US Videos

Why You Shouldn't 'Set and Forget' Your Withdrawals

Christine Benz

Christine Benz: Hi. I am Christine Benz for Withdrawal rates have been a hot topic recently with some recent research suggesting that the old 4% rule for withdrawal rates maybe too high. Joining me to discuss that topic is financial-planning expert, Michael Kitces. Michael, thank you so much for being here.

Michael Kitces: Great to be here.

Benz: I'd like to talk about withdrawal rates, Michael. There has been so much research flying around, and it's a really hot topic among our users. Some recent research has actually suggested that the old 4% rule that some retirees had been using was too high, that maybe they should ratchet it down. I'd like to start by doing a little bit of stage setting. First of all, where did the 4% rule come from and how has it evolved as the years have gone by?

Kitces: It's a great question. So, the whole 4% withdrawal rate has been on a little bit of a strange roller coaster over almost 20 years now since the first research came out. When it was originally done, a planner named Bill Bengen out of California in 1994, his criticism at the time was that everybody was running projections based on markets averaging 10% to 12% a year, running in a straight-line projection, which means no assumed volatility and coming up with these sustainable withdrawals of 6.5%, 7%, 8%. And his objection was, "Well, wait a minute, sometimes markets are little more volatile, bad things happen from time to time, and maybe we should back off of that a little bit and find a lower number." And so his approach was to simply look through history and say, well, how bad does it get when you retire on the eve of the Great Depression or heading into the 1970s? What's the worst-case scenario we can find in history, and let's set our withdrawal rate low enough that even if you live through the worst-case scenario, you will make it through OK. If you live through anything better, you'll either finish with a lot of money leftover or you'll simply ratchet up your spending over time. And when he came out with the research, he was roundly criticized for being ludicrously low.

Benz: Too conservative.

Kitces: How can you withdraw 4% in a world where the S&P 500 compounds 15% a year, which was the environment we were in going through the mid- to late 1990s. And so, ironically, as returns have ratcheted down and down now to the point where we're actually talking about whether the 4% number is too high and not too low, the irony is it was actually built on the same conservative framework all along. We keep moving our expectations based on whatever happens to be going on lately, while the research sort of stays where it is and where it's always been that we start our number at a point that's low enough to defend against bad environments. And if we get good environments, we can always figure out how to deal with too much money at some point down the road.

Read Full Transcript

Benz: So, some recent research, and my colleague David Blanchett was involved in producing some of it, did look at that 4% rule and said, "Gosh, given where bond yields are today and given that bond yields have historically been a pretty good predictor of bond returns, can you really support that 4% rule? Is perhaps a 3% withdrawal rate more sustainable?" Your view is that perhaps in fact the 3% rate is too low, that in fact some people may be able to take their withdrawal rates higher. Let's talk about the things that you look at when you're making such an assertion?

Kitces: Sure. So, it's absolutely true. As I frame it, when we look at something like the 4% rule and where it is, when we're in a great market environment and things are going wonderfully, we're walking along on what is clearly safe ground, maybe there is a cliff somewhere over there, but we're nowhere near it. When we get to the sort of lower-return environment we're in right now, and it certainly is not a pleasant return environment--valuations are a little bit high, if we look at things like Shiller trailing 10-year P/Es and yields are low, that's not a very good combination for income sustainability. But that's not unique. We've been through environments like that in the past. That's actually the whole reason why we ratchet the number down to 4%.

So, as I look at environments like that, I characterize them as instead of walking over here on safe ground, we are now kind of tiptoeing along the edge of a cliff. That doesn't mean we're sure to go off of it; that doesn't mean we're in imminent danger. It simply means, yes, we are a little bit closer to the line than we would have been. If I were doing updated Monte Carlo projections for a client who was retiring late 1920s, and said, "How are you doing half way through the Great Depression?' it wouldn't have looked very good at the time, but they actually made it fine. If I did the same thing in the 1970s, it would have looked a little concerning at the time, but they did OK. And that's not just in sort of the general philosophy of, "Oh, well, markets always eventually get there in the long run," but that literally as market conditions change, that too changes forward expectations. 

So, while we don't necessarily like the returns going forward from here, and we're concerned about things like higher bond rates, which could take a hit to bond prices, that also means we'll then be reinvesting in a higher bond returns in the future. At some point the market may take a hit, but that means going forward we're investing in markets that have better forward returns. And so there is a duality to that effect that means ironically while we look at some kinds of Monte Carlo projections and say, "Well, maybe they underestimate some of the tail risks at the extremes," and we've had a lot of research recently talking about that, my concern over longer-term periods of time is actually the opposite that at some point we fail to account for the fact that there is inherent pricing and valuation in markets. It doesn't always play out in the short term, but it virtually always plays out in the long term. And then simply put if we had some kind of market crash where prices fell 50% from here, the return of stocks going forward from there is better. The return of stocks if we run a Monte Carlo projection says, it's no different buying the S&P at 1500 and a yield of 2% as it is buying the S&P at 750 and a yield of 4%, even though clearly in the real world those are different circumstances.

Benz: So one piece, though, that is a little bit different about the current environment is that that the very low-yield environment that we are in is somewhat unprecedented, correct?

Kitces: So, we've had low-yield environments before. We've had high-valuation environments before. We're in a little bit more of a combination than we usually do. So, when we look at other points through history, usually if valuations are high, the rates are still a little bit higher. The market goes down, valuations improve. Then we cut rates to stimulate a market that just crashed, and so that's sort of the normal cyclicality that plays out. Ironically, it sort of played out over the past 10 years as well, except it just got us from stratospheric valuations that we had in 2000 to merely high ones that we have today, not the most pleasing way to go down that path.

And so, certainly, I still view it as a potential concern environment. If there were one I was going to watch out for, it would be this kind of environment. But I think it's crucial to keep it in context, particularly for someone who is looking at retiring today that to say that we're going to have something worse than 4% in the past that worked through the '70s, that worked through the Great Depression, you're literally talking about environments where the next 10 years have to be a worse Great Depression than the one we had in the '30s, which had ultralow bond yields deflation, and a horrible market crash to actually make things worse. And when we look at that kind of environment, you're literally making a forecast of things like we're at a 1500 today on the S&P 500 and it could be 1600 in 2028 because that's the kind of environment we played through in the Great Depression.

Now, I can't discount that entirely as a possibility, but at the same time I see a lot of clients, and certainly we look at the markets, have some concern about forward returns from here. That's different than saying, "I don't think the S&P is going to do any better than getting from 1500 to 1600 over the next 15 years," which is the kind of returns you're talking about to really break safe withdrawal rates permanently. So, I am a little bit skeptical on that end. I'm actually ironically a little bit more concerned for people who retired in the year 2000 that we may ultimately find the year-2000 retiree doesn't quite work at 4%; they have to be at 3% something because truly we had unprecedented market valuations at that time.

But there is a difference between what happens for retirees today versus what happens to someone who is already 12, almost 13 years into a retirement as of now. The ones who retired 12 years ago I'm a little bit about more worried about than the ones today, and frankly when we look at most retirees who started in 2000, they've already done some moderations to their expenses along the way, just in case, because it has been a really tough decade plus a little.

Benz: Right. So, what I am hoping you can do because I think retirees trying to figure this out are probably pulling their hair out at this point--we are changing things around on them--can you share some best practices for people who are attempting to figure out what is a sustainable withdrawal rate? Are there some things that people should make sure are part of their plans as they attempt to navigate this question?

Kitces: Good question. So, absolutely the starting point is, you do need to be monitoring this. As much as I would like to have supreme blind faith in a 4% framework, I can't quite enter into it that blindly. Certainly I think some level of ongoing monitoring and prudence is just basic and fundamental.

Benz: So, "set it and forget it, take this dollar amount, inflation adjust it," you think you just can't do that.

Kitces: Yeah. And frankly, no one ever has really been doing that. I mean Bengen started that as a framework, because as anyone doing research, you have to start with some framework and assumption, but Bengen himself is a planner. If you go talk to him, not a single client that he is been doing this with in the real world for the past 20 years was ever on a blind "set it and forget it" routine where he never did a review for them going forward. He is planner. So, I think intrinsic in our planning process that maybe he sort of took it for granted that everybody would [monitor their portfolios], and it didn't need saying, but apparently it did.

Certainly monitoring is a starting point and really there are two things that we can look at around monitoring on ongoing basis. The first is simply looking at what is our current withdrawal rate. So, what is our spending now compared with what it was when we started? Now, that number is going to fluctuate a little bit just because markets are volatile and not terrible stable. A 10% move in your portfolio alone can make your withdrawal rate fluctuate around a little, as well. So, we had some great research from financial planner named Jon Guyton in 2004 and follow-up study in 2006 that looked at setting up what he calls guardrails. So, if we start at some withdrawal number, maybe it's 5%, and say, "We're willing to adjust up a little or down a little depending on what happens." We can set guardrails that say we're going to start at 5%. If things are going badly for the portfolio, which means our spending is outpacing our portfolio, our withdrawal rate is rising and we go from 5% to 6% that's concerning. If that happens too quickly, we need to take a spending cut.

Conversely, if we start going the other direction, we start at 5%, but the portfolio is outpacing our spending, we are in a bull market, and our 5% has gone down to 4%, we can take a little bump up in our spending. And so he kind of says, well, "We are going to aim to have spending that sell increases for inflation, but because the portfolio is kind of doing this bumpy thing along the way, we'll set some guardrails that say a lot of fluctuation is OK, [but if there were] a whole lot [of fluctuation] maybe we need to make an adjustment just to be certain we are not wandering too far off track."

Now, your colleague David Blanchett here at Morningstar actually did what I thought was a great follow-up to that research a few years ago and pointed out that perhaps an even better way to do it rather than just updating on current withdrawal percentages is to run updated Monte Carlo analyses and project on that basis. And the big improvement that you get is when you keep rerunning Monte Carlo analyses; A, you can adjust it for current market conditions if you're so inclined; and B, you update where you are in your life expectancy and your mortality.

So, the challenge to sort of the Guyton guardrails, it's great in year two when I go back and say, "Well, I started here and I'm now wandering a little off track from the guardrail." It doesn't work so well when I'm 10 or 15 years in, because frankly my percentage should be a lot higher simply because I'm older and my life expectancy is 10 years shorter now.

Benz: Right.

Kitces: So, running updated projections I think in sort of a Monte Carlo approach lets us keep updating this within the bands of changing mortality and changing market environments. And so it can get us to an approach that says, well, maybe we're not going to set guardrails that say "Cut my spending if my withdrawal rate climbs over 6% and increase it if it falls below 4%." But we might set Monte Carlo probabilities that say things like "If I am climbing above 98%, that's so darn safe, Maybe I can raise my spending a little, and I'll just cut later if it comes back." And if I fall below some threshold, maybe that's 85%, 80%, or 75%--we all have slightly different numbers based on our risk tolerance--then that's a point where I'm concerned that it's careening downward a little bit too much and maybe I need to pull back my spending a little and let my portfolio recover a little. So, we can monitor a little bit more dynamically that way, as well.

Benz: So, the general thrust and obviously some people are using planners, but individual investors who don't have Monte Carlo simulators sitting on their desk, I guess the general thrust of your comment would be when market valuations look high, you want to pull back a little bit on spending, and also as you age you can actually potentially spend a little more. Those are sort of some key takeaways.

Kitces: Correct, with the caveat that the further you ratchet your spending down, the less you necessarily need to cut just because a market is down. The whole point of safe withdrawal rates was how low can we set the spending that even if we do retire in the midst of a catastrophe, we're going to be OK. That's why I've actually written and have been critical of some of the required minimum distribution methods of doing withdrawal rates. While it's great that we are sort of updating for where markets are and where our life expectancy is, if we do that in a real world and we go through an environment like 2008-09--say in 2008 you're living in your retirement home and you enjoy it. In 2009 you update your retirement plan on an RMD method, find out that you have to cut your spending so much because of a market decline, and you sell your house. Two years later the market starts rebounding back and [you think that maybe] you actually could have afforded to stay in your house all along. Maybe you can go to the old people and buy it back from them.

So, we start getting this volatility to our spending that really can materially change people's lifestyles and force them to make permanent decisions for temporary scenarios. So, I am actually less concerned about when we need to cut, particularly if we start at a prudent baseline in the first place, which is what we are trying to get to with a safe withdrawal rate. But simply have a monitoring process to help us understand if it really seems that we're going off the next second Great Depression, we need to acknowledge that there's always some potential that we're going to have to make a change. But keep it in context that in normal bear markets, that's what 4% was made for. If you just wanted to adjust based on current conditions, average returns give you 6% to 7% withdrawal rates. So, we go from 6% to 7% down to 4% in case you do have bear markets. We go from 4% to something lower than 4% if we're forecasting hyperinflations, Great Depressions, things like that. Now, if some of our viewers are in that camp, all bets are off on 4%. But keep it in context of that's the kind of thing it takes to really break 4% safe withdrawal rates.

Benz: Michael, obviously a lot to chew on here.

Kitces: Indeed.

Benz: Thank you so much for the great overview.

Kitces: Absolutely, my pleasure.

Benz: Thanks for watching. I am Christine Benz for