Reduce Stock Exposure in Retirement, or Gradually Increase It?

by Michael Kitces and Wade Pfau

Reduce Stock Exposure In Retirement, Or Gradually Increase It? Splash image

For the past 20 years—due to the growing research on safe withdrawal rates, the adoption of Monte Carlo analysis (a method of considering many simulations), and just a difficult period of market returns—there has been an increasing awareness of the importance and impact of market volatility on a retiree’s portfolio.


Dubbed “sequence of return” risk, retirees are cautioned that they must either spend conservatively, buy guarantees (e.g., annuities), or otherwise manage their investments to help mitigate the danger of a sharp downturn in the early years.

One popular way to manage the concern of sequence risk is through so-called “bucket strategies” that break parts of the portfolio into pools of money to handle specific goals or time horizons. For instance, a pool of cash might cover spending for the next three years, an account full of bonds could handle the subsequent five-to-seven years, and equities would only be needed for spending more than a decade away. This “ensures” that no withdrawals will need to occur from the equity allocation if there is an early market decline.

Yet the reality is that strict implementation of such a cash/bonds/equities bucket liquidation strategy is more than just an exercise in mental accounting; it can actually distort the portfolio’s asset allocation and lead to an increasing amount of equity exposure over time. This occurs as fixed-income assets are spent down while equities continue to grow. Recent research shows that despite the contrary nature of the strategy—allowing equity exposure to increase during retirement when conventional wisdom suggests it should decline as a retiree ages—it turns out that a “rising equity glide path” (where the path of equity exposure ‘glides’ higher year after year) actually does improve retirement outcomes. If market returns are bad in the early years, a rising equity glide path ensures that retirees will dollar cost average into markets at cheaper and cheaper valuations. Conversely, if the markets are good, retirees won’t have a lot to worry about in retirement anyway (except, perhaps, how much excess money will be left over at the end of their life).

Of course, the challenge to utilizing a rising equity glide path strategy as a retiree is that many would obviously be concerned about having more equity exposure during their later retirement years. Yet the research shows rising glide paths can be so effective that exposure to stocks can start much lower to begin with—so low, in fact, that the strategy may actually lead to lower average equity exposure throughout retirement, even while obtaining more favorable outcomes. And ironically, it turns out that for those who do want to utilize a rising equity glide path, the best approach might actually be to implement it as a bucket strategy in the first place, where cash and bonds are spent down over time and equities are allowed to run (higher) for a period of years.

Sequence Risk, Bucketing, and Retirement Success

In the 1990s, William Bengen published his first research on safe withdrawal rates. He made the fundamental point that it doesn’t matter what a retiree’s average returns are to determine if their plan will be successful; instead, it’s necessary to look at the sequence in which those returns occur, as there are numerous historical scenarios where the long-term average may have been healthy, but the sequence meant retirees had to withdraw far less. After all, it doesn’t really matter if returns average out in the long run, if ongoing retirement withdrawals plus bad market returns early on mean there’s no money left when the good returns finally arrive.

Accordingly, the conclusion of the safe withdrawal rate research was that retirees should set their spending targets based not on average returns, but on the worst-case scenarios that have occurred in history1. Spending rates are presumed to be “safe” if they are low enough to survive all of those scenarios (whether due to below-average returns throughout retirement, or reasonable returns but an unfavorable sequence).

Of course, in practice, retirees often notice the importance of return sequencing the moment a bear market occurs as they’ve begun down the retirement path. They see their account balance fall precipitously and begin to worry—sometimes excessively so—about whether they need to adjust their spending or change their portfolio (or both). Accordingly, some retirees have adopted various forms of “bucket” strategies over the years, designed to help them get comfortable with their ongoing volatility. For instance, if there’s a “bucket” with three years of cash, retirees may worry less about a short-term market decline. If there’s a second pool of money with intermediate bonds for the subsequent few years, such that equities don’t have to be touched for eight to 10+ years in total, retirees may be further soothed. If the bulk of ongoing expenses can be covered from direct cash flow sources (e.g., Social Security payments, a pension, annuitized income, etc.), it may be even easier to tolerate the market volatility.

However, the reality is that many of these types of bucket liquidation strategies often are more than just an exercise in mental accounting acrobatics; having alternative fixed-income sources to tap in the early years can help, both mentally and financially, to mitigate the danger of sequence risk. After all, the mathematical reality is that if there are no cash flows (i.e., there are no necessary withdrawals) then return sequencing is no longer an issue. In other words, as long as the retiree truly doesn’t have to take any withdrawals from equities in the early years and can allow time for any early retirement bear markets to recover, the primary sequence risk of the portfolio/retirement scenario can be avoided.

Impact of Rising Equity Glide Paths on Retirement Income

While bucket approaches—from just holding cash reserves, to “time-segmenting” pools of money for use over time, to partially annuitizing to reduce income needs in the early years—can be effective at reducing sequence risk, the reality is that such strategies disproportionately spend from fixed-income assets and let equities grow. As a result, bucket approaches will end up, after a number of years, with “distorted” asset allocations: A shrinking fixed-income allocation and a rising percentage of equities. Of course, such outcomes can be averted by systematically replenishing the cash and fixed-income buckets, but separate research has shown that approach can actually result in less retirement income2. The retiree simply ends up dragging too much into cash/low-yield investments as those asset classes are constantly replenished over time. In fact, as it turns out, the best outcomes are the ones where the short-term buckets are used to mitigate sequence risk in the early years, but are not replenished. For instance, if the fixed-income assets are used to partially annuitize the portfolio, with the remainder predominantly or fully invested in equities (and never reallocated back to bonds later), the results actually improve3.

SPECIAL OFFER: Get AAII membership FREE for 30 days!
Get full access to, including our market-beating Model Stock Portfolio, currently outperforming the S&P 500 by 4-to-1. Plus 60 stock screens based on the winning strategies of legendary investors like Warren Start your trial now and get immediate access to our market-beating Model Stock Portfolio (beating the S&P 500 4-to-1) plus 60 stock screens based on the strategies of legendary investors like Warren Buffett and Benjamin Graham. PLUS get unbiased investor education with our award-winning AAII Journal, our comprehensive ETF Guide and more – FREE for 30 days

In fact, as our research on partial annuitization showed4, it turns out that much of the benefits being attributed to that bucketing strategy were not about the benefits of annuitization and having cash flows to avoid early withdrawals at all. Instead, the benefits of the strategy were actually primarily attributable simply to the path that the retiree’s asset allocation follows as a result of spending down fixed-income assets in the early years and letting equity exposure rise. We label this the “rising equity glide path” throughout retirement. In other words, it was less about “not liquidating stocks in down markets” and more about “letting equity exposure grow over time” instead. Accordingly, in our latest research paper (“Reducing Retirement Risk With a Rising Equity Glide Path,” Journal of Financial Planning)5, we delve further into this rising equity glide path effect and find that in reality, increasing equity exposure throughout retirement can actually enhance retirement outcomes. Doing so is so effective that the retiree can start more conservative and may even end up with less average equity exposure over their lifetime.

For instance, Figure 1 shows the probability of success (not outliving savings) of various stock/bond asset allocations based on a 4% withdrawal rate approach. The figure uses historical returns with an average annual compound real (inflation-adjusted) growth rate of 6.5% for stocks and 2.4% for bonds. The starting equity exposure is on the vertical axis of Figure 1, and the ending equity allocation is on the horizontal axis. Thus, for instance, a portfolio that starts at 60% in equities and ends at 60% in equities (the intersection of the 60% row and column) has a 93.2% probability of success. However, a portfolio that starts at 30% in equities and finishes at 70% in equities actually has a higher (95.1%) probability of success, not to mention a lower average equity exposure through retirement (an average of only 50% in equities instead of 60%).


When viewing the magnitude of potential failures, we see a similar effect in Figure 2. Looking at how many years the portfolio lasts at the fifth percentile (i.e., how quickly we ran out in money in the worst one-out-of-20 scenarios), the 30% to 70% glide path portfolio lasts for 30 years, while the constant 60% equity exposure portfolio is depleted in 27.7 years (or faster) in the 5% worst scenarios. (It’s worth noting that historically, a 60% stock/40% bond portfolio has never actually failed, implying some potential for time diversification6 or mean reversion that we did not specifically model in our Monte Carlo analysis for this research.)

Michael Kitces , CFP, CLU, ChFC, RHU, REBC, is a partner and the director of research for Pinnacle Advisory Group and the publisher of The Kitces Report newsletter and the financial planning industry blog Nerd’s Eye View through his website
Wade Pfau , Ph.D., CFA, is a professor of retirement income at The American College and the 2011 recipient of the Journal of Financial Planning’s Montgomery-Warschauer Award. He hosts the Retirement Researcher blog at


Robert Sadofsky from NY posted 6 months ago:

This approach resembles dollar cost averaging with a cushion/bucket of 30% to 50% depending on the assumed bond returns...since we are in a probable rising rate environment, I would stick closer to last table for allocation criteria. This analysis makes a lot of sense to me.

G Flowers from AZ posted 6 months ago:

The analysis makes sense on paper but there is one factor not mentioned at all in this article. Assuming the 20 - 30 year retirement mentioned, the retiree would be 85 to 95 years old. The amount of analysis and management of the stock portfolio suggested, would be beyond the mental capability of most retiree's of that age. Having lived in a retirement (55+) community for over 18 years, I see that decline on a daily basis. In addition, passing on a stock portfolio of that magnitude to the widow, would be problematic.

Charles S from TX posted 6 months ago:

I agree with an emphasis on longevity risk. But I think a better approach is using (increasing) annuities to balance long lives with early deaths. Financial firms are at small risk of changing mortality, but they can invest at a higher level of equities vs bonds with knowledge of future likely disbursements. A big problem is convincing retirees to relinquish assets when some will die early, but their payoff is overcoming longevity risk.

Tim Fenning from PA posted 6 months ago:

It seems that this approach assumes an extremely large cash balance to begin retirement. Otherwise the early year withdrawals will significantly exceed any growth potential of the portfolio, thus resulting in declining balances in the early years.

John Eterno from TX posted 6 months ago:

I like the article, but think the pre-retirement speculation (2nd-to-last paragraph)is wrong (U-shaped curve). I suspect the trend just-before should be similar to the trend just-after: set the near-term buckets aside, and leave the equity buckets to earn more over time.

If the pre-retirement market is down, there's more time to recover; if it's up, you're ahead of the game.

Roberto Plaja from Switzerland posted 6 months ago:

A great piece, very well thought-out and with solid structural considerations.

Richard Thomas from FL posted 6 months ago:

A better solution is to put 100% the portfolio into strong dividend paying stocks. Those that have a long history of paying increasing dividends for decades. Then live off the dividends. It is not hard to build a portfolio of 20 or so quality stocks averaging 4% in dividend payments. Then let the market fluctuate all it wants and just keep raking in the dividend checks. If you need to spend more than 4% of your portfolio each year then you might seriously consider reducing your expenses during retirement.

G Fischer from VA posted 6 months ago:

Having managed client money since 1987. Using the article distribution rate of 4% is not realistic for most people because it requires an additional 25% more capital in the beginning than a 5% distribution let alone 6%.

The second part I disagree with is the idea that you need your investments to generate in free cash flow 100% of the income you spend. Portfolio management is about total return not just yield. Portfolio withdrawal rates are circumstance dependent and have nothing to do with bond or stock yields.

The third point that is hard for me to accept why would you orient an portfolio to towards defending against a 1% or 5% possibility of occurrence. If the black swan event you are worried about occurs with a 5% frequency it could be 19 years before we see that disaster happen. You can already be dead.

In the last four decades I have been doing this only one thing has been true. At some point every year the stock market will be in net negative territory for the year and this year will not behave like last year.

Keep an open mind in your investing there is always opportunity. For starters look where everyone is selling. Top performing assets class last year was Greek Bonds up 110%. If you are patient that is where you find opportunity. Remember to "Buy when there is blood in the streets." It was as true in 1780's as it is today. People have not changed.

Vern Andrews from CA posted 5 months ago:

After 20 years in retirement and working at the same company for 44 years prior to retirement,I believe the U approach to retirement would be a great approach. Using a high equities Mix prior to retirement, moving some of the equities dollars to fixed or dividend paying securities that provide about 80% of your retirement distribution monthly activity and living expenses at a Mix of 70/30, and as retirement continues depending on the market situation using a flexible Mix strategy will provide a less risking Portfolio. My experience for an IRA Portfolio valued at year end 1999 of 1.15 M$ with Distribution starting at age 70-1/2 over the following 13-1/3 distribution years since year 2000 to 3/2013 provided 2.4 M$ or a gain of about 108.7%. The S/P 500 Index over this same period only gained 6.8%. Unfortunate my kids had numerous financial problems in this time period and I provided no interest loans to them for their achievement of financial independence of over 1/2 of that return that was not planned during the development of the retirement plan in the years prior to 1994. Details are available @ website, if interested.

Donald Griffith from CA posted 5 months ago:

The one thing missing, it seems to me, is the absence of fixed income consideration. EG: social security, pensions, real estate income, & pre-purchased annuities. This income, if applied to the program would skew the Fixed Income Bucket and allow more equity investment. This has worked out good for us. During the early days of retirement (1958) the interest rates were much higher and we chose the certain return instead of the "maybe" return accordingly. We own untouched I bonds paying 4-5% as a result. How's that for a great Bucket? Living in wonderful San Diego, and walking the beach most evenings - but we have devoted a lot of of our free time to charity work, when not traveling.

Donald Griffith from CA posted 5 months ago:

Oops, the 58 was age. The year was 78. Sorry!

Harry Rich from OH posted 5 months ago:

Thought provoking article.

I'd like more of the thinking on applying buckets to the rising glide path. At a
1.75% bond return I see the un-replenished cash + bond bucket emptying at around two decades, with 70% stock allocation coming at about 1.5 decades based on a 5.5% stock return. Perhaps with volatility incorporated the expectation is different.

But using average returns more seems necessary to match the 30 year target. Is that going to be easier than just increasing the stock allocation percentage annually?


FinancialDave from WA posted 5 months ago:

G.Fischer >In the last four decades I have been doing this only one thing has been true. At some point every year the stock market will be in net negative territory for the year and this year will not behave like last year.>

One thing I know is that history applies to all years and not just selective ones.

Unfortunately, your market history was shattered in 2013 when the market was up big on Jan 2nd and NEVER looked back from there!

Dave Gilmer from WA posted 5 months ago:

Very interesting article.

What I found most interesting is the fact that if you stay with a 4% withdrawal rate and stay away from 0% equities, your asset allocation has very little to do with your results.

I did not understand why until I read some research by Jim Otar, which he explained in a post on another site - "asset allocation is very effective in reducing the fluctuations of the vertical component (the dollar amount of the total portfolio) but it has (almost) no impact on the horizontal axes (the longevity of the portfolio)."

thanks Jim!

Ruben Rosales from MA posted 5 months ago:

It is difficult to try and justify a new approach to retirement investments after so many years of a set formula in which bonds took the substantial lead. With the improved life span of our generation I am so glad we are being exposed to more realistic investment programs. In his annual shareholders letter, Warren Buffett indicates that what he will leave his wife will follow the following ratio: 90% invested on the Vanguard 500 Index Fund and 10% in short term bonds to be accessed during bear stock cycles. He phrased it slightly differently, but the general idea is correct. I think he has a bold but meaningful approach.
We should be assuming living 35+ years after retirement, if not ourselves for sure for our wives.
Would like to hear your comments.

Michael Henry from OR posted 5 months ago:

Sooooo complicated.....
We would like an annual 4% withdrawal adjusted for inflation over a 30 year retirement.

But, available evidence leads us to believe that stock and bond returns over the next 20 years or so will be below average, so we are advised to only take 3.4% annually.

But, that is only if we are paying no fees. If, for instance, we are paying a 1% expense ratio on our mutual funds, then we can only take 3.4% - 1% = 2.4%.

But, if we have turned all this over to an investment adviser who is taking 1% of assets annually, then we only get to take 1.4% a year.

But, if all that is true, we have to jump off a bridge.

You need to log in as a registered AAII user before commenting.
Create an account

Log In