Welcome back to our series on **Safe Withdrawal Rates**. This is already the 17^{th} installment! See Part 1 here and make sure you also check out the Social Science Research Network Working Paper we posted on the topic, now with 2,000+ downloads!

In any case, if you have followed the series so far you must have noticed that we are no fans of the 4% Rule and much of what we posted here dealt with the “4%” portion of the 4% Rule. For example, in Part 3 of this series we show that when equities are as expensive as today (Shiller CAPE > 20), failure rates of the 4% Rule have been unacceptably high in historical simulations.

But I think I missed this really important point:

**The only thing more offensive than the “4%” part is the word “Rule”**

That’s because the word “Rule” makes it sound as though the 4% is some sort of a scientific or mathematical constant. But it’s not. It ain’t scripture either, even though it’s often portrayed that way! There is no one-size-fits-all solution for withdrawals in retirement. With today’s lofty equity valuations and measly bond yields, a 3.25% to 3.50% initial withdrawal rate would be much more prudent. But there is another element that creates just as much variation in SWRs: Different assumptions about Social Security and/or pension benefits: The benefit level, the number of years before benefits kick in, how much of a haircut you want to assign to account for the risk of potential future benefit cuts, etc. and they all create so much variation in personal SWRs that the whole notion of a safe withdrawal rate “Rule” is even more absurd. The 4% Rule should be called the **4% Rule of Thumb** because 4% is merely a starting point:

SWR = 4% Rule of Thumb

+/- adjustments for equity/bond valuations

+/- adjustments for idiosyncratic factors, e.g. age, Social Security, pensions, etc.

How much of a difference do these idiosyncratic/personal factors make? A **huge** difference! A prime example is the case study I worked on over at the ChooseFI podcast: a couple in their early 50s expects pretty generous Social Security benefits after a long career and probably wouldn’t have to worry too much about future benefit cuts. If they both wait until age 70 to claim benefits and are able to reduce their withdrawals from their portfolio dollar for dollar once Social Security kicks in, their Safe Withdrawal Rate estimate goes up from a measly 3.5% to somewhere around 4.5% or even 4.75%. Instead of saving 28.6x annual expenses, they’d need only 22.2x or even 21.1x. That’s a difference of several $100k!

How to quickly and easily gauge the impact of future cash flows from Social Security or pensions on the SWR is the topic of today’s post!

### How do we translate future benefits into today’s SWRs?

OK, if you’re not a math wonk you might want to skip to the next section. But for the mathematically inclined, those who didn’t unsubscribe from our email list after the Technical Appendix in Part 8, recall that the safe withdrawal rate can be expressed as:

The terms Ct are used to calculate the opportunity costs of taking withdrawals at different times, please refer to technical appendix for more details. So, it turns out that the impact from future cash flows on the SWR enters **linearly** into the SWR formula. If a future benefit of X translates into an increase in the SWR by Y, then 2 times X will increase the SWR by 2 times Y. That’s very convenient! All we need to calculate is how, say, a $10,000 future benefit (with inflation adjustments) impacts the SWR of a $1,000,000 portfolio and we can easily scale up/down the numbers for $20,000 or $30,000 benefits. Or $23,456 in benefits for a $1,987,654 portfolio.

Also, quite amazingly, the impact on the SWR is independent of the targeted final asset value (FV). Of course, FV enters the formula for the SWR but the two terms for FV and supplemental cash flows are completely separate and additive. How cool is that???

### Introducing: Big ERN’s cash flow translation tool

Let’s use the above formula to calculate the potential marginal impact on the SWR of an additional annual cash flow worth 1% of the current portfolio.

- Cash Flows start at between zero and 40 years after the retirement date, in 5-year steps.
- Cash Flows are adjusted for inflation (though, we will also calculate a table for non-COLA cash flows below).
- Retirement horizons are 40, 50, or 60 years.
- Equity Allocations are 60%, 80%, or 100% in an S&P500 index fund, the remainder in 10Y U.S. Treasury bonds, e.g., the iShares IEF ETF.
- Since I want to get a feel for the
**range of possible outcomes**I calculate the minimum, median and maximum SWR impact over the simulation windows from 1871 to 2015. Please refer to Part 1 of this series for more details on the historical simulations.

The results are displayed in the table below. A few observations:

- If benefits start in year zero you can raise your consumption level by exactly 1%. Makes sense because the supplemental cash flow is adjusted for CPI just like your withdrawals. Technically speaking you don’t really raise your SWR from the portfolio but you raise your sustainable consumption level one for one with the additional cash flows.
- The longer you delay your benefits the smaller the impact on the SWR. In the extreme case where the retirement horizon is 40 years and benefits start in 40 years (month 481 to be precise), you get zero impact on the SWR!
- There is a small but still noticeable impact from the asset allocation. The higher the equity share the lower the impact on the supplemental cash flows. Makes sense because what we do in the SWR formula above is to discount future cash flows by cumulative portfolio returns. Since returns are higher on average with a higher equity weight we discount future benefit flows with a higher rate.

How exactly do we use this tool? Let’s look at a few case studies:

**Case Study 1:**

- Retire at age 30. Portfolio $1m, 80% Stocks.
- Expect $15k p.a. Social Security (=1.5% of today’s Net Worth) at age 70, i.e., 40 years from now. This may seem low but keep in mind that people starting to receive benefits in the year 2057 will already be subject to substantial benefit cuts in Social Security.
- Retirement horizon 60 years.
- The impact on the SWR: Minimum: 0.174% (=0.116%x1.5). Median: 0.282% (=0.188%x1.5). Maximum: 0.461% (=0.307%x1.5).

**Case Study 2:**

- Retire at age 40. Portfolio $1m, 80% Stocks.
- Expect $25k p.a. Social Security (=2.5% of today’s Net Worth) at age 70, i.e., 30 years from now.
- Retirement horizon 50 years.
- The impact on the SWR: Minimum: 0.395% (=0.158%x2.5). Median: 0.673% (=0.269%x2.5). Maximum: 0.980% (=0.382%x2.5).

**Case Study 3:**

- Retire at age 50. Portfolio $1.4m, 60% Stocks.
- Expect $56k p.a. Social Security (=4.0% of today’s Net Worth) at age 70, i.e., 20 years from now.
- Retirement horizon 40 years. Not too different from the case study I did for ChooseFI, by the way.
- The impact on the SWR: Minimum: 1.012% (=0.253%x4.0). Median: 1.668% (=0.417%x4.0). Maximum: 2.188% (=0.547%x4.0).

**Case Study 4 (close to the ERN family situation)**

- Retire at age 44 (Big ERN) / age 35 (Mrs. ERN). Portfolio $3m, 80% stocks.
- Expect $28k in benefits at age 70 (in today’s dollars) = 0.8% of the initial portfolio . This already reflects a hefty haircut to the Social Security benefits.
- Retirement horizon 60 years (due to age difference and longevity in both families)
- Benefits start in 26 years, so we do a linear interpolation: 0.2 times the 30-year and 0.8 times the 25-year row: Minimum: 0.200% =0.8x(0.8×0.262%+0.20×0.205%). Median: 0.318%, Maximum 0.426%

Results so far: The range of the potential impact of Social Security is huge. Only around a quarter percentage point for the very young retiree. Combine that with the 3.25% SWR over a 60-year horizon given today’s high CAPE ratio and lo and behold you land at a 3.50% SWR, exactly the rate we prefer. Even for the ERN household with maximum benefits rolling in only 26 years after retirement, Social Security’s impact on the SWR is only around 0.2%-0.3% (being conservative we consider only the range between the lower end and the median of the historical simulations). That’s because the benefit level is still relatively small compared to the $3m portfolio target!

On the other hand, a couple where both spouses are in their early 50s and both expect substantial Social Security (especially compared to the moderate portfolio size) can expect a massive boost to their SWR (minimum of 1%, median over 1.6%!), see Case Study 3! That would put the SWR into the mid-4% range!

### What if benefits are not adjusted for inflation?

Glad you asked! We can also do the same exercise with future cash flows that are not adjusted for inflation. This would be the case for many corporate pension plans and most annuities.

What I did in this table is to assume that the initial cash flow is again 1% of the initial portfolio value. But then, every month, we decay benefits by an estimated 2% p.a. inflation rate. Notice that even an immediate benefit of 1% of the portfolio value (e.g. an immediate annuity) translates into significantly less than 1% extra SWR. That’s because the benefit will slowly decay by 2% p.a. in real terms.

**A simple example on how to use this table:** A retiree has $1,200,000 portfolio today, expects $20,000 in 20 years from a company pension (in future dollars, not adjusted for inflation). The retirement horizon is 60 years, the portfolio allocation is 80/20.

- Step 1: At a 2% annual inflation rate the initial future benefit is $13,459 in today’s dollars, which equals roughly 1.122% of today’s portfolio value.
- Step 2: We look up the three values from the table: 0.247%, 0.376%, 0.456% and multiply with 1.122. So the impact on the SWR is between 0.277% and 0.511%. with the median at 0.422% over the 1871-2015 simulation period.

### Conclusion

I’m not the Grinch who is against the 4% Rule all the time. Quite the contrary! If someone had retired around the market bottom in March of 2009, any withdrawal rate under 5% would have been crazy conservative considering how undervalued equities were back then. Moreover, even with today’s elevated CAPE ratio, I have recommended 4%, even close to 5% safe withdrawal rates for some case studies I performed! But that’s due to idiosyncratic factors. For the typical extremely early retiree in their late 20s or early 30s, Social Security benefits will have a negligible impact on the personal withdrawal rates. You are compounding three factors that all diminish the usefulness of Social Security:

- You have relatively few years of prime earnings, thus benefits will be lower.
- Benefits are far away in the future and have to be discounted more severely than for retirees in their 40s or 50s.
- There is more uncertainty about future benefit cuts the younger you are.

What I learned from the calculations here and from some of the case studies I have done (one at ChooseFI, one here on my blog, and another one coming next week) is that determining a safe withdrawal rate is more than pulling an arbitrary number from the web just because the Trinity Study or some FIRE bloggers say so. Instead, **determining your SWR is a highly customized exercise!** We don’t all wear the same shoe size, we don’t all wear the same clothes and we all have very different drawdown strategies, see the #drawdownstrategy initiative started by Fritz Gilbert and Physician on FIRE for a sample and our own ERN family retirement plan here. Then why should we all use the same withdrawal rate? We can certainly start with a 4% Rule of Thumb, but we have to be cognizant of macro-factors (equity valuations and bond yields) and, equally important, all the different idiosyncratic and personal parameters of the individual retiree. The path to wealth may be simple, but the withdrawal phase is a lot more complicated than many in the FIRE community want to admit!

And just for the record, even this customized SWR is just the starting point. In practice, of course, we’d have to revisit the withdrawal amounts, for example with a Guyton-Klinger or CAPE-based dynamic withdrawal method to temporarily scale up or down the withdrawals, if the market goes through a bull or bear market, respectively. See Part 9, Part 10 and Part 11 of this series for additional information.

### We hope you enjoyed today’s post! Please weigh in and leave your feedback below!

- Part 1: Introduction
- Part 2: Some more research on
**capital preservation vs. capital depletion** - Part 3: Safe withdrawal rates in different
**equity valuation**regimes - Part 4: The impact of
**Social Security benefits** - Part 5: Changing the
**Cost-of-Living Adjustment**(COLA) assumptions - Part 6: A case study: 2000-2016
- Part 7: A
**DIY withdrawal rate toolbox**(via Google Sheets) - Part 8: A
**Technical Appendix** - Part 9:
**Dynamic**withdrawal rates (Guyton-Klinger) - Part 10: Debunking Guyton-Klinger some more
- Part 11: Six criteria to grade dynamic withdrawal rules
- Part 12: Six reasons to be suspicious about the “
**Cash Cushion**“ - Part 13: Dynamic Stock-Bond Allocation through
**Prime Harvesting** - Part 14: Sequence of Return Risk
- Part 15: More Thoughts on Sequence of Return Risk
- Part 16: Early Retirement in a low return environment (The Bogle scenario!)
- Part 17: Why we should call the 4% Rule the
**“4% Rule of Thumb”**

[…] Part 17: Why we should call the 4% Rule the “4% Rule of Thumb” […]

LikeLike

[…] Part 17: Why we should call the 4% Rule the “4% Rule of Thumb” […]

LikeLike

[…] Part 17: Why we should call the 4% Rule the “4% Rule of Thumb” […]

LikeLike

[…] Part 17: Why we should call the 4% Rule the “4% Rule of Thumb” […]

LikeLike

[…] Part 17: Why we should call the 4% Rule the “4% Rule of Thumb” […]

LikeLike

Great post! Thanks for showing in-depth analysis of the 4% rule.

LikeLiked by 2 people

Thanks for stopping by! Geez, you are fast this morning. 🙂

LikeLike

[…] Part 17: Why we should call the 4% Rule the “4% Rule of Thumb” […]

LikeLike

[…] Part 17: Why we should call the 4% Rule the “4% Rule of Thumb” […]

LikeLike

[…] Part 17: Why we should call the 4% Rule the “4% Rule of Thumb” […]

LikeLike

[…] Part 17: Why we should call the 4% Rule the “4% Rule of Thumb” […]

LikeLike

[…] Part 17: Why we should call the 4% Rule the “4% Rule of Thumb” […]

LikeLike

Great analysis! I had noticed that, too: On the one hand you propose 4%+ SWRs for some of the case studies but in this series you always use caution and recommend 3.25-3.5%. Thanks for the clarification!

One other surprise: Are you now recommending the Guyton-Klinger Method? Or how do I interpret your last paragraph? It sounded like you criticized GK pretty badly in part 9 and 10. Thanks, ERN!!!

LikeLiked by 1 person

Great question! I think that GK is a decent method for adjusting withdrawals in response to market reality. It’s passive and straightforward. Good fit for most people in the FIRE community. Though, personally, I’d prefer the CAPE-based method.

I took offense to the whole notion that GK can somehow, miraculously increase your sustainable initial WR. That’s a myth!

LikeLike

GREAT post (again!), thank you! My question is a double-check of my understanding re: the mechanics of using the 2 tables:

For someone who expects to receive BOTH future Social Security and corporate pension (non-COLA adjusted) payments, I assume it is correct to add these future cash flow impacts onto our SWR%? For example, if our scenario expects an equivalent of 0.2% via SS and 0.2% via corporate pension, then we could consider adding 0.4% to our selected SWR to account for these 2 independent cash flows? Didn’t want to assume 🙂

LikeLiked by 1 person

Great question! You should be safe to aggregate the median. Technically speaking only the mean can be aggregated that way but the median will be close enough. If the starting dates of the benefits are close enough then you should be able to aggregate even the min/max. One way to do a precise analysis is to go to Part 7 and do your own study with the Google sheet, though. 🙂

Good luck!!!

LikeLike

Two thumbs up for this 4% rule of thumb post!

The combined math for Mr and Mrs PIE (51 and 44 respectively at FIRE), longevity of 50 years for each, SS for both at age 70 provides a 0.98% total bump as median. Let’s call it 1% for simplicity. My pension, starting to pay out at FIRE, will pay an annuity of 0.82% of our portfolio size until we both pass and is NOT inflation adjusted, like many corporate pension schemes. That provides a median bump of 0.57%, using your second table.

Jeez, think of all those jet-skis, high-end Burgundy wines, Armani suits and Gucci dresses we can can now spend more money on. Magical! Living high on the hog indeed! Who cares about the 4% rule and SoR risk? Let’s go bananas with 5.6%! Yeah, yeah….you know I am messing with you……

Seriously, an excellent article on the math behind how you can be more intentional about your spending in retirement and make informed decisions based on future projected income streams rather than guessing with your finger (thumb?) in the air.

Hope all is well in ERN land and the search for a new ERN landing spot at FIRE is proceeding smoothly!

LikeLiked by 1 person

Ha, I was thinking of you guys when writing this. You are so lucky with your company pension! You’ll do great in retirement with so many layers of safety!

Yeah, all is well here in Ern-land. Secretly plotting my early departure! Best of luck and thanks for stopping by, Dr. PIE!!!

LikeLike

When paying down my mortgages I was shocked at how small extra principal payments at the outset of the mortgage dramatically shortened the term of the mortgage. Fast forward three decades and I had a surprising revelation: the curve for mortgage principal decline is shaped the same as for a decumulating portfolio. Thus “sequence of returns risk” is mathematically equivalent to those extra principal payments i used to snowball/avalanche my way to freedom from debt.

Ergo, deferring Social Security benefits do little to impact SWR because they are applied when the system is least sensitive to them. Thus given a choice between taking an early SS benefit vs selling equities in a down market, the former may be the safer course of action since it preserves principal.

LikeLiked by 1 person

I personally don’t recommend accelerated mortgage payments. It would exacerbate the sequence of return risk because you lower your initial equity investments and then concentrate flows into equity investments once you have all the extra cash flow when the mortgage is paid off.

In the top panel (minimum) the multiplier you apply to delaying SocSec by 5 years is close to 0.7. That’s exactly the 0.7 multiplier of claiming SocSec at age 62 vs. 67. So in the bad scenarios, you should be about indifferent. At the median, the multiplier for delaying is much higher than 0.7. Around 0.85. So, from an SWR perspective, it’s worthwhile to delay by 5 years (or even 8 years all the way to age 70).

LikeLike

I think accelerated mortgage payments are primarily a risk mitigation strategy nowadays. During the early ’80s when rates were higher, and Roth accounts did not exist, I did not know how to earn an after tax rate that was higher than the interest rate saved. Pity I didn’t know about John Bogle.

LikeLiked by 1 person

This may be my favorite part of the series yet. I’ve tried to estimate a SWR added for my SS and pension in the past by using a cash flow NPV analysis and adding that to my assets, but that was a clunky way to do it. Your method is so much simpler and elegant. Even better it gave me a result similar to mine further increasing my confidence that I truly am FI. Now I just need to get over the fact that my pension NPV increases by ~ $500K overnight in a about 6.5 years when I turn 50 so I can RE now instead. I know, first world problems…

LikeLiked by 1 person

Oh, Thanks! That made my day! Thanks for the compliment and thanks for stopping by!

Make sure you also check out Part 7: A Google Spreadsheet you can completely customize the future flows and see the impact on the SWR. 🙂

Cheers!

LikeLike

The need for proper asset allocation and making adjustment for inflation is very important. I may need to spend a lot of time on your site for me to fully understand the thread as I just stumbled on your site from the link I got from other personal finance website. You are doing a great job here.

LikeLike

[…] the safe withdrawal calculations are best performed on a one-by-one basis. As we pointed out in our post last week, a withdrawal rate strategy should respond to market factors like equity valuations and bond yields […]

LikeLike

Why do I always feel like I’m back at University doing math and statistics courses when I read your posts….. But hey, always interesting data and conclusions! Still trying to determined to which degree the 4% rule of thumb applied to our situation with a portfolio of ETF’s, Dividend stock and Real Estate. Probably not really, it’s going to be mostly cash-flow driven. Still, find these posts on the 4% rule rather interesting.

Thanks again Big ERN.

LikeLiked by 1 person

Thanks CF! I don’t see a reason why dividend stocks should be treated any differently from equities in the simulations.

Real estate is a different story. I started dipping my toes in RE and like the results so far. It’s a high return and roughly CPI-adjusted every year.

For the SWR exercise, it would probably be prudent to calculate the withdrawal amount from the financial asset portfolio and RE separately. So, retirement income = FinancialSWR x Financial Assets + Rental income.

LikeLike

The latter makes sense and is actually what we have tried to roughly do. It’s not unthinkable that we will increase the RE portion of the investments, got certain tax advantaged over here.

LikeLiked by 1 person

NIce! Thanks for sharing!

LikeLike

Thanks for in-depth knowledge of 4% rule/rule of thumb. At the same time, it is important that one saves enough for the future

LikeLiked by 1 person

That is so true. There is no substitute for saving in the first place! Thanks for stopping by!

LikeLike

[…] written about this topic extensively. Forget about the 4% Rule! It’s should be called the 4% Rule of Thumb! Specifically, withdrawal rates should actively respond to economic fundamentals (in addition to […]

LikeLike

[…] Part 17: Why we should call the 4% Rule the “4% Rule of Thumb” […]

LikeLike

[…] the rate of return I use for my own projections. It’s also one of the reasons I’m skeptical about the 4% Rule: How can you withdraw 4% every year if equities return less than that and bonds hardly return more […]

LikeLike

Hi. Brad or Jonathan at ChooseFI were recapping some points you made on their podcast, including (if I understood them correctly) that you said a 3.3% withdrawal would preserve the initial capital (and with inflation adjusted dollars I think), implying that if one didn’t have legacy concerns a higher withdrawal rate would be possible. I don’t remember you saying that a 3.3% withdrawal rate assumes preserving the initial portfolio amount. Please clarify and many thanks!

LikeLiked by 1 person

Great question!

Over very long horizons (60 years) it doesn’t make a huge difference in the SWR. We currently target 50% final capital (in today’s dollars) for a bequest and charitable causes. But going from 50% preservation to full capital preservation would only reduce the SWR by about 0.1-0.2%. Likewise, planning capital depletion we could increase the SWR by 0.1-0.2%.

Cheers!

LikeLike

Thanks. What about a 30-35 year time horizon (retiring at ~63 1/2)?

LikeLiked by 1 person

Over a shorter horizon, asset depletion vs. preservation can make much more of a difference. I suggest you look at the Google Sheet (Part 7: https://earlyretirementnow.com/2017/01/25/the-ultimate-guide-to-safe-withdrawal-rates-part-7-toolbox/) and play with the numbers there. I found that going from depletion to 50% preservation you have to knock off 0.5-0.6% from the annual SWR.

Also, the more you want to preserve the more you want to be in equities. Probably target 80/20 instead of 60/40…

LikeLike

[…] any of her assets and what accounts to draw down first? So many questions! As I pointed out in Part 17 of the Safe Withdrawal Series, a safe withdrawal rate calculation has to be a highly customized affair and that’s what […]

LikeLike