Discussing Retirement Bucket Strategies with Fritz Gilbert – SWR Series Part 55

January 25, 2023 – Welcome to another part of my Safe Withdrawal Rate Series. Today’s topic: Bucket Strategies in retirement. As you know, my blogging buddy Fritz Gilbert has written extensively on this topic at his Retirement Manifesto blog, for example:

And likewise, I have written about my skepticism of bucket strategies in Part 48 of the series: “Retirement Bucket Strategies: Cheap Gimmick or the Solution to Sequence Risk?”

Fritz’s most recent post on the Bucket Strategy started a lively back-and-forth on Twitter, and it seemed appropriate to pursue a more detailed discussion with more than 280 characters per answer in a “fight of the titans” blog post. So if you haven’t done so already, please check out our awesome discussion over on Fritz’s blog:

Is The Bucket Strategy A Cheap Gimmick?

The response was overwhelmingly positive, and we decided to craft a follow-up post here on my blog. We came up with two new questions, and we also need to address two major themes from the comments section in Part 1, specifically, the role of simplicity and behavioral biases in retirement planning.

So, let’s take a look…

1: How important is simplicity in retirement planning?

Fritz: 

Let’s start with “how important is retirement planning?”, to which I’m sure we’ll all agree the answer is “Very.”  The transition from the Accumulation Phase to the Withdrawal Phase is, perhaps, the largest transition you’ll make in your financial life, and it’s far too complex to jump into it without a plan.  (See “Our Retirement Investment Drawdown Strategy” for our master plan, with links to 23 other bloggers’ drawdown strategies, including Big ERN’s as the #10 link). 

That said, retirement planning is complex.  What should your Safe Withdrawal Rate be?  What should you target for your Asset Allocation? Should you do Roth conversions?  Which accounts should you draw from first? How do you cover health insurance?  Etc. etc. etc.

Therefore, any simplification that can be done brings significant value.  My goal has always been to simplify wherever possible without losing the critical elements of the plan.  Finding ways to simplify the implementation of a strategy increases the likelihood that the strategy will be followed.  A sound strategy, implemented poorly, equates to a poor strategy.  Secondly, the ability to explain a complex strategy in simplified language can be extremely helpful when communicating your plan to the ones you love.  Typically, most marriages include one partner who loves the financial detail and one who doesn’t want to be bothered with the details.  Being able to simply communicate your plan increases the odds of alignment with the strategy.

Bottom Line:  Simplification matters, primarily in the fact that it increases the probability that the strategy will be properly implemented when it matters the most.

Karsten: 

One issue that came up in the comments section in Part 1 is that readers prefer the method that’s simpler to manage. I certainly have sympathy for that. If we assume for now that both methods are about equally useful but also equally susceptible to Sequence Risk, then who wouldn’t want to pick the simpler of the two? Intriguingly, most commenters on Fritz’s blog preferred the Bucket approach. Like this one:

I was puzzled because I was under the impression that a simple Strategic Asset Allocation (SAA) seems to have far fewer moving parts than a bucket strategy, where you must constantly decide what bucket flows where and when to replenish and rebalance the cash bucket; effectively a form of market timing that is anything but simple. So, SAA is the equivalent of “remove old bulb, install new bulb” when the bucket strategy needlessly complicates the whole process.

Then how can readers claim that my method is too complicated even though it’s objectively much simpler than a Bucket Strategy? I came up with an analogy. Imagine there are two teams of portfolio managers, one working for an actively managed fund and one running the VTSAX at Vanguard, a passively managed total US stock market index fund. The active managers claim that they can outperform the index fund. The index fund managers take on that challenge and provide many pages of research results showing how active managers have trouble consistently beating an index fund. And they thought that would settle the argument, right? Boy, were they wrong because you know what happened? Readers of this exchange now say that they prefer the simplicity of the active managers because index fund investing seems way too complicated. And the index fund managers wonder what the heck just happened here! That’s how I felt after reading the comment from “Jeff in MN.” When comparing the two approaches, people got sidetracked by the necessary quantitative and analytical considerations. But the average retiree should focus on what method is simpler in practice. You don’t have to replicate the analytical work on my blog to run a simple SAA strategy. I do the work here, so you don’t have to! It’s like driving a car; you don’t have to understand all the mechanical and engineering details. The mechanics and engineers did that for you, so you can enjoy the simplicity of driving your car.

So, in a nutshell, the initial step, figuring out the target asset allocation, is the same for both methods. But the simple SAA approach requires no additional tactical asset allocation and market timing; therefore, it is simpler to implement than the Bucket Strategy. It’s another reason I prefer my approach.

A casual reader might wonder now whether my philosophy has changed. Isn’t Big ERN supposed to be the math wizard who overthinks and overcomplicates everything? Not at all. I want simplicity when that’s all we need. If you recall my 2021 post “When to Worry, When to Wing It,” I go through a laundry list of possible complications to my simulation setup. Some of them we can safely ignore because, in the grand scheme, they make very little difference in historical simulations.

As Fritz states above, very correctly, some financial planning challenges are more complex in retirement than in the accumulation phase. I have spent years researching safe withdrawal rates. The asset allocation in retirement, especially the tactical deviations from the SAA, would be one part of retirement planning that we can and should keep simple. Focus instead on the withdrawal rate and how we need to adjust it in response to idiosyncratic parameters and market valuations.

2: How should retirement planning address behavioral biases?

Karsten: 

Another common thread in the comments section on Fritz’s blog was the issue of behavioral biases in personal finance. I’m the first to admit that most people, myself included, are inflicted with behavioral biases that often draw us away from making optimal financial decisions.

Several commenters pointed out that the presence of such behavioral biases justifies applying the bucket strategy to get jittery retirees over their fear of withdrawing money in retirement. I was intrigued that two of the proponents of the bucket strategy are financial planners (see the comments from “Liz” and “Eric“).

The bucket strategy falls into the behavioral bias called “mental accounting.” I wrote a post in 2016 about the topic, and back then, I defined this bias as follows:

Mental Accounting: Intentionally or unintentionally creating different buckets of money and ignoring the fact that money is fungible; displaying different degrees of risk aversion and/or different propensities to consume out of different buckets.

It’s almost like I knew in 2016 that I would have this discussion with Fritz in 2023! In any case, my response to folks who justify the bucket strategy to address behavioral biases is that if you’re a financial planner, it’s your job to explain to your clients what you’re doing. Why resort to crutches like the bucket strategy, then? The truth is liberating, so please repeat after me…

… A balanced stock/bond portfolio, regularly rebalanced, would have withstood all the past market volatility, including the Great Depression of the 1930s and the Great Inflation of the 1970s/80s, and will likely survive whatever the future will bring.

There you go; with those 37 words, you can explain to the client that the simple SAA is really all you need. Maybe add a few charts and tables to enforce the massage.

In fact, if I were malicious, I would believe that the reason for the popularity of the bucket approach among financial planners is slightly sinister. Is it possible that the industry needs a selling point for its services to distinguish itself from DIY investors? Is it possible that CFPs falsely insinuate that they can do the tactical asset allocation timing better than the average Joe investor with an SAA approach? I would find it much more ethical if the industry tried to educate people about the dangers of mental accounting instead of reinforcing this potentially dangerous behavioral bias.

The discussion reminds me of my old post, “Good Advice vs. Feel-Good Advice.” Sometimes financial planners and financial celebrities on TV or the internet market feel-good advice that may not be as good as it sounds and feels. Granted, at least the bucket approach is just hit-or-miss with zero average impact, while some of the behavioral biases highlighted in my old post are certifiably stupid and mathematically inferior. So, I’m not going to lambast the DIY investors using a bucket approach if they feel that’s the right thing to do. Fritz and his fans are totally on the safe side. But it’s my job as an educator of sorts to point out that a simple SAA works just as well, on average. Apparently, I’m doing a job that the average financial planner can’t or doesn’t want to offer.

Fritz: 

All of us will face a bear market during our retirement years, most likely several. It’s just the nature of the beast (pun intended).  Having a pre-determined plan for how you’ll handle the emotional response to “losing” 20% of your net worth when you’re dependent on those assets to fund your retirement lifestyle is critical. It must be addressed in your retirement plan.  The worst thing a retiree can do is liquidate stocks after a major downturn due to panic, which turns Sequence of Return Risk from a concept into a reality.  

By highlighting the importance of maintaining short-term liquidity to fund your next three years of retirement spending, The Bucket Strategy forces the retiree to think about the reason “Why,” and highlights the genuine prospect of having to live several years without selling stocks during a downturn.  By forcing a decision on the size of Bucket 1 (1 year?  3 years? 5 years?), the strategy requires a retiree to consider their risk tolerance and pre-determine their defense against an inevitable bear market and the resulting SORR.  

Further, many retirees fail to rebalance and, in a Bull Market, could see their risk exposure unintentionally increase beyond their preferred tolerance.  Getting caught up in the euphoria of a Bull Market is another behavioral bias that must be addressed with an effective plan.  Again, the Bucket Strategy, with its emphasis on systematic “refilling” of Bucket 1 based on Asset Allocation movement, provides some protection by automating a “Sell The Winners” approach. It’s also important to note that these rebalancing moves are not driven by market timing, but by a systemic review of asset allocation during the quarterly refill process.

Importantly, The Bucket Strategy also provides a simple means to explain to a spouse what you’re doing with your investment portfolio given the current market dynamics, and why you’re doing it.  It’s not only the behavioral bias of the one managing the investments that matters but also the biases of other people in the relationship who may have anxiety when they see the news headlines and worry if they’ll run out of money. In my case, my wife takes comfort in understanding the broader strategy and my simplified explanation of actions we’re taking in our portfolio.  I’ve heard the same from many of my readers.

Another behavioral factor is the reality that many retirees struggle with giving themselves the freedom to spend in retirement.  After a lifetime of diligently saving, it is a difficult adjustment to change one’s habit and learn to spend in retirement (within your SWR limits).  By establishing an “automated paycheck,” many find it comforting to know they can spend whatever is flowing into their checking account.  In the event of under-spending the checking account balance grows, which is a helpful reminder that the retiree is “safe” in increasing their spending to consume the surplus (or, perhaps, to donate it to a cause they believe in).  Also, the annual review process allows a simple methodology to implement flexible spending rules in conjunction with the bucket strategy, which has been proven the most effective means to ensure you don’t outlive your money.

The Bucket Strategy provides a simple solution that protects against many of the common behavioral biases using an easy-to-understand methodology.

3: How Do You Effectively Manage The Bucket Strategy?

Fritz:  

I’ve written two posts on how I manage The Bucket Strategy, one written in a Bull market and one in a Bear Market.  I’ll summarize the key elements of managing the buckets below but would encourage you to read those posts for more details. 

First, it’s important to touch on the starting point, as discussed in Post 1 of this discussion with Big ERN and in the first post of The Bucket Strategy Series on my site. The starting point dictates the Asset Allocation, as determined by the size of each bucket.  From Post 1:

For the sake of an example, let’s assume you hold 3 years of cash (Bucket 1), 6 years of bonds (Bucket 2), and everything else in stocks (Bucket 3).   If your portfolio equals 30 years of spending, the asset allocation becomes:

  • Cash: 3 Years (10%)
  • Bonds: 6 Years (20%)
  • Stocks: 21 Years (70%)
  • Total: 30 Years (100%)

I maintain Bucket 1 cash in a stand-alone CapitalOne360 money market account and set up an automated ACH transfer every month (my “Retirement Paycheck”).  By simply comparing my balance over time, I can quickly determine my spending level.  For example, I can subtract the 3/31/22 balance from the starting position on 1/1/22 to determine my retirement spending in Q1 22.  As part of my refill process, I also check my current Asset Allocation using Personal Capital and incorporate rebalancing considerations into my refill decision.

In a Bull Market, I will refill the amount of spending each quarter, selecting either stocks or bonds based on which has outperformed in the quarter (using Asset Allocation as the guide). For the sake of simplicity, I’ll provide examples assuming refills are conducted only on 12/31 of each year.  Here is an example of how I would refill in a Bull Market:

In this example, you can see Bucket 1 has been refilled to 3 years of spending (+$40K), but the cash allocation declines to 9% given the overall portfolio’s growth from 30 years to 33.5 years of spending. Stocks have been rebalanced from 75.2% to 70% ($70k rebalanced, with $30k to bonds and $40k to cash), and bonds have been increased from 18.8% to 21.0%.  

In essence, maintaining Bucket 1 at the 3 years of spending in a Bull Market results in a reverse glide path approach, with the cash bucket falling as a % of the AA and the bond/equity portion increasing over time.  Assuming the growth of the portfolio exceeds the rate of inflation, this phenomenon would also be exhibited if spending increases at the rate of inflation, though at a slower rate than shown in the example above.

In a Bear Market, I will review my portfolio to see if there are any holdings that have had a positive return from the date of the last refill.  I compare Asset Allocation vs. target, but allow actual allocation %’s to float in a range with a “floor/ceiling” approach. Depending on the results, I’m content to skip the quarterly refill and draw down the cash in Bucket 1.  This avoids selling positions in a bear market and provides some protection against SORR.  The following is an example. 

In this case, since both stocks and bonds have declined but remain near the targeted AA %’s, Bucket 1 is not refilled and cash is drawn down from 10% to 8.4%.  The entire portfolio has declined from 30 years to 23.9 years of spending, but no stocks or bonds have been sold to fund retirement spending (providing some SORR protection).

To complete the example, following is what Year 3 would look like assuming a market recovery (Stocks up 15%, Bonds up 5%).  For comparison, I’ve included in the right-hand columns an example using a strict SAA approach (assumes the same starting point, but maintaining 70/20/10 allocation throughout, as I believe Karsten would recommend)

By comparing the two approaches, you can see The Bucket Strategy actually results in a slightly higher balance over the three-year period ($1,027,000 vs. $1,025,940), given that no stocks or bonds were sold during the downturn with The Bucket Strategy approach.

In reality, there is very little difference between the two approaches. In my mind, it simply comes down to the preference of the retiree and which concept is more easily understood in their mind, along with the comfort of knowing you can target whatever size cash bucket best suits your risk tolerance.  I find it easy to explain the bucket concept to my wife (as do my readers, based on comments received), and that’s a yardstick that matters to me.  In essence, The Bucket Strategy is simply a modified SAA approach, with an increased focus on the management of the cash balance in Bucket 1.

On a side note:  starting in mid-2022, I also redirected any interest/dividend payments to be automatically transferred into cash (instead of automatically reinvesting) in my After-Tax accounts, which reduces the amount of selling required to refill the bucket at quarter-end.  I continue to automatically reinvest dividends in my Roth and IRA’s.

Karsten:

I don’t use a bucket strategy, so I have little to manage. In fact, I firmly believe that you can’t effectively manage a mostly ineffective strategy. Thus, I like to take the opportunity to rephrase the question into this:

“Why is the Bucket Strategy Mostly Ineffective?”

The answer to that question is that in historical safe withdrawal simulations, a bucket strategy would not systematically hedge against Sequence Risk. For some historical cohorts, a bucket strategy indeed outperforms a simple fixed weights SAA. But in other cohorts, a bucket strategy would lag the SAA. Thus, the bucket strategy can potentially even exacerbate Sequence Risk.

I also found that the bucket strategy’s relative performance is often susceptible to small parameter settings changes. Also noteworthy, a glidepath systematically outperforms both the SAA and the bucket strategy.

Showing all my detailed simulations would go beyond the scope of this Fritz vs. ERN discussion because it would involve a lot of charts and a minimum of 3,000 words on its own. Watch out for a detailed future post in this SWR Series. So, in today’s post, let me give you the 30,000-foot view. Let me outline the intuition for why the bucket strategy would have had such inconsistent performance in simulations. 

Here’s a numerical example that’s rich enough to convey the intuition but simple enough to handle and not lose the forest for the trees:

  • The initial portfolio value is $1,000,000
  • Annual withdrawals are $40,000.
  • There are two assets, one risky (e.g., stocks) and one safe (e.g., bonds/cash).
  • The target weights are 70%/30% for the risky and safe assets.
  • We take the first withdrawal proportionately at the end of year 0, i.e., $28,000 from the risky and $12,000 from the safe asset.
  • In subsequent years, we start withdrawing from the asset bucket that’s above its target weight. If drawing the entire annual budget from that asset still leaves the asset weights away from their targets, there is no further rebalancing. Specifically, suppose the risky asset goes into a deep enough bear market. In that case, we’d withdraw the entire $40,000 annual budget from the safe bucket and let the risky asset weight slip below the 70% target to avoid withdrawing at the bottom of the bear market.

And that’s it. There are no other bells and whistles like shuffling around dividends or interest payments. That is all just a gimmick, anyway. And in the worst case, it may even hurt the investor – more on that in a future post.

In any case, there’s your Bucket Strategy. And we can now check if and how this approach would help us during a bear market and the subsequent recovery.

Let’s assume that the safe asset gives us a 3% return every year, and the risky investment goes through a 3-year bear market and 5-year recovery, as in the numerical example below. If you think the bucket strategy would perfectly hedge against this Sequence Risk, you’d be very much mistaken. Quite the opposite, after eight years, the bucket strategy lags the SAA fixed weight asset allocation by over $3,000. Not a significant difference, but considering that a bucket strategy is often misleadingly marketed as the panacea against Sequence Risk, even a “SoRR Insurance,” this is very disappointing. Let’s see in detail why the Bucket Strategy performs so poorly:

  • First, notice that after the mild drop in the risky asset in year 1, the portfolio is still rebalanced back to the 70% weight simply by withdrawing about $28k from the safe bucket and about $12k from the risky asset. Bucket Strategy proponents, Fritz included, to my knowledge, often stress that small enough equity fluctuations should not yet trigger the bucket strategy, so I replicate this feature here.
  • However, after a second and more significant 15% drop in year two, even withdrawing the entire annual budget from the safe bucket and leaving the risky asset bucket untouched, we are left with only 69.2% in the risky bucket at the end of the year after the withdrawal. But notice that, so far, the bucket strategy hasn’t added any value: both portfolios, SAA and Bucket Strategy, end year two with the same value of $769,116.
  • The lower equity weight at the end of year two will help you when the risky asset further unravels in year 3. After the drop of 25% and the withdrawal coming entirely out of the safe bucket again, the bucket strategy is $1,633 ahead of the SAA. The risky bucket now stands at 66.2% at the end of year three.
  • That underweight in the risky asset now hurts us in year four because we miss out on the stark reversal in returns (+40%). The bucket strategy is now over $6,000 behind the SAA model. Now withdrawals will come entirely out of the equity bucket because the cash bucket is well below its 30% target.
  • For the remainder of the simulation, even after withdrawing the entire $40k from the risky asset, the risky weight stays above 70%. Now we can milk this positive momentum in the risky investment and gradually recover some of the prior losses. But even at the end of the simulation, the Bucket Strategy is still more than $3,000 behind the simple SAA.

Update 2/5/2023: As requested by a reader below, I posted the Excel Spreadsheet here in case people want to check the underlying math and play around with their own return assumptions.

The outperformance of the Bucket strategy relative to SAA is all tied to how future returns correlate with your current asset weight deviations from the fixed 70/30 SAA. And your current tactical asset weight correlates with past asset returns. Does that sound familiar? The bucket strategy functions like a momentum strategy. As I outlined above, that’s a form of TAA. And TAA based on such a naïve, crude, and purely backward-looking rule, i.e., scaling back the risky asset weight after an extended drawdown, is always hit-or-miss.

Sometimes you get it right, and the stock market continues to fall, so the momentum strategy pays off. But if the market recovers quickly, like after the Great Depression bottom in 1932, the 1987 meltdown, or the GFC bottom in March 2009, you got caught on the wrong foot; the Bucket Strategy vastly underperforms during the recovery period.

It’s nice to see that the intuition and mechanics of this simple numerical thought experiment are also present in my more complicated simulations with three assets (stocks, bonds, money market) and all the other bells and whistles, like monthly vs. quarterly. vs. annual withdrawals, occasional rebalancing, “buying the dip,” upper and lower bands on the asset weights, transfers of dividends and interest income, etc. (stay tuned for a future post on my blog). All those other ingredients are just clutter and distract from the essential mechanism of the Bucket Strategy: asset return momentum. Sometimes it works, and sometimes, it backfires. It’s the main reason why the Bucket Strategy cannot consistently beat the simple SAA assumption of fixed asset weights.

And I want to stress when I say that the Bucket Strategy is ineffective, I don’t mean that it is so bad that you shall never employ it. I could have reshuffled some of the annual returns to construct an example where the BS outperforms SAA by $1,000, like in Fritz’s example above. The bucket approach is unable to beat the SAA reliably and consistently. You can still use it, but don’t expect miraculous results in hedging against Sequence Risk. And absolutely don’t expect the Bucket Strategy to offer full SoRR Insurance.

The entire discussion reminds me of the misleadingly labeled “Yield Shield.” Raise your dividend yield to about 4%, and you now have a perfect “shield” against Sequence Risk. Only you don’t. It’s a hit-or-miss strategy, too. Sometimes you do better than a plain old index fund strategy. Sometimes you do worse, most recently during the Global Financial Crisis and in 2020. I pointed out that issue in Parts 29, 30, and 31 of my series. I’m not saying that you consistently underperform the passive index fund approach with either the Yield Shield or the Bucket Strategy. But you cannot consistently beat the passive approach with such gimmicks, either. And even in the cases when momentum works in your favor, the impact on your portfolio is so minor (e.g. just $1,000 in Fritz’s example) that you can’t really claim victory over Sequence Risk. This goes back to the point I made in Part 1 about how the sizing of the TAA bets is insufficient to make a difference in your safe withdrawal rate, even if you generate a little bit of TAA momentum alpha.

Also, just like the Yield Shield proponents, Fritz declares that his strategy is “working.” But Fritz is vague about what he means by “working.” If “working” means he hasn’t run out of money yet, that’s a very low bar. That’s not what our discussion was about. We need to set the bar much higher: the comparison should be, is the bucket strategy better than a simple SAA approach? Fritz didn’t provide any simulations to present how he would have personally fared with the much simpler SAA approach since 2018. And granted, I presented only one numerical example and had to defer the detailed simulations to a later post.

But if you are a regular reader of my blog, you will remember Part 39 of my SWR Series; I performed very detailed simulations to study a related issue: how changing the rebalancing frequency would have changed the experience of historical retirement cohorts. It’s the same hit-or-miss experience: sometimes, the asset weight drift helps you when you can milk that asset return momentum. Sometimes the drift hurts you when asset returns go through extreme whipsaws. This lack of reliable alpha relative to SAA is true in “made-up” numerical examples and historical cohorts. And I will show in a future post that the same holds for the bucket strategy.

4: If The Bucket Strategy is a gimmick, what’s the better strategy, and how do you manage it?

Karsten:

The way to manage Sequence Risk is to a) acknowledge that it exists and b) understand when Sequence Risk is more likely and less likely. If you retire while equities have been in a long bull market, you likely want to start with a lower initial safe withdrawal rate. But on the flip side, if equities have already fallen by enough, for example, in 2022, we can also afford to raise that safe withdrawal rate; see my recent post in the SWR Series on this topic.

But of course, there is relatively little we can do to insure against Sequence Risk fully. The best we can hope for is to hedge against Sequence Risk partially. The one method I’ve mentioned in our exchange that appears to be consistently beneficial during all past bear markets is the glidepath model. And indeed, I have marginally raised my risky asset share over the last four years. Post-retirement, I now have two bear markets under my belt. Knock on wood; those downturns were mild enough not to threaten my finances. Quite the opposite, our portfolio is up significantly despite the market volatility. We can now afford to take slightly more risk. But to new retirees, I still recommend using caution right around the retirement date. Start with a somewhat larger safe asset bucket and ease yourself back into risky assets as time progresses.

Another route to enhance retirement success is to think outside the box, i.e., employ asset classes outside the spectrum of assets in your standard retirement calculators. Real estate would be one option. But it’s not for everyone. I certainly don’t want the hassle of managing the day-to-day operations of a rental property portfolio. So, we have shifted about 12% of our investments into private equity real estate funds. If interested, please get in touch with Reliant Capital – accredited investors only, $250k minimum investment. We have less control over the investments, but we also don’t have to waste time dealing with tenants and plumbing problems on Christmas Eve. I also like the broad diversification over different regions and multiple properties. Moreover, these large multi-family properties offer enough diversification over idiosyncratic tenant risk. So far, we are happy with our investments and will likely shift more of our portfolio into this asset class.

Another approach involves an “alpha strategy” with a better prospect for adding excess returns than an unreliable TAA momentum approach. I have been running an options trading strategy with a neat track record. The strategy involves selling put options on the S&P 500 index. You trade derivatives on margin; thus, my strategy doesn’t require me to shift any of my existing assets; rather, I trade the options strategy on top of my current portfolio. The idea is to enhance the returns of my existing 75/25 portfolio and add about 1.5-2.0% additional returns with only small correlations to my existing asset classes; see the efficient frontier diagram below. The details of the strategy would go beyond the scope of today’s post, but here’s a link to my most recent write-up about this strategy.

You cannot shift the efficient frontier that far with a hit-or-miss TAA strategy based on backward-looking momentum signals and applied to just 10% or so of your portfolio. But that said, I don’t recommend any options trading strategy unless you have extensive knowledge in derivatives trading and risk management. It’s probably a bridge too far for most folks in the FIRE community. But I plan to offer this “short-put yield” strategy to a small number of high-net-worth clients at a future date, so stay tuned!

Fritz:

I like and support the concept of the Glidepath model and believe it’s actually a concept that supports The Bucket Strategy being a sound strategy.  As I mentioned, it’s my belief that The Bucket Strategy naturally leads to a Glidepath model, assuming the growth of the portfolio exceeds the SWR over time.  If one keeps Bucket 1 at a maximum of 3 years and the market outperforms the SWR (as it should over time), the retiree’s risk allocation will, by definition, increase.  In fact, the only time the Bucket Strategy would not lead to increased exposure to risk assets is if the market were underperforming, which is the time you would be pleased to not have the additional risk exposure.

It seems to me that The Bucket Strategy is perfectly aligned with the Glidepath strategy, given that it will lead to an increase in risk assets (assuming a rate of return > withdrawal rate).  In a longer-term timeframe in which growth exceeds the SWR, a retiree using The Bucket Strategy would see an increasing risk exposure with time, which is exactly what the Glidepath model dictates.  

An interesting side note:  This phenomenon could lead to drastic changes after the retiree begins drawing Social Security, assuming the buckets were “sized” based on pre-SS figures.  Using the asset allocation example from earlier, and assuming the same portfolio size, let’s assume a retiree now has Social Security income that covers 50% of their spending (and their spending stays the same).  By definition, this would lead to a 50% reduction in the annual portfolio withdrawals.   Theoretically, there could be a huge increase in equity exposure as the portfolio grows from 30 years of spending to 60 years of spending (a simplified example to make a point):

  • Cash:  3 Years (5%)
  • Bonds: 6 Years (10%)
  • Stocks: 51 Years (85%)
  • Total: 60 Years (100%)

I realize that’s an extreme example (to make a point) and many are already using Social Security and their portfolio to cover their spending.  If they’re already drawing SS and using a SWR of 3.3% (1 year of spending from a 30-year portfolio) to cover their spending, they obviously wouldn’t see the impact shown above.  Rather, I point it out for those retirees who have calculated their Bucket Strategy on their pre-SS scenario and haven’t thought about how the introduction of SS changes the numbers.

The Bottom Line:  The Bucket Strategy is not a cheap gimmick.  It’s a sound strategy, with some defense against SORR, elements of the often-referenced Glidepath strategy, and ease of execution for DIY retirees.  The strategy allows a retiree to cover their retirement spending with a minimal amount of management and stress and will allow them to sleep soundly through all but the worst of the inevitable bear markets we’ll face through our retirement years.

PS – Finally, I should stress that my goal with The Bucket Strategy is not to “achieve Alpha.”  Rather, it’s to have a strategy that allows me to sleep well, enjoy living in my retirement years, and know in advance what changes should be made to my portfolio based on market performance. To me, the Bucket Strategy achieves that goal, and knowing it should also lead to a Glidepath allocation (which has been proven by Big ERN to be the most effective model) gives me the assurance that the returns will be sufficient to support my lifetime spending needs.

5: Final final thoughts?

Karsten and Fritz: 

We hope you enjoyed our exchange. This wasn’t exactly the “celebrity deathmatch” that some readers expected or feared. We really agree on most issues. We are good buddies, and we respect each other’s work. We also frequently refer to and link to each other’s blogs. One of the commenters in Part 1 put it best; see below. Well, ThomH, your comment is a worthy final word that made us warm and fuzzy, too. And we hope to see you and your wife in Ecuador!

Thanks for stopping by today! Please leave your comments and suggestions below! Also, make sure you check out the other parts of the series; see here for a guide to the different parts so far!

Picture credit: pixabay.com

118 thoughts on “Discussing Retirement Bucket Strategies with Fritz Gilbert – SWR Series Part 55”

  2. “And the index fund managers wonder what the heck just happened here!” I totally get your reaction on this one but it’s totally your fault Karsten, with your overly complicated math and excessive analysis. Something that should be simple becomes a monster. Just go back to your SWR series, would someone find that simple? You’d need a math PhD to understand half of that…so I’m sorry but you have a great deal of fault on this one.

    Reply

    1. I completely disagree with this comment. The SAA approach could hardly be more simple. The most difficult aspect (applicable to either strategy) is selecting how much of one’s portfolio is to be allocated to risk assets. After many months of dithering, I’ve settled on 70%. I now carry out a rebalancing once a year and that’s pretty much all that’s required!

      Reply

    2. Not sure if this is a troll/sarcasm comment. But you comment is a slightly less civil version and a lot less thoughtful version of the ‘Jeff in MN” comment.
      Math Ph.D. material reads like “theorem, proof, theorem, proof, lemma, proof,…”
      On my entire blog, I’ve never published anything even remotely similar to Math Ph.D. level. This is all just basic applied math and personal finance.

      Reply

      1. I’m sorry, I must be an idiot then because I find your analysis and calculations extremely difficult to understand. But again, I work in construction and I don’t have a college degree which seems to be a prerequisite to become FIRE. I’m sorry

        Reply

        1. OK, understand.
          I don’t need to understand every step you are taking when doing your work. But I’m happy there are people that specialize in construction.
          And I appreciate the work my buddy Gary is doing when he helps me with construction projects around our house.

          I also hope that folks without a college degree can benefit from the insights in the FIRE movement. The formula is the same (save, invest in index funds, rinse and repeat and watch your portfolio grow), but I concede that it may take longer if your income is lower.

          Reply

      2. Karsten, Many of us appreciate that the points you make are backed up by worst case analysis based on math and history, instead of just hand waving with one example. Many times while reading your articles, I think of some worse case scenarios, “what if…”, and your next point addresses exactly that, with details to demonstrate the answer. So your articles are well thought out, with simple points for those who are not interested in the math, and also the details for those of us who find it difficult to bet our retirement on someone else’s hand waving.

        Reply

    3. I disagree. Even though I don’t fully understand the math behind (I’m a chemist) I feel that seeing all that goes behind Karsten’s results gives me full confidence in his conclusions. That’s the only blog to my knowledge that goes so deep in the analysis and that’s why I will follow his strategy.
      Thanks for the excellent work Karsten. Hope you can give us your opinion on the state of the economy. Also for your international readers like me, if you could sometimes make simulation with world index instead of S&P500 it would be very appreciated 🙂

      Reply

      2. Thanks Sebastien! That’s a sentiment I hear a lot: the math helps people get over their fear of early retirement. It helps us quantitative/logical/scientific people in a way that some of the psychological mumbo-jumbo just can’t.
        I’m glad I’m not the only one thinking that way! 🙂

        Reply

  3. “there is relatively little we can do to insure against Sequence Risk fully. The best we can hope for is to hedge against Sequence Risk partially ”
    This is not reassuring at all ! It scares me to a point that I prefer to keep working than having to face this.

    Reply

    1. Uuhhhm, that’s not my intention. I initially thought that there is no safe withdrawal rate for retirement lengths 50+ or even 60+ years. Turns out, with a SWR slightly less than 4% you can safely retire. Even with the the sequence risk.

      Reply

  4. Thanks for both your work. What you do is important to many who, due to limited financial experience (like the commenter above), rely on financial advisers blindly. This should give them comfort – whether they understand or not.

    Switching gears here, do you have any resources that compare the “37 words” assumptions you make for the last century (2B pop in 1900 to 8B today) to a century with zero population growth? What should long term stock multiples be with zero secular growth at various rates of inflation?

    Reply

    1. Thanks!
      Great question. There is some concern that demographic change (slower population growth, more old people vs. young) will have an impact on asset returns. I doubt that this effect is that large for stocks. Demographic change is slow enough so that the capital stock will also grow at a slower rate. And mean expected rates of return will not fall off a cliff as some people predict.
      But there is a concern that larger debt burdens and more government involvement will lessen our productivity.
      But there’s also the hope that some technological advancements like AI, Machine learning, etc. will boost productivity and make up for the losses.
      So, overall, I think it’s not a crazy assumption to use the same return distribution going forward.

      Reply

  5. “It’s also important to note that these rebalancing moves are not driven by market timing, but by a systemic review of asset allocation during the quarterly refill process.”

    Fritz – “Quarterly review – refill process” – I don’t think so! What about we have few years slide down – then what!!! You are propossing to sell equities at the bottom? The answer is naturally, no, but then what!

    I don’t believe in bucket strategy because it will make focus on something that I shouldn’t be focused too much on.

    At the end of any day, I am looking foward in growing my entire “net worth”. Take out money when I need it and move on on fun things in life.

    BTW nice article, but let’s focus on “one bucket” management and make sure it is going in the right direction.

    Reply

    1. Yes, that’s exactly my concern: Once you draw down the safe bucket too far, you might make an emotional decision and rebalance just at the wrong time – the market bottom. Too much of a whipsaw risk.

      Reply

    2. “Take money out when you need it” could also lead to having to sell at the bottom, IMHO. I like having the quarterly review process to keep an eye on things (I ignore it at all other times), and often find that there are counter-cyclical positions which allow refill even during a bear (as I did in 2022 and mentioned in the “bear” post linked above). There’s always a risk of having to sell in a downturn, but I believe the 3 years of cash and a formal review process mitigate the risk to the extent possible. I’m also a believer in a widely diversified portfolio, including 10% to Alternatives.

      Reply

  6. ERN I think you got it right when you said that the bucket strategy is a means of complicating the issue of asset allocation that one may believe they need a financial planner to execute!

    Reply

  7. Interesting, I haven’t heard of this “buckets” strategy before – I usually associate the term “buckets” with different tax-treatment accounts (IRA vs Roth vs Taxable).

    Karsten’s SAA approach seems more mathematically rigorous, as well as simple, in my mind. However, I think I can see part of the appeal of the buckets strategy Fritz recommends – it seems best suited for overcoming psychological difficulties with spending money in retirement, especially by having a regular “paycheck” that most folks are accustomed to.

    It would be very interesting to see how the average and variance of the two approaches differ for a wide range of historical market returns, as well as the impact on SWR. That could help settle the “which is better mathematically on average” question, even if some folks still find the “buckets” approach more comfortable. Unless you’ve already done that in a previous post Karsten? Perhaps SWR Series Parts 48 and 25? I need to read those more closely.

    Reply

    1. The problem with the bucket strategy: There isn’t the bucket strategy. There’s only one way to do SAA with monthly rebalancing, but you can model so many different bells and whistles. For what it;s worth, I did a careful analysis for rebalancing your SAA less frequently than monthly See Part 39.
      Not really earth-shaking.
      Main Results
      But you get a very small advantage on average when you rebalance less frequently. It’s possible that a Bucket strategy might go in the same direction.

      Reply

      1. I think I’m more of a never rebalance guy. Although taking withdrawals “off the top” of the tilted allocation makes some sense. I’m mostly in taxable. I just never seem to see how the different rebalancing styles affects someone in taxable (please consider blogging for both taxable and Roth/IRA portfolios when doing the math). I’d certainly be more open if it was in an IRA without the tax drag.

        Reply

        1. Depends on the return patterns and your portfolio.
          If you have the “right” account mix, you can certainly do the asset shift only in the IRA/401k/Roth and keep the taxable accounts free from rebalancing and tax drag.

          Reply

  8. Buckets are just an asset allocation until you address the most important question: How are the buckets refilled? That’s where the devil is in the buckets strategy. Too few talk about it so I’m glad Fritz did. In the end, it’s a market timing strategy but perhaps not a bad one as your asset allocation gets more aggressive as expected returns climb (and risky asset values fall).

    Reply

    1. Thanks, Dr. Dahle! That’s exactly my concern about bucket strategies as well. Too much of a headache to decide when and how to refill the buckets.
      PS: Your comment was initially filtered out as spam, not sure why! Sorry about that!

      Reply

  9. Thank you for this. I love this blog, but then again I did study math a while ago and enjoy google sheets. I like following your logic as you come to a conclusion as to the best way to do things in early retirement. It does seem to me, anyway, like you’ve arrived at a conclusion (at least for now).

    To summarize, use the ER spreadsheet and take into account SS and other planned income and expenses. Use the CAPE ratio on an ongoing basis to determine how much you can spend for any given month. Use a glide path to the retirement date (100% equities to 80%/20% over 5 years) and then after (80%/20% ->100% equities over 5 years). Consider alternative investments like Real Estate (10-15% range). Use SAA to keep the equity percentages at your target.

    Reply

  10. I am far less experience than either Karsten or Fritz but I plan to execute elements of both.

    Buckets allow me to compartmentalize and let’s me stop my emotions from causing me to do rash things during wild market swings. If I maintain a fixed cash bucket (which includes ultra short term bonds) that can fund 3 years of essential expenses, this makes it easier (psychologically) for the rest of my portfolio to stick to an SAA strategy. So I will always maintain this floor level of expenses in a safe “cash” bucket to keep my sanity. The rest of my portfolio will follow SAA rules (except it will be more aggressive than 70/30).

    I understand this is not totally rationale (money is fungible), but a constant minimum cash bucket provides psychological comfort. I’ve paid off my mortgate early even though I would have increased my net worth more if I borrowed at around 3% and invested in the market because it feels good not to owe anyone. I keep a separate set of assets to dabble in individual stocks even though I am lucky to match the market (with higher volatility so it’s not optimal to do so) but it scratches my itch to play the market so the bulk of my portfolio sticks to a SAA strategy. I’ll likely overweight dividend stocks, REITS and real-estate syndications because I like the dividend cash flows better than selling appreciated stocks to meet my immediate spending needs because spending dividends just feels more like responsible spending.

    At the end of the day, having a reasonable, yet sub-optimal strategy that I’m comfortable sticking with is better than an optimal strategy that causes stress or even worse, may not be consistently followed.

    Finally, since this is a “fight”, my “split decision goes to Karsten” but I’ll still practice some of what Fritz advocates.

    Reply

    1. UK reader, two years into my retirement journey.

      I started out holding a couple of years’ living expenses in cash, purely for psychological comfort, and despite accepting that over the long term, this would be a drag on growth.

      However, my portfolio is invested in global trackers and is very liquid, and I can sell small amounts to access cash within a couple of days. Accordingly, I have weaned myself off this “cash-heavy” position; I no longer need the cash comfort blanket!

      Once again many thanks Karsten for your wonderful articles (and comments). I’ve been following you for several years now and you have quite honestly changed my life.

      Reply

      1. Thanks Joe, for the kind words!
        Yes, that’s also exactly my thinking: back the old days it might have taken a phone call, a hefty brokerage fee and several days for you liquidate your equity holdings into cold hard cash. But today I can log in to Fidelity and transfer money out of a brokerage account just as easily as out of a MM account.
        If people want to hold cash, there can certainly be an asset allocation reason, but liquidity is no reason anymore.

        Reply

    2. Thanks a lot Phillip. Great summary! I can live with people using bukets as a psychological help, to sleep better. But I’m glad that you also recognize the math vs. psychology aspect and don’t expect miraculous results form the bucket strategy.

      Also, I have about 1% of my portfolio in dividend stocks. Play money. Nothing wrong with that either!
      Thanks for sharing your honest opinion!

      Reply

    3. Phillip, you’re explanation of “a blend” is actually a pretty good summary of how I manage my buckets. In essence, it’s SAA with a focus/priority on the cash allocation. I also agree there are psychological benefits to having the cash cushion. I enjoyed your comment until the last paragraph. 😉

      Reply

  11. Thank you, Fritz and Karsten for this post!

    Assumptions / Theses:
    – People are different and that is a good thing.
    – People act mixed from logical and emotional reasons.
    – But the range goes from logical Vulcans to emotional divas.
    – Supposedly 20% are rather/majority reason-driven and 80% rather/majority emotion-driven.
    – People do not like uncertainty, but look for reliability, predictability (even if only apparently).
    – People do not like not understanding something. It gives them bad feelings. They are than more afraid
    – People are masters of displacement and postponement
    – People do not change if they do not have to / strongly want to.
    – People are imperfect, poor creatures, driven by fears, greed and emotions.
    – We all need redemption 😉

    Deductions:
    That is why the majority prefers the more emotional bucket strategy and the minority the more logical SAA strategy.
    For SAA you have to be able to calculate a little bit, you have to have a feeling for market histories (or believe Karsten) and you have to make a decision for every single withdrawal (stocks or cash) and implement it (broker), so you have to be rather affine to capital markets.
    For BuckStrat one has the illusion, the first 3 years nothing can go wrong (cash buffer), the next 6 years actually also nothing (bond buffer) and at the latest in 10 years the stock market “will stand somehow higher”. Perceived risks in the future are never as important as perceived risks in the present. The idea of seemingly having to decide less often is appealing.
    The BuckStrat is vividly understandable (Buckets), and the group one can relate to are normal people (instead of math geeks).

    Solution:
    It doesn’t matter if you use the SAA or the BuckStrat or if you think it’s better, the reasons are in your personality. The main thing is that everyone respects each other, because everyone is equally valuable.

    What do you think about this explanation approach?

    Liebe Gruesse
    Joerg

    Translated with http://www.DeepL.com/Translator (free version)

    Reply

    1. A thoughtful comment, Joerg, and I tend to agree. Different personalities are drawn to different solutions for different reasons. The beauty of “personal” finance is that each can do what best fits their personal style, so long as their reason for doing so is based on knowlege of the tradeoffs.

      Reply

    2. This is an awesome summary! I cater more to the math geeks, so hence my bias. But I am not going to claim that the BuckStrat is worse in the outcomes.
      Apparently you typed that in German and translated it through an app? Pretty neat! Modern technology is amazing!

      Liebe Gruesse zurueck nach Deutschland! (oder Oesterreich oder Schweiz!)

      Reply

  12. I think at the end of the day you should follow whatever strategy lets you sleep well at night, but……..

    Professor Estrada at the IESE did a study on static allocation versus bucket strategy from a global perspective.(for those of you that need more math) https://blog.iese.edu/jestrada/files/2019/01/BucketApproach.pdf

    I think his final paragraph says it best:
    “In short, the results discussed here suggest that financial planners should strive to explain to clients the benefits of static strategies relative to those of bucket strategies. They should explain that satisfying the behavioral need of mental accounting imposes a cost in terms of performance. And they should attempt to convince retirees that however plausible, comforting,and easy to implement the bucket approach may be, a static strategy with an appropriate asset allocation would be just as easy to implement and would ultimately make them better off.”

    Reply

    1. That’s a nice paper! Thanks for the link! Glad that there is more empirical evidence to support my approach.
      I really think that financial advisers should spend that time to ease people’s concerns and explain that the SAA has all of the same features of the BuckStrat, like buy low, sell high, rebalance, buy the dips, etc.

      Reply

    2. I agree, this is a great paper. Exactly the kind of analysis I was wondering if someone had done. The final section conclusions are right in line with my intuition as well. Thanks Alex.

      Reply

  13. Fritz, I think you made an error in your comment which followed the side-by-side comparison of Bucket v. SAA. You wrote, “By comparing the two approaches, you can see The Bucket Strategy actually results in a slightly higher balance over the three-year period ($1,027,000 vs. $1,025,940), given that no stocks or bonds were sold during the downturn with The Bucket Strategy approach.” Your error, if I’m reading it correctly, is that Karsten’s portion of the side-by-side encompassed only 2 years, whereas yours was 3. Yes, your total after 3 years was higher than his after 2, but if you’d allowed his stocks/bonds to improve during the post-bear recovery like yours were able to do, I believe the math would tip in Karsten’s favor.

    Reply

    1. Good catch, Casey. I looked it over – the problem is the yellow highlighted title in Year 3, which erroneously reads “Year 2”. The figures in the table are correct, only the title is incorrect. Thanks for paying attention to the detail. Apologies for missing it during my editing.

      Reply

      1. OK, that’s a good catch then. It’s not a bog deal because I have generated cases where the BuckStrat worked in your favor. It’s relatively easy to generate a $1,000 difference in outcomes.

        Reply

    2. Well, you can come up with lots of different assumptions and return patterns and some will favor the SAA and some will favor the BuckStrat. Fritz used different returns and came up with a different final outcome, so I trust that his calculations were correct.

      Reply
  14. Pingback: Discussing Retirement Bucket Methods with Fritz Gilbert – SWR Collection Half 55 – Early Retirement Now - Get Invest USA

  15. So what about the “never-refilled cash bucket” as mentioned in Part 25?

    There you wrote:

    “So, for the record, let me state that this cash bucket strategy seems to work pretty well, despite my previous doubts!

    It’s relatively inexpensive insurance against Sequence Risk!

    Think of it as a mini-glidepath during the first few years of retirement!

    And it ‘only’ takes the flexibility of getting to 27.5x instead of 25x annual spending!”

    I realize it means saving extra (27.5x versus 25x) but you never have to worry about refilling the bucket from your regular portfolio.

    Seems cheaper & easier than paying the price since my portfolio is nearly all in taxable to move from my current allocation (80/20) to 100/0 then to 80/20 then back to 100/0.

    Reply

    1. As I stated: its a form of a glidepath: You phase out the cash bucket and leave it empty after that.

      My only problem with the BS approach is that people falsely believe that they can market time themselves through a bear market and end up with the SAME asset allocation again as before.

      Reply

      1. What market timing?

        I’m just pursuing portfolio survival with a less complicated method than repeatedly changing asset allocation.

        With fewer transaction costs for people like me whose portfolio is nearly all taxable.

        Reply

  16. Hello ERN and thank you for all the great information on your blog!

    2023 will be the first year that I live off of my accumulated assets. I’ve spent quite a bit of time working with your Google Sheet establishing an SAA and safe withdrawal rate, so I feel comfortable with my spending budget for 2023.

    Looking forward, is there a place on your blog where you walk through the basic mechanics of adjusting the spending amount for subsequent years? I’m content to take the simplest strategy possible. I plan to rebalance back to the SAA, so my question is merely what are the mechanics of establishing a spending budget for each subsequent year.

    Reply

    1. If you are already comfortable with ERNs spreadsheet, just reuse it every year subtracting one year from your assumptions. This will get you an SWR you can apply for your budget.

      Reply

    2. Not directly related, but every year-end I update my Net Worth, subtract “non-spendable” assets (e.g., cars, home), and multiply the remainder by 3.0%, 3.25%, 3.5%, 3.75% and 4.0% to give me a range of spending options from my new balance. It’s a good way to keep your spending somewhat variable. I target the 3.0 – 3.25% of the range. I wrote about my annual review process here:

      https://www.theretirementmanifesto.com/a-step-by-step-guide-to-your-annual-financial-update/

      Reply

      1. What if the portfolio dropped by 50% during the last year? Should someone then decrease the withdrawals accordingly or stick to the previous year’s result?

        Reply

        1. I don’t think there will be a “last year” because your life expectancy moves with you. And by the time your LE is down to 1 year, you’re 113 years old. You can tell your “kids” that their inheritance will be a little bit leaner at that point. They will understand. 😉

          Reply

          1. By “last year” I meant previous year, not the last year of your retirement. Lets look at this example:

            Lets assume a 40 year old applied a safe withdrawal rate of 3%. He has a portfolio of 1 Million and can therefore withdraw 30k each year.

            One year later he wants to update the figures. Unfortunatley the portfolio is down to 500k due to a recession. What is now the new annual budget?

            a) 500k x 3% = 15k

            or

            b) still 30k, since he initially applied a safe withdrawal rate of 3%, which is derived based on severe market downturns and should therefore still be ok even there was a 50% drop?

            What do you think?

            Let us ignore the change in life expectancy here, it is in this context of minor relevance (we ignore the fact that the remaining time for withdrawals decreases from lets say 60 to now 59 years (which means a initial planning horizon of 60 years until age of 100), which is perfectly fine for such long maturites).

            Reply

            1. Gotcha! Sorry, was overthinking here.
              First, your portfolio will likely not decline by 50% if you have a diversified portfolio, somewhere between 60/40 and 80/20.
              Second, even if your portfolio is down, you should not reapply the historical SWR to that decimated portfolio. Those historical failsafe rates are calibrated to when the market is right at the peak. Once the market has fallen, you can afford a higher rate. $30k/year (6% now relative to the portfolio) might still be totally safe. See Part 54 of the series.

              Reply

              1. Thank you very much! I understood that someone should only reajust the absolute withdrawal amount in case the value of the portfolio increases over time. But in case of a decrease of the portfolio value, the withdrawals will remain unchanged, since the safe withdrawal rate was initially applied.

                Reply

                1. Yes, I mostly agree. Of course, if your portfolio is down even worse than what we’ve observed in the past, you might still have to adjust.
                  Another way you would adjust up and downward would be by using a CAPE-based withdrawal rate. It allows you to withdraw more initially but also has the risk that you might have to lower the withdrawal amount occasionally.

  17. Pingback: Weekend Reading - More dividends and 2023 Predictions edition - My Own Advisor

  18. For those of us who are DIY in mostly balanced funds and not separate bonds, stocks, how to rebalance? We have held these funds for years so it seems to me we’d be fine to take money out even in a bear market if needed and just let the funds rebalance themselves. I plan to build 3 years of expenses in cash or Treasuries, then more or less keep that full by pulling out of the balanced accounts each year (or not) based on how the market did. Is that kind of a bucket strategy? Kirsten are you saying I don’t need that much cash to start out and it is an unnecessary drag on returns?

    Reply
  19. Pingback: Discussing Retirement Bucket Methods with Fritz Gilbert – SWR Collection Half 55 – Early Retirement Now - Best Investment Shop
  20. Pingback: Weekend Reading - More dividends and 2023 Predictions edition - Breaking Pro News
  21. Pingback: The Sunday Best (01/29/2023) - Physician on FIRE

  22. Another excellent column. Question for Big ERN: How do you classify private real estate funds? As you noted previously, they are somewhere between public equity and debt on the risk spectrum. So if you have 12% of your portfolio in such funds, does that imply 6% less in each of public equity and debt? Put another way, if I would target 70 / 30 without private real estate funds, then is the correct allocation 64 / 12 / 24 including 12 in such funds?

    Reply

    1. I would view the PE funds as effectively 100% equity. THey don’t fluctuate as much in value but that’s just because the folks running the funds are not too eager to marking down the assets during a recession.
      So, if you have 70/30 S/B plus 12% RE, I’d consider that (70+12)/112=73% equities

      Reply

  23. Thanks a lot to both of you for presenting your views and arguments!

    For most of the FI community, it is not easy to figure out which strategy is the best to achieve FI and then navigate during RE. Even this morning, I had a discussion with a friend and we could only agree that there is no single path to FI depending on the country you live in. So even more paths when you are a retiree.

    My wife and I calibrate our withdrawals based on the SWR formula based on the CAPE ratio. Our portfolio includes ~1 year of cash, and we apply a passive glidepath strategy for the rest. The 1-year of cash gives us the peace of mind that we have enough money to worry about, and we remain unaffected by the short-term market performance. The passive glidepath consists of rationalising our portfolio, and selling Reits and bonds, or taking the opportunity that some of our bonds were called. Thanks Karsten for all your analysis and simulations!

    To add some alpha, I am diving into your selling put option strategy, and will play with a paper account before putting any real money into it. Needless to say that I was ecstatic when I read that you were planning to offer your “Short Put Yield” strategy. I always thought that it would be wonderful for people to access your strategy out of the box.

    Thank you again, and can’t wait for your next post!

    Reply

    1. It’s hit-or-miss. That’s the whole point. I simulated a bucket strategy and sometimes it outperforms and sometimes it underperforms.
      A glidepath will do consistently better during the bad retirement cohorts. But it will also lag behind a static SAA in cohorts that are not hit by SoRR.

      Reply
  25. Pingback: Discussing Retirement Bucket Strategies with Fritz Gilbert – SWR Series Part 55 – Early Retirement Now - Online Special Finance

  26. hey guys great debate!!! to each retiree his own it seems between SAA vs. bucket strategy. Are you guys familiar with Wade Pfau and the RISA (Retirement Income Style Awareness)? Seems that the RISA matrix would hold a place for both retirement strategies, where SAA would be for “total return” folks while the bucket strategy would be more for the “time segmentation” retirees. do you guys agree?

    Reply

    2. Nice looking tool. Fritz has written a post and he seemed to like it. I’m not convinced that this is the right tool for me. For example, it doesn’t seem to take into account my specific future cash flows and the current market valuations (PE, Shiller CAPE, bond yields, etc.). But it’s certainly visually appealing.

      Reply

      1. Seems I am a year late to the party – just saw these and read through both blog posts and all comments, so sorry for the late comment.
        Karsten, I encourage you to take a much closer second look at RISA, because based on your criticism (calling it “visually appealing” is, likely unintentionally, actually dismissive of seminal work by the creator of the RICP (Retirement Income Certified Professional) certification, Dr. Wade Pfau, even if you didn’t mean it to sound dismissive), I believe you misunderstood its primary purpose, which is actually behavioral finance. RISA stands for Retirement Income Style Awareness assessment, and its primary purpose is to determine a retiree’s preference and comfort for different retirement income sourcing strategies based on two axes: safety vs probability and optionality vs commitment. The probability/optionality quadrant is total returns, which what you call SAA falls squarely within, and is acknowledged by Dr. Pfau and the other RISA developers to be the one with the highest expected return. It is clear that you would score far into the total returns quadrant if you were to take the RISA. However, note that 1/3 of respondents fall in this quadrant, meaning that 2/3 fall in other quadrants. I believe this is the underlying reason some readers are resistant to SAA – because it does not fit with their own behavioral finance tendencies.

        I was going to make a similar comment as Rikki made before I saw Rikki’s above comment. I think RISA is the missing piece in the SAA vs bucketing debate you and Fritz had here. Fritz alluded to it earlier when he responded that bucketing wasn’t necessarily generating alpha, but may matter more if it allows someone to stick to an allocation strategy more of the time. This actually applies more to income protection (Social Security, annuities, pensions) vs total returns, where having a reliable income floor can make it easier for someone to ride out a bear market and long recovery because they are less harmed by portfolio losses.

        But the point is, your perspective is to teach what is the financially optimal outcome without regard for behavioral finance, but it is well established that people usually achieve lower returns than the investments they hold because behavioral finance gets in the way and they buy and sell at times that undermine portfolio performance. So behavioral finance matters and may play as big or bigger role than making the optimal financial choices. So which strategy a retiree picks and more importantly can stick with depends as much on their RISA style as on financial optimization considerations.

        Reply

        1. The “behavioral finance” reference doesn’t help. In fact, if you view the bucket strategy as anything different than a simple SAA setup to diversify (sequence) risk, you already lost me because you fall into the “mental accounting” bias.
          There are two ways to deal with behavioral biases:
          1) explain to the client why the SAA makes sense.
          2) keep the truth from the client and keep him/her/them dumb and uneducated and reinforce irrational behavior.
          It sounds like RISA is choosing path 2 (at least for some people), which is making it less appealing to me. See my post here: https://earlyretirementnow.com/2017/06/14/good-advice-vs-feel-good-advice/

          And my other concerns are also still out there:

          it doesn’t seem to take into account my specific future cash flows and the current market valuations (PE, Shiller CAPE, bond yields, etc.).

          Is there a behavioral finance way to nudge people to get people interested in asset valuations?

          Reply

          1. We seem to be speaking past each other a bit, though I have reason to believe this is simply a small miscommunication as opposed to an fundamental disagreements.

            First, I agree that bucketing can be technically equivalent to SAA if it is consistently rebalanced to SAA sufficiently frequently and consistently. Kitces provided a case study showing exactly that in one of his blog posts (which I think was referenced in the comments).

            However, bucketing can be any number of different things, some of which can really mess up portfolio success odds, depending on the specific replenishment rules. And the RISA folks agree with you that without careful consideration towards asset allocation, you can really mess up by using bucketing. I believe Dr. Pfau wrote it up elsewhere, but what I found after a quick search is in the RISA podcast (Retire With Style), here: https://risaprofile.com/episode-101-uncovering-how-time-segmentation-actually-works/, how anything less than careful replenishment rules diligently executed can cause portfolio failure or at least much worse outcomes. So I don’t think they’re hiding the truth. In fact, they’re making a point of educating on it.

            Wade describes an alternate bucketing replenishment strategy in the above podcast that he favors, which he referred to as critical path. It’s basically a rising equity glidepath when your portfolio is down, and you can replenish only when it’s up from a starting benchmark. That pushes allocation higher when the portfolio is down and needs the boost from higher equity expected returns, rather than static SAA targets. You, Wade, and Michael Kitces have all identified a rising equity glidepath as a useful tool for the toolbox.

            Also, mental accounting can be neutral or negative from an optimal finance perspective (if it results in more optimization, I don’t think either of us would call it mental accounting any more). For instance bucketing to an SAA is neutral. Bucketing to a sub-optimal asset allocation is negative. If it’s neutral, and especially if the retiree understands how to make sure it remains neutral or nearly so, then it does little to no harm. But it may be important for the following reason…
            The point the RISA folks, Fritz and I have made, that you don’t seem to be acknowledging is that behavior finance matters to the success of any retirement income strategy, and in some cases, may actually matter more than optimizing the numbers, because if a retiree fails to follow the optimal strategy due to behavioral finance issues (too much perceived risk causing someone to sell near the bottom, for instance), then no amount of theoretical optimization will help. And simply explaining why they’re wrong (e.g. that the other method is more optimal) won’t magically cause people to completely abandon behavioral finance biases, because they’re often emotionally/fear based, so logic can’t necessarily overcome them. Bucketing, or more generally, time segmentation may serve a role for some retirees if executed in a way that is close enough to optimal to not reduce the success likelihood of their retirement portfolio too much while allowing them to be comfortable/feel safe enough to stick with it.
            My interest in time segmentation is more around a Social Security delay bridging strategy with a rising equity glidepath (e.g. spending down fixed assets to delay Social Security claiming), being most conservative with asset allocation around the age of retirement, combined with a healthy post retirement emergency fund, so I don’t think I’m succumbing too badly to mental accounting (though I know the model isn’t 100% optimized). But I also recognize that to sleep at night, I would never be OK with 100% equities and little to no reliable income and/or emergency fund, regardless of what the odds say, because I fall on the safety first (vs. probability based) side in RISA. In other words, I’m willing to give up some upside to provide a larger (CPI indexed) reliable income floor that will last my lifetime. Because if my portfolio does run out while I’m still alive, my income floor is higher.

            Regarding your question about behavioral finance nudging people to be interested in asset valuations, I think it’s primarily a communication issue rather than a fundamental disinterest issue. Your writing approach is deeply technical, because you are a technical expert, but a significant portion of your potential audience for above isn’t, and really only wants to know “what is a SWR in my specific circumstance and at this time?” and “whom can I trust to provide me that answer?” You’ve established a prominent reputation, at least in FIRE circles, which hopefully should already take care of the second. The reason the first doesn’t gain more traction, I believe, is due to the voluminous nature of your writing style/blog. Most people aren’t looking for a 60+ part blog series on SWR to read through (though having a lot of these available for on demand reference/education/curiosity is great! and I’ve read several of them myself). They’re just looking for their SWR answer. I think you’re on to something with the [Start here] tab, but I would reword the SWR references on that page, to start with the SWR toolbox spreadsheet (part 28), with instructions, and a very brief explanation of why and how to use it. I’m assuming this would satisfy your above concern, since it includes success probability tables incorporating most or all of the factors you mentioned. People who want to learn more can dive into all the detail they want with the full series (it took me a long time before I was ready to dive into the series, because frankly the breadth and depth of it was daunting), but I think most just want to know what is my SWR (or SCR) now with my portfolio characteristics, retirement plan, and current market conditions. You have all the pieces, but it requires a lot of reading and rummaging around to find it.

            I think if you were to steer people directly to the SWR toolbox and spreadsheet with very simple and plain instructions on how to use and what it means, it will get a lot more visibility and use. As to how this relates to behavioral finance: easy and automatic are key behavioral finance factors, and above would make it much easier and accessible to more people.

            Reply

            1. I don’t think even Fritz would argue that his bucket approach can overcome behavioral biases. I don’t even think the questions in the RISA approach are capable to really trickle out the degree of behavioral bias anyway.

              I agree with the glidepath approach. But nothing in the RISA approach utilizes the GP. Even worse, not only did the RISA make no recommendation about an SWR, there was also no asset allocation recommendation.

              You seem to be very into this approach. No idea why anyone would spend so much time defending the shoddy analysis that’s RISA. Do you work for that company?

              Reply

              1. No, I don’t work for the RISA folks. I have found it useful however.

                The reason I strongly advocate for RISA is only related to their specific advisor & investor psychology/behavior finance bias aspects (I’m not claiming, nor are they that RISA can overcome all behavioral biases), and I am *not* advocating for RISAs implementation recommendations. Your primary criticism seems to be the implementation recommendations, and I’m neither advocating for nor defending RISA’s implementation strategy recommendations. I agree that you have some valid criticisms of those. Christine Benz @ Morningstar explains the part of RISA I am advocating for better than I can, here: https://www.morningstar.com/retirement/whats-your-retirement-income-style.

                The primary problem RISA solves is the issue of an investor going to a financial advisor or reading/listening to financial media steering them to or away from total returns (probability & optionality) or conversely increasing reliable income (safety first & commitment) – via SS delay strategy, annuitizing pensions, purchasing annuities, etc – when the investor’s behavioral finance characteristics are a better fit for the opposite strategy. RISA provides a new tool for investors planning for their retirement to first figure out what retirement income sourcing approaches they’re comfortable with and then seek assistance as needed (or do it themselves) consistent with their own investor psychology and priorities (for instance safety first or probability based) to implement something consistent with the psychology and approach in a way that has sufficiently good odds of success.

                In the case of safety first/commitment (reliable income), it’s a conscious informed tradeoff to reduce the downside risk (higher contractually obligated income) with the potential for an expected reduced upside (reduced or even exhausted portfolio income). Still behavioral finance assisted by RISA, but not mental accounting, because in this case, the tradeoff is explicit and intentional and simply optimizing for a different attribute (safety first over probability). It’s best thought of as an insurance decision vs an investment decision.
                How much is an investor willing to pay for insurance against the worst case vs not pay for that insurance and run with the probabilities for a higher outcome?

                Reply

                1. Thanks for the link. I like Christine Benz at Morningstar and the article does a good job explaining the 4 quadrants. It also explains why nobody in the Bogleheads and FIRE movement takes this very seriously because, as Benz states, most of us are in the total return preferred quadrant and don’t really need this other stuff. But it’s good to check if you’re really a Total Return kind of retiree.
                  So, for determining what flavor you like, sure, use the survey.
                  I should also state that I’m working on a post to go through the pros and cons of “Safety First” so I’m not generally opposed to that idea.

        2. I actually took the RISA and wrote a review on it. I agree it’s a valid contribution to the retirement planning space, as it helps customize the best solution based on one’s psychological approach to money and security. Karsten, it’s worth taking a look – the RISA offers solutions to the retirement income puzzle by matching the solution to the personality. Want more security? Consider an annuity, etc. While the various solutions may not be absolutely optimal mathematically, I agree with John that their value lies in the incorporation of one’s existing preferences.

          Here’s the link if interested:

          https://www.theretirementmanifesto.com/my-evaluation-of-a-powerful-new-retirement-tool-risa/

          Reply

          1. Fritz, thanks for the response and pointer. I read your above post a couple years ago and forwarded it to a couple people and a mailing list back then to let people know about and try RISA.

            I re-read it after seeing your above response and was surprised to see that you fell in the total returns quadrant and not time segmentation (though you were near the border between the two) given your affinity to the bucket strategy. I’m curious if you have any insight on that apparent (minor) paradox?

            Reply

            1. Interesting question, John. While I’m a proponent of the Bucket Strategy (time segmentation), my thinking is that it allows me to ensure I keep a cash buffer on hand for shorter term volatility, but my main focus is on keeping Bucket 3 cranking out those inflation-beating returns for the long term. To me, it fits with my being “on the border” with my RISA score – I like time segmentation to protect my near-term spending requirements, but I favor total return to offset the longer-term risk of inflation.

              Reply

          2. Thanks for weighing in, Fritz.
            I did the RISA survey and I’m not impressed:
            1: Mine did not spit out any SWR recommendation. Only the relatively generic, boilerplate output that tells me that I’m not benefitting much form hiring an adviser.
            2: Even if it had spit out the SWR, it doesn’t seem that the SWR is based on anything resembling asset valuations (i.e., CAPE ratio, interest rates, etc.) or future cash flows (Social Security, pensions, etc.). That fact that for you, Fritz, the program spits out 3.75% is the glaring proof that this program is useless. For folks who retire in their 50s, I would routinely generate SWRs in the high-4s or even 5% and 6% range due to expected future cash flows.
            3: The questions were really repetitive. Sometimes multiple questions all with the same flavor on the same page and on consecutive pages.

            Again, I’m happy people find this useful. But it doesn’t impress me at all.
            And now I’m getting emails from the provider offering me help, even though I clearly answered on the survey that I don’t want any.

            Reply

            1. Karsten,
              I few thoughts on your experience trying the RISA survey.
              1) Based on the title (Retirement Income Style Awareness assessment), what I’ve read and heard about it from the founders and multiple reviews, and having used it myself, my sense of what is important and groundbreaking about RISA, is not the specific plan recommendations, but rather the RISA Style Matrix (the safety first/probability, and commitment/optionality) axes, the resulting implied retirement income sourcing strategy, how you prioritize retirement concerns (longevity, lifestyle, liquidity, legacy) and advisor usefulness/preferred relationship (if any). This is important because typically people fall into a retirement income style based on the biases of their advisor or whom they’re listening to (you appear to be total returns, an insurance agent will be income protection, etc) rather than an approach that fits with their own investor psychology and then finding an implementation approach they can live with for the long term and any resources/people needed to help based on that knowledge. This tends to result in people abandoning/changing their retirement income strategy after it was “forced on them” or causing them to not implement the plan at all or consistently because it’s a mismatch, for instance with their risk tolerance.
              2) For someone who is self-directed like you, the report says: “Do-it-yourselfer:
              For DIYs, the RISA™ Profile serves as an effective way to get a baseline review of strategies and tactics that resonate with you and should be further reviewed. In addition, it will point out potential issues that you may need to overcome related to your personal preferences and the behavior hurdles of self-managing your financial situation into retirement.” These are not ready to implement, robust plans. Rather starting points for consideration only. I agree with your criticism that they’re not very robust, but I think the RISA folks would also agree and say they weren’t meant to be.
              3) Whether it spits out SWR recommendations or not would depend on which quadrant and sub-quadrant in the RISA Style Matrix you ended up in.
              4) I’m confident that your own SWR analysis is far more comprehensive than what is implemented in RISA per #2 above (heck, your own SWR toolkit and SWR series is probably the most comprehensive in the world!), though I suspect the RISA one might take into account how you answer balancing lifestyle vs longevity risk (front vs back loading).
              5) I’m a bit surprised by your comment on 3.75% being out of line, as using your latest SWR toolkit, I’m seeing exactly 3.75% (in the upper table) come out as the 0% failure rate with CAPE>20 with the assumptions I put in for a similar circumstance to the one you describe (FYI, I filled it out and got that answer before you had replied here). The lower table solving for 0.00% failure rate shows 3.9%. Not seeing high 4s-6% range and I include SS/pension estimated income streams. Let me know if you’d like to diagnose this further offline. If so, I could email you a filled in SWR spreadsheet at the address you list on your contact page.
              6) Regarding repetitive questions, they’re a necessary evil for internal consistency measures in survey statistical analysis. They ask basically the same thing a few different ways to increase confidence that the answer isn’t skewed for some respondents by the way the question is asked. If you ever took a professional personality style test (like Myers-Briggs, or Margerison-McCann, or Big Five), you may recall they do this also. It’s standard practice.
              7) As to you answering that you aren’t interested in help (e.g. self-directed) and still getting contacted, that’s likely for two reasons:
              a) RISA and the affiliated Retirement Researcher provide educational support to self-directed investors, so it makes sense they would ping for a follow-up. In fact, it’s a distinguishing characteristic of their offering, as most companies only offer advisors or brokers (sales).
              b) I’d be very surprised if they connected the survey result logic to the marketing email logic. If you use their free tool, they likely unconditionally email you with the option to unsubscribe, regardless of how you answered the survey.

              Reply

              1. 3: that’s disappointing. But again, since the SWR in Fritz’s case is wrong anyway it’s probably better to get no recommendation than a totally wrong one.
                5: you shouldn’t be surprised. 3.75% is the SWR for someone without any supplemental flows. Someone like Fritz has large supplemental flows around the corner: pension and Social Security, so the initial SWR is much higher.
                6: yeah I suspected that, but I took a screenshot of 6 questions in a row that are essentially all asking the same.
                7b: yeah, agree. The marketing folks are different from the research folks.

                Reply

                1. You misread what I said re SWR. I said “Not seeing high 4s-6% range and I include SS/pension estimated income streams.” In other words, my circumstances are similar to Fritz’ regarding supplemental SS and pension, which I included in the SWR spreadsheet and it came up with 3.75%/3.9% in the 0% failure values for CAPE>20. So, it is outputting inconsistent with your assertion that it should be high 4s-6%.

                2. I can’t comment on what you put into your sheet. I am referring to Fritz’ situation. 3.75% is very likely far too low for him.

                  Some observations:
                  In my 10-part case study series, there was only one case that ended up with 3.75% (much younger than Fritz, special situation with special provisions for large expected health expenditures later). Most were above 4%.
                  But the 10-part case study series would be way too conservative because they were all far younger than Fritz. Fritz was 55 in 2018 when he retired, so he should be 60 or 61 yo today in 2024.
                  I’m not quite sure whether Fritz can go all the way to 6.7% as I recommended for Becky and Stephen but I would estimate that Fritz might land somewhere in the middle, between the 4%+ for the 10 case study volunteers and the 6.7% for Becky. So it’s guaranteed above 5%, thus, far higher than the 3.75% recommended by the RISA program.

                  But you should start your own blog and do case studies for people like Fritz if you think that my estimate is wrong. 😉

                3. Thanks for the pointer to your Becky and Stephen case study. That was very helpful. After carefully reviewing it, I was able to fix my filled in SWR toolbox spreadsheet, and it is now much more in line with expectations and consistent with other modeling I’ve done using entirely different methods (90% success Monte Carlo simulation, and present value Funded Ratio).

                  P.S. I responded to my higher level comment rather than your response, as there was no reply button on your response with the case study pointer.

                  Thanks again. The SWR toolbox spreadsheet is a great tool! The combination of both forward and backward looking SWR analysis, offering 0% failure values (I know that doesn’t mean it can’t fail, just that failure probability with both forward and backwards analysis rounds to 0%), and custom cash flows make this tool very robust and useful.

                4. OK, great to know that the SWR is more in line with what I expected.
                  I still use Monte Carlo sometimes for comparison. It’s best to use a 90% success rate as you state (even though in actual life I would find that too low). The lower success rate compensates for some of the quirks of MC that potentially overestimate the failures. Especially the absence of valuation/mean-reversion.

  27. Hi Karsten and Fritz: very interesting discussion. I thought I am pretty good with excel, however I can’t recreate Karsten’s spreadsheet comparing the SAA method to the bucket strategy right after the paragraph below. Maybe that is why Fritz’s approach is just more intuitive. Would it be possible to post the actual spreadsheet so that others (or at least I) could follow the math?

    For the remainder of the simulation, even after withdrawing the entire $40k from the risky asset, the risky weight stays above 70%. Now we can milk this positive momentum in the risky investment and gradually recover some of the prior losses. But even at the end of the simulation, the Bucket Strategy is still more than $3,000 behind the simple SAA.

    Thanks, Hannelore

    Reply

    1. This is exactly the fallacy I described in my post here: the Vanguard vs. Active managers analogy. People favor the convoluted bucket strategy over the simple SAA approach. Because people get hung up on my attempt to show a simulation where the SAA beats the TAA. Let me ask you back: were you able to recreate Fritz’s spreadsheet? Can you explain why Fritz’s bucket approach worked better than the SAA in that specific example? In any case, for complete transparency, here’s a link to the Excel sheet I created:
      https://docs.google.com/spreadsheets/d/1lfrdwgbMYO1r0T-YYtqcrHpV7CiurTUi/edit?usp=sharing&ouid=110384362952922098781&rtpof=true&sd=true
      And please appreciate that the SAA approach takes just one single column to compute/simulate (Column F), while the bucket strategy needs all of the additional messy computations along the way. Don’t confuse my attempt to compare SAA with TAA and the complications involved with the amazing simplicity of the SAA approach!

      Reply
  28. Pingback: T.G.I.F. Friday: Volume 89 - Accidental Fire

  29. Maybe I didn’t understand this right because I think this SAA method is really really simple? One example I might implement:
    1. Start retirement with a 70% stocks – 30% bonds portfolio plus one year of living expenses in cash.
    2. At the end of the year, sell enough stocks and bonds to fund next year expenses such that portfolio is balanced again towards 70% stocks and 30% bonds.
    3. Repeat Step 2 until the end.

    I could vary the numbers, frequency, etc. but the steps are pretty much the same? And add on a glide path to adjust stock/bond ratio starting 5 years before retirement and ending 5 years after retirement to protect a bit again SORR?

    Reply

    1. Thank you, I appreciate seeing the numbers in this and other posts! I’m looking forward to your future post about, if I heard it right, the sensitive parameters of a cash bucket glidepath/BS. I think it will help me synthesize a number of your posts. Specifically, I gathered from your post 48 that given a 30-year retirement horizon with 75% asset depletion, a 52.5/17.5/30 stock/bond/cash allocation gliding to 75/25/0 over the course of 10 years was about as good as a 60/40 stock/bond allocation gliding to 100/0 that was detailed in your post 19. However, the reason I hesitate is that post 19 also showed that a 60/40 gliding to 80/20 underperformed a 60/40 gliding to 100/0 in the context of a 60-year retirement horizon with 100% asset depletion. But perhaps that underperformance was rescued by the change in weighting parameters of the 52.5/17.5/30 and its inclusion of cash?

      Or perhaps 52.5/17.5/30 being on par with 60/40 is conditional on being at market peak? Or an environment with inflation and rising interest rates? I think you gave a nod to both of these in post 48.

      The reason I ask is, in line with the advice in post 43 on pre-retirement glidepaths and in post 42 of working ~1 more year, I “retired” my portfolio to 75/0/25 around market peak late 2021 but kept the job, with future salary stored in cash. In combination with a recent shift into bonds, this has effectively left me with a cash cushion allocation of ~52.5/17.5/30. I now need to decide whether to glide to 75/25/0 or shift this to 60/40 and glide to 100/0 – in the context of a long retirement horizon with 100% asset depletion. Any pointers are greatly appreciated!

      Reply

      1. Note that the GP comparisons in Part 48 are all case studies. The more cash-heavy allocation did well in the 60s-80s.More bonds did the best in 1929.
        Given your current allocation I wouldn’t recommend making a drastic change like moving everything to 60/40/0. Plan a GP to 75/25/0. And maybe 100/0/0 later in retirement if that feels right. But there’s no need to make that decision now.

        Reply

        1. ERN (for some reason I still prefer typing that than Karsten, which is more personal),

          You taught me to love the glide path, and I’m in the process of implementing it, but I always cringe when you say to *actually* glide all the way to 100/0. As you’ve pointed out, there is always SoRR, and so the only way that 100/0 is “safe” is when the retiree never increases their withdrawals even after their portfolio has come up *so* much that their “current” WR is at or below 2.75%, and probably 2.5% (and that’s for a 30 year retirement, for a 60 year it’s probably a bit lower still). How many people are likely to actually do that?

          Since in practice virtually no one is going to maintain the original spend rate when their portfolio has gone up *so* dramatically in value (and of course in the downside cases, virtually no one is likely to be permanently willing to keep their allocation at 100/0, even for the ones who are wise and maintain it for several years after a bad bear market), it seems to me to be much more responsible to advocating “re-retiring” and restarting the GP if in fact you get past 80/20 or certainly 90/10 and your portfolio has done well. Or if you want an even simpler message/method, just tell ’em to stop at 80/20 except in the case where they’d lived through a bad early bear market. I know this lowers the effective SWR by a few hundredths of a percentage point, but it’s way more realistic – and in all probability *safer* when implemented.

          Reply

          1. Good point!
            The full transition to 100/0 is only necessary if the SoRR event is long enough, so the shift to equities catapults you out of the portfolio drawdown. In all other cases, you’ll likely be OK with 80/20 or 90/10. So, consider that 100/0 scenario only a worst-case one.

            Reply
  30. Pingback: February 2023 Article Roundup - Engineering Fire
  31. Pingback: Discussing Retirement Bucket Strategies with Fritz Gilbert – SWR Series Part 55 - Investment Club Live

  32. Loved the exchange and the only winner I can award is for blog/webpage layout. I find Fritz’s setup much easier to read.
    ERN is a bit heavy on the intrusive ads and definitely makes the reading process more difficult. I can’t blame him though b/c I’d want to monetize the amount of work I’d have put into this website too.
    Having said that, I did sign up for ERN newsletter and am looking forward to reading the SWR series!

    My initial approach could be called a bucket strategy. I plan on having xx years of $$ easily available (cash equivalents) that could be bucket one and bucket two are equites that throw off dividends that continually fill bucket one.

    Reply

    1. I decluttered the content a little bit and took out some of the annoying ads. I think there are about 10 ad blocks in a 3,000 word post at this point. Should be worth it. If you still don’t want to scroll through ads, get the “Brave” browser.

      Reply
  33. Pingback: Accounting for Homeownership in (Early) Retirement– SWR Series Part 57 – Early Retirement Now

  34. I’m late to the party on this, but I feel the need to weigh in and echo almost exactly what Bob J wrote on Fritz’s blog in Part 1:

    Bob J: “Karsten’s approach is actually simpler to execute – just maintain the desired asset allocation – but it is harder to understand thoroughly why it works and have full confidence in it.”

    Admittedly I come at this as an ERN reader and not so much a TRM reader, but it is patently obvious that the ERN SAA approach is *much* easier to implement and execute than the TRM Bucket Strategy. It’s actually kind of embarrassing that there’s *any* discussion at all on that point.

    But the easier to understand/intuit point *is* more valid, and that’s the place where “different strokes for different folks” can reasonably apply. The ERN approach *is* harder to understand and have confidence in, until/unless you take the time to go through it *and* have at least *some* mathematical inclination. I share ERN’s concerns that people who feel confident in the TRM Bucket Strategy probably shouldn’t feel quite as confident as they do, but if it’s more intuitive to them, ok.

    So IMO what *both* of you *should* be saying (Fritz might spin it a bit differently) is:

    1) The ERN SAA approach is without question simpler to execute, and will do at least as well on average, and *if you stick to it* results in fewer chances for error.

    2) But if the TRM Bucket Strategy makes you more comfortable – you think you’re more likely to stick with it during bad markets – *and* you’re willing to deal with somewhat more complexity in executing it, then so long as you use a reasonable WR and reasonable bucket sizes (initial asset allocation) *and stick to it,* you will do about as well on average.

    Reply

  35. Hi Karsten, What do you think of bridge strategies to get to social security? Do you view that as similar or kind of another form of bucket strategy? For example say you retired at age 52 with a 3m portfolio of investment assets in total and you were expecting 40k annually(in age 52 dollars) in social security beginning at age 67. So you would divide your portfolio with 600k(I’m simplifying by ignoring interest/inflation) in ‘safe’ assets like Tbills, bond ladders, CD’s, money market etc to fund as a bridge” to social security. And then the withdrawal beginning at age 52 would consist of the 2.4m remaining at say 3% withdrawal rate so 72k plus the 40k annually from the bridge so 112k combined in first year. That will continue plus inflation adjs until age 67 when actual social security kicks in. I don’t recall you writing about this before but what do you think of this type of strategy? I’m guessing you would view similar to bucket strategy as just another form of mental accounting that although maybe not causing real harm doesn’t provide any real benefit either but would that be your view? Thank you.

    Reply

    1. That’s exactly a strategy I’d propose. It’s a form of a glidepath and it will underperform in the case of good sequence risk (i.e., if the stock market performs well early in retirement) but also hedge against a bad market early in retirement.

      Reply

  36. I am new to the game, but applying Karsten’s “forest for the trees” algorithm to my anticipated retirement situation worked well for even the worst historical periods. I then looked at the effect of adding a rule that “in a year when the withdrawal comes from the safe pot, also transfer X% of the safe pot to the equity pot”. I thought this would be a good way to bump up equities when they were down without trying to time the market, but the examples I examined were not consistently positive. Is this rule fundamentally flawed? Are there any situations in which it is useful?

    Reply

    1. It’s a conjecture that needs some additional evidence. Do you have simulations to show that this worked better than a simple fixed asset weight? Would this have worked better in 1929-1932 and 1968-1982?

      Reply
  37. Pingback: Trading Options: A Primer (Options Series Part 11) - Early Retirement Now

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.