Friday, March 7, 2014

The Downside of Upside

Run those Monte Carlo simulators at brokerage house websites and you will see that it was historically possible (though not probable) to spend 4.5% of your retirement nest egg each year and accumulate a huge amount of wealth over 30 years of retirement with a systematic withdrawal (SW) spending strategy.

We talk about using those stock market gains as a means to perhaps increase our future standard of living with certain retirement strategies. But, what are the probabilities that our portfolio will grow enough to increase spending by a meaningful amount before we're too old to really enjoy it?

The calculations in this post were made using a "safe withdrawals" spreadsheet downloaded from the Retire Early (RE) Homepage, which uses market data from 1871 to 2003 from the Robert Shiller website. All figures are in real (inflation-adjusted) dollars.

Here are the terminal portfolio values (TPVs) from that spreadsheet for a retiree spending 4.5% of his initial $1M nest egg annually for 10, 20 and 30 years with a 60% stock portfolio.
As the table shows, the best outcome was a legacy of $5.279M. 90% of the outcomes were less than $2.738M, but half were less than $835,685.

You can see from the frequency distribution of 30-year TPVs in the table below, that in spite of those big numbers at the top of the columns in the table above, most periods (about 56%) ended up with a decline in portfolio value from the original $1M, or an increase of less than 10% and would not have created significantly more spending. The really big numbers are long shots.
Let's look at ten-year periods of this SW model's performance because money that's going to be spent in less than about 10 years shouldn't be invested in stocks, anyway, and money invested for 20 years doesn't pay off until we reach age 85 at the earliest. Let's further assume that we would want to live at least five years to enjoy the additional spending after earning it.

The following pie charts show the percentage of rolling time periods for ten, twenty and thirty years in which the retiree's portfolio would have grown enough to support additional spending (a 20% or more increase in wealth, shown in green), about the same spending (gray), or a loss of spending (a 20% or greater loss of wealth, shown in red).

So, risking your existing standard of living to improve it by at least 20% over the next ten years of retirement works out for you a little less than half the time. Based on historic market returns, about 22% of the time, instead of improving your standard of living after 10 years, portfolio value would have declined at least 20%, lowering your standard of living.

These pie charts show only part of the bet you're making with SW strategies. They show the impact of risk on your standard of living. Several studies have shown that when you lose that wealth is critical. Losing the bet early in retirement not only impacts your spending, it has a disproportionate effect on the survivability of your portfolio.

Personally, I don't give a rat's hind quarters about increasing my consumption when I am 95. Or 85, for that matter. Having a lot of money at the end of my life to leave to my children is appealing, but betting current spending in hopes of creating even more future spending is quite a different proposition than betting current spending so that I might leave my heirs a bundle.

While there are a number of reasons a lot of money would be nice to have late in life, like paying for long-term care insurance or leaving a legacy, the 20-year and 30-year columns in that first table don't interest me that much when I think about increasing my standard of living.

I think it is unlikely that I will reach even 85. About 41% of males will and, if I'm one of them, I don't expect that a lot more money to spend will make me much happier. How about the big numbers at 95? Perhaps I could finally afford that classic '62 Porsche convertible I always wanted and drive it until the state takes away my license!

Of course, I'll have to be around at age 95 to buy those wheels and a male turning 65 in 2014 has only a 6% chance of reaching that age.

Furthermore, safe withdrawal studies like the one on Retire Early Homepage assume that a retiree spends a constant dollar amount throughout retirement. If we really did increase our spending as we accumulated more along the way, say during our seventies, we wouldn't have accumulated those big TPVs later in retirement.

Some retirement spending strategies trade off secure future income for the chance to improve our standard of living over time if our stock investments perform well. Purchasing a life annuity, for example, locks in a secure stream of future income for life, but with no potential to increase spending. Investing in a stock and bond portfolio and implementing a systematic withdrawals strategy, on the other hand, offers the potential for increasing future spending but provides no minimum “floor” below which spending cannot drop.

How likely are we to both increase our standard of living through equity appreciation after retiring and to be around to enjoy the additional consumption?

In order to achieve both, we need three things:
  • Time for our stocks to grow significantly in value,
  • Time after those stocks grow to enjoy the additional spending they can generate, and
  • A desire for a possible higher standard of living that exceeds our desire to preserve our present standard of living. In other words, a continuing willingness to bet.
Generating more spending capacity in our eighties and nineties probably won't interest most retirees, because at those ages we will have little time remaining to spend the additional money, we will have an increasing interest in not going broke, and because spending typically declines as we age, anyway.
(Also see Blanchett, You'll Spend Less As You Age.)

The most economically efficient way to spend would be to match our income stream with our expected expenditures. You may recall the following chart from Untangling Retirement Strategies: Systematic Withdrawals. Spending from a SW portfolio would match expected retirement spending best in the downside case, when market returns are worse than we had hoped for.
When the market performs well for a retiree (the green line), wealth actually moves opposite the desired income stream, providing the most income when we need it the least.

(Interestingly, a life annuity without an inflation rider also matches this spending stream pretty well. This EBRI study estimated a decline in spending of 2.3% per year  —  about the average rate of inflation.)

The following table, calculated using Vanguard's life expectancy calculator, shows the probability that a person will live 15 more years beginning at age 65, 70, 75 and 80. Remember, that's 10 years for portfolio growth and 5 years to enjoy it. I'll stop at 80 because that would put the retiree at age 95 after the assumed 15 years. The table below shows the probability that a male or female would live 15 more years at various ages and the last column shows the probability that at least one of them will live that long.
There is about a 45% probability that a retiree's portfolio will grow at least 20% over any 10-year period (the first pie chart above) and for a male aged 65, a 63% chance that he will live 15 years to see his portfolio grow and enjoy the additional spending for at least five years. The probability that both of these statistically independent outcomes will occur is about 28%. So, at age 65, a male retiree using this spending strategy has about one chance in four of increasing his standard of living over the next 10 years by investing in equities and living at least five years after that to enjoy it.

A female retiree has a longer life expectancy, so her chances of spending more are a little better at 32%, or about one in three. The odds for at least one surviving spouse (usually the wife) being around to enjoy the additional spending are better still, about 41%.

The following table shows the probabilities that retirees at the four different ages will both grow their portfolio by at least 20% and be around for another five years after that to enjoy the additional consumption.
What this table does not show is our health and spending demand at each phase of retirement. Both decline with age. Extra money to spend at age 75 is better than the same amount of money at 95 both because we typically need less as we age and because at 95 we have less time to spend it.

A recent paper by Wade Pfau and Michael Kitces shows that, statistically, early retirement is a time for low equity exposure to minimize longevity risk, while later retirement is the safest time to take more stock risk. That is probably true if one seeks to maximize lifetime spending and minimize the magnitude of losses that do occur, and if one lives a long time, but that strategy doesn't provide the additional spending earlier in retirement when most retirees will want it.

It's a complex bet, really. Stock-heavy strategies with their sequence of returns risk are most vulnerable in the first decade of retirement when the utility or “usefulness” of more spending is greatest. In other words, the time when it is most useful to bet on the stock market increasing your future standard of living is the same time that losing that bet has the greatest risk of depleting your savings prematurely. Making that bet later in retirement is safer, but the rewards are significantly less interesting.

Sometime between age 70 and 75, betting our current standard of living to maybe improve our future standard of living at 85 or 90 if we are still around and active starts making a lot less sense. After age 80 or so, stock investments are more likely to improve our heirs' future spending than our own, a thought that should particularly give pause to those of us who have no heirs.

Maybe you're willing to risk some of your existing standard of living at age 65 in hopes of spending even more when you're 75, but are you willing to bet your standard of living at 80 in hopes that you can spend more at 90?

When planning retirement, you need to look not only at how much money you might have, but when you might have it. You're best shot at increasing your standard of living after you retire and being able to enjoy the additional consumption is between the ages of about 65 and 75, even though that is the riskiest time to invest. After that, the expected returns diminish significantly along with your life expectancy and health.

(If this is a lot to chew on, there is a lighter explanation at Diminishing Returns.)

Stock-heavy retirement strategies make a lot more sense very early in retirement when your goal is to increase spending over time. When a large bequest is one of your goals then, by all means, party on, Garth. Likewise, continue to invest in stocks if you plan to use your upside potential to pay health care and long-term care costs or simply to improve your emergency fund.

But, if your goal is to improve your own standard of living, you'll only have about ten years after retiring to grow your portfolio and begin spending more before reaching age 80. It may pay to switch retirement strategies away from upside and more toward a floor after no more than 10 years of retirement and spring for the convertible '62 Porsche while you can still enjoy cruising around town with a breeze blowing through the place where hair used to be.


  1. Dirk, I usually have to read your blogs 2-3 times before it sinks in, and this one was no different. Excellent blog, and confirms my asset allocation at this stage of my life. Thoroughly enjoyed the blog. Brad

    1. Thanks, Brad. I sympathize. I read them at least 20 times before posting.

    2. Thanks, Dirk, for another thought provoking post.

      So, let me get this straight. A 65-year old male with a 60-40 stocks-bonds portfolio has about one chance in four of increasing his standard of living over the next 10 years and surviving another 5 years to enjoy it. He also has (according to the first pie chart) a similar probability - about one chance in four - of experiencing a 20 percent or more decline in his spending due to his 60-40 asset allocation.

      Sounds like a sucker's bet to me. Not at all attractive when you consider the risks involved.

      No thanks. I think I'll invest my retirement portfolio in short-term Treasury notes instead and preserve my standard of living.

  2. Dirk,

    Always a pleasure. As you know, I am not your regular $500,000 plus pension United States of America citizen. That being said you also know that I have more at stake since I have no heirs but I do have enough to get me a modest but thankfully enough (I hope) to continue to have a safe place to live and adequate healthcare (no heroic measures please) in my retirement years...which are either three years away (with some part time work) or six years away if I can continue to work at a decently paid job.

    My question. I see the market is hot hot hot and I do not not not want to lose it in what I see as the coming bubble in the equity market. Right at this moment I am 1/3 equities funds, 1/3 bond funds and 1/3 Series I Bonds and cash in equal measure with another 1/3 in an emergency fund that would hold me for two years.

    Common sense tells me that once the equities market bursts the cost of owning the equity mutual funds I am in (Vanguard) is going to go down in price per share big time.

    I am about to do something some would think I was crazy to do. That is to move my equities into bonds. Get out while I can. Take my profit and run.

    Do you think there is any chance of keeping SOME (like a 50% of the equities portion) still in the mutual fund and move the rest to all bond positions?

    OR, leave the market entirely.

    I find it telling that Vanguard Wellington Admiral shares are trading at roughly $65 a share with the same payout as the Investor shares at roughly $30 a share. That is more than twice as much per share just to get 7 points shaved off what Vanguard takes for it fee.

    This does not make sense!

    Personally, I like the hybrid approach that you discussed in your last post, however, at the valuations in the equities markets I just do see how I am going to make a profit going forward. There is just no where to run.


    1. What you are describing is an attempt to time the market. There is no evidence that anyone profitably times the stock market, though there is evidence from Morningstar that market timers tend to lose about 3% a year on average when the stock market gains 8%. I agree that stock prices appear high right now, as do bond prices, but I have no idea when that will change, nor does anyone else. It could happen tomorrow or in five years.

      I recommend that you look at your investments in a totally different way. Think about the largest percentage loss your portfolio could sustain in a bear market and not make you feel like you had to sell. For me, that number is about 15%. I don't want to lose more than 15% in a market crash. Ever. A 40% stock allocation with 60% bonds is unlikely to lose more than 15% in the worst bear market, so that is the portfolio allocation I use.

      (30% stocks might lose 10%, and 20% stocks might lose 5%. You can find these numbers on p. 144 of *The Intelligent Asset Allocator *by Bernstein.)

      Once you find the portfolio allocation that lets you sleep at night, you no longer need to worry about market timing. If the market crashes, you know you are unlikely to lose so much that you will panic sell.

      It appears from your comments that your stock allocation is much higher at present than you are comfortable with. Your present 33% equity allocation could lose about 12% in a market crash. If that is uncomfortable for you, then sell stocks and buy bonds to lower your stock exposure to your point of comfort, but do it because you have the wrong stock allocation for your risk tolerance and not because you think the market is too high.

      After you change your allocation, if you get it right, you shouldn't feel a need to sell again because you think stocks are overpriced. You will be betting the amount of money that you'd be OK losing.

    2. I am answering your Vanguard Wellington question separately because it is not directly related to the allocation issue. I believe you may be confused about the price difference. These are actually the same fund, holding identical investments,except that the cost structure is different. They have different minimum investment amounts and different expense ratios (Investors is 0.25% and Admiral is 0.17%.) Admiral shares are 0.08% cheaper because you have to buy more of them (consider it a bulk discount).

      If you go to []( and click on “Compare” after selecting VWELX and VWENX, you will see that they are otherwise identical.

      I assume by “payout” that you mean yield. The SEC Yield for Admiral shares is 2.3%and for Investors 2.2%. The difference is the expense ratio. If you invested $50,000 in both funds (the Admiral minimum) you would buy 750.075 shares of Admiral and 1,295.337 shares of Investors at 3/8/2014 prices. The yield, however, would be 2.3% of $50,000 for Admiral shares, or $1,150 and the yield for Investors shares would by 2.2% of $50,000, or $1,100.

      The yield would be $50 more for the cheaper Admiral shares, not double.

  3. Great blog. which I stumbled upon and will continue subscribing.

    If I have no heirs then based on these charts should I just got for SPIA so I get more money than with a bond portfolio but the security of a floor irrespective of the stock markets behavior? Any data how this woould compare to stock/bond portfolio

    Also what is the best age/interest rate scenario to purchase a SPIA?

  4. Thanks, John.

    Economists love SPIAs for retirees "without a bequest motive". I'm not as fond of them, but perhaps that's because I have three children. Theoretically, a SPIA would always pay more than bond ladders because insurance companies also invest in bonds and additionally enjoy a mortality premium.

    The primary reason to go with a bond ladder instead of a SPIA would be to retain control of your capital. I have frequently read that around age 70 is the "sweet spot" for life annuity purchases, but I believe there is now some question about that.

    Personally, I wouldn't buy a SPIA today, or a TIPS bond with duration greater than about 5 years, because interest rates have nowhere to go but up, in my opinion. As I mentioned in Hybrid, you could invest in intermediate TIPS and wait for SPIA payouts to improve.

    You might want to read The Annuity Advisor by Michael Kitces before making a purchase.

    Good luck and thanks for reading!

    1. Just came across this excellent blog. In my own case (widowed female, 66), SPIAs have been helpful. I used about 1/3 of my portfolio to purchase a straight immediate annuity through Vanguard, having a payout rate of 7.2%. No inflation rider: too expensive.

      I do have small, part-time self employment, but it may evaporate at any time, and I am not going out to replace it! Between Social Security and the annuity income, I have enough guaranteed income to live my modest life (don't care about cruises). Meanwhile, the rest of my 50/50 portfolio can hopefully grow enough to make up the inflation premium I'll need later on-- and I don't need to take any withdrawals until 70 (RMD).

      There's a lot of peace-of-mind in this mixed strategy. And, as
      Dirk points out, it actually gets me more to spend now than I feel I safely could. More money today is better!

    2. Thanks for writing. I'm not a big fan of SPIA's, personally, but I do believe they are ideal for some situations and yours sounds like one.

      A couple of thoughts. Having no inflation protection on your SPIA might not be so bad unless we have 70's-like inflation because retiree spending tends to decline with age at about the same rate as normal inflation, around 3% a year.

      Also, RMDs are a tax event. You will need to take the money out of a tax-deferred account and pay any taxes owed. But then you can re-invest them immediately in a taxable account (though not a Roth).

      Sounds like you're happy with your retirement plan. Good for you!

  5. Thanks Dirk for explaining what appears a forever expanding lack of knowledge on my part. That being said, that is why I appreciate so much the information I gain from those like yourself that, yes, shows me I have much to learn and shows me I am at the right places to learn it.


    I moved the Wellington into Wesselly (sp) and combine it with Target Retirement 2010 to help me sleep better at night. Unfortunately this still puts me at 50/50 mutual fund stock and mutual fund bond allocation. My 1/3 Series I Bonds I see as a different animal and do not think of this as something to live on until my health deteriorates and I can no longer use a part time job to supplement my income. Which I believe I have in substitute teaching.

    Thanks again, and I will not try to time the market.

  6. Hi Dirk,
    You said:
    "I don't want to lose more than 15% in a market crash. Ever. A 40% stock allocation with 60% bonds is unlikely to lose more than 15% in the worst bear market, so that is the portfolio allocation I use."
    Just wanted to point out that a 40% stock 60% bond portfolio (Vanguard Target Retirement 2010 Fund -VTENX ) lost 30% during the financial crises.

  7. You bring up a few important issues.

    According to Yahoo! Finance, VTENX actually declined over 33% from October 8, 2007 to March 9, 2009. SPY (S&P 500 ETF) declined over 54%. Vanguard Total Bond Market ETF (BND) grew 7.165%

    A portfolio of 40% SPY and 60% BND had a weighted loss of -17.6%.

    That is a greater loss than Bernstein’s estimate of 15%, but it isn’t far off. Bernstein presented these as guidelines given past market performance prior to 2007 and the 2007 crash was unprecedented in recent S&P history.

    As you noted from my post, I said “unlikely to lose more than 15%". There are no certainties with stocks. To make absolutely certain that you won’t lose more than 15% in the stock market, don’t buy stocks.

    I cannot explain the performance of VTENX. The obvious guess is that they didn’t actually have a 40/60 stock/bond allocation at that time, since the majority of the variance of returns is explained by portfolio allocation, but that is purely a guess.

    Thanks for writing.

  8. A book was recommended to me by a Vanguard advisor last year before I made my claiming decision (I took survivor benefits). It is called "Social Security Strategies" by William Reichenstein and William Meyer, and is very comprehensive. Written primarily for advisors, it is detailed and quite helpful for thoughtful individuals.

  9. This comment has been removed by a blog administrator.