Monday, February 4, 2019

Honey, What's Our Retirement Plan?

Retirement planning is a complex problem we can simplify by first focusing on what matters most. The key is getting the big decisions right.

[An earlier version of this post appeared at on January 25, 2019.

The "big picture" retirement finance problem definition is straightforward. How can we work for perhaps 40 years to pay our household's living expenses and simultaneously save enough that, when combined with Social Security benefits and pensions, we can maintain our desired standard of living for one to perhaps 35 more years?

Straightforward, but daunting, right?

The most important step, often given too little consideration, is defining the goals and challenges for your unique household (see The Retirement Plan I Would Want - Part 7). Identifying and agreeing your retirement financial goals with your spouse and planner is a critical first step before the financial strategy even comes into play.

Sometimes, one spouse wants a total return investment plan and the other just wants a guaranteed monthly check for life. On occasion, I even find a client who has self-conflicting goals of his or her own, like wanting to maximize retirement spending and leave a large bequest (a perfect example of wanting to have your cake and eat it, too, by the way). It's difficult to solve a problem when no one agrees what the problem is.

Once the goals are resolved, we can attempt to meet them financially. There are many factors we might consider but some are far more important than others.

The most important factor in determining retirement outcomes is how long we will be retired. Nearly anyone can maintain their standard of living throughout a retirement that lasts a year or two but far fewer households could fund one that lasts 40 years.

We can't predict how long a healthy retiree or retired couple will live, what we call "longevity risk," so the safest bet is to plan for a long, expensive retirement. But, you may not want the safest strategy. Perhaps you're willing to take a little more risk hoping to spend more. This should be clearly evident from your agreed goals. If it isn't, your goals need more work. Regardless, there is one inescapable fact: if you spend more you will have more risk.

The key to retirement planning is getting the big decisions right.
[Tweet this]

If you're with me so far, then the most important decision of retirement planning—the first "big decision"— is how to deal with longevity risk and that is largely determined by the funding strategy we choose.

Many funding strategies have been proposed in the research literature so even this first step can seem intimidating. Wade Pfau and Jeremy Cooper identified eight proposed strategies ranging from safety-first (expensive but safe) to probability-based (less expensive but riskier)[1]. If you haven't heard of most of them, there are good reasons. Some are too complicated for retirees and advisors to grasp, some are challenging to implement, and some are not broadly palatable, like using a triple-leveraged risky portfolio for the Floor-leverage rule.

How should we choose from this extensive menu? Actually, I suggest that you don't.

I'm going to make a claim that may sound a bit outrageous: there is only one grand retirement-funding strategy. That strategy is to allocate some amount of retirement plan resources to generate a floor of safe lifetime income, to invest the remaining assets, if any, in a risky aspirational portfolio, and then to decide how to spend the risky assets throughout retirement. The correct balance will depend on how willing you are to risk losing your standard of living for the chance of having an even higher one.

We can allocate zero dollars to the floor portfolio, in theory at least, and have a purely total return strategy. In reality, nearly all Americans are eligible for Social Security Benefits or a public pension and will, consequently, have some floor whether they want it or not.

This grand strategy may simply sound like the strategy we call "floor-and-upside" (see Unraveling Retirement Strategies: Floor-and-Upside). But, choose to allocate nothing to the safe income floor portfolio and to spend 4% of the initial value of the risky portfolio and we have a "sustainable withdrawal rate" strategy. Change the "4% of initial portfolio value" spending rule to "4% of remaining portfolio value" or RMD-spending, for example, and we have one of Pfau and Cooper's variable-spending strategies. They're different takes on the single grand strategy.

Retirees who fund part of their spending needs from a risky portfolio will also need a spending strategy. For them, this will be important decision number 1(b).

By making the important decisions first instead of selecting from a list of strategies, you can simplify the planning process. Decide how much risk you are willing to take with your standard of living in exchange for the possibility of improving it and allocate retirement resources to your floor and risky portfolios accordingly. If you decide to fund a risky portfolio, then also decide on a spending strategy that equally fits your risk tolerance.

These decisions will have a far greater impact on your outcome than say, tweaking your equity allocation 5% or worrying about whether equities get safer the longer you hold them. You may find that this process provides a better high-level understanding of your retirement plan. You may even be able to describe the important parts of your plan in a couple of sentences.

"Honey, what's our retirement plan?"

"Glad you asked! Here's the big picture in two sentences. . ."

How sweet would that be?

Identify your goals, get the big decisions right, and your plan will be 80% to 90% of the way home.


[1] The Yin and Yang of Retirement Income Philosophies, Wade Pfau and Jeremy Cooper.

[2] The Retirement Plan I Would Want - Part 7, The Retirement Cafe.


  1. [The following comment was emailed to me because the author wanted to remain anonymous. Although readers can post anonymously and most do, I request permission from the commeter to re-post for them so others can see my response.]

    There are several questions here, so I’ll respond by paragraph:

    I'm leaning toward selling $50-75k from my portfolio in order to build a small 2-3 year CD ladder. I'm about to retire in a month (age 63 1/2) and am concerned about sequence risk, especially until collecting social security at age 70. If the market tanks I'd have these CD's (and money in a savings account for this next year) to supplement portfolio withdrawals, thereby reducing my withdrawals until age 70 from 4+% to 2%. If the market was doing OK, I'd just roll the CDs over as they come due.

    Having some cash on hand for whatever reason is a good idea. Sequence risk becomes dangerous about a decade before we retire but better late than never. Sequence risk never goes away but it will obviously decline once you collect Social Security benefits because you will be spending less from your volatile portfolio. The point is that you might not want to get rid of that cash bucket in three years.

    What you are describing is a type of “bucket” or “time segmentation” strategy, though what we think of as bucket strategy is more involved. The advantage of a bucket strategy is largely behavioral — to help you avoid panic-selling at a market bottom. As Wade Pfau points out in the “Yin and Yang” paper referenced above, “ . . . it must be emphasised that on a theoretical level, income bucketing cannot be a superior investing approach relative to total returns investing.” Moshe Milevsky has pointed out that a bucket strategy won’t always outlast a market downturn. If it doesn’t, you would end up selling stocks at even lower prices after the bucket is depleted. Still, if your primary concern is avoiding panic-selling, it might be the strategy for you.

    My question is whether to sell stocks or bonds (or some of both) to buy the CDs. It seems to me that it'd make sense to sell stocks and lock in gains. I'd incur $10-15k in cap gains which would be taxed at 0% but which would reduce the amount of Roth conversion I could do next year (I had been planning to max out the 12% bracket with Roth conversions).

    I’m not sure why you couldn’t max out Roth conversions next year whether you sell or not this year and I’m not sure how you can be more efficient than 0% taxes. You’re comparing a certain sale with no taxes this year to uncertain tax savings in the future. The idea of “locking in gains” suggests you believe you can time the market (you can’t). Locking in gains on Bitcoin in 2017 would have worked out beautifully; locking in gains on Google stock in 2000, not so much — you would have missed out on 20 more years of gains.

    My rep at Vanguard is more inclined to sell bonds. "As a portfolio manager I am always attempting to keep us close to the allocation target, 45% stocks / 55% bonds. However, tax efficiency is a competing goal. So if I was attempting to raise $50,000 I would raise the cash from bonds because the tax efficiency of that sale is very high, however even after that size of a sale, bonds are still relatively close to target (approximately 52%).

    It is not possible to know whether you are better off with a 52% or a 55% bond allocation, or a 50% or 60% bond allocation for that matter. It all depends on how stocks will perform in the future relative to bonds and no one knows that. Your rep is trying to manage your portfolio to an unknowable optimum. 3% is noise you should ignore.

    (continued below)

    1. (continued from above)

      I'd appreciate your thoughts. Doesn't it make sense to sell stocks when they're high, and lock in gains?

      “Buy low and sell high” is fantastic advice, or it would be if we could actually know when stocks are "high" or "low."

      Just one follow-up: selling bonds (to set up a CD ladder) would also mitigate market losses in the next few years, correct?

      Not exactly. As I explained above, creating a bucket mitigates panic selling not market risk. If you don’t sell any of your equities, your market losses will obviously be the same whether or not you have cash. Having cash on hand might enable you to avoid realizing those losses, assuming the downturn is shorter than three years.

      But you'd be inclined to sell the stocks instead because of the 0% tax on the gains?

      No, I’d be inclined to sell stocks if I were you because you are describing anxiety from owning too much equity. The cure for that is to own less equity.

      My target allocation in VG now is 45/55. Should I have VG just rebalance to that or would you recommend that my VG allocation move to slightly higher % equities since I’ll have $75,000 in CDs (and possibly MM) (fixed income)?

      I wouldn’t rebalance at all if I were 3% off a target allocation that is purely a SWAG to begin with. But, if I did, I would treat the cash like bonds so that my cash plus bonds were 55% of my portfolio. If the reallocation is free (no fees and no taxes) it will probably neither hurt nor help.

      One last comment: I think you’re working too hard at managing your plan. (I think most retirees do.) Lower your equity allocation to something that you can stomach in a severe bear market. Keep some cash on hand as a reserve for any emergency, not just to attempt to avoid panic-selling at a market bottom. And, if your equity allocation is off by less than 5% to 7% in either direction, shrug and take in a movie.

      Thanks for writing!

  2. Hi Dirk,

    I caught your post in Fortune from January 11th - Take Your Best Shot at a Retirement Plan - and thought I'd ask you if you had any thoughts one way or the other on utilizing a Roth feature in a 401(k) and if so, what you thought a good allocation would be between traditional 401(k) and the Roth. My feeling is that we will see continued rising federal tax rates into my retirement (I'm 37 going on 57) and I'm considering beginning to contribute to that vehicle.

    Any thoughts or wisdom are appreciated.

    1. Jon, I don’t think it’s possible to know whether your federal taxes 30 years from now will be higher or lower than today. Consequently, I split my savings between Roth and non-Roth accounts 50/50. If you feel more strongly than I do that your taxes will be higher, put more in the Roth. I don’t think you could go terribly wrong going 100% either way — the important thing is that you're saving. Also, I try not to let taxes dominate my retirement decisions because they can change with political whim and if you get the big things right, you’ll be fine. I don’t think taxes are one of the “big things."

      Thanks for writing!

  3. I have the following approach to our Retirement Distribution or Withdrawal strategy. We have an emergency fund equal to one year of fixed expenses in a FDIC savings account earning 2.35%.

    I estimated our essential expenses which includes everything except medical expenses not covered by insurance and LTC. Bucket 1 is covered by Social Security, a pension, and a ladder of 2 deferred fixed annuities to cover inflation and any increases in taxes over what I’ve estimated. The annuities are QLACs and represent a small percentage of our portfolio.

    I estimated our discretionary expenses which includes our hobbies and a moderate annual vacation. Bucket 2 is covered by dividend income (Dividend Aristocrats), but can be supplemented periodically from Bucket 4 if needed, for example, to pay additional taxes on the dividend income.

    I estimated the one-time replacement costs of our home appliances. Bucket 3 is covered by a small cash account and CDs. The cash account should cover 1-2 appliances being replaced in a given year and the cash account can be re-funded when a CD comes due. We just moved into a new home, so this bucket is covers future rather than short-term replacements. It’s surprising how much home repairs add up. Major home repairs like a new driveway or roof would be funded by Bucket 4.

    Medical expenses not covered by insurance and LTC are the biggest unknowns. Fidelity estimates that a couple will need $280,000 to cover healthcare expenses. Bucket 4 is covered by a diversified portfolio of ETFs. I also have an HSA that is invested in stock funds. Fortunately, my husband and I enjoy biking and hiking, as well as gardening, so we keep active. This bucket covers major home repairs, medical expenses, and LTC. We will be doing some Roth conversions over the next 5 years.

    For the ‘normal’ expected market down turn, I believe we should be able to ride it without too much impact to our dividend income and we can always trim back our expenses.

  4. The large company I work for just offered the ability to add an additional $19,000 after-tax contribution to my 401K (beyond the $25K I can add to either a regular or Roth 401K). Is there any downside of doing this versus placing these extra funds in a non-401K savings account?

    1. The only risk I see is that under some circumstances you might be subject to a penalty if you need to withdraw funds but that shouldn't be a showstopper.