Safe Withdrawal Rates: What Do We Really Know?

Journal of Financial Planning: October 2012


Recently, the Journal of Financial Planning interviewed prominent experts on safe retirement withdrawal rates to discover where they agree, where they disagree, and where the discussion is going. Participants included William P. Bengen, CFP®; Jonathan T. Guyton, CFP®; and Wade D. Pfau, Ph.D., CFA.

Kitces: Let me start right out the gate: with all the research that’s come out over the years, at this point with everything that you all have taken in, what do you consider to be the safe withdrawal rate?

Bengen: You have to make a lot of qualifications, [whether] it’s for a portfolio of 30 years’ duration that’s tax-deferred, and so on. I’m still using 4.5 percent because I haven’t seen anything that would endanger that yet, although the possibility of that occurring may be in the future…. But based on historical performance, it appears that that 1969 retiree who generated a 4.5 percent number was the unlucky one so far.

Guyton: You know, Michael, I was remembering that the last time I got asked this question was by you in a session that you and Bill and I did in the fall of 2008…. My answer was that if you were looking at a static, always adjusting for inflation, 30-year time frame, that rate was between 4.5 percent and 5.5 percent, depending on the market valuation at the time.

And if you were—or a client was—willing and able to implement more dynamic withdrawal policies that allowed for some adjustments along the way, it was between 5.5 percent and 6.5 percent, depending on the market valuation. Given where we are today, I would be at the low end of both of those ranges, depending on whether or not we were using static or more dynamic withdrawal policies.

Pfau: I’m not really thinking about a safe withdrawal rate any more. It’s more like the lifetime satisfaction maximizing safe withdrawal rate, or something like that, that allows for the chance of wealth depletion or of potentially making cutbacks later on, like Jonathan was saying, with having the rules in place to have a dynamic withdrawal rate and to cut spending if we did have bad market returns….

I think we really have to consider now that the bond yields are so low, and that with bonds being a big component of someone’s retirement portfolio, I wouldn’t be that comfortable going much higher than 3.5 percent at the current time.

Kitces: There’s a great deal of concern over whether all the safe withdrawal rate research is or is not still valid in today’s world and, really, I think that that seems to cover two different environments.

One is, should we have a different safe withdrawal rate from here going forward, and the other is, are we in the middle of some different safe withdrawal rate that maybe started 12 years ago in 2000 and we’re living through now—which might be a problem for past retirees, but not necessarily a problem for current ones? How do you view the current environment in particular? Is that changing the number you’re using or the conversations you’re having?

Bengen: In my conversations with clients I tell them it’s very much up in the air whether the period we’re going through will differ significantly historically from past periods, which included the Great Depression and the world war, and the 60s and 70s. I don’t know. So I advise them, in the face of that uncertainty, be conservative. If you can get by on less than 4.5 percent, let’s do it. But I don’t instruct them to change the 4.5 percent yet. I’m just kind of issuing early warnings, hurricane warnings, and revisit the issue with them every year.

Kitces: Is that more an issue for people who are already several years into retirement, as opposed to the ones who are coming to you today to start retirement anew?

Bengen: I suspect the people who are starting retirement anew are probably going to be in better shape than the people who started back in 2000, because they’ve already had an extended period of very low returns. And if this continues for the rest of the decade, which is entirely possible, that could be a deal breaker for [those who started in 2000]. I think the people starting now, if we have maybe five to seven years of an uncomfortable period, it might be followed by another, secular bull market, which will help them recover to their proper withdrawal rate. But it’s hard to call.

Guyton: It’s funny to think back to the person who retired in 2000, because the person who retired in 2000 had so much more money than any reasonable and even optimistic projections would have given them, had they been done just five years before they retired. I think that’s important to have in mind, that we were in the mother of all overvalued markets, which meant that, even taking the low end of your safe withdrawal rate range, that would have generated an income level that would have been quite a bit above what their planning would have called for just five years sooner.
That being said, it’s this question about “is it different this time” where I think it’s actually a trap for financial planners, because it leads to the assumption that you then have to make a prediction about the long-term future and that you have to make that prediction before you can render advice…. To the degree that we have a more dynamic process that allows us to systematically make adjustments along the way, if the client is okay with that, then essentially we don’t have to make that call.

The Importance of Asset Allocation Decisions and Diversification

Pfau: Bill Bengen had a really excellent article recently looking at the progress report for the 2000 retiree, and I took a look at that, too, and verified what he was saying. But then a person at a discussion board pointed out that the problem is using bonds. The intermediate-term government bonds performed very well since the year 2000, and [this person] said that when you look back at 2000 people entering retirement, they may have been looking at a bit higher stock allocation and they may have been looking at using cash instead of bonds.

And so I ran those same numbers, like in large-cap stocks and with 90-day Treasury bills, and that really wipes out the 2000 retirees—they’re in the worst position than anyone who’s retired thus far. The current withdrawal rates may be above 10 percent, especially if you add in a little bit of a fee aside from the single-index returns. If a 2000 retiree had a healthy dose of bonds in their portfolio, then they may be in shape to have the 4 percent or 4.5 percent still work. If they had used just stocks and cash or divided their fixed income between longer-term bonds and cash, they may not be in as good a shape.

At first I was convinced by Bill’s article that, okay, the 2000 retiree is on course to making it, but I’m not so sure about that anymore. It really depends on what kind of asset allocation they held and, of course, also depends on them staying the course and not panicking and selling their stocks after any of the market drops in the early 2000s or in 2008….

And looking at retirees today, now looking at TIPS, and having negative yield bond TIPS return, we can’t even necessarily be assuming that the real return’s going to be 0 percent on the fixed income; it could be even negative, and that really makes it hard to sustain a 4 percent withdrawal rate over a 30-year period.

Guyton: I think Wade makes a good point about if a retiree had had a fixed-income component that was not reflective of the way all of us who have run studies have modeled fixed incomes—Michael, you are included in that—that basically said other than whatever their cash component was, the vast majority of their bond allocation was intermediate-term Treasuries. And to me, one of the things that matters most is that that particular choice for the fixed-income component is correlated exactly the way we want it to be with equities at exactly the time we wanted it to be that most. In other words, when equities go down, they go up.

And I think that negative correlation at exactly the time we want it most is one of the factors that has allowed withdrawal rates to be at the level they are. If that wasn’t true in the scenarios that put the most stress on the sustainability, they had a different bond portfolio, then that withdrawal rate wouldn’t have worked. To the degree that practitioners have bond portfolios that look substantially different from what the research has modeled, I think they should have less confidence in how their portfolios are going to stand up because their portfolios haven’t really been tested.

Bengen: The excellent point you gentlemen are making probably speaks to the fact that asset allocation really deserves a lot of attention in today’s world, and that we just can’t simply assume that stocks and bonds and cash will represent fixed portions of the portfolio. Perhaps it would make sense, over the next several years, that a portfolio very heavily invested in dividend-paying, high-quality companies might be a better option than a heavy exposure to fixed-income investments, and that might get you back to your 4.5 percent withdrawal rate.

Kitces: At least part of the takeaway to me on this is … if you’re going to rely on the safe withdrawal rate studies that are out there, you better make sure you are sticking to diversified portfolios and the kinds of allocations that were studied in the research. If you’re going to materially deviate from that, you’re stepping into uncharted territory, if not harmful territory.

Bengen: I don’t know. I think it may be necessary to make changes in this environment from what perhaps the portfolio would have had 30 years ago, because obviously, after a 30-year bull market in bonds, we’re at a much different place than we were in the early 80s with respect to bonds. Remember, stocks, relatively speaking to bonds, offer much better returns going forward. I’m in favor of taking a more dynamic view of things, and I think the greater risk is sticking to a fixed asset allocation in this environment, rather than changing it.

Pfau: Another limiting assumption of the safe withdrawal rates that could be helpful for maintaining a higher withdrawal rate is just the limited asset classes—it’s mostly just large cap or, Bill Bengen’s very fond of the small-cap stocks [laughter], and some kind of bond component and cash. But the international diversification could be another avenue to help lower portfolio volatility or boost returns. Well, probably not boost returns, but lower the portfolio volatility, which is an important component of supporting a higher withdrawal rate. So, broad international diversification or including REITs or other asset classes could be another avenue to help support a higher withdrawal rate.

Applying Withdrawal Rates to Real-World Client Situations

Kitces: Is something as simple as a 4 percent withdrawal rate relevant for complex real-world client situations?

Guyton: I think it’s one of the best tools we have, although I’d qualify it by saying that every retiree has a core element of their spending that needs to be maintained and sustained throughout the length of their retirement. We can segment a portion of their portfolio that is designed to fulfill that spending need and, given that it matches (as Bill summarized the research early on) that it’s at least a 30-year distribution period, it’s inflation-adjusted, etc., I think this is the best work that we have on that.

Now, there may be other periods of time, too. We may be advising a client to wait to start taking Social Security, so we have to use a portion of their assets to bridge them to that aspect of things. They may have other chunks of money that they have earmarked for other purposes that would not follow the kinds of safe withdrawal policies that we’re talking about. But for me, just because clients’ spending is more complex than that doesn’t mean that the safe withdrawal research doesn’t apply—far from that.

Pfau: I would agree with that point, that what you get from the safe withdrawal rate is some basic idea about what’s going to be feasible. And then you can have a spreadsheet where you build in the additional complications, like it may be very beneficial to delay Social Security to age 70, and then it’s okay to spend higher than 4 percent until age 70 with the idea you’re going to cut back dramatically after age 70. And other issues, too—if you’re still paying off the mortgage or if you’re going to help with a child’s wedding, and so on.

Bengen: After thinking about it for many years, I’ve concluded that 4.5 percent will kind of treat your portfolio as a cow. It’s an animal and you’re trying to figure out how much milk you want to get out of this cow. [If] it’s just one person and one cow, it’s a fairly straightforward problem to come up with a number.

But if you’re dealing with a client, he really consists of a large number of cows, in the sense that he may have other sources of income besides his portfolio: Social Security, a fixed pension plan, an annuity, an inheritance. And there may be more than one person needing milk in the sense that the expenses that the client has during retirement are not simply increasing with inflation every year. There may be lump-sum expenses like to buy a house, buy an SUV. It can be quite complex.

I, myself, have clients whose initial withdrawal rate has ranged from 2.5 percent to 7 percent…. In practice I find that the actual number in the first year could be vastly different than the 4.5 percent, even though the 4.5 percent rule is not invalidated by that.

Financial Planning Software and Tools

Kitces: What sorts of factors do you use to adapt that 4.5 percent number further?

Bengen: Where the client has any degree of complexity to their retirement situation
at all, I find I must have financial planning software to incorporate all these factors. Even a spreadsheet, and I love to use spreadsheets, just doesn’t handle the complexity with the surety that I’d like. The financial planning software is essential to blending all these elements and coming up with a withdrawal rate and the use of Monte Carlo, and so forth—I just find that essential.

Kitces: One of the criticisms out there is that there really isn’t a safe withdrawal rate software package that models all this stuff into a financial planning package for anybody who wants to. So, how does the average practitioner try to figure this out from software?

Bengen: I use a particular financial planning software, which I’m pretty happy with, which I think handles all the variables and produces a reasonable model for retirement. It doesn’t generate a number or percentage; it’s whatever works, whatever gets the portfolio through retirement without extinction determines what the first year withdrawal rate will be. I’m not totally convinced that the Monte Carlo analysis—it’s not complete yet—all the complexities of the choices that retirees have is not fully addressed. Sometimes, the probability of failure may be overstated, the degree of failure is not represented. But I think we’re getting there.

Guyton: I do not share Bill’s satisfaction with at least the commercially available planning software that’s out there. When we run Monte Carlo simulations, those simulations assume that we literally wind up the thing and let it go, and there’s no ability for any kind of adjustment at any point along the way. And then it tells us how we do, assuming that’s the way we would have handled things.

The thing that I think would be a real advance would be software that allows for some dynamic adjustment within the simulation runs that continue the scenario, assuming that you had applied whatever policies you wanted to assume. And then you could really see that if we intervened in the way in which we intend to, this is the result we would have had. I really find planning software to be relatively useless in terms of actually doing projections on which we base advice.

The Impact of Fees and Taxes

Kitces: We have a couple of different threads here: ways you adjust safe withdrawal rates for clients, how do you model that given the lack of standardized safe withdrawal rate software?

Pfau: Regarding the issue of fees, I like to look at basically calling it a fee and/or portfolio underperformance. And to get at the idea with the safe withdrawal rates, we’re using the index returns … and clients could either underperform those indices or they could outperform them…. I think we should be looking at the case where clients may be underperforming the indices, either because they’re paying fees or because the mutual funds they choose have a high turnover rate that ends up being an implicit fee due to the taxes taken out of the mutual fund, that are going to lead the client to not earn a return that matches the index returns used in the research.
It can really make a big difference, just a 1 percent fee. Not only because of the actual fee amounts paid, but once you start taking the money out of the portfolio to cover the fees or to cover the accrued taxes on capital gains or interest … then you lose the ability to earn the compound interest from that money, and it can have a really big impact on your wealth level after 30 years…. With a 1 percent fee you may have more than a third less wealth after 30 years than you would have without that 1 percent fee.

Kitces: So, for people at that baseline level who are trying to keep it safe, how far down are you reducing the withdrawal rate for fees? What kind of adjustment should someone make? We talked about 3.5 percent, I think, from you, Wade, 4.5 percent from Bill as sort of starting numbers, tax-deferred accounts, even before we account for some of these other pieces. What kind of numbers should someone be thinking about if they do want to work in fees and taxes?

Pfau: I haven’t really looked at the tax issue myself, though I was reviewing in Bill’s 2006 book he looked at taxes and found that like a 20 percent effective tax rate had pretty much the same effect as a 1 percent fee that reduced his SAFEMAX from a 4.15 percent down to around 3.67 percent—an 11 percent, or 12 percent reduction in spending power. So I suppose with taxes you’re going to have to do something like that and then, for the fees issue, I haven’t really thought much about what the real clients may be investing. I just think in terms of my own situation, and what I would suggest to other people would be low-cost index funds.

And then if you can get your fees down to in the neighborhood of 20 basis points, then you don’t have to make that big of an adjustment for the fee part. So in total with fees, with low-cost index funds, and if you have taxes—or part of your wealth is in tax-deferred accounts, part is in taxable accounts—you may need to cut off another 0.5 percent on the withdrawal rate.

Bengen: You know, the whole point of fees is very valid. But, I prefer to treat it as an expense that a client is going to incur during retirement and should be budgeted for. I don’t know if I necessarily want to net the—although we could, theoretically—net the management fees against the returns that are earned in the portfolio, for a couple of reasons.

One, who knows what the client would have earned without the presence of the adviser? Many clients on their own do very poorly. You see the results from returns they earn from mutual funds, which are far below what the theoretical returns from the mutual fund has earned. And the other, there are some clients who want to pay me from sources other than their investment accounts. So, I think you have to be sure that you account for the management fees or the expenses of being paid during retirement, but I don’t necessarily net them against the 4 percent withdrawal rate because I think that can be distorting.

Pfau: I’m thinking more of the fees in terms of what’s being taken out of the mutual fund. And I know that financial planners may be helping greatly to earn their fees 10 times more by getting somebody into a proper asset allocation in the first place, so that they’re actually following the baseline assumptions and the research….

The financial planner can get them to a situation where they can be looking at a 4 percent, 4.5 percent withdrawal rate, and then it’s the matter of what are the fees from the account doing after that. I don’t think you can treat it as just an expense. The mutual fund takes out its fee at the end of the year, and the client may not even be aware of what that fee was. So that fee is gone from the account and it’s not going to earn the interest anymore—there’s this double-whammy effect. Then, not only are you paying the fees, but your wealth is not going to grow as much as it would have, assuming you’re getting positive returns.

Bengen: Negative compounding, yeah?

Kitces: Jon … obviously you’re in the role as practitioner that charges the fee as well as giving advice about withdrawal rates, so how do you bring this together?

Guyton: I think you can look at it a couple ways. You pointed out, Michael, in a newsletter piece, that an asset-based fee of 100 basis points seems to have an effect of reducing the safe withdrawal—whatever withdrawal it is—by about 35 basis points. So that’s one way you could look at it. Bill’s offered another way to consider it.

I think the tax piece, when people ask about how to account for that, I think that’s almost a moot question, because several clients that have the same-size portfolio could have such a different makeup of assets from a tax perspective. And, to some degree, the planner may have had some ability to shape that mix of assets—that tax diversification.
And when people get into retirement, there is so much that can be done with year-on-year tax planning. We focus on the pre-tax withdrawal rate amount, and then we put on our financial planner hats from the standpoint of doing yearly tax planning and tax management. To simply have a number that we reduced the withdrawal rate by, I think, doesn’t really do justice to the different situations that we find.

Spending and Triggers for Change

Kitces: Is a level, inflation-adjusted spending assumption that we have really a reasonable assumption for clients? How do you monitor this stuff on an ongoing basis? When you’re sitting down with clients, what are you reviewing to figure out whether things are going okay?

Guyton: One of the things around this question that I find fascinating is that most of the time the behavioral heuristics that make up our humanness actually prompt us to do exactly the wrong thing at the wrong time, financially. Buy high and sell low is the classic example of this….

But as it comes to spending, people’s inclinations are to pull back at times when a policy-based approach might, in fact, be tending to say to do the same thing, and I think that’s a tailwind that is really important to note.

Regarding the monitoring and the tracking, we go into every client meeting knowing what their withdrawal rate is. In fact, we have trained them to look at this, and that this is one of the most important measurements or metrics that they can track.

To the degree that we see their withdrawal rates trending up and we’re getting to a point where we might recommend that they make a spending adjustment—because we meet with our clients every six months—we will prepare the way for that. In 2008 and early 2009 when we made those recommendations, we primarily got two reactions from clients. One was: Thank you, I wanted to be sure. I always wanted to know if you would tell me to do something like that, because I got the sense that this would be the time. The other reaction was: Is that all? Shouldn’t we be doing something more drastic? To which our response was: well, maybe, but we don’t have to assess that further for another year. So, we really found that that worked pretty well if you are tracking and training your clients to really keep their eyes on that ball.

Kitces: Is there a particular trigger point that you’re looking for? When does the current withdrawal rate lose so much that it is time for an adjustment; how is someone supposed to make that assessment?

Guyton: You have to track the withdrawal rate, of course, to be able to make the assessment, but from my perspective, when it has risen by about 20 percent from where you started, that would be a trigger point.

Kitces: So, if you’re at 4.5 percent, 20 percent of that is 0.9 percent, so you’re adjusting it at 5.4.

Guyton: Yes, although, if you intended to follow that, the research that I’ve been involved in would indicate that you could start at a higher level because it assumes you would make such a change at such a time.

Bengen: I’m just curious. There’s a 20 percent [trigger] based on the starting withdrawal rate? Because over time, it’s not unusual for withdrawal rates to rise with age because of a shorter life expectancy. Your safe withdrawal rate at age 65 might be considerably less than what it is at 75. So is it 20 percent against the original withdrawal rate or against an adjusted withdrawal rate, adjusted for age or life expectancy?

Guyton: Well, like you, we’ve only looked at a lengthy period. We were looking at a 40-year period, so it was really based on that notion of a full distribution period. But you’re right. My sense is, when you believe that 20 years is the remaining time period, the withdrawal rate over that 20-year period would definitely be higher than if you were looking at 30 or 40.

Bengen: I like the 20 percent figure Jonathan came up with as a trigger; that reverberates with me.
I do very similar to what they do. I review every client’s withdrawal rate every year with them, and also with a Monte Carlo analysis—what’s the success rate. And if these start deteriorating in alarming fashion, I’ll raise warning bells. And I’ve been doing that in the last few years, telling clients, we’re in a low-return environment and we don’t know how long we’ll be here, so I think it makes sense to pull in your oars a bit if you can afford to do so without undue pain. Most of them have responded to that positively.

Pfau: There’s a variety of reasons to think that it’s not really optimal to plan for constant inflation-adjusted spending. There is the issue that as people age they tend to spend less on discretionary expenses—they may travel less or go to restaurants less. But then there’s this 800-pound gorilla with health care costs that needs to be planned for in some way. They’re uncertain … and also long-term care things that may rise dramatically with age.

I like a lot the contributions from Jonathan about the dynamic withdrawal rates; that it allows for a higher initial withdrawal rate because you’re ready to make adjustments as you move along and if you get a bad sequence of market returns.

And the basic idea about survival probabilities is: your probability of living to age 95 or 100 is very low, so you have to be incredibly conservative or incredibly averse to outliving your wealth to want to plan in advance to spend the same amount when you’re 65 as when you’re 95, because you have a much lower probability of still being alive at 95. The way I think about it is, well, if I’m lucky enough to live to 95, I think I’d be willing to live a much more meager existence at that time. Well [laughs], when you’re 95, it’s harder to plan what you’re going to do.

Bengen: [laughs] Harder to remember what you’re going to do!

Pfau: And then just one more issue related to that is, the IRS has their minimum required distribution numbers based on their estimates of remaining life expectancy…. A 65-year-old has, let’s say, a 3.2 percent withdrawal rate, a 70-year-old has a 3.7 percent withdrawal rate, an 80-year-old has a 5.4 percent withdrawal rate. It’s like 8.8 percent for a 90-year-old, and so on. And I think that could be a good way—you’d just use those numbers of your remaining assets as you go along in retirement. That could be another way to think about your planned withdrawals, and you can adjust those, too. Those basically assume you have a 0 percent real return on your wealth. You can make those numbers more aggressive and withdraw more than that if you’re thinking that you can earn higher than a 0 percent real return and so forth.

But that could be another way to smooth out your lifetime spending…. I think there’s a lot of reasons why we should be thinking beyond just always assuming constant inflation-adjusted withdrawals, and looking at some of the other possible ways to push that.

A Role for Annuities?

Kitces: Is there a role for annuities around the safe withdrawal rate discussion?

Bengen: I definitely think there’s a role for them. Particularly in this environment where the investment risk is high. If you want to transfer some of the investment risks to a third party—gosh knows where they’re going to get the returns from to pay their annuity payments in the future, but many insurance companies are going to stay solvent—it does make sense. And I have been having more discussions with my clients about that than ever in the past—particularly clients who I think are marginal in terms of the withdrawal rates.

The resistance point, of course, is basically transferring part of your wealth to an insurance company that you might have been planning to give to heirs. But if it comes down to putting food on the table when you’re 80 years old, I think the choice is pretty clear—and taking stress off the investment portfolio.

Guyton: I think Bill makes an important point when he quips about Lord knows where the insurance company’s going to get the return from, because they are handling that money in the very same environment that our clients are investing.

My concern is that when you look at the scenarios that could put so much stress on the safe withdrawal portfolios that we’ve been talking about, and literally so much stress that it could sink that ship, I just worry that it would have the very same effect on the annuitization income that a client relied on. Also, in the environment that we’re in right now, those annuitization rates are so markedly different than they were even five years ago, certainly 10 years ago, that I’m just reluctant to lock those in, given where we see them.

Kitces: Yeah, one of the themes I’ve written about on my blog is the unfortunate correlation between the scenarios that can cause safe withdrawal rates to fail and the scenarios that can cause insurance companies to fail.

Guyton: It’s easy to talk about guarantees as though all guarantees are the same, just because you make one or you have one. And it’s not the same guarantee as, let’s say, the guarantee, for whatever it’s worth, that’s behind a Social Security payment.

Pfau: I do think that annuitization can play a role as well. Moshe Milevsky came up with the idea of product allocation and, basically, retirement income strategies about deciding how much you’re going to allocate to what he calls systematic withdrawals—that’s basically like the safe withdrawal rate area, the withdrawing from your wealth—how much to allocate to immediate annuities that are either nominal annuities, fixed amounts, or adjusted for inflation, and then how much to allocate to the variable annuities that have the guarantee riders to them. I think that’s a useful way to think about things.

Paula Hogan had the article in the June Journal of Financial Planning about the life-cycle finance and how it’s very important that with your essential needs you want to have them funded by reliable—you were just discussing that guarantees may not be the true guarantees—but kind of guaranteed income sources that may be exposed a bit to credit risk, but that are not really relying on market performance….

I don’t know a lot about how the insurance companies do things, but I’ve been hearing from actuaries who have some experience that the concerns about the blow-up of insurance companies are a bit overdone. That basically they’re building bond ladders, and they’re heavily regulated. It’s not like they’re mostly relying on the stock market and that they’re going to collapse and not be able to pay their beneficiaries.

And what the annuitization does is that it provides the mortality credits. So when you’re planning your own systematic withdrawals, you have to plan for that 30-year horizon or whatever the case may be, because you have to plan for a very long life. But the insurance companies, by having many customers, can pool that risk and pay you at a rate for something much closer to your life expectancy. So the annuity company can pay you like you’re going to live 20 years and then if you only live one year, those assets are transferred to the people who end up living 40 years. But that way they’re able to pay you something more than you necessarily can achieve on your own through systematic withdrawals.

With those things considered, yeah, I do think that we have to treat a role for annuitization, at least as a part of this product allocation, [as] part of this income strategy for retirement to ensure that at least some of your basic needs are going to be covered.

The Next Horizon

Kitces: I’d like to wrap up just talking a little bit about what all of you would foresee as the next horizon around safe withdrawal rates. What’s the next step?

Bengen: I think advisers are going to have to address the asset allocation issue.

Kitces: And what’s that issue, that they may need to be more dynamic with it?

Bengen: Yeah, I think so. I think that bonds are probably in the final stages of a huge bull market, which may go on for several more years, we don’t know. But at some point, that will reverse, and bonds are going to become more dangerous places to place investments.

I could see in five years that even some of my older clients might have portfolios consisting of 75 percent or 80 percent high-quality, dividend-paying equity companies, and it may be 20 percent cash, and perhaps no bonds at all to speak of—which is somewhat different than the asset allocation I’ve advocated in past research. But, this is where it’s going, I think. I think you have to be prepared to deal with something radically different.

Guyton: I think one of the big items, and again, Bill, you touched on it, was the impact of a dynamic approach to the various things that you have to do when you’re taking money out of a portfolio—the asset allocation, the withdrawal amount itself. There’s been work in the last 10 years that has isolated those. I’ve done some things around, what if you’re dynamic in the withdrawal amount. Michael, you and Wade have done some work around, what happens if your asset allocation is sensitive to the market valuation level.

We’ve not yet had a study that puts the two of those together and explores whether the improvements that each brings separately are additive or to what degree they complement each other. I would love to see something in that regard.
And I also think that we could do with taking a closer look at, to what degree the value of the bonds in the portfolio is related to the yield they produce and to what degree their value is also in that they rise in value at the very time we want to be keeping our hands off the equities. It matters less what their real yield is, and it matters more that we have a substantial part of our portfolio that is negatively correlated with equities at exactly the time that we most want it to.

Pfau: I love the concept of SAFEMAX that Bill Bengen developed, and that’s just the worst case scenario in history for a sustainable withdrawal rate. And a few years after that the Trinity study came along and defined the portfolio success rates and the natural corollary, the failure rates—what’s the probability that the portfolio runs out of wealth—and I’ve never really liked that concept. I think we basically have to forget about failure rates, because failure rates don’t account for when the failure happens or what the implications of failure are.

I think the future of safe withdrawal rates is about incorporating more clearly the client goals. Do they have a minimum spending level in mind that they think it’s going to be very bad to fall below that level? Do they have a lifestyle goal in mind, basically their desired spending level? So they’ve got the desired spending level, they’re going to be willing to make some cuts in a very bad situation that could switch to a minimum acceptable spending level, and then they may have a maximum spending level—that’s if things work out good.

And then it becomes an issue of how do we best meet those goals; how do we make sure that there’s a very low chance they fall below the minimum level while also giving as much chance as possible to meeting that desired level? And how can we do that by combining systematic withdrawals and those that Bill and Jonathan made very good points about—the dynamic strategies for the withdrawals and for asset allocations. How can we combine that side along with the product allocation? What role can annuitization play and variable annuities, and so forth?
How do we put together a package, a retirement income strategy, that’s going to best meet the client’s lifetime spending goals—accounting for the fact that there’s this diminishing enjoyment as we spend more, that there are these basic needs you have to meet, [and] there are desired lifestyle goals that you want to meet. How can we best match that?

And then I think we move away from the idea of a safe withdrawal rate. You may spend more with the idea you may run out later, but that’s going to be okay because you have other resources to fall back on. If there are bad market returns you’ve got Social Security, you may have a company pension, and so on and so forth. I think that’s the way the research is going, and that basically, safe withdrawal rates is an important issue that’s just one piece of a much bigger puzzle for a retirement income strategy.

Michael E. Kitces, CFP®, CLU, ChFC, RHU, REBC, is a partner and research director at Pinnacle Advisory Group in Columbia, Maryland. He publishes The Kitces Report newsletter and Nerd’s Eye View blog through his website Kitces is practitioner editor of the Journal of Financial Planning.

Roundtable Participants

William P. Bengen, CFP®, is an investment manager and financial adviser based in Chula Vista and La Quinta, California. His seminal article “Determining Withdrawal Rates Using Historical Data,” published in the October 1994 Journal of Financial Planning, laid the groundwork for the “4 percent rule” and the “SAFEMAX” concept. His subsequent articles and book, Conserving Client Portfolios During Retirement (2006 FPA Press), inspired numerous researchers to build on his work.

Jonathan T. Guyton, CFP®, principal of Cornerstone Wealth Advisors Inc. in Edina, Minnesota, writes a retirement column for the Journal and is a former winner of the Journal’s Call for Papers competition. Guyton’s 2004 paper, “Decision Rules and Portfolio Management for Retirees: Is the ‘Safe’ Initial Withdrawal Rate Too Safe?” published in the Journal, proposed that safe initial withdrawal rates could be higher than previously supposed, depending on the retirement investor’s allocation to equities. In 2006, Guyton and co-author William J. Klinger introduced the capital preservation rule and prosperity rule as “financial guardrails.”

Wade D. Pfau, Ph.D., CFA, published the contribution “Safe Savings Rates: A New Approach to Retirement Planning over the Life Cycle” in the May 2011 Journal, which won him the Journal’s inaugural Montgomery-Warschauer Award. That paper challenged much of the research on safe retirement withdrawal rates, proposing instead a system that treats the accumulation and decumulation phases as an integrated whole. A frequent contributor to the Journal and other scholarly publications, he also explores the efficacy of the 4 percent “safe withdrawal rate” outside the United States.

Retirement Savings and Income Planning