When it comes to early retirement the most important (and difficult) thing you have to grasp is your safe withdrawal rate.
FIRE bloggers rave about “the shockingly simple math behind early retirement,” but they almost never talk about the shockingly un-simple math behind safe withdrawal rates.
So this week, I invited Karsten Jeske, PhD – a former professor, Fed economist, quantitative finance researcher, and early retiree – to the podcast to share insight on how to estimate your safe withdrawal rate in retirement.
This is the most important financial planning concept early retirees must grasp to stay retired and guarantee they never have to go back to a j-o-b.
What you’ll learn:
- The fatal flaw of the “4% rule” and why you might easily run out of money in retirement if you follow it blindly
- The importance of analyzing the conditional probability of failure based on the actual year you retire
- Why sequence of returns risk and safe withdrawals rates are inextricably linked
- Why the business cycle in the year you retire is crucial to your safe withdrawal analysis, plus how to use the Shiller CAPE ratio to estimate your safe withdrawal rate
- How Social Security factors into your safe withdrawal analysis
- Why sequence of returns risk is front-loaded for early retirees but back-loaded for wage earners
Have you analyzed your safe withdrawal rate? Do you think it is more or less than 4%? 3%? Let me know by leaving a comment when you’re done.
Don’t miss an episode, hit that subscribe button…
If you liked this episode, be sure to subscribe so you don’t miss any upcoming episodes!
I need your help, please leave a listener review 🙂
If you liked this episode, would you please leave a quick review on Apple Podcasts? It’d mean the world to me and your review also helps others find my podcast, too!
Links mentioned in this episode:
- Early Retirement Now Safe Withdrawal Rate Series
- Trinity Study (4% rule)
- Asset allocation: How to use a bond tent to reduce sequence of returns risk (HYW068)
- Asset allocation: Does the “yield shield” really protect against sequence of returns risk? (HYW069)
- Asset allocation: Is rental real estate a safer type of “yield shield”? (HYW070)
- Schedule a private 1:1 consultation with me
- HYW private Facebook community
Read this episode as a post:
Andrew Chen 01:23
For today’s episode, I invited an economist to come talk to us about safe withdrawal rates and sequence of returns risk in retirement.
Our conversation was so substantive that I ended up deciding to break this up into two episodes: part one today, which covers a lot of the theory and underpinnings behind these two concepts, and part two next week, where we cover mitigation strategies for reducing our exposure to sequence of returns risk.
I hope you enjoy the conversation as much as I did. Here it is.
My guest today is Karsten Jeske, who also goes by the nickname “Big ERN.” Karsten has a PhD in economics and is a chartered financial analyst. He’s also an early retiree who FIREd a couple of years ago at age 44.
But before that, he taught economics at Emory University. He worked at the Federal Reserve Bank of Atlanta and also at the Bank of New York Mellon in the asset management group in San Francisco.
Since early retiring, he’s been traveling a lot with his wife and daughter, and he writes extensively on his blog, Early Retirement Now, about financial planning for early retirement with special emphasis on analyzing safe withdrawal rates in retirement.
I invited Karsten today to come and share insights about both safe withdrawal rates and sequence of returns risk, which are perhaps the most foundational financial planning concepts that early retirees must grasp to plan for a successful early retirement where your retirement portfolio doesn’t get depleted to zero before you pass away.
Karsten, thanks so much for joining us today to share your insights and analysis about these important topics.
Karsten Jeske 02:55
Thanks. Happy to be here.
Andrew Chen 02:57
I’d love to start out just by learning a little bit more about your background. You have this really interesting nickname “Big ERN.”
Where does that come from? Why do they call you “Big ERN”?
Karsten Jeske 03:04
I started my blog in 2016 and I was still anonymous at that time. And Early Retirement Now, so I would sometimes sign emails with ERN as a signature. So the name ERN was already around, and somebody made it “Big ERN.”
If you know the movie Kingpin (Woody Harrelson and Billy Murray), Billy Murray is the villain. He’s called Big Ern in that movie.
I really love that movie. I like Bill Murray. So the name stuck.
And then on top of that, I’m a big guy. I’m 6’6”.
Andrew Chen 03:40
Oh, wow.
Karsten Jeske 03:40
Just under 200 pounds. That’s why it stuck. And I like the name.
Andrew Chen 03:46
Awesome. In terms of your background, how did you get into becoming an expert on early retirement finance and on the topics we’re going to discuss today: safe withdrawal rates and sequence of returns risk?
Karsten Jeske 03:57
I had always planned to retire early, even before this was an issue on the web. The funny thing is it was already an issue on the web and I hadn’t just found it yet. I found the whole thing, Mr. Money Mustache and all these guys in 2016, and I started my own blog.
But I had already been working on my own plans basically since I got my first job, which was in 2000. And I realized that you have the challenges of accumulating money, which in some way, it’s mostly a psychological challenge. You just have to stick to the plan and not lose your nerve when the market goes down a little bit.
But the funny thing is that the math behind accumulating assets is actually relatively simple. Mr. Money Mustache has this simple math.
It’s not just simple math. It’s the shockingly simple math of achieving retirement. And what I found was what’s not so shockingly simple is then the withdrawal math.
What makes accumulation relatively simple is that you have two forces that go in the same direction. You put money into your account, and then on average, at least you should have positive growth in your account. So they both pull in the same direction.
If sometimes you have a little bit of return volatility, returns are low. You still put money in. You use dollar cost averaging.
But I realized that on the way out, when you start withdrawing money, then the two forces pull in opposite directions. You hope that your portfolio starts adding some money from capital gains and interest and dividends, and you hope that you don’t run out of money before you die. And that is mathematically and financially a much more complicated problem.
And I thought, “Well, before I pull the plug,” and again, I pulled my plug 20 years early. So I basically made a decision that impacted 20 times annual salaries.
You think that buying a house is your biggest financial decision in your life, which is maybe 3x your annual salary. In the Bay Area, it’s more. But for most people, it’s somewhere between 2x-5x your annual income.
And that is your most impactful, most extraordinary financial decision you will ever make in your life. But retirement, now we’re talking about 20x.
So you want to do your math before you make such a wide-ranging decision in every aspect, socially, psychologically, but also financially.
So I thought, “Well, there has to be some guidance on the internet.” So I started searching and I found some information. I found the “Trinity study” and Bill Bengen’s work.
And then also some of the things that the other FIRE blogs were writing about the safe withdrawal math. But it was usually treated as something of an afterthought.
While the biggest challenge is getting there, and then, by the way, once you’re retired, you just wave your hands and you just wing it basically. There’s the 4% rule and off you go.
And I thought, “Whoa. Wait a second.”
It’s actually the accumulation part that was the mathematically easy part. Withdrawing money and ensuring that you don’t run out of money, that’s the complicated part.
Long story short, it occurred to me that nobody else had done it in a way that I would like it done, so I had to do it myself. And that’s how I ended up writing a lot on my blog about safe withdrawal method.
By the way, I write about other topics too. It’s not the only topic, but probably about 50% or so of what I write on the blog is about the safe withdrawal method and then some other stuff too.
But I would definitely say that’s a bit of the claim to fame of my blog that I think about safe withdrawal method and safe withdrawal strategies in a way that is scientifically sound and well thought through.
Andrew Chen 08:20
Before we jump into the details of safe withdrawal rates and sequence of returns risk, just to level set with our listeners, you’ve alluded to this already in a few places. In retirement finance, there’s this concept called the 4% rule, which you alluded to.
It’s a heuristic that was formulated in the ‘90s by a financial planner named Bill Bengen, which you alluded to, later popularized in an academic paper known as the “Trinity study” that came out of Trinity University in Texas from a couple of finance professors there.
What the 4% rule essentially says is that on a typical 30-year horizon, if you as a retiree withdraw 4% per year from your portfolio, you can essentially safely retire without having to worry about outliving your money. And that works greater than 95% of the time historically. So 4% is this heuristic of a safe withdrawal rate.
Now, there’s been a lot of debate and even updates from the authors themselves to the theory behind the 4% rule in the years since it was first published. And on your blog, you talk in various places about how the 4% rule has a lot of problems.
What are the biggest issues with the 4% rule that FIRE adherents should be thinking about?
Karsten Jeske 09:32
I actually wrote a blog post on that: “Ten Things the ‘Makers’ of the 4% Rule Don’t Want You to Know,” which is a non-mathematical summary post. I think it’s Part 26 of my series.
The biggest problem is that you cannot use a one size fits all rule for everybody. I think it’s a pretty good guide post. If you are still more than five years away from retirement and you want to set something like a target, absolutely, do the 25x annual expenses, which is equivalent to the 4% rule.
It’s a little bit like, say, you want to drive from San Francisco to the East Coast, and how do you even start driving? I guess you would start to drive on I-80 to the east, but then at some point, you have to make up your mind where exactly you’re going. Are you going to Boston, to New York City, to Washington, D.C., or to Miami?
The closer you get to that target, the more you have to ignore the simple one size fits all rules, and the more you have to start customizing it to your personal situation and then also to market fundamentals.
By the way, I’m not even trying to ding the researchers in the “Trinity study.” The “Trinity study” researchers had to make up an academic sample study, and they had to say, “A retiree has a 30-year horizon, a flat spending profile.”
If the “Trinity study” had started with “We’re going to start publishing a paper and it’s a 37-year horizon. This retiree has a non-flat spending profile, has a crazy-looking curve that goes up and then it goes down, and then it levels off and then it goes down more,” then people would have said, “Do all your results come from just using these crazy assumptions?”
So you had to come up with the most generic assumptions possible: 30-year horizon. You look at different asset allocations from zero/100 to 100/zero. A flat spending profile.
But that’s not how most people in the real world are going to operate.
If you are a traditional retiree, you could argue that you start Social Security and your withdrawal plan at the same time, you get $30,000 Social Security, and then you ask yourself, “How much do I need to fund the gap between that $30,000 Social Security income and my spending needs?”
But for example, for early retirees, I retired at age 44, so that means I can start drawing on Social Security at age 62. That’s a few years in retirement.
Or I could wait ideally until age 70. And then at age 70, I would actually get a pretty meaningful Social Security benefit.
That means even if my spending profile is flat, it means my withdrawals are no longer flat. I will start withdrawing more initially, and then I can reduce my withdrawals one for one when Social Security kicks in. I also get a small company pension at age 55.
So there’s so many idiosyncratic factors in my personal situation that, just based off of these idiosyncrasies, some people start retirement at age 30. They all have to wait until age 70 to draw Social Security. That’s so far into retirement that you can almost ignore that from a computational perspective.
But I’ve done a lot of safe withdrawal rate analyses of people. They retire in their mid-50s and they are getting company pensions and the spouse gets a government pension and they get Social Security.
I told them, “You’d be crazy to withdraw 4%. You can withdraw a lot more.”
Because you have this two-faced problem where initially you can withdraw probably 7-8%, then your benefits kick in, and then you can reduce your withdrawals.
And then other people, if you retire at age 28, you probably don’t want to be all the way up at 7% or 8%. You probably don’t even want to be at 4%. So there’s such a wide range of withdrawal percentages just based off of the idiosyncratic differences in people.
And then on top of that (and this is probably something where I would want to ding a little bit even the “Trinity study”), even if we all accept that we all start with a 30-year horizon of flat spending profile, you still have to account for different market valuations.
You can’t just say every retirement cohort, when they retire, they have the same probability of running out of money. They don’t.
If you had started your retirement in 2000 at the peak of the market, you had very different chances of your retirement going through than somebody who retired in 2003 at the bottom of the market.
So you have to take into account equity valuations. Also, in some way, bond market valuations.
But obviously, equity market, that’s the portion that has the most volatility. So it’s mostly explained by equity valuations.
If you are finding yourself after a very long bull market and you retire then, you probably want to have a lower safe withdrawal rate than somebody who retires relatively early in the bull market or even at the bottom of the bear market.
So it’s both the idiosyncratic differences across different people and then also the market valuation aspect across time that should all play into your safe withdrawal math. Ignore it, but you would ignore it at your own risk.
So these are some of the things that I’m writing about on my blog. And this is why I said the withdrawal math, that’s the complicated part. And that’s why I have 37 or 38 parts in my safe withdrawal rate series already.
Andrew Chen 16:19
Awesome. Well, I definitely want to dive into a lot of the themes that you just identified, including retiring at the top of the bull market versus at the bottom of a bear market, etc.
One thing I wanted to just pick up on is you mentioned your withdrawal math is a function of your own idiosyncratic circumstances: do you have a pension, when can you expect Social Security, as well as corporate valuations in the stock market where you’re retiring.
On the former, some people will get a pension. Others will not. But most people, as long as they’ve worked the requisite number of years, can count on some Social Security benefit.
But you mentioned that the younger you are when you retire, the more that you should ignore that benefit. And I was curious, when does that inflection point occur where you should bake it into your safe withdrawal math versus just ignore it, given that Social Security, the base fund is projected to deplete in something like 2035?
And then something has to happen: Either it’s just tax revenues that continue to fund Social Security or everybody gets a haircut or some other way to raise more money if benefits are not going to take a hit.
And since the farther you go out, the harder it is to predict these things, when is that inflection point for you when early retirees should just disregard relying on Social Security at all?
Karsten Jeske 17:41
There’s two aspects here. One is the mathematical aspect of just purely looking at the time value of future benefits.
If you are one year away from Social Security, you have to bridge only one year. So the time value of that is pretty much one for one.
And then if you are 40 years down the road, it’s almost zero. And in between, it’s exactly a sliding scale.
So you could say that there’s no inflection point where “At this horizon, you have to ignore it. At this horizon, I’m going to factor it in.” No.
You should ideally factor it in all the time. But what you will notice is that Social Security 30 years down the road is probably not going to have such a big impact on your safe withdrawal rate.
Without any benefits, your safe withdrawal rate might be 3.3%. And with Social Security, you may expect something like a 1% point or a 2% point worth of Social Security 30 years down the road.
But guess what. That might only be worth 0.1% or 0.2% or 0.3% in your withdrawal. Again, it’s not trivial because we’re talking about 0.1% of your nest egg.
You have a withdrawal rate of 3.3% and you can raise it to 3.6% in light of Social Security. That’s actually a 10% bump in the money that you can withdraw. It’s only a 0.3% point in terms of your net worth.
My rule is that you factor it in either way. It’s just going to have a very small impact if it’s that many decades in the future.
And I think there is a little bit of a cutoff. So the rule of thumb for you is if you look at the political risk.
Andrew Chen 19:43
Yeah. That’s actually what I referred to.
Karsten Jeske 19:45
Right. When will politicians pull the rug under your feet and start telling people, “Hey, by the way, your Social Security benefits will be lower in the future.”
And I think the rule of thumb is that if you are 55 years old by the time some transition starts taking place, politicians will be very careful about protecting you, because they don’t want to tell people, and certainly they don’t want to tell people at age 65, “Hey, by the way, we’re just going to reduce your benefits next year.”
That’s not going to happen. I don’t think they’re going to take away benefits from existing retirees.
Basically, what they will do is they will say, “At some date in the future, if you are 55 and older, you are protected. And if you are under 55, we’ll start.” And even that has to be on a rolling scale.
You can’t tell people, “If you were born up until December 31st of this year, nothing happens to you. And then next January 1st the next year, you have to take the full force of this reduction of benefits.”
It will be a sliding scale where they say, “For every year you are born after that cutoff date, we’re going to delay your full benefits by another month or another two months.” And then it will be a sliding scale.
And I think you have to make it to age 55. In my case, it’s going to be 2029, so I hope and I cross my fingers that nobody is going to dare touch Social Security by then. And if you’re not much younger than 55 at that point, it’s probably going to be a sliding scale.
My hope is that politicians are going to kick the can down the road for as long as they can until the problem becomes really serious and then they have to do a reform and recalculate benefits. That would be my baseline assumption.
I think there is a small risk, and this will be how they could still hit existing retirees, where they say, “We’re going to change the formula of how we adjust benefits every year.” So there’s a cost of living adjustment of the benefits.
And there could be some way of messing with that formula where they say, “Well, guess what. We are no longer going to do the full adjustment.”
“We’re going to take away 0.1% or 0.2%,” or “We’re going to take a different inflation index to adjust that.” So there could be some ways of messing around with that.
But again, when they do that, it would also be a very slowly phased-in approach where every year you lose a fraction of a percent of your benefits in real terms. And if the problem gets serious enough, that might be in store for us.
Andrew Chen 23:05
Okay, cool. Got it. Let’s shift gears to sequence of returns risk.
Before we dive into the details, what is sequence of returns risk?
Karsten Jeske 23:17
Sequence of return risk, that is the number one reason why people run out of money in retirement.
Because if you look at the average retirement horizon, it’s maybe 30 years, and for example, over the very long term, equities have returned over 6% in real terms.
Bonds, even if right now bonds are yielding only zero in real terms, you take a 75/25 portfolio, even a 60/40 portfolio, your portfolio return of a diversified portfolio over any 30-year horizon, on average, should be high enough to sustain a 30-year retirement with capital depletion.
On average, nobody should run out of money if we just had the average return. The only reason really why people run out of money in retirement is that they had what’s called sequence of return risk. That means they had low returns early on.
And even though you have a pretty decent return over a 30-year horizon, I’ve done some calculations where over a 30-year horizon, you had a 5% average return in your portfolio but you still run out of money. How is that possible with a 4% withdrawal rate?
It means that, early on, your portfolio took such a severe beating and then you started withdrawing from the portfolio during the trough, during the bear market, and you depleted your portfolio so severely that even the strong rally that followed afterwards, even the strong rally that set the average return over that 30-year horizon to 4% or 5%, was not enough to ever rescue your portfolio.
It basically means that after two years, your portfolio is already down by 70% and you still withdraw 4% from that initial amount.
So your actual withdrawal rate, your effective withdrawal rate at that time, what you withdraw as a percentage of your portfolio value is much higher than 4%. It could be 8%. It could be 12%.
And even if the market starts rallying after that, it’s no longer enough to take you over the finish line.
This is something that really has to sink in because we are so used to looking at “buy and hold” investments. A lot of calculators you see on the internet, you see “If you had invested x dollars in the stock market at this year and you let it run for 30 years, how much money would you have?”
If you are a “buy and hold” investor, you don’t have to worry about sequence of return risk. All that matters is what was the cumulative average compound growth rate over that horizon when you kept the money in.
It doesn’t matter if you had zero returns for 30 years and then all the returns came only in the last year, or they came only in Year 17 and you had zero returns everywhere else. It doesn’t matter as long as the total average return was 3% or 4% or 5%. You have the same final value no matter what.
But then once you start making withdrawals or you put in contributions, that’s when the final outcome changes. It’s no longer dependent only on the average return. It also depends on literally the sequence of return.
Did you have high returns first and low returns later, or vice-versa? Or was it just a flat return, the same return every year? And that would then be the average return obviously also over the 30 years.
That is basically the biggest headache for retirees. It’s the sequence of return risk.
Are you unlucky and have low returns early on during your retirement, or are you lucky and the market rallies? And even if you have a bear market late in your retirement, it’s no longer sufficient to wipe out your capital then.
Andrew Chen 27:43
Yeah. I guess this is what you were alluding to earlier about the danger of retiring at the top of a bull market and then it drops in the early years of your retirement.
And you keep withdrawing, eating into principal during those down years because then when the good returns finally show up again, it’s too late. The portfolio has been ravaged so badly. There’s just not enough capital left to restore your savings to outlast you.
And then you called this the poison to your retirement finances on your blog.
Karsten Jeske 28:09
Right. Yeah.
Andrew Chen 28:11
Is it fair to say that during your retirement years, on average, each year’s portfolio returns are more important to your ultimate retirement outcome than the subsequent years’ returns in terms of the predictive power to forecast your ability to outlast your savings?
Karsten Jeske 28:31
There are multiple ways of proving that. For example, you could calculate your safe withdrawal rate over a 30-year window, or you could calculate what is your internal rate of return for a certain fixed withdrawal rate.
You look at the initial capital value, you take regular withdrawals, and you look at the final value, then you calculate an internal rate of return out of that. And then you run a regression on either the internal rate of return or the safe withdrawal rate. You run a regression as explanatory variables.
Maybe you don’t have to use 30 explanatory variables. You do it in buckets of five years. Say you split the 30-year retirement horizon into six five-year windows, and then you run a regression.
Out of these five-year windows, which ones are the windows that have the most explanatory power to your retirement success? Whatever way you want to evaluate that.
And sure enough, the first five-year window has the biggest explanatory power, and then it dies down. And then the last five years has essentially almost zero explanatory power.
I think I ran a calculation like that in Part 15 of my series. I think that’s what you’re alluding to.
And that’s exactly the problem with safe withdrawal rates and safe withdrawal calculations. It’s the initial returns that pose the biggest risk, and the later returns have less impact.
Whereas for a “buy and hold” investor, every five-year window has exactly the same explanatory power on your internal rate of return. And for the retiree, it’s very front-loaded.
And by the way, for somebody who is the opposite of the retiree, you start with zero dollars and you start making contributions to a retirement portfolio, then it’s the other way around. Then the later windows, they are the primary source of the risk of how much money you’re going to have in the end.
Andrew Chen 31:02
Gotcha. The “Trinity study” suggests that if you hit your 25x annual expenses and you withdraw 4% a year roughly, you should be fine more than 95% of the time.
But when you factor in all these nuances, I think in your blog, you call that calculation, the Trinity calculation, an unconditional probability of randomly retiring over the last century. Whereas if you retire conditional upon reaching that savings target but you do it after the long bull market run, it would actually be very risky.
Not 95% of the time you’re successful. It could be much lower.
And I was just curious, just so people can really sink this in. That 95% success rate could flip to order of magnitude what in the extreme cases?
Karsten Jeske 32:00
There are different ways of conditioning it. What the “Trinity study” does is they would just do the completely unconditional probability.
And you could condition it on “Imagine you retire at the peak of a bull market.” So you look at what is the S&P 500, and is it at the highest point up to that point in history.
And you could potentially take your failure probability from 4% to 15%. Not 50%, but 15%.
Instead of looking at relative valuations (I call that relative valuation because relative to the most recent history, we’re at the most expensive S&P 500), you could also look at it in terms of earning, so it’s relative to earnings. You look at the price earnings ratio or the Shiller CAPE ratio.
Same thing. If you condition on the CAPE ratio being above 25 or being above 30, your failure probability could go up from 3-4% to 15-20%, depending on how you slice and dice the data.
And again, the reason for that is if you look at the historical failures of the 4% rule, they all have one thing in common. They happened right around the peak of a long bull market. They happened right at the time when you had very expensive equity valuations.
A CAPE ratio normally above 20, the really bad ones even above 30, and then 2000 was even above 40. These would be the stereotypical cohorts that face a much higher failure probability.
I always use the analogy if you want to calculate the probability of getting into a traffic jam. The way the “Trinity study” calculates the probability of a traffic jam is that they send 24 people off to drive over the Bay Bridge and they do it at the top of the hour, at midnight, 1:00 a.m., 2:00 a.m., 3:00 a.m.
And then people figure out, out of the 24, six got into a traffic jam: three somewhere in the morning rush hour and three in the afternoon rush hour. So it’s only a 25% probability of getting into a traffic jam on the Bay Bridge in San Francisco.
But if you already know in advance, you have to drive every morning at 7:00 a.m. or 8:00 a.m. over the Bay Bridge one way, and in the afternoon at 5:00 p.m. the other way. You will be in a traffic jam 100% of the time.
So forget about the “Trinity study” calculation, 25% probability of a traffic jam. No. Your probability is going to be 100% conditional on having these unpleasant and unattractive starting conditions.
It’s the same thing with the safe withdrawal math. If you already know that you start at the peak of the bull market, you probably want to be a little bit more conservative.
You don’t have that 3% or 4% probability of failure anymore. It might be much higher. I don’t think it’s going to be 100%, but it will be a higher failure probability than what the “Trinity study” sometimes wants to allude to.
Andrew Chen 35:45
That’s a really powerful analogy. I love the traffic jam analogy because it’s very internalizable.
I’d love to understand, given that no one really can predict when the top or bottom of the market is, but there are indicators. I think you mentioned some statistical indicators like the CAPE ratio, etc.
What to your mind is the most useful indicator to use to assess, relatively speaking, how expensive or cheap we are in the market cycle? And what specific threshold do you personally use as a rule of thumb in evaluating this?
Karsten Jeske 36:27
I think the CAPE ratio is a pretty good starting point. And what I like about the CAPE ratio is that it’s accessible. You can check it essentially daily.
Even if Professor Shiller hadn’t updated his Shiller CAPE for two weeks, all you have to do is look at what he uses as the S&P 500 value. You just compare it to today’s S&P 500 value.
You could use his number. It’s probably close enough. If you want to be super precise, you would adjust it from his S&P value to the actual S&P value, and you do a reweighting there.
Andrew Chen 37:05
And just to level set, could you define CAPE Shiller ratio for the audience?
Karsten Jeske 37:09
The CAPE Shiller ratio, it’s a price earnings ratio. You look at how many multiples is the S&P relative to annual earnings.
But instead of looking at just one year earnings and the most recent year earnings, you look at the 10-year rolling average of inflation-adjusted earnings. So that makes it a little bit adjusted.
That’s why it’s called CAPE. The C stands for cyclically adjusted price-to-earnings ratio.
Andrew Chen 37:44
That’s to smooth out volatility.
Karsten Jeske 37:46
Exactly. That’s to smooth out not the volatility in the stock price, but the volatility in earnings, because we could be at the bottom of a recession and earnings are really low.
And you say, “On average, earnings should grow roughly in line with GDP.” It could be growing maybe a little bit faster than GDP because there’s some share buybacks and some technicalities.
And then you say, “If I look over a long enough period, 10 years, I’m going to filter out a lot of the business cycle volatility because 10 years is probably long enough that I filter out the recessions versus expansions.”
There are obviously some weaknesses of the CAPE ratio because it’s backward-looking. First of all, if I look at only one year backwards looking, that’s problematic because that might be very volatile. So let’s look even further back.
So you’re looking in the rearview mirror and you’re trying to drive forward. And in order to have more precision driving forward, we not just look in the rearview mirror. We look in the rearview mirror with binoculars or something.
We’re looking really far back, not just one year, but 10 years. So I can see that some people would be a little bit troubled by that.
Personally, when I used to work in the industry, we would never look at the CAPE ratio. We would look at forward-looking estimates of earnings.
The problem with forward-looking estimates is it seems scientifically more proper to look at forward earnings and not backward-looking earnings. But of course, you also have the estimation error with that. So there’s pros and cons for everything.
What I like about the CAPE ratio is that that is something that is readily accessible to even lay men. And the main reason is it actually has a very strong correlation with forward-looking earnings.
And again, it doesn’t have a very strong correlation with tomorrow’s stock return or next week or next month or next year even. But if you look over long horizons, say 5-10 years, you start seeing a very strong correlation between the CAPE ratio and forward-looking stock returns.
So it is a predictor of future stock returns. And it has to be because there has to be some sort of a mean reversion of the stock market. If the stock market is really expensive this year, it has to adjust over the next 10 years.
Now, I can’t tell you exactly when and how it will adjust, but in order to get back into some sort of an equilibrium, it probably has to return a little bit less than historical averages over the next 10 years.
Whereas if stocks are really cheap and we’re at the bottom of a bear market, then chances are that we might have gone a little bit overboard with the correction and the bear market, and we could even expect a little bit above average returns.
So there are some both theoretical and empirical backing to the CAPE ratio. Anyway, that’s a very long-winded answer. But you could definitely do something like you look at the CAPE ratio and tie that to your safe withdrawal math.
Now, the problem is the CAPE ratio doesn’t really move all that much on a day to day basis. And in the CAPE ratio, there’s not one fixed threshold where you could say, “If it’s below 20, everything is hunky-dory, and then if it’s above 20, hair on fire, panic, run for the hills.”
It’s again a sliding scale. And I don’t want to get too far into the details, but I’ve written on my blog and I’ve looked at what happens if you set your initial withdrawal rate as a function of the CAPE ratio.
Basically, what you want to do is, instead of the CAPE, you do one over the CAPE because the CAPE is a price earnings ratio, so it’s higher when it’s expensive and it’s lower when it’s cheap.
So you do the inverse, and then you actually get something that is expressed as percentages. Then you have an earnings yield.
If the CAPE ratio is 25, then $100 worth of stocks would have given you roughly $4 worth of earnings over a 10-year average horizon.
And it doesn’t mean dividends. It’s earnings. But it’s close enough.
So you could argue that your safe withdrawal rate should probably move as a function of the CAPE ratio.
And again, I don’t want to get into the weeds here, the different formulas and different specifications of how you translate a CAPE ratio of, say, 27 into an actual safe withdrawal rate. But there are some ways of translating the CAPE ratio into a withdrawal rate, and that gives you something really powerful. So now you have a valuation-based safe withdrawal rate.
And then, by the way, the beauty of that approach is that imagine your portfolio goes down by a certain percentage. What do you do then? You look again what is the CAPE ratio at the bottom of the bear market because the CAPE ratio probably went down.
The stock market dropped. The CAPE now looks more attractive, so it’s down. That means one over the CAPE, which is the earnings yield, has gone up.
Now you have a higher earnings yield. So it also gives you some sense of “If I am in the recession and I retired before the recession and I’m at the bottom of the recession, at the bottom of the bear market,” now you can actually increase your safe withdrawal rate, but it’s the safe withdrawal rate of that diminished portfolio.
So you have two effects. Your portfolio is down, but your safe withdrawal rate is up. So they might not perfectly cancel each other out, but it might be a way of having a more flexible withdrawal strategy where you don’t insist on a fixed amount of money every month and every year.
So you are able to roll with the punches a little bit. You’re at the bottom of the recession. You lower your withdrawal amount a little bit because your portfolio is down, but the withdrawal percentage is up.
So it’s still a net loss for you. You withdraw a little bit less. But then eventually, your portfolio is going to recover again.
And the beauty of that is it’s valuation-based. And then on top of that, it cushions a little bit your sequence of return risk because you will withdraw less when your portfolio is down, and then you start scaling up your withdrawals again once your portfolio recovers.
So it’s very effective and also robust. It’s a concrete way and it’s a systematic way of changing your withdrawal amounts to roll with what financial markets throw at you.
Andrew Chen 45:35
Yeah. That’s a really powerful framework.
You mentioned that if you were to use some heuristic like one over CAPE, you will find that ratio to increase at the bottom of the market, but from a diminished portfolio, and vice-versa. It will decrease at the top of the market, but you have an expanded portfolio.
At least at a high level, if possible, how can an investor who is looking at this situation determine how they can monitor and keep on track their portfolio glide path so that they are making wise rebalancing choices and wise drawdown choices?
How can they actually take these offsetting effects of portfolio size versus one over CAPE ratio offsetting each other and translate that to an actual number at the end of the day that they have to go tell their stockbroker, “This year, I want to withdraw x%”?
Is there a high level framework of sequence of steps that the lay person, not the PhD economist, but the lay person can grasp to make a reasonably good choice?
Karsten Jeske 46:46
There are different ways. One would be that CAPE-based withdrawal rate.
I think it’s Part 18 in my safe withdrawal rate. It’s flexibility and the mechanics of the CAPE-based rules.
And it’s actually not exactly one for one. You don’t use exactly one over the CAPE as your safe withdrawal rate. There’s some adjustment factors to that.
I’ve done some simulations to see how that worked out in the past, and it would have worked actually quite beautifully during the 2000 recession, also the 2007 recession. So that would be one thing.
If you have the flexibility to roll with the punches and you say, “I would like to withdraw $70,000 in my first year. But if the market tanks and next year I have to withdraw only $60,000, I can sustain that too. That’s fine.”
A different approach would be, first, when you initially do your safe withdrawal analysis, you ask yourself, “What would have been the worst thing that history would have ever thrown at me in the past?”
Depending on your personal parameters, it’s probably 1929. Everybody knows that.
It was the Great Depression. Very nasty bear market. Very long.
Another very bad episode for investors and retirees would have been if you had retired in the mid to late ‘60s, you had this very mediocre market until the early ‘70s, and then you had basically one recession, one bad event after the other.
I think in 1970 or 1971, you had a recession. In 1973 to 1975, there was a recession and bear market. 1980 and then 1981 to 1982 were more recessions and more market drops.
Either one of the events was not so bad, but the whole sequence of these events was actually so bad. For some parameterizations, that’s actually worse than the Great Depression because it was longer. So from a sequence of return risk perspective, it was even worse than 1929.
Anyways, long story short, you look at what was the worst thing that history had ever thrown at you, and you set that as your safe withdrawal rate.
Now, imagine you find yourself in March of 2020 and the market is down by 33%, and you ask yourself, “Do I have to make adjustments now?”
“Do I have to take my diminished portfolio now and apply that historically worst safe withdrawal rate to that newly diminished portfolio again? Or can I keep withdrawing what I initially withdrew?”
My recommendation was that as long as what happened in 2020 is not worse than what happened in 1929, you should still be safe. From a historical point of view, the 33% drop was not as bad as the 80% plus drop between 1929 and 1932.
In that sense, if you had been really conscious and really conservative, and say you had retired in December of 2019 after the big run-up, and then 2020 throws a little bit of a curveball and the market sinks by 33%, at that point, you can probably say, “Hey, I don’t have to do anything yet.”
Because this is bad. This is obviously worse than all the corrections we have seen in the 2010s, including 2018. We almost had a bear market that was a drop by about 10% top to bottom in the fourth quarter of 2018.
But you could make the case that you don’t have to make any adjustments yet. Because if you look at what would have been the safe withdrawal rate in 1929 or 1930 or whenever the S&P 500 had dropped by 33% during that event, by definition, that initial safe withdrawal rate was still safe because we just determined that it was safe during the 1929-1932 event.
At some point during that event, the market had also dropped by exactly 33%. In hindsight, that initial safe withdrawal rate that you picked was still safe.
So you could make the case that as long as the most recent history is still in line with the previously worst events, and as bad as 2020 has been so far, it’s nowhere close to 1929 and 1930.
In that sense, basically what you want to do is you would determine your initial withdrawal rate, and then every once in a while, you do a check. Is this worse than what we have seen in the past? And if it’s not, you’re probably still going to be safe.
Leave a Reply