Backtesting or Studies on Withdrawal Strategies versus DCA-out

[HI Bill]

I read somewhere that taking quarterly (or monthly) distributions is better than once a year (more money is invested longer in the markets). If once a year, I think they recommend end of December.

For my peace of mind, I usually take at least 50% of my annual spend out as early in the calendar year as possible, when I perceive that markets are up (my only market timing). Yes, I may be losing out on that year's returns, but I have more comfort that I won't need to reduce my spending that year.

 

[rmcelwee]

Re: the classic withdrawal study, the creator of the 4% study, Bill Bengen, originally said a 50/50 domestic portfolio has always sustained 4% withdrawals for 30 years, with an inflation adjustment.

He’s now advising 65% stocks and a 4.7% safe withdrawal rate, with an inflation adjustment.

Here’s a recent interview with him. He Invented the 4% Rule | Bill Bengen on Why He Now Thinks 5% Works

I find it very interesting that he changes the rule and gets lots of press right when his new book is coming out. If he writes another one you can expect him to change to 70% stocks and 5.2% SWR.

As far as "Are there any studies which compare DCA-out to other withdrawal strategies like buckets, bond/CD ladders, etc?". I tend to discount any studies by people who haven't travelled here from the future. Who could possibly know which is the best way to do your withdrawal (meaning what the market is going to do in the future)? Throw all those studies, interviews, and books in the trash. Just live your life. You are doing fine...

Our withdrawal this year was over $1M. We had budgeted for $60K. No one really knows what is going to happen.

 

[TheWizard]

I find it very interesting that he changes the rule and gets lots of press right when his new book is coming out. If he writes another one you can expect him to change to 70% stocks and 5.2% SWR.

As far as "Are there any studies which compare DCA-out to other withdrawal strategies like buckets, bond/CD ladders, etc?". I tend to discount any studies by people who haven't travelled here from the future. Who could possibly know which is the best way to do your withdrawal (meaning what the market is going to do in the future)? Throw all those studies, interviews, and books in the trash. Just live your life. You are doing fine...

Our withdrawal this year was over $1M. We had budgeted for $60K. No one really knows what is going to happen.

Can you explain that please?
60k vs 1000k

 

[Markola]

I read somewhere that taking quarterly (or monthly) distributions is better than once a year (more money is invested longer in the markets). If once a year, I think they recommend end of December.

For my peace of mind, I usually take at least 50% of my annual spend out as early in the calendar year as possible, when I perceive that markets are up (my only market timing). Yes, I may be losing out on that year's returns, but I have more comfort that I won't need to reduce my spending that year.

One can play with different withdrawal timings on Portfolio Visualizer. I’ve never seen that it makes much difference but YMMV.

 

[rmcelwee]

Can you explain that please?
60k vs 1000k

We decided this was the year we were going to start our new house build and sold a bunch of stock to pay for part of it. The plan was to start this in 2022 but we missed it by 3 years. Once we are in the new house our budget will go back to the normal $60K. I actually have no idea how we will spend that much but we always just round way up to $60K for planning.

FWIW, we've spent around $800K of it so far and our net worth has only dropped by $50K. No amount of planning, spreadsheets, Firecalc, etc would ever predict this. It is a waste of time (but sometimes fun) to do any planning at all.

 

[Theophrastus Bombastus]

If DCA-in is the best way to invest, is DCA-out the best way to withdrawal?

Any other withdrawal strategy seems to either be a form of market timing or risk mitigation at the cost of possible future gains.

Are there any studies which compare DCA-out to other withdrawal strategies like buckets, bond/CD ladders, etc?

My withdrawal rate is low enough that I believe that I can weather downturns of the normally expected types

Additional personal info if it’s relevant:

I’m over 5 years into retirement and have the common story: it’s all great, more money than when I retired, etc.

I’m currently 55YO, married and at a 2% annual withdrawal rate. All our retirement is self funded other than future SS. Healthcare is via ACA

I currently pay our monthly expenses by pulling money from a large non-retirement standard brokerage account. I pull money every month or so. We also have retirement accounts that we’ll use later. All accounts are invested in mostly equities and a three fund (primarily) type of portfolio

I don’t post often here but am reading posts often. The advice from folks on this website helped get me where we are financially

Try tinkering with this:

https://www.portfoliovisualizer.com/backtest-asset-class-allocation

 

[jmb1855]

We decided this was the year we were going to start our new house build and sold a bunch of stock to pay for part of it. The plan was to start this in 2022 but we missed it by 3 years. Once we are in the new house our budget will go back to the normal $60K. I actually have no idea how we will spend that much but we always just round way up to $60K for planning.

FWIW, we've spent around $800K of it so far and our net worth has only dropped by $50K. No amount of planning, spreadsheets, Firecalc, etc would ever predict this. It is a waste of time (but sometimes fun) to do any planning at all.

We’ve been thinking about a similar move. I need to figure out how to best do it. It would be a timing issue with ACA, 401k, and a conventional brokerage account.

I’ve been floating along for 5 years simply managing for our annual budget and ACA MAGI. The next phase of my ER education needs to focus on the more complex longer term tax implications

Those are good “problems” to need to figure out

 

[jmb1855]

I think I’ve finally got my head around a withdrawal optimization plan that works for me. I just recreated the wheel and convinced myself to use a bucket approach.

When I’d been using the DCA-out term, what I was trying to understand was the sequence of return risk of monthly withdrawals from equities and how to manage that for my risk tolerance

I won’t be able to implement this yet because of the tax implications that I want to avoid.

My optimization plan for an aggressive/high equities portfolio:

1) Sell off assets to get to an around 2 year cash reserve. Place into lower risk investment “buckets”

2) Make withdrawal from these buckets monthly

3) Refresh the buckets annually if the market is in an up cycle. If not, don’t refresh the buckets until the market recovers

This reduces risk in a way that I have some control over and still keeps a higher upside

ETA: I’d still like to find a way to back test this to see how much all the extra work matters. I’ll look at PV as recommended

 

[Markola]

I did a bunch of that kind of modeling this spring using Perplexity AI, and the “wait until things recover” route is indeed psychologically appealing. There’s ways a tradeoff, sadly, and in this case, it’s both 2 years of cash drag, and the missed opportunity to rebalance into assets while their prices are depressed. I’m no professional but the AI made a compelling case to me that regular, annual rebalancing has long term benefits. It takes guts to rebalance into the teeth of a bear market, no doubt, though.

 

[jmb1855]

Those types of things are what keeps me with the simple option:

Selling equities every month or two regardless of what happens in the market in order to pay my monthly expenses

 

[jmb1855]

I spent some time with ChatGPT and found a way to back test this over longer time horizons. My take aways for optimizing withdrawal frequency, asset allocation, and withdrawal rates are:

1) Without having cash reserves, monthly withdrawals are better than quarterly withdrawals, which are better than annual withdrawals

2) At a low enough withdrawal rate, high equity asset allocation and no cash reserves is optimal. As withdrawal rates increase, cash reserves and lower equity asset allocation can come into play.

With my low withdrawal rate, I am doing the right thing with monthly or so withdrawals, very low cash reserves, and high equity asset allocation

Here are some of the snippets out of ChatGPT that are interesting:

At a 4% withdrawal rate, monthly withdrawals are decisively superior.

• ✔ Higher median ending wealth
• ✔ Lower failure probability
• ✔ Less damage in bad decades
• ✔ No downside other than trivial complexity

With cash reserves, equities are replenished only when markets are up or flat

For a 4% withdrawal rate:

• Minimum: 12 months cash
• Optimal: 18 months cash
• Upper bound: 24 months (beyond this, diminishing returns dominate)

Once you do this:

• Monthly vs quarterly withdrawals no longer matter
• Simplicity wins

Bottom line

Cash buffer beats frequency.

• Monthly > quarterly without cash
• With 12+ months of cash, frequency becomes irrelevant

One-line practical rules

• ≤ 2.5% withdrawal:
  Ignore cash buffers; maximize growth
• 3–4% withdrawal:
  12–18 months cash is optimal
• 4–5% withdrawal:
  18–24 months cash strongly recommended
• ≥ 5% withdrawal:
  Lower spending first; buffer second

Best practice

• 12–24 months cash
• Keep equity-heavy portfolio intact
• Use buffer as a shock absorber

This combo (high equity + buffer) often beats higher bond allocations historically.

Across retirement-length simulations:

• 80/20 + 18mo cash
  • Often outperforms 60/40 without cash
  • Lower left-tail risk
  • Higher median outcomes
• 
  
• 60/40 + 12mo cash
  • Very robust
  • Near-minimal failure rates at 4%
• 
  
• 40/60 + cash
  • Underperforms due to return drag
  • Inflation becomes the dominant risk

• 80/20 + buffer can behave like 65/35
• But still earn higher long-term returns

Optimal combinations by withdrawal rate

3% withdrawal

• 80/20 + 6–12mo cash
• Or 60/40 with no buffer

4% withdrawal (classic)

• 70–80% equities
• 12–18mo cash
• Bonds mainly for rebalancing, not spending

5% withdrawal

• Bonds + buffer both help
• But spending pressure dominates
• Guardrails become more important than allocation

Bottom line

• Bonds and cash buffers overlap but don’t substitute perfectly
• Best results historically come from:
  • Higher equity
  • Moderate bonds
  • Targeted cash buffer
• 
  
• Cash is most valuable when it replaces forced selling, not when it replaces bonds

 

[Al18]

I have more than 5 years in CD’s and MYGA’s. I withdraw or sell a year’s worth of expenses at the beginning of the year, put it in a high yield saving account and have automatic monthly transfers setup for my checking account. My spending has no correlation on how the market is doing, because I’m retired.

 

[audreyh1]

I read somewhere that taking quarterly (or monthly) distributions is better than once a year (more money is invested longer in the markets). If once a year, I think they recommend end of December.

For my peace of mind, I usually take at least 50% of my annual spend out as early in the calendar year as possible, when I perceive that markets are up (my only market timing). Yes, I may be losing out on that year's returns, but I have more comfort that I won't need to reduce my spending that year.

To me I have so much that remains invested even after the annual withdrawal that I don’t worry about it.

 

[Out-to-Lunch]

With cash reserves, equities are replenished only when markets are up or flat

Is that really what you (or ChatGTP) meant to say?

 

[jmb1855]

Good catch. It should say “cash is replenished when markets are up or flat”

I can’t see any way to interpret it as anything other than an AI mistake?

Below is a screen shot of the text capture that I took from ChatGPT in this section

 

[Out-to-Lunch]

I can understand people making that kind of mistake. It surprises me that an algorithm would - except perhaps that it was trained on people!?!

Not to put too fine a point on it, but does that make you at all skeptical of the conclusions of the bot's disquisition?

I enjoyed your ChatGTP discussion, and may even have learned something or at least it provoked some thought. However, my biggest complaint on bucket strategies, including the bot's, is that it never seems all that different from having an AA with cash, and rebalancing. If you spend from cash and then replenish cash from equities, it is identical to spending from equities (i.e., rebalancing if equities were up). If equities are down, and you spend from cash, but don't replenish, it has about the same effect as rebalancing.

 

[jmb1855]

I was never going to take ChatGPT as a direct robo-investor but the analysis does make sense to me. It also matches well with what I expected the answer to be — so that means it has to be accurate, right?

It also matches what you say in your last paragraph. It’s just different ways of arriving at the same point

It did get me thinking differently about bonds/lower risk investments and how that relates or compares to holding cash for downturns

 

[JohnnyPHX]

I was never going to take ChatGPT as a direct robo-investor but the analysis does make sense to me. It also matches well with what I expected the answer to be — so that means it has to be accurate, right?

It also matches what you say in your last paragraph. It’s just different ways of arriving at the same point

It did get me thinking differently about bonds/lower risk investments and how that relates or compares to holding cash for downturns

If you want to go down the rabbit hole further copy and paste ChatGPT response in another Ai model like per Perplexity. Say something like: ChatGPT suggests this withdrawal optimization strategy, please share your thoughts on it. The feedback from another AI model can be quite interesting.

 

[DEC-1982]

It's worth noting that if one needs to come under the 400% FPL ACA subsidy cliff, one should keep some withdrawals for the end of the year. Just in case ...

 

[COcheesehead]

The guy has a great series of podcasts on retirement. He did one on reverse dollar cost averaging.

#536 - Can I Do a Reverse Dollar-Cost Averaging for Withdrawals in Retirement? — The Retirement Answer Man®

 

[jmb1855]

Did he ever talk about what strategies he’d use if someone is underfunded in retirement?

If they’re underfunded then there isn’t a solution other than get more income from somewhere. I’m assuming that spending less isn’t an option or else the theoretical person would have already pulled that lever

“Whereas if you're constrained and definitely underfunded, it could be a recipe for disaster.”

He’s talking about people retiring under these circumstances along with their 6% withdrawal rate. Scary

 

[Out-to-Lunch]

If they’re underfunded then there isn’t a solution other than get more income from somewhere.

The other classic "cure" to underfunding is an annuity. Yes, that has a host of other negatives, but sometimes it is the only way to get enough to eat from a limited nest egg.

 

[jmb1855]

Thanks - That helps explain what might be meant by underfunded.

An annuity would remove risk but at a cost. It depends on what level of “underfunding” is used in each scenario

 

[mrfeh]

Thanks - That helps explain what might be meant by underfunded.

An annuity would remove risk but at a cost. It depends on what level of “underfunding” is used in each scenario

Jim Otar wrote a book in which he grouped retirees into 3 categories; don't remember what he named them, but let's call them red, yellow and green, where red is someone who is definitely underfunded, yellow may or may not have enough, and green is likely overfunded.

His conclusion was that annuities may make sense for someone in the yellow group, but not for either of the other groups.

He wrote it many years ago; I don't know if his views have changed.

Otar's Retirement Page

Landing page

www.retirementoptimizer.com

 

[Out-to-Lunch]

Thanks - That helps explain what might be meant by underfunded.

An annuity would remove risk but at a cost. It depends on what level of “underfunding” is used in each scenario

Jim Otar wrote a book in which he grouped retirees into 3 groups; don't remember what he named them, but let's call them red, yellow and green, where red is someone who is definitely underfunded, yellow may or may not have enough, and green is likely overfunded.

His conclusion was that annuities may make sense for someone in the yellow group, but not for either of the other groups.

He wrote it many years ago; I don't know if his views have changed.

Otar's Retirement Page

Landing page

www.retirementoptimizer.com

If you (@jmb1855 ) are interested in pursuing this line of thinking, Wade Pfau is a good source. The typical complaint here on this forum is that he is kind of in the annuities camp, but I think he gives good advice on those for whom annuitization makes sense (and those for whom it does not).