SWR failures from history

[FREE866]

Caveat that this ama is a few years old now, but I found it to be a valuable thread with lots of interesting questions and answers. I learned a lot that I did not know.

https://www.reddit.com/r/financiali..._bill_bengen_and_i_first_proposed_the_4_safe/

The first reply is that Bengen states that he later found out that it should actually have been a SWR of 4.5% vs. the 4% he initially reported. For someone with longer retirement plans over the traditional 25-30 years he offers alternative SWR rates.

I was shocked that he felt one could live in perpetuity on a 4% w/d rate.

He does also point out what his research indicated is the sweet spot of an AA to maximize the SWR. This which was 45-55% equities and the remainder in bonds/cash ("According to my research, I would have a well-diversified portfolio with probably 50% equities, 40% fixed-income, and 10% cash")

Thx for this. Read through that thread; interesting he says 55% stock is optimal equity exposure. I have never and could never hold such a low equity holding.

 

[goingtotravel]

Inflation does have a huge impact in my model if I carry a mortgage into retirement. My 2.75% mortgage is basically free money if inflation is 2.75%. That's one of the big reasons I am seriously considering keeping the mortgage for a full 30 years. If inflation goes above 2.75%, I am making money on my mortgage. And if I get 2% real return on my portfolio with 2.75% inflation, the NPV of my mortgage is less than the mortgage balance.

Just refinanced our mortgage to 2.75% with the intent to carry it into retirement. The payments are just over 10% of our budget at the start of retirement and we also have pensions that have a COLA feature. Our WR from our taxable accounts includes an 'extra' 1% for the mortgage. I have treated the mortgage expense in Firecalc as a non-inflation adjusted expense and still get to 100% success.

Back to WR's. Our current FIRE model is 4.5% WR for the first 5 years; 5% for the next 5 years; and then 3.5% after we both claim SS (67). Reading some threads makes me nervous when people are advocating a 3% perpetual WR but our numbers seem to work and we have flexibility in our budget if needed. Biggest worry at this point for us seems to be SORR, right?

 

[flyingaway]

Just refinanced our mortgage to 2.75% with the intent to carry it into retirement. The payments are just over 10% of our budget at the start of retirement and we also have pensions that have a COLA feature. Our WR from our taxable accounts includes an 'extra' 1% for the mortgage. I have treated the mortgage expense in Firecalc as a non-inflation adjusted expense and still get to 100% success.

Back to WR's. Our current FIRE model is 4.5% WR for the first 5 years; 5% for the next 5 years; and then 3.5% after we both claim SS (67). Reading some threads makes me nervous when people are advocating a 3% perpetual WR but our numbers seem to work and we have flexibility in our budget if needed. Biggest worry at this point for us seems to be SORR, right?

I am having trouble to understand this argument. How about if your portfolio return, especially the fixed income part, is below 2.75% annually?

 

[USGrant1962]

...

Back to WR's. Our current FIRE model is 4.5% WR for the first 5 years; 5% for the next 5 years; and then 3.5% after we both claim SS (67). Reading some threads makes me nervous when people are advocating a 3% perpetual WR but our numbers seem to work and we have flexibility in our budget if needed. Biggest worry at this point for us seems to be SORR, right?

You have discovered what I call the "two step withdrawal rate". Trinity, Bengen, SWR, etc. all assume you are spending just your portfolio for the rest of your life. This results in a single, very conservative SWR. Yet most of us at least have SS (or 75% of SS if you prefer).

FIRECalc nicely illustrates this if you input SS or a small pension, often resulting in something like a 4.5-5.0% followed by 3.0-3.5% WR.

Managing SORR is best done with your AA, in my opinion. See for example Kitces' discussions of bond tents: https://www.kitces.com/blog/managing-portfolio-size-effect-with-bond-tent-in-retirement-red-zone/

 

[goingtotravel]

I am having trouble to understand this argument. How about if your portfolio return, especially the fixed income part, is below 2.75% annually?

Maybe I'm not stating things properly (new to this whole world of talking about finance TBH). Our portfolio is a smaller part of our overall retirement income. Around 65% of our expenses are covered by a pension and SS. Around 20% of our entire budget is flexible (travel, extra things like dining out more, hobbies) so we can reduce out WR easily if needed. If DH gets his VA disability claim approved our portfolio is now reduced to fun money and nothing else.

To pay off our mortgage before retirement (this year) we would have to pull from our taxable investments and that would push us into a higher tax bracket. If we wait until we retire we will still be in a higher tax bracket due to the pensions. I can't find a scenario where holding a mortgage at 2.75% hurts us but, I do understand it's more than likely that the fixed income portion of our portfolio will return less than 2.75% for a long time.

Does that make more sense than my original statement? Am I missing any other financial consideration (cost) to hold the mortgage? It isn't keeping me up at night so the only consideration is financial.

 

[goingtotravel]

FIRECalc nicely illustrates this if you input SS or a small pension, often resulting in something like a 4.5-5.0% followed by 3.0-3.5% WR.

That is exactly how I ran FIRECalc and came up with the WR's. Since we will be federal retirees we have multiple income streams on top of the amount we can withdraw from our portfolios. FERS pension + FERS supplement + TSP from 56-62; FERS pension + 1 SS +TSP (DH draws first) from 62-67; FERS pension + 2 SS +TSP from 67 onwards.

Will need to do some more reading on SORR and AA's. Thanks for the link.

 

[TechLead]

Just refinanced our mortgage to 2.75% with the intent to carry it into retirement. The payments are just over 10% of our budget at the start of retirement and we also have pensions that have a COLA feature. Our WR from our taxable accounts includes an 'extra' 1% for the mortgage. I have treated the mortgage expense in Firecalc as a non-inflation adjusted expense and still get to 100% success.

Having mortgage when retired makes it much harder to qualify for ACA. You need income to make mortgage payment. That extra income may put you over subsidy cliff and cause couple to pay 20k+ for Health Insurance instead of 4k-6k.

I would rather consider having reverse mortgage when retired. (Income that is not part of MAGI) Which I would repay as soon as we get Medicare.

 

[OldShooter]

... Managing SORR is best done with your AA, in my opinion. See for example Kitces' discussions of bond tents: https://www.kitces.com/blog/managing-portfolio-size-effect-with-bond-tent-in-retirement-red-zone/

Thank you for the link. I have always felt that SORR was over rated and fairly easily handled by an appropriate AA. I have also made a fuzzy sort of argument that equity percentage could increase as we have fewer and fewer years of SORR risk. But the Kitces article wraps those up nicely.

My only criticism of his concept is that we don't spend percentages; we spend dollars. So while his percentages may make sense for some "average" portfolio, a smaller portfolio may need more bonds in the tent and a larger one definitely does not need as large a percentage as he suggests.

 

[tb001]

I do the same thing, but I was frustrated when I ran my data through Firecalc. I use real returns in my model (2%) and there is no input for inflation. My pension is COLA'd so this works great. But when I run Firecalc with my numbers, inflation has a huge impact because of the way Firecalc uses inflation. The returns on my portfolio don't change in Firecalc if I change inflation. This makes sense because Firecalc uses nominal return and simply subtracts inflation from the returns. This has the effect of reducing probability of success for me even though I have supposedly have no inflation risk.

This really got me thinking about my model and using real returns with no inflation input. So I did some research on returns vs. inflation. What I found was the market does not do well during inflationary periods. So the way Firecalc models it seems to reflect reality. But then I look at my model and it does not. The fix I put in was to add an inflation input and then drop my real returns as inflation goes up. Right now I drop the real returns 0.25% for every 1% of inflation. This is completely counterintuitive, but I felt I needed a way to account for inflation and the reality of what the market does during inflationary periods. Maybe not the best solution, but it works for planning.

Inflation does have a huge impact in my model if I carry a mortgage into retirement. My 2.75% mortgage is basically free money if inflation is 2.75%. That's one of the big reasons I am seriously considering keeping the mortgage for a full 30 years. If inflation goes above 2.75%, I am making money on my mortgage. And if I get 2% real return on my portfolio with 2.75% inflation, the NPV of my mortgage is less than the mortgage balance. We'll see how long I hold onto that, though. Best laid plans and all.

We went through a similar analysis regarding the mortgage. It’s a decent chunk of our spend and it was striking the impact it had on success rates. I’m normally a no debt ever kind of person, but over 30 yrs at these rates, it was a hard argument to make.

 

[Montecfo]

Having mortgage when retired makes it much harder to qualify for ACA. You need income to make mortgage payment. That extra income may put you over subsidy cliff and cause couple to pay 20k+ for Health Insurance instead of 4k-6k.

I would rather consider having reverse mortgage when retired. (Income that is not part of MAGI) Which I would repay as soon as we get Medicare.

The OP appears to be funding the mortgage from taxable. If you do that it may have little impact on MAGI.

 

[goingtotravel]

Having mortgage when retired makes it much harder to qualify for ACA. You need income to make mortgage payment. That extra income may put you over subsidy cliff and cause couple to pay 20k+ for Health Insurance instead of 4k-6k.

We have health insurance in retirement through the FEHB so we aren't worried about ACA. Our projected HI costs are $450 per month through FEHB based on 2020 rates. That is a good point though for folks that have to buy through ACA.

 

[Retch The Grate]

However, I'm not really familiar with rigorous published analyses that show how the regular or reverse glide path approaches affect SWR with clear charts documenting the comparisons. Anyone have anything on this?

BigERN from Early Retirement Now, did some analysis:
https://earlyretirementnow.com/2017...e-withdrawal-rates-part-19-equity-glidepaths/
Context for the following graphs is in the article linked.

 

[corn18]

Very cool article. Thanks for that.

 

[HI Bill]

Our current FIRE model is 4.5% WR for the first 5 years; 5% for the next 5 years; and then 3.5% after we both claim SS (67). Reading some threads makes me nervous when people are advocating a 3% perpetual WR but our numbers seem to work and we have flexibility in our budget if needed. Biggest worry at this point for us seems to be SORR, right?

You're making me feel much better about my current plan. Buying a retirement home, and will have a 2.5% WR prior to travel resuming, when we hope to bump it to 4.4%. We're taking on loan at 2.875% 30-year fixed... I do intend to pay off the property in 5-10 years....

 

[Ed_The_Gypsy]

I put together a graph of historical 30 year 60/40 SWRs vs the P/E 10 for the S&P 500 (a measure of how pricey the market is). The data is from a portfolio that has small cap and international stocks, so its lowest historical SWR was 4.5%. At a P/E 10 of 30 (about what it is today), it suggests a SWR of 4.05, with a 50% chance of success. If you want an 84% chance of success, the SWR drops to 3.3%. There is reason to believe that even this withdrawal rate is too high because we have an aging population and slowing growth rate in the US, which indicates we will have slowing increases in GDP going forward. The bogleheads seem to be correct: a 3% fixed withdrawal rate seems to be reasonable for a 30 year retirement with fixed withdrawals.

Nice work! Thanks.

 

[TechLead]

BigERN from Early Retirement Now, did some analysis:
https://earlyretirementnow.com/2017...fe-withdrawal-rates-part-19-equity-glidepaths
[/IMG]

I was also thinking to post pointer to BigERNs analysis.
Those are best withdrawal analysis I had ever seen.
https://earlyretirementnow.com/safe-withdrawal-rate-series/

Most point to 3.0-3.5% WR.

 

[goingtotravel]

You're making me feel much better about my current plan. Buying a retirement home, and will have a 2.5% WR prior to travel resuming, when we hope to bump it to 4.4%. We're taking on loan at 2.875% 30-year fixed... I do intend to pay off the property in 5-10 years....

I'll post what the FA says about our plan when we meet with him. I keep reading that the current consensus is 3% but I think with a flexible budget, 2 pension payments with a COLA feature, and possibly a new income stream from the VA (that won't be taxed) I'm not too worried. We can drop all the way to 2.5% WR except for the 5 years between our FERS supplement ending and me claiming SS at 67.

 

[audreyh1]

I'll post what the FA says about our plan when we meet with him. I keep reading that the current consensus is 3% but I think with a flexible budget, 2 pension payments with a COLA feature, and possibly a new income stream from the VA (that won't be taxed) I'm not too worried. We can drop all the way to 2.5% WR except for the 5 years between our FERS supplement ending and me claiming SS at 67.

The 3% consensus is simply due to some folks believing that the future is going to be worse than the past, plus longer retirements.

 

[Andre1969]

I always used to think 1973 was a failure cycle, because of that graph on the front page of the FireCalc website. But just recently, I noticed it's using a 4.67% withdrawal rate, not 4% ($35K per year/$750K). I swear, I've been looking at that picture for probably 15 years now, but only recently, did I actually do the math.

 

[Out-to-Lunch]

What was the interest rate on your 40%? Most of the 70’s, the CD interest rate was 2X your SWR. Who in their right mind would not adjust holdings when guaranteed money is more than withdrawal rates.

I realize this post is a few days old. I think you may not understand what "the 4% SWR rule" actually is. The rule is that you are supposed to increase your annual withdrawal to keep up with inflation, i.e., a 4% real withdrawal rate, not nominal.

I don't think that in the '70s, CD rates were 4% higher than the rate of inflation, let alone 8% higher.

 

[Retch The Grate]

I was also thinking to post pointer to BigERNs analysis.
Those are best withdrawal analysis I had ever seen.
https://earlyretirementnow.com/safe-withdrawal-rate-series/

Most point to 3.0-3.5% WR.

yeah the whole series is really well done. It's a big part of why I settled on 3% as an indefinite duration SWR.

 

[Dtail]

We are at 2.93%WR average through our first 3 years. Would be willing to go to 3.25%, but not higher yet.

 

[GravitySucks]

A related table Figure 1 in page 2 from Bengen. Pretty amazing how few times 6% failed since 1926.
https://www.fa-mag.com/news/is-4-5---still-safe-27153.html?section=47&page=1

Since I retired I keep seeing threads here and Bogelheads and online articles that the 4% rule is dead. I don't believe them. I don't think we're living in the worst of times of the last 100 years. Look at the 2000 retiree, they have gone through the dotcom bust, the great recession, and the Covid shut down. With only 10 years left on a 30 year plan they'll probably be fine.

I like Bengen's term SAFEMAX as it implies this is a guardrail on your spending. If you want to live with this much risk, you need to stay at or below that level spending.

Personally, I did a strict 4.5% withdrawal the first 5 years of retirement, and spent it almost every year. Readjusted at year 5 to a 5.5% max of the balance at that date. Almost spent that much last year with the new roof, new fence, tree removal, etc. etc., Not going to come close to it this year. I plan on readjusting every 5 years

 

[W2R]

I am in my 11th year of retirement right now. I calculate my withdrawal rate as the percentage of the previous December 31st balance of my portfolio.

During those first 10 years, my average WR was 1.7%, or 2.4% if you include the costs of buying my new home in cash and moving, back in 2015.

Either way, it is less than I had planned. According to my written financial plan which I devised during the run-up to retirement, my spending limits were going to be 3.5% (or my total dividends, whichever was smaller). But I have found that since I retired I just do not want or need to spend that much in order to be happy and content. Plus, recently I have SS and a mini-pension to rely upon too.

At this rate, if my WR ever fails I'll be in a state of shock! But you never know what the future may bring.

Or maybe I'll change my lifestyle and develop a burning need for a high end Tesla, a yacht, a private plane, an RV, and a castle-like second home on a cliff overlooking the Mediterranean, in southern France. Probably not.

 

[thepalmersinking]

Caveat that this ama is a few years old now, but I found it to be a valuable thread with lots of interesting questions and answers. I learned a lot that I did not know.



https://www.reddit.com/r/financiali..._bill_bengen_and_i_first_proposed_the_4_safe/



The first reply is that Bengen states that he later found out that it should actually have been a SWR of 4.5% vs. the 4% he initially reported. For someone with longer retirement plans over the traditional 25-30 years he offers alternative SWR rates.



I was shocked that he felt one could live in perpetuity on a 4% w/d rate.



He does also point out what his research indicated is the sweet spot of an AA to maximize the SWR. This which was 45-55% equities and the remainder in bonds/cash ("According to my research, I would have a well-diversified portfolio with probably 50% equities, 40% fixed-income, and 10% cash")

I’ll point out that AWSHX on the back test model using a starting point of 1985 and a starting balance of $1million and a 7% annual withdrawal had an ending balance thru last month of $3.3million for a CAGR of 3.3% after withdrawal. I am not a fan of this mutual fund due to expenses but it is a true real life test that survived through the 1987, 2000,2008,2020 meltdowns. It’s an old fund so the Calculations are not hypothetical.