The Safe Withdrawal Rate is one of the most debated concepts about retirement. Why? The human being is a creature that needs certainty. Assurance. Certitude.
We cannot accept that what happens to us is simply…random. The existence of humanity itself happened by chance. And yet we created religions, God(s), philosophy, to give ourselves a justification. What’s the purpose of life? Why are we alive?
Do not look at physics, do not look at the universe, otherwise you might come to the conclusion that a curious series of random events brought us here, a “lucky roll of dice” that would be otherwise meaningless for the remainder of the infinite space around us.
In this context, it does not take too many mental steps to go from asking “why am I here?” to ” when can I stop working?”. I mean, since we are here, let’s try to have as much fun as possible.
Safe Withdrawal Rate
I do not remember when I learned about the theory behind the Safe Withdrawal Rate but I still feel the peace that it brought into my life. I am an engineer, I find comfort in quantities. The fact that existed a figure, a target amount, THE FUCK-YOU NUMBER that would tell me “now you can spend 4% of X and be fine” sounded great.
Look, my father opened for me a saving account when I was really young. The bank branch was on the other side of the road from his office. There was no internet, online banking and shit. I had this folded piece of paper, in a plastic cover, where the bank teller was writing down each deposit I was making and how much interest accrued. (At the time, inflation in Italy was double-digit but I did not know, what I knew where the high interest I was getting!). Each trip to the bank was a happy event.
It didn’t take me too long (or at least I feel so now) to ask my father “ok, when I can start to spend some of this without breaking the toy?” It is nice to see these numbers growing but…that’s not the point of it, innit? My father told me that, if I wanted to keep the money tree alive, I could only spend the interest generated by the interest. Now you can understand why the 4% Rule felt so generous to me when I read about it.
[As usual, I am not going to explain here what the SWR or the 4% Rule is. If you are still reading, do not have a clue what I am talking about and didn’t reach for Google yet…it is because you are still pissed at me when I said that the god you believe in does not exist. Get. Over. It.]
All Models Are Wrong But Some Are Useful
If you are not the guy who wrote the paper about the 4% Rule, your path to fame (well, relatively speaking since many know the rule, few know the author) is to write a paper to disprove it. As I said, I wanted to believe in the rule because the reality of it would have been so cool. But the moment you sit and spend even just 10 minutes to think about it, you realize how reality is so complex, multi-facet and RANDOM that it would be impossible to have anything other than a rule of thumb, a back-of-the-envelope calc that just brings you in the right ballpark. Which is better than my father’s theory 😉
Unfortunately, a multitude of critics of the rule, instead of trying to improve it, decided to go down the rabbit hole of preciseness and unnecessary complexity. I want to address these critiques first and then propose two positive ways to build on top of the SWR theory.
Rain on your parade
I would like to first address a weakness of the model, to highlight how difficult is to effectively represent reality and its multitude of possible outcomes. The original SWR model, the 4% one, tested financial portfolios over a 30-year horizon, to mimic the life expectancy of a ‘standard’ retiree. The saving pot generates cashflows to fund retirements in two ways: returns on the invested pot and withdrawals from the pot itself. It is easy to understand that cashflows from investments are heavily impacted by the total amount invested: once the retiree starts to meaningfully deplete the saving pot (say more than 10% of the starting amount), there is no way back…unless the retiree aggressively cut their spending for years.
If we look at year 20 in retirement and the saving pot is 80% of the original amount, the model considers this scenario as a success. Something crazy has to happen in the financial markets to bring the pot to 0 in less than 10 years. BUT. If you are a retiree in year 20, you do not know when your finish line is; most likely, you wish you have more than 10 years left…and act accordingly. That’s why seeing your pot 20% depleted would not make you feel like a success.
That’s at least how I see it as a non-retired 45yo. This morning I was reading this interview on Monevator with a couple that is early retired.
“With our spending heavily front-loaded so that we can make the most of retiring early – and with the tough market conditions since 2020 – that hasn’t always been easy. But we’re not too far off plan!“
It makes sense to spend more when you are ‘younger’. I am not sure how you can model it, given that the biggest risk these guys are facing is the sequence of returns. Front-loading spending means using part of the principal (unless you plan to die with more than you had when you retired), but there is so much you can safely touch before putting yourself at serious risk if the future takes an inconvenient path.
Americans did not invent pizza
Using only returns from US stocks and bonds was obviously an error in that 4% study. But what would you expect from the guys that call football a sport that is played 95% of the time with hands? The guys that can spend hours debating where you can find the best pizza and then the only options they allow are Chicago or NYC?
What I do not understand is why later studies looked at other countries…but switched from one type of home-country bias to another flavor. Why Italian investors should own mainly Italian stocks and bonds? In other words, why every investor cannot be exposed to the same global market portfolio, adjusted for the relevant currency? In 1950, an Italian investor could not access US stocks and vice-versa, but the point here is to simulate what today’s investors might face in the future, not to draw conclusions on what past investors should have done. Today, many European investors own Apple and Treasuries, for very good reasons; why the model cannot reflect that? Can we assume that today’s investors learned the relevant lessons from the past…or that paper would not generate enough clicks? I do not know…
Who wants to live forever?
I have the feeling that, once there was enough consensus around the fact that 4% might be a too high SWR, a competition started to see who could drive it down the most. Up to silly conclusions.
0.8%? Dude, for real? I understand the goal of not providing unrealistic expectations but here the authors went 180 degrees. The paper does not provide a split by country other than the US sample but I suspect that 0.8% is driven down by places like Argentina, Chile and Czechoslovakia. Do you think that today, Argentinian investor has the majority of their saving in Argentinian assets? And if so, should we really take them as a reasonable benchmark? If 0.8% is the “real number” then investors would be better off playing a different game (working longer) or taking their chances and hoping they will not be facing the 5% of scenarios that led to ruin while spending 2.26% (maybe their family will help them? the Government?).
Morgan Housel just wrote a nice article titled The Art and Science of Spending Money. As much as part of the population has the problem of spending too much, there is a section that is afflicted by the reverse: the inability to spend enough. Stating that the SWR is 0.8% does not only mean that a retiree can spend less during retirement, it means they have to save more, and therefore spend less, while they earn. Are these papers actually bad for society? At the so irresponsible withdrawal rate of 4%, more than half of the total population dies with more money than they had when they retired!
Should we really alienate everyone like this? What’s the mortality rate for crossing the street in front of your house? Maybe you should stay home all the time, just to avoid the risk 😉
My Solution
What about the Belfers, a family that managed to lose money investing in Enron, Madoff AND FTX? What’s their SWR??
Instead of focusing so much on what can go wrong, the PhDs authoring those papers should look at what can be done to improve that %…like not investing is stupid shit or not selling at the wrong time 🙂
What really drives safe withdrawal rates is the sequence of returns, in particular what happens in the 10-15 years before and after retirement. A streak of bad returns in that period can drive the portfolio to levels so low that it would not be able to recover once conditions improve. A model that takes into consideration the particular path an investor is facing and acts accordingly, would probably be too hard to codify and explain. But in real life, in real-time, I can adjust my decisions.
Let’s say I decided to retire at the end of 2022 and the stock market lost c20%. I could postpone my retirement for a couple of years, so that I do not have to sell stocks at depressed levels (at least, not yet). I can do some random jobs and live out of that income for a year. I can postpone that trip I was planning to have and live on a slimmer budget. Or I could have planned this moment a la Target Day fund and start to fund my retirement by selling only short-term bonds. Nothing forces me to spend 4% today if I feel it is not the right decision.
The closer I am to the retirement date, the more options I have. 15 years into retirement I would not be able to start working again but…who cares because the sequence risk is passed by then. If in the last 15 years the market went up, I would have more money than I need; if it went down, provided I did not do anything stupid, the chances I will see positive, above-average, future returns are way way higher. If I managed my spending during the down years, I would have more funds to participate in the recovery than the traditional 4% model would suggest.
This is to say that retirees have more options than the 0.8% SWR paper would suggest, especially if they planned for a comfortable retirement where their budget is higher than their basic expenses (food, shelter and so on).
This is not market timing. For a person that retired in 2014, and saw the stock market grow 15%/year until the end of 2021, 2022 was just a one-year step-back in price. In other words, including 2022, the stock portion of their portfolio grew 10%/year, still an above-average result. For a person that retired in 2021, 2022 was a non-negligible hit to their portfolio. What is important is not what they both forecast for 2023, but their idiosyncratic situation, where their specific portfolio balance sits compared to their budget. If the 2021 retiree is entering a path that the 4% withdrawal would consider ruinous, adjusting to 3% or better 2% will improve his chances substantially; one ‘wrong’ year of spending at 4% is not going to matter.
Someone last year correctly tweeted something like “FIRE people mistook a bull market for a retirement strategy”. If you retired in your 30s, or even 40s, on a tight budget because I am frugal anyway, then you might have a problem. The longer you lengthen the horizon your savings have to work for, the slimmer the margin of safety in your budget, and the tighter the rope you are walking on. Just a reminder that yes, you can play with those models…but up to a point.
A Permanent Portfolio
The solution that is closer to my heart is to build a retirement portfolio that is ‘resistant’ to drawdowns (without sacrificing too much of the upside, ça va sans dire). In 2022, TheItalianLeatherSofa Model Portfolio lost 13.3% vs 15.9% for a 60/40 and 18.6% for the S&P500. This means that a retiree using the MP with a 2.6% SWR will have their 2023 budget ‘for free’, compared to a retiree using the 60/40.
In other words, it would take a 15.3% gain for the MP to go back to the end of 2021 level, 18.9% for the 60/40 and 22.8% for the S&P500. These figures make a difference, especially if the retiree faces multiple, consecutive negative years.
The best way to illustrate this concept is to check that incredible tool that is PortfolioCharts.com
One of the things we can test for each portfolio is the Withdrawal Rates: for a set retirement length, what is the safe WR based on that portfolio’s worst case? Sure, the analysis here is less sophisticated than what the previously mentioned papers did (PC only test past, real periods and not possible, MonteCarlo simulated ones) but what I want to show you is more the direction, than the precise SWR value.
Look at the results for the following 4 portfolios: the Total Stock Market, the Classic 60-40, the Pinwheel and the Golden Butterfly.
The SWR goes from 4% to 5.8%! This is because each subsequent portfolio brings the retiree closer to the risk/return efficient frontier, where risk is defined as the Ulcer Index and return as the Baseline Return. Controlled drawdowns and stable returns are the ideal ingredients to defeat our enemy, the sequence of returns.
The Golden Butterfly has lower expected returns than the Total Stock Market portfolio, but that’s not relevant when the retiree can anyway only spend the SWR; the goal is not to grow the portfolio at the fastest rate, it is to spend and avoid the risk of ruin.
Lastly, a piece of bad news for non-US-based investors. Their SWR would most likely be lower compared to their American friends because of the currency risk:
- if they decide to not hedge, their portfolio will have fatter tails: the probability of facing a big drawdown immediately after retirement increases
- if they hedge, their portfolio will generate lower returns (hedging costs)
Here you can find the Golden Butterfly for a EUR investor (same ETFs as above, so the difference is pure FX rates…if I am reading PC correctly):
The real-life example would be slightly different and better, because instead of a 100% USD portfolio converted to EUR, the retiree would have a mix of USD, EUR, GBP and so on assets (the market cap portfolios for stocks and bonds).
Maybe to close on a brighter note, I will quote Kris Abdelmessih: “Luck is not a strategy. But it exists.”. If you retire and hit, maybe not the best paths, but even the average ones….SPEND THE FRIGGIN MONEY!!!
What I am reading now:
Follow me on Twitter @nprotasoni