Bucket Approach

[liftlock]

Apr 14, 2024 0:37:39 GMT mba123 said:

For anyone who wants to share, I'm curious: Do you use a bucket approach or an income withdrawal approach (which, to me frankly, seems easier)? What are most people doing or talking about doing?

I have one portfolio consisting of 4 accounts (Treasury Direct (100% iBonds), Taxable account (96% equity, 4% MM), Roth IRA (100% equity), T-IRA 96% equity, 4% MM).  I don't think in terms of buckets, except in my T-IRA where I harvest dividends and capital gains to maintain a cash position equal to 1 years RMD.  RMDs in cash are taken for QCDs, Tax Withholding, and spending needs.  The remaining balance of RMDs are taken with in-kind shares. Since Pension and Social Security generate 60% of my income, I have minimal amounts invested in bonds.   

[Mustang]

The purpose of the bucket approach is to prevent panic selling. It is more of a structural approach than an income approach. Structural as how to set up allocations, not how to withdraw from them. Harold Evensky, who created the bucket approach in 1985, simply bolted on a cash account to his total return portfolios. He had only two buckets. He said more buckets just added complexity. So basically any portfolio that has two to three years of cash or cash equivalents to prevent selling in a down market is technically using the bucket approach. Modern advocates have complicated it.

I'm not voluntarily withdrawing anything yet because my pension and social security cover all our spending needs. I'm taking RMDs from my traditional IRA but reinvesting the after tax amount.

RMDs are a variable withdrawal method. A government mandated table is used to determine the percent withdrawn. RMDs don't care about portfolio structure. Any asset allocation can be used in an IRA. But, there are big penalties if less than the proper percent is taken.

Variable withdrawals are specifically designed to protect the portfolio's value. The year after the market is up, the dollar withdrawal is higher. The year after the market is down, the dollar withdrawal is lower. The same percent is used but the previous year's ending balance is either higher or lower. The market could completely crash and a withdrawal will still take place. It would just be significantly lower. It isn't possible for the portfolio to go to zero using a variable withdrawal method. Withdrawals may get too small to live on but they will keep coming.

I read a Vanguard study that said variable withdrawal methods provide less than desired income 48% of the time. Vanguard advocated floors and ceilings to limit swings in the dollar withdrawal but even with floors and ceilings less than desired income occurred 45% of the time. Guyton/Klinger Guardrail method is similar to floors and ceilings. I read a report by a FIRE (Financial Independence Retire Early) blogger where if the G-K approach was used during the Stagflation years the retiree's income would be lower than planned for an entire decade.

Variable withdrawal approaches protect against sequence of return risk which is low returns during the first 10 years of retirement. This in turn protects against longevity risk, the risk of running out of money during the retiree's lifetime. The retiree might not have enough to live on but he will never run out of money.

The nice thing about traditional IRAs is that you can use a different withdrawal method, like the 4% Rule. You just can't take less than the minimum and you cannot apply more taken one year to the minimum the next year. Each year stands on its own merit.

The 4% Rule is a fixed dollar withdrawal approach with inflation adjustments. It protects the retiree's income not the portfolio's value. Because of this it can be a draw down method during a prolonged bad economy like the Stagflation years Draw down methods might spend principal as well as returns and care must be taken to not run out of money. The sequence of returns is important. But, studies show that taking a 4% initial withdrawal has a 99-100% success rate and will most likely leave a large ending balance. Average return for the market is something like 6.5%. During the last ten years or so it has been near 10%.

But if a worst case scenario occurs you cannot count on a 10% or even a 6.5% initial withdrawal. The worst time to retire in history was 1966. If that retiree used a rate higher than 4% their retirement portfolio would run out of money before 30 years.

I believe that a portfolio should be divided according to purpose: Retirement and all other. All other includes money for emergencies, health care and heirs. That way you can plan and budget for your expected income. I believe that a steady, reliable income source is needed for necessities and that a variable income source can be used for wants. There are overlaps between the two. That is when the retiree must reduce spending during market downturns.

I also believe in simplicity. The simpler the retirement portfolio, the easier it is to manage and transfer to your successor. A written succession plan is also a good idea.

I will be using multiple methods. Since most of my retirement income is from an inflation adjusted pension and inflation adjusted social security I can tolerate a little volatility. I have simplified my traditional IRA to one fund, American Funds American Balanced Fund (65% stock). Monthly RMDs are calculated by American Funds. They take out federal and state income taxes and send me the difference. To have multiple funds would require me to pick which fund the RMD is taken. I don't want to do that.  It goes against my goal of simplicity. The variable withdrawal and the balanced allocation of the fund is enough protection. After 4 years of RMDs during a highly volatile time 99.5% of the RMDs have been taken from income (dividend and capital gains are reinvested), The number of shares owned has decreased only 0.5%. That is not much of a draw down and it could easily be recovered.

My pension ends when I die. I have purchased an annuity that will replace half of my income for my wife. The other half of will be covered using the 4% Rule. I could use one fund like I did in my IRA but I thought it was too much risk. I tested Vanguard Wellington using a 1968 retirement and it ran out of money after 25 years, not 30. It had a big sequence of return failure losing money in 1969, 1973 and 1974. Because of this I paired it with a more conservative fund, Vanguard Wellesley.

Research using historical returns has shown that the asset allocation should be between 50% and 75% stock. Since my IRA is 65% stock I chose equal portions of Wellington and Wellesley which should result in an asset allocation of 50% stock. The dollar withdrawal will be from the fund with the highest previous end of year balance.

Only one time in history has both Wellington and Wellesley lost money at the same time two years in a row (1973 and 1974). Because of this I put two years of withdrawals in a money market fund so I wouldn't have to sell in a down market.  I guess technically that is the bucket approach.  But, I'm not sure you could say Wellesley is bucket 2 and Wellington is bucket 3.  It is more like Evensky attaching a cash bucket to a total return portfolio.  The two funds will never be re-balanced. I tested this during the Stagflation years and the the two funds automatically re-balance after to or three years.

I'm not a financial planner and I'm still learning every day. These approaches fit my needs. I have no idea if they would fit anyone else's.

[liftlock]

Mustang, Excellent post!

Bud Hebeler used to write extensively about retirement topics after he retired from Boeing.  He advocated using the IRS like RMD method of for determining safe withdrawal rates that would reduce the risk of running out of money.. His formula was to subtract your age from 100 to determine how long your portfolio might need to last.  That would be 25 years for someone age 75.  Divide the number of years by 1 to arrive at a safe withdrawal rate.  1/25 = 4% at age 75.  One can play around with different time frames.

www.analyzenow.com

I also found another one of Bud's safe withdrawal rate methods here:

howmuchcaniaffordtospendinretirement.blogspot.com/2015/01/bud-hebelers-autopilot-withdrawal-rule.html

[saratoga]

Apr 14, 2024 1:32:19 GMT saratoga said:

mba123: For anyone who wants to share, I'm curious: Do you use a bucket approach or an income withdrawal approach (which, to me frankly, seems easier)? What are most people doing or talking about doing?

My soft budget is pension + SS + RMD that should meet most of my normal expenses.
(RMDs are taken mainly from PRWCX and TIAA Traditional)
So, I do not need a complicated income withdrawal plan.

I may construct a trust for my beneficiaries with two bucket distribution plan, however.
It could be an x% withdrawal of the 3 year average of investment asset values, like what universities often do with donated funds.
Withdrawals could go to the second bucket invested in safe assets.

Distributions can be made from this safe asset bucket, but I have not figured out the details.

One possibility is to use a payout fund such as T Rowe Price TRLAX for tax deferred or Roth money.  It provides equal monthly payouts that amount to 5% annually of the trailing 5 year asset average value.   Its portfolio has 51% equities.   Another possibility is to create two income streams, one based on TRAIX, another on VWENX or global Wellington, each paying out 5% of 5 year asset value average annually.   In both cases, there is no second bucket.   However, there are complications with RMD and Roth rules.

[mba123]

Mustang,
That makes much more sense - thank you!! Structural.

We have no pensions—even though my husband worked for a large corporation, the pensions were stopped about 15 years ago and are only honored for folks who started work x number of years ago - ahhhhh that would make things sooooo much easier. I often wonder now with the rareity of pensions, what retirement will look like for the average American. I can’t even imagine to be honest.

I’m not quite old enough to start thinking about RMD or SS (57 yo), but it’s very interesting to hear that the RMD is variable. I’m sure I’ll be more interested in that as time goes on. Using one fund makes a ton of sense to me.

We estimated using a 4% withdrawal rate from our entire portfolio and even discussed it with a Fidelity advisor, and I think we are fine, but I have to be honest, it’s still so scary. Especially because we have kids who are adults (21, 23)—undergrad is paid for, but we really wanted to help them with graduate school, weddings, and their first home—now it feels awful that we might not be able to do those things.

Your thinking about fixed expenses vs. variable (wants/needs) expenses is interesting. Now, I might have to revisit the idea of not withdrawing 4% of the total since you’re right—we will need an emergency fund, health care money, etc. But I think we did budget for healthcare premiums.

It was interesting (and scary) to read about the Wellington Fund. I probably would not have thought to test it that way, but it’s smart.

Your thinking is very strategic, MUCH more so than any FA has ever done for us. So, I am very eager to learn all I can. And I believe no one cares more about our money than we do.

Since we (potentially) may be early retirees, I may also need to join a FIRE discussion board (if such a thing exists?) to ensure that I am considering issues unique to a long and early retirement. I think a part-time job with healthcare benefits may be in order!

[Mustang]

If you are thinking about retiring early you may want to read some of these articles. At the bottom he provides links to specific topics. Part 48 discusses the bucket strategies. earlyretirementnow.com/safe-withdrawal-rate-series/

P.S. A 4% initial withdrawal rate was found to be 99-100% successful for a 30 year payout. That takes someone 65 to 95. I still use it because my wife is younger and women live longer. Early retirement will most likely have a payout period of 40 years or more. For a longer payout period a 3% initial withdrawal rate was shown to be successful.

Note: The stress test I mentioned concerning Wellington happened with a 1968 retirement start date. It was a sequence of return failure. The 1969 loss set the stage and the losses in 73 and 74 slowly finished it off. But returns were only part of the problem. The 70s were the worst economic period in history. A little understanding of Stagflation is needed. There was very high, often double digit inflation. The initial withdrawal on a $500,000 portfolio was $20,000. Withdrawals are adjusted for inflation. The 1992 annual withdrawal was over $81,000. Four times higher just to keep the same purchasing power.  I hope we never see an economic period like that again.

Looking back a lot of people really do not understand how bad the 1970s were. But there are ways to make 4% work. Take 1 point less than inflation increases. With that simple modification Wellington lasted 30 years with a small ending balance.

Since I paired Wellington to Wellesley I did a side by comparison but used a 1971 start date (the year Wellesley was created). Using the original 4% Rule Wellington lasted the entire 30 years with a $1.4M ending balance. This is how important sequence of returns are. Without the 69 loss the fund was able to recover from the 73 and 74 losses. Wellesley was still better than Wellington though. Wellesley's heavier bond allocation allows it to prevail during bad times but Wellington has more growth during good ones.

[mnfish]

mba123,
"Anyhow, we are mostly in 401Ks, little cash (maybe 100000)"    What are you going to live on until you can draw from the 401ks? I had (have) a lot less than $2.8m but if yours is all in 401ks, how do you withdraw without a penalty?

I quit work at 55 but had a largish taxable account to draw from, usually just dividends, to make up for the shortfall from my beneficial IRAs. My kids were grown, married and out of school. I have gifted to them annually for about 10 years.

I used a High Deductible ($7,500) MnSure plan, basically ACA, for health insurance and saved thousands in premiums compared to when I was self-employed and contributed to an HSA which helped keep my income below the threshold for the ACA tax credit. MY current SS plan payments exceed what I spent on MnSure premiums, but my deductible is far less.

[yogibearbull]

There is a special rule for 401k/403b - one can withdraw without 10% penalty between ages 55-59.5 from the current 401k/403b on job separation. It doesn't apply to old 401k/403b, or T-IRA.

Any retirement withdrawal plans should account for required RMDs that start at 3.65% at age 72 (using prior yearend balance for T-IRA) and then gradually go up annually. Be sure to use new RMD tables.

[marg]

yogibearbull, One can withdraw more than what RMD table requires?

[yogibearbull]

marg, sure, RMD is the minimum amount required.

In fact, many retirees draw RMDs first, and then more for (charitable) QCDs, or more for expenses if needed.

[archer]

One qualm I have about the bucket approach is the opportunity loss of holding safer assets. I can see how it makes sense for preventing sequence of return loss starting retirement in 2000, or 2008, but that is only 2 episodes in 23 years. The other downturns have been minor or compensated by substantial rebounds. I might err more on the cautious side if my expenses were more. Its just that when I run asset class combinations through PV, most start years come fare better with a more aggressive PF.

That said, I do keep about 50% of my PF about 50/50 allocation and the other 50% in another account (roth) for trading stocks.

[liftlock]

Apr 15, 2024 13:08:40 GMT yogibearbull said:

There is a special rule for 401k/403b - one can withdraw without 10% penalty between ages 55-59.5 from the current 401k/403b on job separation. It doesn't apply to old 401k/403b, or T-IRA.

Any retirement withdrawal plans should account for required RMDs that start at 3.65% at age 72 (using prior yearend balance for T-IRA) and then gradually go up annually. Be sure to use new RMD tables.

I had a 401K plan that did not permit partial elective distributions after job separation.  The plan would only allow partial distributions for RMDs. This forced former employees to rollover their 401K to T-IRAs to gain access to funds for partial distributions.  Before I retired I did a partial rollover of my 401K to an T-IRA to give me more flexibility in accessing funds upon retirement. 

[yogibearbull]

liftlock, that is strange, but plans may have stricter rules than allowed by the IRS. Many delay 401k/403b rollover to T-IRA to 59.5+ (if they rollover at all) just to be able to tap into this feature.

But if one rolls over 401k/403b to T-IRA between 55-59.5, one can use the tricky 72T to tap the T-IRA early.

[liftlock]

Apr 15, 2024 15:27:04 GMT yogibearbull said:

liftlock , that is strange, but plans may have stricter rules than allowed by the IRS. Many delay 401k/403b rollover to T-IRA to 59.5+ (if they rollover at all) just to be able to tap into this feature.

But if one rolls over 401k/403b to T-IRA between 55-59.5, one can use the tricky 72T to tap the T-IRA early.

yogibearbull,  Part of the company I originally worked for was sold to / acquired by the company I subsequently retired from.  The terms of the old companies 401k plan were more favorable than the terms of the new companies plan.   As I recall the old company plan allowed access to funds upon reaching age 55 whereas the new companies plan delayed access until age 59.5.   The new company grandfathered old company plan rules for 401K funds acquired while employed by the old company.

Regardless, it always seemed a bit short sighted for an employer to not allow partial distributions from a plan once a employee terminates employment. Doing so would seem to encourage rollovers to T-IRA which reduces 401K plan assets under management which reduces leverage to obtain less expensive asset management fees. However, someone may have thought that it was best to avoid the extra cost of administrating partial distributions from the plan.   

[win1177]

Excellent post Mustang! I am planning on using a “modified bucket approach”, where we have a cash/ short term bucket, intermediate bucket (mix of fixed income, safer equities), and long term equity bucket. Retired one and a half years ago, have a pension from S.C. that has only a very small inflation adjustment. Also will have S.S. when I start it, but am waiting for now since it goes up ~8% per year. Probably will start w/in next 1-2 years.

Currently living on income (dividends, interest) investments, plus my pension which provides about 1/3 of needs. Our portfolio is pretty large, probably WAY more than we will need, so we look to be “safe” no matter what. Looked at LTC coverage, very expensive, so decided to “self insure”. I account for that in our longer term equity bucket. Right now, we are living on 2-3% per year, so we’re fine. Much of what is withdrawn goes to Govt. For taxes.

Wife still working as realtor, she will be turning 65 later this summer. I’m already 65, have some health issues so I anticipate I’ll be first to go. Slowly building up bucket 2- intermediate which is a mix of fixed income and core safer equity. Fixed income now ~6% of total portfolio, so have a nice chunk of income there, especially when combined with dividends. It’s enough to allow us to live on w/o tapping much principal at all. Anticipate leaving much of portfolio to heirs/ charity. Good place to be! May do a conversion of IRA, (rollover from my work) between now and age 70. Already have a good sized ROTH IRA for myself/ wife, which we will touch “last”.

Win

[anitya]

Apr 15, 2024 13:08:40 GMT yogibearbull said:

There is a special rule for 401k/403b - one can withdraw without 10% penalty between ages 55-59.5 from the current 401k/403b on job separation. It doesn't apply to old 401k/403b, or T-IRA.

Any retirement withdrawal plans should account for required RMDs that start at 3.65% at age 72 (using prior yearend balance for T-IRA) and then gradually go up annually. Be sure to use new RMD tables.

I was not following this thread but try to read as many of your posts as possible so this caught my eye.

Thought the info at this link might be useful to others -

www.irs.gov/retirement-plans/plan-participant-employee/retirement-topics-exceptions-to-tax-on-early-distributions

Yes, plans can have a higher early retirement age.  Plans also amend and restate to change this while staying within IRS / DOL requirements.

[mba123]

Apr 15, 2024 13:08:40 GMT yogibearbull said:

There is a special rule for 401k/403b - one can withdraw without 10% penalty between ages 55-59.5 from the current 401k/403b on job separation. It doesn't apply to old 401k/403b, or T-IRA.

Any retirement withdrawal plans should account for required RMDs that start at 3.65% at age 72 (using prior yearend balance for T-IRA) and then gradually go up annually. Be sure to use new RMD tables.

Yes, we would plan on using his current 401K and probably also should stash more in cash. I knew nothing about this rule until last month and just converted an old solo 401K to an IRA but it sounds like I wouldn't have been able to use it anyway since I'm no longer involved in that business. I have a new one that provides just a small annual gross income. 

[Mustang]

Apr 15, 2024 18:14:25 GMT win1177 said:

Excellent post Mustang! I am planning on using a “modified bucket approach”, where we have a cash/ short term bucket, intermediate bucket (mix of fixed income, safer equities), and long term equity bucket. Retired one and a half years ago, have a pension from S.C. that has only a very small inflation adjustment. Also will have S.S. when I start it, but am waiting for now since it goes up ~8% per year. Probably will start w/in next 1-2 years.

Darrow Kirkpatrick wrote, "Many retirees will use systematic withdrawals from an investment portfolio for retirement income. I’ve done new research into the best retirement withdrawal strategies. History shows that your success can vary widely using the same portfolio and the same overall withdrawal rate, without changing your investments. It all depends on how you withdraw from different asset classes like stocks and bonds. The secret is keeping it simple and using a consistent, value-driven approach. The payoff could mean extracting millions more income from your nest egg over the course of a long retirement." www.caniretireyet.com/new-research-the-best-retirement-withdrawal-strategies/

He tested various withdrawal methods using the 4% Rule and a 50/50 SP500 and 10 year treasury allocation. In his update he expanded it to include other asset allocations. He received a lot of comments from his first publication. “Given their lagging performance I dropped the three momentum-based strategies (Last Year, 3-Year, 7-Year) from consideration. I also renamed the “Equal Withdrawals” strategy to “EqualTarget,” reflecting that this strategy is really about withdrawing in proportion to the target asset allocation, which was no longer constrained to 50/50.”

“Next, by reader request, I added a new withdrawal strategy for consideration: “Proportional.” This is likely the withdrawal strategy you’d be using if you weren’t thinking about this issue at all. It simply withdraws in the same proportion as your portfolio’s current asset allocation, however it has grown. So, if your portfolio is at 65/35 stocks/bonds, so is your withdrawal.”

And he further clarified re-balancing, “A withdrawal strategy executes during the year, and specifies the logic for how much or in what proportions to withdraw money from your portfolio for living expenses. A rebalancing strategy executes at the end of the year (or years), after your portfolio has grown, and specifies the logic for transferring money between holdings solely to adjust their proportions.”

Note: I was never a fan of the momentum strategies. They had significantly lower probabilities of success and ending balances. Even though the CAPE Mean strategy had better results it was more complicated than I wanted. Besides he said it didn’t work well with balanced funds. www.caniretireyet.com/the-best-retirement-withdrawal-strategies-digging-deeper/

The 4% Rule research doesn’t have a cash bucket, just stocks and bonds so how does this research relate to the Bucket Strategy? It could be used to refill bucket 1.

I believe in buy low, sell high. That would be Withdrawal Rebalancing not Annual Rebalancing. And withdrawing from a balanced fund is automatically a proportional withdrawal. You have no other choice. Here are the success rates and ending balances:

     Asset             Withdrawal         Proportional              Proportional
Allocation         Rebalancing                                    W/Annual Rebalancing
80/20                 93.1% $7.4M     91.4% $7.8M             93.1% $6.7M
60/40                 93.1% $4.5M     93.1% $5.2M             93.1% $3.6M
50/50                 87.9% $2.7M     87.9% $4.2M              91.4$ $2.5M
40/60                86.2% $1.5M      87.9% $3.2M             87.9% $1.5M
20/80                55.2% $0.1M      72.4% $0.9M             55.2% $0.1M

The lower success rates of bond heavy portfolios are completely in line with other studies. Annual Rebalancing is a double edged sword. The portfolio has to be very stock heavy to gain any benefit and it significantly lowers success rates and ending balances in bond heavy portfolios.

While the cash bucket can prevent panic selling the question is how to refill it. Annual Rebalancing the Bucket Approach can still result in forced selling during a down market. If bucket 2 does not provided enough income some of the refill is from bucket 3 regardless of market performance.

Survey the data for yourself. The most likely asset allocation between buckets 2 and 3 would be around 60-65% stock. For that asset allocation Withdrawal Rebalancing, Proportional Withdrawals, and Proportional withdrawal with annual rebalancing all have a 93.1% success rate. But adding annual rebalancing to the mix, which all writers do, significantly lowers the ending balance.

P.S. How have I used this information?

Half of my retirement portfolio is two balanced funds (overall 50/50 allocation) with a cash bucket that is refilled annually in January. Taking from the fund with the highest end of year balance is Withdrawal Re-balancing. My tests have shown the funds automatically rebalance every two to three years. The withdrawal from the specific fund is by its very nature proportional. Half the time the withdrawal will be from Wellington (65% stock) and half the time it would be from Wellesley (65% bonds). For a 50/50 portfolio both withdrawal rebalancing and proportional withdrawals have an 87.9% success rate.  This is different from the 4% Rule studies (Bengen, Trinity and Wade Pfau's update of the Trinity Study) but Kirkpatrick is using his own mathematical model and the others studies are just based on historical data with a pass/fail criteria. How the cash is specifically withdrawn from the portfolio is unknown.

[mba123]

This was very interesting! Honestly also sounds a lot more straightforward that the bucket approach, which (as someone else also said) involves keeping a lot more than I think I'd ever imagined in fixed rate stuff). And I never realized that until recently. I suppose it depends on how much you have in your PF and how much you plan to spend.

I feel like inflation is always eating away at those fixed rate returns so I'm not convinced that keeping a lot in bonds etc is the way to go for us anyway but I'm sure I'd regret that in bear markets.

I'll have to read your post (and the info in the links) Mustang a few times because I'm very unfamiliar with the CAPE ratio approach and that seemed to have the safest outcome at ~96% success rate. Interesting. 

I feel like we will probably go with a fixed withdrawal rate when and if the time comes using the lowest possible percentage. And hopefully agree on an asset allocation. 

[mnfish]

"I suppose it depends on how much you have in your PF and how much you plan to spend."

Bingo!