Several years ago I discovered the concept of financial independence, and the idea was a revelation. Life too often becomes a series of developed habits, and the process of questioning the basic assumptions behind those habits and envisioning a system where your investments can do the work for you to fund your lifestyle in perpetuity was a life-changing exercise.
As an engineer with a math minor, I was especially taken by a few prominent retirement studies on the subject of safe withdrawal rates. They famously came to the conclusion that for a traditional mix of stocks and bonds one could have retired with a 4% safe withdrawal rate, adjusted their expenses for inflation each year, and had a successful 30-year retirement with a high amount of certainty historically. The process they used for back-testing retirement scenarios was fascinating, and the results form the backbone of the vast majority of retirement advice today.
So I dove in and explored the data and assumptions and engaged full-force in various early retirement communities online. I played with a few of the prominent retirement calculators out there and tinkered relentlessly. Anyone else who has done the same can understand how intoxicating it all can be.
So much so, that after building my own models I realized a lot of people totally misinterpret the conclusions and get it all wrong!
To explain, let me ask a relatively simple question:
What’s the safe withdrawal rate of the Permanent Portfolio?
The Permanent Portfolio consists of 25% stocks. Referencing the handy Trinity Study table, a portfolio with 25% stocks survived a 30-year retirement with a 4% SWR 71% of the time. With a 3% SWR, it survived 100% of the time. So a reasonable conclusion might be that the safe withdrawal rate for the Permanent Portfolio is somewhere between 3 and 4%.
It seems pretty straightforward on the surface. However, I’ll give you a sneak preview – I plan to provide data that shows that the Permanent Portfolio has supported a safe withdrawal rate closer to 5%. That’s a massive difference that flies in the face of conventional wisdom. How can that possibly be true?
The answer lies in a little-understood base assumption of not only the Trinity study but also other retirement studies and most retirement calculators you find online. They almost all assume that you only have two investing options — a United States S&P500 fund and a single broad bond fund. The bond fund chosen varies by study, but the key element is that they narrow it down to only one for simplicity.
Let’s look again at the Permanent Portfolio asset allocation. Yes it is 25% stocks, but the percentage of stocks does not tell the whole story. More specifically, it uses Total Stock Market, Long Term Treasuries, Total Money Market (T-Bills), and Gold.
The first sign of a problem is gold. It’s neither a stock nor a bond. So right off the bat it’s easy to see that a study or calculator that only looks at stocks and bonds might not model the Permanent Portfolio very well.
Next, let’s look at how the performance of Long Term Treasuries and T-Bills compare to the Total Bond Market:
(Last 5 years, data courtesy of Yahoo Finance)
Even though they can all be classified as bonds, there’s really very little similarity in how these funds act in the real world. And why should we expect any different? Different indices track different things, and you’re probably researching portfolios to begin with because you know this to be true and are looking for ways to augment your returns. Why should that not also affect retirement withdrawal rates?
The answer is that the asset allocation and index you choose absolutely affects retirement scenarios, and the difference is significant. Back to our question, let’s compare the calculated safe withdrawal rates of the Permanent Portfolio – using its actual specific asset allocation – to a simple 25/75 broad stock/bond blend:
As one would expect from reading any retirement study, the SWR for the traditional stock/bond blend trends towards the 3-4% range the longer out you look. (Astute observers can read about why the specific numbers may differ a bit from your favorite study here). What you probably didn’t expect is that another portfolio with 25% stocks trends closer to 5%. Clearly not all stocks and bonds are created equal, and asset allocation has a huge effect on retirement performance just as it does on compound returns in accumulation.
Now that I hopefully have your attention, let me offer a tool to help you explore this idea in more depth – I’ve created a new calculator to allow you to make the same charts for your own portfolio. It’s called the Withdrawal Rates calculator. For detailed information on how it works, be sure to read the Methodology.
In addition to showing the Safe WR for your portfolio over a variety of retirement timeframes, you’ll also notice that the new chart has a second data point called the Perpetual WR. That’s my reaction to a second base assumption in most retirement studies and calculators – that having $1 to your name at the end of the studied period constitutes a retirement “success”. If you’re like me and hope for a long happy retirement for more than 30 years or you plan to leave an inheritance to your family, that possibility does not sound reassuring in the slightest. To address that issue, the PWR is the maximum withdrawal rate that would have ended with the same inflation-adjusted principal as when you started. Why spend down your life savings when, managed properly, it can not only fund your own retirement to a very old age but also be passed to your children and grandchildren?
Just like the moment when I first heard about financial independence, I was initially skeptical of the data and I imagine you might be as well. If your developed habit is to reflexively quote a 4% SWR, questioning that assumption may even feel a bit heretical. But the more I studied the effects of asset allocation on retirement performance the more eye-opening it all became, and once you think beyond the limiting assumptions of classic retirement studies the possibilities become empowering. Are there portfolios that have supported higher safe withdrawal rates than 4%? Are there portfolios with good withdrawal rates that maintained principal even over short 10-year timeframes? The answer to both is a resounding yes! The Portfolios section now also includes withdrawal rate information, so feel free to browse for a few ideas.
Play with the calculator for a little while and watch how the rates change, and you’ll start to see withdrawal rates in a whole new light. The effects of diversification beyond lumping all “stocks” and “bonds” together will quickly become clear, and hopefully you’ll find that realization as interesting and exciting as I do. Perhaps the withdrawal rates for your own portfolio will surprise you. Or maybe a portfolio you like but were skeptical of for retirement – simply because it doesn’t fit the mold of a “traditional” retirement portfolio – suddenly looks extremely appealing. Good investment opportunities are out there for people willing to put in the effort to find them.
As you go off to explore, I do want to encourage you to read the Withdrawal Rates Methodology for much more information in the methods, assumptions, and limitations. (Update: I’ve also started a FAQ to help answer some of the most common questions, and written a follow-up post on How Safe Withdrawal Rates Work). The more you understand how it works under the hood, the better. The last thing I want to do is shake you from your old misplaced conclusions only to latch you to new ones. Please don’t take my word for it, don’t assume that just because something was awesome in the past means that it will continue to do so in the future, and definitely don’t run off and change asset allocation tomorrow because of what you see. Study the topic for yourself, and challenge your assumptions at every step.
You might be surprised what you discover, and it may unlock retirement possibilities you didn’t even know existed.
UPDATE: The original version of this article used the Swedroe Min Fat Tails portfolio as the working example. New and better data cast doubt on some of the specific Swedroe numbers, causing me to temporarily remove the portfolio from the site. I replaced the example with the Permanent Portfolio to illustrate the same idea without confusing readers who stumble across this article for the first time. The portfolio may be different, but the overall message is exactly the same.
Also, the chart has visually changed quite a bit with lots of additional data. Read this for a walkthrough.