing is that it supports a higher overall stock allocation, which makes bucketing look good in the post-1926 US data.
Hi Michael. Sorry I missed this earlier. We use a bucketing strategy called Timelining(r). Investment wise it normally is a 60/35/5 strategy internally but is presented as a time line (read asset liability matching using forward looking std dev. Modeling)
We have had great mental accounting success with the timelining metaphor and kept our clients invested thru 08/09.
It’s the visualization that makes it work!
The real world assumed success of the bucket approach as I understand it is based on the utilization of the Evensky 2-3 year short duration bucket’s allowing the rest of the portfolio to avoid any liquidations in down markets. Since the worst imaginable market disaster (so far) lasts 3 years or less, the success of the approach is thought to be self evident. The other buckets (or bucket) make an added contribution (to traditional allocation/withdrawal strategies) only to the degree that they contain other than traditional assets, including annutization, or assets which receive an illiquidity premium. So “bucket” is a bad term and “alternative sources of income” may be a better way to think about it.
Alternative source of CASH FLOW or LIQUIDITY. not income?
Ray,
Or viewing it another way – if your goal is to hold up to 25% of the portfolio in cash to have something else to liquidate in down markets, why not just hold a more conservative portfolio in the first place, since that’s effectively what’s being re-created anyway.
Michael,
Except for the study you discussed here, which I already indicated before I have my doubts about, all of the pro-bucket tests I’ve seen have been based on historical simulations.
This is a matter than is rather in need of a more systematic analysis based on Monte Carlo simulation…
Wade
Hi Mike,
As the “original creator” of the bucket model, I can assure you that the psychological benefits have been extraordinary during this past 10 yrs. I first designed my model in 1984 and have used it religiously in my practice ever since. My model has been featured in Kiplinger Personal Finance, Bottom Line and on PBS Nightly Business Report. Yes, it is more labor intensive than a systematic withdrawal, but more reliable, predictable and intellectually superior. Thanks for your comments.