Christine Benz and John Rekenthaler: How Much Can You Safely Spend in Retirement?

Listen Now: Listen and subscribe to Morningstar's The Long View from your mobile device: Apple Podcasts | Spotify | Google Play | Stitcher

On this week's special episode, we'll be chatting about a recent study on retirement withdrawal rates that Christine and I authored with our colleague John Rekenthaler, who joins us for this conversation.

As listeners probably know, this is an especially challenging time for retirees trying to figure out how much they can safely withdraw in retirement given lofty stock valuations, paltry bond yields, and uncertain inflation. With that in mind, the study assessed potential retirement withdrawal rates, projecting what spending rate would be successful over an assumed 30-year retirement horizon after taking the market outlook into consideration. In this episode, I'll be asking Christine and John about the study and key takeaways, including the finding that withdrawal rates will likely have to be lower going forward than they've been in the past. For reference, you can find a link to the study in the show notes to this episode.

Morningstar Retirement Research

"What's a Safe Retirement Spending Rate for the Decades Ahead?" by Christine Benz and John Rekenthaler, Morningstar.com, Nov. 11, 2021.

"The State of Retirement Income: Safe Withdrawal Rates," by Christine Benz, Jeffrey Ptak, and John Rekenthaler.

Other Retirement Research Referenced

"Determining Withdrawal Rates Using Historical Data," by William P. Bengen, Journal of Financial Planning, October 1994.

"Decision Rules and Maximum Initial Withdrawal Rates," by Jonathan Guyton and William Klinger, Journal of Financial Planning, March 1, 2006.

"Asset Valuations and Safe Portfolio Withdrawal Rates," by David Blanchett, Michael S. Finke, and Wade D. Pfau, July 2013.

"Exploring the Retirement Consumption Puzzle," by David Blanchett, Journal of Financial Planning, May 2014.

"Experts Forecast Stock and Bond Returns: 2021 Edition," by Christine Benz, Morningstar.com, Jan. 20, 2021.

"Bill Bengen: Revisiting Safe Withdrawal Rates," The Long View podcast, Morningstar.com, Dec. 14, 2021.

"Wade Pfau: The 4% Rule Is No Longer Safe," The Long View podcast, Morningstar.com, April 29, 2020.

"Michael Finke: Here's What Makes Retirees Happy," The Long View podcast, Morningstar.com, Oct. 2, 2019.

"Jonathan Guyton: What the Crisis Means for Retirement Planning," The Long View podcast, Morningstar.com, June 17, 2020.

"David Blanchett: If You're Retiring Now, You're in a Pretty Rough Spot," The Long View podcast, Morningstar.com, Sept. 18, 2019.

10-Year Breakeven Inflation Rate, Federal Reserve Bank of St. Louis

Transcript

Jeff Ptak: Hi, and welcome to The Long View. I'm Jeff Ptak, chief ratings officer for Morningstar Research Services.

On this week's special episode, we'll be chatting about a recent study on retirement withdrawal rates that Christine and I authored with our colleague John Rekenthaler, who joins us for this conversation.

As listeners probably know, this is an especially challenging time for retirees trying to figure out how much they can safely withdraw in retirement given lofty stock valuations, paltry bond yields, and uncertain inflation. With that in mind, the study assessed potential retirement withdrawal rates, projecting what spending rate would be successful over an assumed 30-year retirement horizon after taking the market outlook into consideration. In this episode, I'll be asking Christine and John about the study and key takeaways, including the finding that withdrawal rates will likely have to be lower going forward than they've been in the past. For reference, you can find a link to the study in the show notes to this episode.

John, welcome back to The Long View, and Christine, thanks for letting me turn the tables and pose questions to you and John.

John Rekenthaler: Glad to be here. And today is indeed a long view.

Christine Benz: Thanks, Jeff.

Ptak: So, why don't we start out big picture, and Christine, I'll turn to you for this first question. What makes deciding how much can be safely withdrawn in retirement such a difficult problem?

Benz: I think it's arguably the most difficult problem in financial planning, because almost every single variable is an unknown. So, there are all kinds of known unknowns. You have unknowable market returns over whatever period the person happens to retire into, so that's stock and bond returns; unknowable inflation, which I think we're all thinking about as a bigger force to be reckoned with right now. The retiree may not know the trajectory of his or her own spending. There may be healthcare expenses or long-term-care expenses later in life. And finally, and probably the biggest unknown of all is just the duration of the spending is unknown. So, we don't know how long we will live. And that is the fundamental challenge of this issue of arriving at what is a safe withdrawal rate.

Ptak: And maybe just so we orient some of our listeners who are less familiar with the research that's been done in this space, what are some of the cornerstone pieces of research that have really influenced thinking on how to manage withdrawals and retirement spending in an orderly way? What would you say?

Rekenthaler: Well, the signature piece was written in 1994 by a financial planner named Bill Bengen called "Estimated Withdrawal Rates Using Historical Data." As Christine mentioned, there are a lot of known unknowns, I think, was the phrase. And so, Bill fixed a few of those. He said, "Let's make the time period 30 years, let's make spending plan fixed but adjusted for inflation," and so forth. And if we make these some key assumptions, and we need to understand that those assumptions exist. But still, we can learn more about this problem, right? We can make it something we can actually research as opposed to throwing our hands up and saying, "Well, there are four different moving parts, who knows what we get." And when Bill did that, he found that over time, starting in 1926 through the early 1990s, when he did this study, that retirees could always safely withdraw at least 4% from their portfolio, assuming that it used a 50% stock, 50% bond and a balanced portfolio. So, that really was the signature piece. That was the stake in the ground. And most of the work that's been done since then, including our own research, has used Bill's paper as a starting point. So, we've adopted many of the same assumptions, or at least we've started with those assumptions, and then maybe we modify them. But that has been the main paper.

Benz: I would say that there are a couple of other papers that I view as really influential in this area. One would be Jonathan Guyton's work on what he calls the guardrails method, and this is some work that he collaborated on with computer scientist William Klinger. But the basic idea is how the retiree might tether his or her portfolio withdrawals to how the portfolio has behaved. And that arrives at when we step back and look at it--and I know we're going to delve into this more in the conversation--but it's a more efficient means of consuming the portfolio. So, I think of that as a really important piece of work.

More recently, I think it was in 2013, there was some work done by Wade Pfau and Michael Finke and our former colleague, David Blanchett, that looked at the 4% baseline but came to the conclusion that given the market environment that prevailed at that point, and that still prevails to this day, their view was that new retirees ought to be more circumspect with respect to their withdrawals; that they need to be cautious.

And then, finally, I would say it's not a withdrawal rate piece of research at its heart, but David Blanchett's research on how retirees actually spend, I think is super important in this area. So, knowing that retirees don't necessarily use a fixed real withdrawal pattern, I think is an important finding that when we think about this problem we ought to incorporate.

Ptak: And I would mention, I think every single one of the researchers you mentioned whose work has influenced some of what we've done for this study, have been past guests on The Long View. So, we invite listeners to go back if you're interested, and take a listen to those episodes. John, you took a close look at what withdrawal rates would have been supportable in the past. What were the key takeaways and how did it vary by asset allocation?

Rekenthaler: Well, the key takeaway was, although people tend to think of bonds as being appropriate for retirees, and often, I think a starting point for maybe more novice investors is just to simply think of a bond ladder, where you're locked in and guaranteed income, particularly if it's Treasury bonds. Bond portfolios, or portfolios that did not have equities, consistently underperformed over virtually every time period, balanced portfolios and all equity portfolios. The contest was a lot closer between portfolios that were just pretty much composed solely of stocks than those that were balanced. But we know that bonds alone, they haven't made it. It hasn't worked. And the reason being is because of the inflation adjustment.

On the surface, it looks pretty good. It doesn't look so great on the surface right now because bond yields are so low. But in the past, Treasury yields were often 5%. You say, “Well, 5%, that's great. I get 5%, I'm going to spend 4%, all is good.” But if you're spending 4% each year and you're adjusting for inflation over a 30-year time period, that 4% could easily grow to be 10% of the initial value. So, then, you're getting late in the time period, you're getting 5% and spending 10%. And that's not the right direction to be in. So, that was the main overarching takeaway as well as the lesson on asset allocation.

Ptak: I think you may have touched on this concept earlier, but one convention for talking about withdrawal rates that might not be familiar to people who aren't steeped in this is that 4% doesn't necessarily refer to 4% in perpetuity. Can you explain that, Christine?

Benz: I think it's really important to set the stage when you're talking about this topic. John referenced Bill Bengen's seminal research on withdrawal rates. And what he was assuming was that the retiree would use a fixed real withdrawal pattern. So, to use a simple example, let's say you have $1 million, and your starting withdrawal rate is 4%. You'd get $40,000 in that first year of retirement, and then you'd inflation-adjust that dollar amount thereafter. So, if it's 3% inflation in year two, your withdrawal would be $41,200 and on down the line. So, it's not 4% in perpetuity. That would be way too volatile for most retirees. That's the thinking behind this fixed real withdrawal system that most retirees do want some version of a reliable income stream in retirement similar to what they had when they were working. So, that's sort of the baseline system. And that was the system that we use throughout this research as the baseline system for taking withdrawals.

Ptak: John, back to you. What did you find was the best 30-year period to have retired in history, and what would have been the worst roughly?

Rekenthaler: Well, let's start with the bad news. The data set starts in 1926 for this study. And so, the first time period we review since they run for 30-year periods starting every five years--1930 to 1959, 1935 to 1964, and so on. So, the worst period was the very first period starting in 1930, which is no surprise since that was The Great Depression. Over that time period, we found that the sustainable withdrawal rates were less than 4% no matter what the asset allocation is. They ranged from 3.3% to 3.9%, which actually isn't all that bad, considering you had The Great Depression. You were jumping right into the heart of it.

The best time periods, depends on your definition. 1975 to 2004, that 30-year time period had the highest withdrawal rate that we came up with, with our approach, which is 6.5% for an all-equity portfolio. And the lower end for all bonds was 4.5%. That wasn't bad either. And then, 10 years later from 1985 to 2014, the equity number wasn't as high. It wasn't 6.5, it was 5.8%. But the bond number was even higher at 5.0%. So, no matter what allocation you chose, an investor chose, they got over 5% real withdrawal rate over the next 30 years safely achieved.

I will say the way the last 10 years have been going, there's a possibility that the time period starting in 2010 could be the best. Of course, we need to have half-decent returns for the next 20. But it's certainly off to a bang.

Ptak: Having looked back, why don't we shift our gaze forward, and I'll turn to you for this question, Christine. Why do you think it's a mistake to take past history and run with it with respect to withdrawal rates?

Benz: It's the same reason that we don't just take fund returns and run with them or select the top-performing fund because we know that that past performance isn't predictive of the environment that might prevail over our specific drawdown period. I do think it's worthwhile, and this is really what John, and we, endeavored to do with this research, it's worthwhile to look into the future using some view of equity valuations and our expectation of what bonds might return and what inflation might be to determine what sorts of variables might prevail over the next 25- or 30-year period.

Ptak: How did you go about trying to ascertain a safe withdrawal rate for the future? What kind of assumptions did you make?

Rekenthaler: Well, it's important to recognize that when running through this exercise, we're not using the simple approach of saying, here are the expected returns for the marketplace, here's expected inflation, and so forth, and just plugging these in and assuming that each year is the same, or there's a consistent pattern. What we're doing instead is, we're using an expected return, so number for expected return of various different asset classes that we roll up to form a portfolio, so we'll have a separate expected return for large-company growth stocks and large-company value stocks and long high-quality bonds and international stocks and so forth. We'll assume that people own diversified portfolios, and we'll roll all those up into expected portfolio performances. And then they vary by year randomly.

We simulated 1,000 different trials, or 1,000 different investor experiences, each of which simulated 30 years of during retirement because these are 30-year time periods we're looking at standardly. So, that's 30,000 forecasts that we're shuffling in random to see how retirees would function with effectively the same levels of returns overall, but different orders of the returns. Sometimes in some of the simulations the markets start off very badly, a la The Great Depression over the first few years. Sometimes they start off very well. Those have quite different effects. So, the study tried to incorporate a variety of possible scenarios, not get locked into one particular path or another. And then, we looked for withdrawal rates that had at least a 90% chance of success per the methodology of when we were running these numbers. And what we call the safe withdrawal rate is the highest withdrawal rate that's exceeded in at least 90% of these trials.

It's important to recall, once again, with these trials all the spending is fixed. There aren't any tools being used to improve the results for the investors--simple things like after a couple of really bad years maybe cut back on spending some and so forth. Those are common sense things that people would almost certainly do in real life. But in the trials, they were not done. So, in that sense, this is a conservative estimate. It's also relatively conservative in terms of the asset class forecast, too, which we'll get into.

Ptak: That's where I wanted to go next. We got the capital market forecast, the asset class forecast, from our affiliate Morningstar Investment Management. A question for either of you, maybe you can give a flavor, whether they are optimistic, pessimistic? What would one see if they took a look at the assumptions that we made for stock, bond, and cash returns over the horizon we were examining?

Rekenthaler: Most people would consider the equity forecast to be pessimistic. The stock portfolio has an arithmetic average expected return of 8% per year, which translates into a geometric average--well, those are fancy words--but geometric average is what people normally see when they see quoted total returns, average annualized return. The geometric average is in the high 6s, like 6.8%. And most people think of equities as making more than 6.8% per year--historically, they've done about 10% in a geometric. So, the equity forecast is quite a bit lower than in history.

On the other hand, currently, interest rates are very low, our inflation estimate is quite low at 2.2%. So, the real returns on stocks is still pretty good. It's near 5%. And that's closer. It's still below the historic norm, but it's not as pessimistic as it first seems. And our bond estimates are for about portfolio making in the high 2s, about 2.7%. Well, that doesn't seem so unrealistically low, does it, because you can't get high-quality bonds that are even making 2.2% today. So, that actually assumes that interest rates and bond yields will be rising a little bit, and then, inflation at 2.2%. So, on the lower end, but I think it's all reasonable given where we're at right now with market prices. It would be quite unrealistic to plug in 6% returns for bonds, say, and quite unrealistic to plug in 10% plus returns for stocks, given how low interest rates and bond yields are. That would imply an abnormally high return for stocks relative to those other securities.

Ptak: You mentioned the 2.2% inflation forecast. I'm curious, how are we feeling about the durability of that forecast, given the fact that we're coming off a period of rapidly rising prices? And maybe you can zoom out a little bit and talk about that variable. If inflation ends up being higher than we projected, or 2.2%, what impact would that have on our estimates of sustainable withdrawal rates? Would that sustainable withdrawal rate have to come down markedly for instance?

Rekenthaler: You had two questions. It's hard for me to follow two straight things. But I think the first one was, what do we think about that 2.2% inflation estimate? We were working on this paper since earlier in the year, does that now seem outdated? My response is the bond market certainly doesn't think so. Thirty-year Treasury yields are 1.86%. So, bond investors are already saying, "Maybe inflation won't even be as high as 2.2% over the next 30 years, or if it is that high, then I'm willing to take an outright loss in real terms on my investment." Certainly, the bond market is not expecting anything significantly higher than 2.2% over a 30-year period, otherwise yields would be higher than they are.

Now, what does this imply? Well, there's obviously a large conflict and disagreement between what bond investors are doing and where inflation is now and what the headlines say, and how the average consumer thinks about inflation. I'm just pointing out that I would hesitate to contradict the bond market too strongly. It tends to be more right than wrong over time. And there's a wisdom of the crowd. There's enormous amount of money and thought that goes into setting these consensus prices. So, that's my answer for what do we think about inflation. We'll see. But at this stage, I'm reasonably confident that that's still a reasonable and realistic estimate.

As for your second question, actually, for more equity-heavy portfolios this would cause less damage than people would think, because if inflation rises, stock prices are going to rise too. They may get hurt--they will get hurt a little bit, maybe even more than a little bit. They will get hurt in the short to intermediate term with inflation surprises, and that has a damaging effect for equity prices, and stock multiples tend to come down. We saw that in the ‘70s, for example. But relatively quickly, companies pass along their extra costs to customers, and they grow their earnings along with inflation. Now, bond portfolios would be in more trouble. So, we're not recommending really, or our numbers don't really push people toward portfolios that are overwhelmingly in bonds. But, if so, yes, that would be a problem if our inflation forecast is too sanguine and too optimistic. But I think the news ultimately for the portfolios with 60% equities, say, the numbers would actually hold up pretty well.

Benz: I've been thinking a lot about this inflation piece with respect to our research. Jeff and I had such an interesting conversation with Bill Bengen over the past couple of weeks. And he is truly concerned about inflation with respect to withdrawal rates because his point is that it has an impact on obviously how the retiree spends. So, if inflation is like 5% in year two of retirement, well, then your withdrawal is elevated, and then all subsequent withdrawals are elevated from that level. So, I think that that's the big concern in my mind with inflation, especially if it elevates those early retirement withdrawals, that would necessitate higher withdrawals throughout the drawdown period. So, I think that's the really concerning piece of it, like when the inflation occurs, and also, how persistent it is.

Rekenthaler: That's a fair point. It's related to a lot of issues with regard to retiree spending and retiree investment performance, which is that which happens early in retirement will have a larger effect upon results than that which happens later in retirement, because there's many more years for that to play out.

Benz: Right. It's making me feel like this is a dimension of sequencing risk that doesn't get enough play. We talk a lot about how equity returns might be in the early years of retirees' lifecycle. But what about inflation? It seems like it's important too.

Rekenthaler: Add it to the list for what will go in the second version of this paper, version 2.0.

Ptak: I wanted to turn back to the study and what was one of the more striking conclusions that we came to, which is that, for people who want to take fixed real withdrawals, they're going to need to be more conservative than the 4% withdrawal rate that's been enshrined in a lot of ways. Christine, I'll turn to you for this one. Can you talk about where we think people should plan to spend? What will be the sustainable withdrawal rate over a 30-year horizon if they're taking fixed real withdrawals from the portfolio?

Benz: I'm stealing from John's portion of the research because he did all the heavy lifting here. But the headline number and when subsequent publications wrote about our research, 3.3% was the starting withdrawal for a 50-50 portfolio or really any sort of balanced portfolio over a 30-year horizon, which is obviously quite a bit lower than the 4% guideline. So, that was the headline figure that we came up with, that John came up with in his research.

Ptak: And John, how did that change if we shifted the mix toward a more aggressive equity-heavy asset allocation? Did it get closer to 4%, or did we see otherwise?

Rekenthaler: We saw otherwise, which is unusual. Typically, as I said, adding more equities tends to improve the results or improve the withdrawal rates historically. The problem is, it's an interesting issue. Remember, this is all based on the future projections from Morningstar Investment Management that we talked about. And it's not that those projections are particularly pessimistic about equities. As I said, that equity results are lower than in the past, but so are the projected bond yields and interest rates and so forth. So, the real returns are not bad on equities. The problem is, as equity returns go down, or the level of returns go down, then the volatility associated with equities becomes more important and more damaging. Maybe the simpler way of putting it is, if stocks are averaging 10% or 12% per year with an 18% standard deviation, they're not going to hit you with as many double-digit losses as if they're averaging 6% or 7% a year with an 18% standard deviation.

It hurts stocks more as the level of returns go down. And with bonds and fixed income, yes, the level of returns has also come down, but we don't project them as being particularly more volatile than they were in the past. So, that's why the equity-heavy portfolios don't work as much magic as they did in the past. It's really because the volatility really starts to kick in. As I think anybody who has looked at this problem knows, the real danger when withdrawing from portfolios where more money isn't going into the portfolios, as in the case of a retiree, is spending down after a big loss. And there's just greater likelihood of a big loss when the level of returns is lower, when the level of the stock market returns is not as high as it once was.

There are some advantages, and we can talk about those too, or some reasons why one might want to start with a higher equity portfolio. But a higher equity portfolio that's absolutely locked into a fixed 30-year plan as in the exercise that we ran for this study--yeah, not so much. That's not a winner in this by these assumptions.

Ptak: We're going to talk about flexible withdrawal strategies in a minute, and Christine, I'll turn to you for that. But John, sticking with you for a minute, we did look at some other levers that a retiree could pull in order to try to get that withdrawal rate a little bit higher--every single one of them involves a trade-off. But can you talk about some of those things, some of those variables that we looked at, which are ways for a retiree who perhaps isn't satisfied with 3.3% over a 30-year horizon, where they can get that withdrawal rate a little bit higher?

Rekenthaler: There are a number of things that one can do and get that number up actually fairly dramatically, like, say, to the 4% level. Most of these things don't just give you one tenth or two tenths of extra to get you from 3.3% to 3.5%. They move you up more dramatically. One, not everybody may want to do, which is work longer. But if you work three or four years longer, that's three or four years shorter time during the retirement period, that's three or four years less that the money has to last--although we didn't model it that way. We still modeled a 30-year period. But nevertheless, that is the case, and three or four more years to put more money into the portfolio and three or four more years where the portfolio has a chance to grow. So, that's a pretty powerful effect. Not everybody is in a position to work longer or wants to do so, but definitely recommended.

Another strategy, and this ties in with what Christine was talking about, and David Blanchett's work on spending is, if retirees tend not to continue to keep up their spending along with a growth rate of inflation, that is, in real terms, they spend less over time as they get older--absent healthcare costs, which admittedly could be an issue--they tend to do fewer things and they're buying fewer things and consume fewer things, then maybe after the first 10 or 15 years in retirement, don't try to keep up fully with inflation. If inflation rises by 4%, grow your spending by 2%, and so forth. That has a big effect as well.

Our assumptions are at 90% success rate. Well, if you look at the numbers with a 90% success rate of having enough money over a 30-year period, in many cases, that's very conservative. I mean, by definition, nine times out of 10 in the simulations, you succeed, and for many of the simulations, people end up, after the 30-year period, having more money than they started with. So, potentially, targeting a somewhat lower success rate and then adjusting, this gets into flexible spending, and I think our timing is good, because I imagine you're going to be going there shortly. But there is some logic for starting with a number that is a little bit more aggressive and optimistic in the sense of having, say, an 80% success rate by the projections, with the idea being that most of the time that's going to work out and if it doesn't, you can pull back and lower that initial spending, at which case I will turn to Christine.

Ptak: Let's talk about that next. We devoted an entire section of the study, and rightly so, to flexible spending strategies. Christine, that was a big focus for you. Much of the research about withdrawal rates over the past decade plus has focused on flexible strategies as a means of elevating starting and lifetime withdrawal rates. What's the thesis behind being more variable versus taking fixed real withdrawals, which was what we were discussing earlier?

Benz: The reason that a variable system would tend to support a higher starting withdrawal rate is that the agreement, if you embark on flexible withdrawals, is that you are going to rein in spending if your portfolio drops. And that's really the name of the game with all of the variable strategies that we tested. They all employ downward adjustments after periods of portfolio losses. And then, there are other strategies that aim to enlarge lifetime withdrawals. So, this is the guardrail strategy that I referenced earlier, where the idea is, not only are you taking less in down markets, but you're also taking more and giving yourself a raise in certain up markets. So, those strategies tend to enlarge not just starting withdrawals but also lifetime withdrawals because the retiree is able to take those periodic raises. So, in a 2020-21-type period, those would typically allow for higher paydays.

Ptak: You probably already alluded to them earlier, but maybe you could talk about some of the flexible strategies that we evaluated in the study? Which ones did we look at?

Benz: Just as a means of stage setting, I like thinking about this as sort of a gradation. So, at the far left, you'd have the fixed real withdrawals, the 4%-style guideline that we talked about, and then at the far right, you would have a 4% in perpetuity, so just taking the same percentage year in and year out. And then, in between, you have a variety of strategies that either fall closer to those fixed real withdrawals or are very, very variable. So, we tested four all together.

The two, I would say that we started with, are just simple tweaks to a 4% guideline. So, one would be simply forgoing an inflation adjustment in the year after the portfolio has experienced a loss. So, you just don't take the inflation adjustment in that year, but then you resume business as usual. And then, another strategy that we tested was just taking a 10% downward adjustment in the year after the portfolio had experienced a loss. So, those are just simple tweaks. They don't result in a lot of changes from the baseline 4%-style guideline. And then, we tested the guardrail strategy, which is the strategy that I referenced that Jonathan Guyton and William Klinger came up with. It's a bit more complicated to employ, but the basic idea is that you are able to take upward adjustments in good markets and you have to take downward adjustments in bad ones. And then, we tested a required minimum distribution-style system, which was simply updating the withdrawal annually based on how the portfolio had behaved in the year prior as well as the person's life expectancy. So, with life expectancy declining as the years go by, that tends to elevate the withdrawal a little bit. But as the portfolio balance changes, that can affect the withdrawal accordingly.

Ptak: When we looked at these different approaches one by one, did we find that being variable helped improve starting and lifetime withdrawals?

Benz: They did. All four methods tended to look better from the standpoint of starting and lifetime withdrawals. As you might imagine, a simple tweak like forgoing inflation adjustments after a losing year, that delivered just a modest upward adjustment in terms of starting withdrawals, so whereas our baseline recommendation for the balanced portfolio was 3.3%, that was closer to 3.8%, and guardrails method got a little closer to 5%, initially, in terms of starting withdrawal rates. So, I would say that's an appreciable lift. And the same goes for the lifetime withdrawal rates, especially for those more what we consider efficient withdrawal strategies. So, guardrails over a lifetime, the withdrawal rate was roughly 4.1%, and for the RMD, it was 4.6%. And I think you have to remember there with the RMD method that the idea is that you will consume your portfolio or come close to consuming it. So, that is why it tended to deliver the highest lifetime withdrawal rate. It's kind of an intuitive finding.

Rekenthaler: I think one thing that listeners should keep in mind with these things is that the higher numbers are great. And for those who can afford to be flexible, afford psychologically to be flexible, that's definitely the way to go. But there is quite a bit of uncertainty and variability around these strategies. For example, Christine mentioned with the guardrails method, I think we found a safe starting withdrawal rate of 4.8% and a lifetime withdrawal rate of 4.1%, sound about right, Christine? Somewhere in that area.

Benz: 4.7% starting, 4.1% lifetime.

Rekenthaler: Yes, 4.7%. The model says, well, there's a 90% chance of success if you start with 4.7%, but the 90% mark for what you actually achieve is 4.1%. So, there's quite a bit of ratcheting down, on average, that people need to do. Now, that's not bad, because as we talked about before, people tend to spend more early on anyway. So, if you're going to have a more aggressive or a higher starting point than, say, a finishing point or a middle point, that's a pretty good tactic. But it's just something that people need to keep in mind. The beauty of the starting assumption that Bengen used and the baseline case for this paper is, things are fixed, and people don't have to make at times what will be unpleasant, potentially unpleasant, decisions or changes. The numbers do move up for those who are willing to accept uncertainty.

Ptak: Uncertainty, cash flow variability is definitely one of the trade-offs involved with the flexible strategies that we researched. Christine or John, maybe you can talk about other trade-offs. Christine, one that comes to mind is bequest motive. You mentioned some of the methods boast efficiency, that is, they do a good job of spending the capital that's available to the retiree over the horizon. And so, that would seem like it's not probably something that somebody that has a strong bequest motive would want to necessarily consider because they'd want to leave more at the end. Can you talk about what some of those other trade-offs that these flexible methods entail, what those are?

Benz: Sure. And Jeff, I should say here, I'm completely referencing a portion of the paper that you worked on. But it is a really important point that the ending values tended to vary significantly based on the withdrawal system. So, at the one extreme would be the RMD method, which inherent in it is that you are consuming your portfolio on an ongoing basis. So, don't use that method if your goal is to have a lot of money left over at the end, because we found that that would tend to have the lowest residual amount at the end of the 30-year period. On the other hand, the baseline Bengen method, or the modest tweaks to it, like forgoing inflation adjustments after a losing year, those tended to result in the highest amount of money left over. Same goes with this simple strategy of forgoing or taking 10% less in the year after the portfolio has a losing year. That one delivered actually the highest ending value, median ending value after 30 years of $1.4 million. So, it's an important variable. I can't overstate the cash flow volatility issue as well, that I just think that that's been such an underdiscussed part of this problem that all of the research has converged on this idea. Yes, be variable, vary your withdrawals with how your portfolio in the market has behaved. But what our research looked at is, well, there are trade-offs to that. There are trade-offs in terms of people's quality of life and their plans and what they might want to achieve, and those need to be part of the mix, too.

Ptak: I wanted to go back to how people spend. And I think you referenced some of the research that David Blanchett has done looking at the pattern of spending over a retirement horizon. And it's not necessarily what one might assume if you were to go and model this in some of the more standard conventional ways; it can vary. So, a related issue in addition to incorporating market movements in annual spending is that retirees they're living their own lives and their spending is often fluctuating. So, when researchers have looked at how retirees tend to spend, what are the big patterns that have jumped out?

Benz: We've referenced this at a few points in this conversation. But the research generally shows, and I think a lot of people listening will find that this resonates with their experience with older adults, the research shows that spending does tend to trail down, especially in the middle years of retirement. So, the early years tend to be the go-go years. In fact, a financial advisor did some research that looked at this, and he called the segments of retirement, the go-go years, the slow-go years, and the no-go years. So, the early years are the pent-up demand period in retirement where people are generally in good health. They have a lot of things that they want to do, they might be doing a lot of traveling, maybe helping adult children paying for weddings, whatever that might be. Then that spending tends to trail down in the middle years of retirement to the extent that they're traveling, it might be more in the U.S. versus outside the U.S. They might make changes like selling a second home that they're using less, or something like that, or going down to one car from two. And then, in the later years of retirement--there's been some conflicting research on this--but I would point out David Blanchett's research as being definitive where he does see a little bit of elevation in the later years, and that is largely related to higher healthcare expenditures later in life. And particularly for people who have uninsured long-term-care costs, that expenses might tend to flare up later in life.

So, I think it's important to think about these patterns when thinking about retirement. And certainly, I think retirees can do some work at the front end to help forecast what sorts of changes do they expect in their own lifestyle and how might they expect them to affect their spending. So, if they are someone who plans to maybe sell a second home, well, that might bring some money into the portfolio, which would reduce withdrawals and would also reduce household maintenance costs. So, really spending some time looking at a forecast of how spending might change and also how cash flow sources might change throughout their retirement lifecycle.

Ptak: Before we get to additional portfolio strategies for maximizing lifetime income, back to you, Christine--something that we touched on in the paper, but probably is a candidate for deeper research in the future, is this idea of smart sourcing of withdrawals, which is a topic that you've written about on Morningstar.com and talked about elsewhere. Can you describe what smart sourcing withdrawals--what that is and what it means?

Benz: Sure. So, the idea with this is that the retiree on a year-to-year basis isn't just taking pro rata withdrawals from the portfolio across all of the securities, which is probably not how most retirees do it anyway. The idea is trying to be a little bit tactical in terms of where those withdrawals come from. And so, if you're using simple rebalancing, that would help inform where your withdrawals would come from. So, in a good equity market year, like 2021, for example, you'd be pulling from U.S. growth stocks, most likely, and use those to fund your cash flow needs maybe for the next year or two. But in a down equity market, you would most certainly not touch your depreciated equity holdings. You would let them be, and instead, take your withdrawals from cash or fixed-income holdings.

So, I do think that this has been an underdiscussed portion of the withdrawal problem. And I suppose it lapses into market timing a little bit, but I think it's just commonsense. And I think that retirees who have gotten used to doing rebalancing during their accumulation years, probably can help elevate their lifetime withdrawals if they're a bit thoughtful about where they go for those withdrawals on a year-to-year basis. And this gets a little bit into the bucket strategy that I often write about where I think that sort of strategy where you have your portfolio siloed by asset class/time horizon can help inform where you go for cash on a year-to-year basis. There's not any one single answer as you embark on retirement about how you'll source your withdrawals. You'll take a step back each year and look at what makes sense to trim. And you can simultaneously optimize the portfolio by pulling those withdrawals from the most appreciated asset classes and spending from those.

Ptak: So far, we've focused the discussion on withdrawal rates, but let's talk about some other strategies for lifting retirement income. What sort of steps could people take at the portfolio level to help improve safe withdrawal rates? And John, we'll turn to you on this one, maybe it's tax, maybe it's cost, what do you think?

Rekenthaler: Well, certainly, tax and fund expenses or portfolio expenses in terms of costs as well as government expenses, in terms of tax costs--those are always helpful. A lot of ways of achieving those, particularly on taxes, and that could be and probably has been a whole separate discussion. But I'll step back and just make the broader observation that volatility matters, or more specifically, reducing volatility matters. We tend to think of risk as psychological or at least often the case in investment analysis--and it does apply for those who are accumulating assets for savers before they're in retirement. The danger tends to be if historically--it's not just historic, probably almost certainly going forward, too, even with our more reduced equity forecast--if somebody holds a diversified volatile asset, like equity market portfolio, stock index, they can survive through. If they don't have to touch that money and they are comfortable with the losses, let it ride and they'll do fine over enough time--10, 20, 30 years whatever the time period. The danger is that the volatility causes them to take an action that doesn't work for them.

Well, it's different during retirement. As I mentioned, volatility, it doesn't cause you to take an action because you already have to take that action, you already have to withdraw from the portfolio. So, just the advice that we give routinely to investors, particularly novice investors, saying, don't sell after a market decline. Don't sell your assets on the cheap. Well, that's what you need to do when you're withdrawing from a portfolio. You need to sell; you need to withdraw assets on the cheap. So, any kinds of strategies that reduce portfolio volatility, for example, potentially the so-called low beta strategies, where per some research, lower beta portfolios can have competitive returns with the higher beta portfolios. If you're an accumulator, that doesn't really matter much. If one bounces around a little more than the other, you end up at the same place over time. It's just sort of a matter of taste. But as a retiree, if a low beta portfolio or lower-volatility portfolio can get you to roughly the same place in terms of expected returns, or eventual returns as a higher volatility portfolio, do it. Anything that could lower portfolio volatility while still mostly preserving the returns is going to improve withdrawal rates.

Ptak: What about looking beyond the portfolio at things like Social Security claiming or annuities? Where do those fit into this? Christine, I'll start with you on that question.

Benz: I think that whole set of tactics should be job one really. Actually, before someone even thinks about withdrawal rates, my bias would be to spend some time thinking about, well, how do we make the nonportfolio income as big as it can be? And so, you referenced, Jeff, Social Security-claiming strategies, a lot of the research points to delayed filing being impactful, especially for the higher-income earner in the family. That's a good starting point, looking at some of those strategies. Annuities, I think, could be a fit as well, especially for the many baby boomers who are retiring without the benefits of pensions. I love the idea of trying to get those nonportfolio income sources to at least match the fixed spending going on in the household. So, if you can obtain coverage through those nonportfolio income sources for housing expenses, for utilities, maybe even for food bills, if you can get those baseline expenses covered, it seems like then you have a lot more wiggle room, you have a lot more appetite to employ some of the variable strategies that we explored. It seems like you'd be much more comfortable with the variability if indeed your baseline fixed expenses were covered through nonportfolio income sources.

Ptak: Last question. What's next? You're envisioning this as an annual study. What sort of topics do you plan to delve into for the 2022 version?

Rekenthaler: I like the idea of an annual study, by the way, because a large majority of the next year's paper has already been done. But that should give us room to take on quite a bit more topics, right, Christine and Jeff? So, no excuses. The main thing on my plate is just what Christine talked about, which is annuities--should be able to help push up the safe withdrawal rate for retirees, judicious use of immediate or even deferred annuities. And that's something that I want to add to the model and explore the effects that it has on safe withdrawal rates.

Benz: I am super interested in this topic of fixed versus discretionary expenses, because I do think that it matters a lot whether the spending is fixed or discretionary. So, I think that that's probably been an underdiscussed part of retirement withdrawal rates. Then another thing that we did not touch on at all in the paper, but nonetheless, I think is important is mortgages and the fact that we have a lot of retirees with a lot of housing wealth, and whether reverse mortgages might play a role for some retirees, especially for those with tighter plans. I think that that's an important part of the discussion. If you look at the retirement problem with fresh eyes, you really see, well, you've got a lot of seniors dying with a lot of untapped housing wealth, and could that be more efficiently deployed during their lifetimes? And certainly, a lot of seniors, and talking to seniors, they say, "No, that's sacrosanct; I want my kids to inherit my house." But I think if you were to ask the kids, "Do you want your parent to spend less than he or she could in an effort to give you this money?" I think that many kids would say, "We'd rather our parents maximize their spending, even if it means tapping some of that home equity." So, I think that's been an underexplored part of this discussion to date.

Rekenthaler: All these items fit into a broad group of outside of traditional investments. We began with the investment question of looking at the asset allocations over time and in the future. And that's a pretty traditional investment-specific issue. But for this problem or question of spending in retirement, so many more things are relevant. The use of annuitizations, spending patterns, reverse mortgages potentially, so many other items. You talked about Social Security, delaying Social Security, and separating necessary spending from discretionary spending. So, I think that's really where we're going. We're going to be broadening the analysis and this paper is going to be even larger, which is hard to believe.

Ptak: Well, John, and Christine, it's been a really enlightening conversation. Thanks for sharing your insights, and congrats again on the paper. It was a fun collaboration.

Rekenthaler: Yes, it was.

Benz: It was great to work with both of you.

Rekenthaler: Thank you much.

Ptak: Thanks again.

Benz: Thank you.

Ptak: Thanks for joining us on The Long View. If you could, please take a minute to subscribe to and rate the podcast on Apple, Spotify, or wherever you get your podcasts.

You can follow us on Twitter @Syouth1, which is, S-Y-O-U-T-H and the number 1.

Benz: And @Christine_Benz.

Ptak: George Castady is our engineer for the podcast, and Kari Greczek produces the show notes each week.

Finally, we'd love to get your feedback. If you have a comment or a guest idea, please email us at TheLongView@Morningstar.com. Until next time, thanks for joining us.

(Disclaimer: This recording is for informational purposes only and should not be considered investment advice. Opinions expressed are as of the date of recording. Such opinions are subject to change. The views and opinions of guests on this program are not necessarily those of Morningstar, Inc., and its affiliates. Morningstar and its affiliates are not affiliated with this guest or his or her business affiliates unless otherwise stated. Morningstar does not guarantee the accuracy, or the completeness of the data presented herein. Jeff Ptak is an employee of Morningstar Research Services LLC. Morningstar Research Services is a subsidiary of Morningstar, Inc. and is registered with and governed by the U.S. Securities and Exchange Commission. Morningstar Research Services shall not be responsible for any trading decisions, damages or other losses resulting from or related to the information, data analysis, or opinions, or their use. Past performance is not a guarantee of future results. All investments are subject to investment risk, including possible loss of principal. Individuals should seriously consider if an investment is suitable for them by referencing their own financial position, investment objectives and risk profile before making any investment decision.)

(Disclaimer: This recording is for informational purposes only and should not be considered investment advice. Opinions expressed are as of the date of recording. Such opinions are subject to change. The views and opinions of guests on this program are not necessarily those of Morningstar, Inc. and its affiliates. Morningstar and its affiliates are not affiliated with this guest or his or her business affiliates unless otherwise stated. Morningstar does not guarantee the accuracy, or the completeness of the data presented herein. Jeff Ptak is an employee of Morningstar Research Services LLC. Morningstar Research Services is a subsidiary of Morningstar, Inc. and is registered with and governed by the U.S. Securities and Exchange Commission. Morningstar Research Services shall not be responsible for any trading decisions, damages or other losses resulting from or related to the information, data analysis or opinions or their use. Past performance is not a guarantee of future results. All investments are subject to investment risk, including possible loss of principal. Individuals should seriously consider if an investment is suitable for them by referencing their own financial position, investment objectives and risk profile before making any investment decision.)