Wednesday, May 24, 2017

Sun Tzu or a Rubik's Cube?

The ancient Chinese military strategist, Sun Tzu, wrote, "Know your enemy" around 500 years before the birth of Christ. A better translation is something along the lines of "Know your enemy and know yourself and you need not fear a hundred battles."

I readily confess that I have not read The Art of War, or at least don't remember reading it, though I have read many things written about it. I stumbled across such a reading this week.

I would paraphrase the "general" idea (pun intended) behind "know your enemy" as recommending that the more you know about your opponent in a strategic game with uncertain outcomes[1] the better your chances of success. Retirement finance is such a game. Knowing a lot about your enemy doesn't guarantee success in warfare or in retirement finance.

The "enemy" I have in mind is the loss of one's standard of living in retirement. It helps to understand this enemy but understanding it doesn't guarantee you won't succumb to it.

This brings me to the Rubik's Cube I noticed in the back of my closet this week. I could only remember how to solve one face but I recalled that there are well-defined algorithms for solving the entire cube and I googled one[2]. Follow these directions and you will solve the Rubik's Cube every time.

Retirement finance is harder to solve than a Rubik's Cube.
[Tweet this]

There is no equivalent set of algorithms to solve the retirement spending problem with certainty. Sure, we can fund retirement solely with Social Security benefits and fixed annuities but that only guarantees the income side of the equation. The expense side will always be uncertain and our goal in retirement isn't really to secure income but to secure non-negative cash flow.

Warfare and a Rubik's Cube are both complex problems to solve but unlike warfare, the Rubik's Cube problem has no opponent – neither man nor nature – to introduce uncertainty.

The problem is that many of us hope to approach the retirement planning problem as a Rubik's Cube instead of a battle, a stochastic game against nature, or a visit to a roulette wheel[3]. Buy an annuity, time-segment your portfolio, invest in index funds, minimize your taxes. Surely there is some list of directions on the Web like those for solving a Rubik's Cube that more or less guarantees a successful retirement, right?

Unfortunately, there is not. Retirement finance is intrinsically fraught with risk. We can know our enemy well but the main thing we know about that enemy is its uncertainty. We can do all the right things and fail. We can do all the wrong things and succeed.

My favorite example of doing the wrong things and succeeding is my friend Gerry's blackjack hand. I once watched Gerry split a pair of fives in a blatantly rookie move, turning a pretty good hand of ten into two bad hands of fives . . . and win both hands.

Sometimes the force is with you.

If you retired in 1982, according to a column written by Wade Pfau[4], and lived 30 years you would have had a hard time depleting your savings with a reasonable sustainable withdrawal rates strategy. Pfau figures a 9.8% annual withdrawal rate would have avoided portfolio depletion. The same retirement beginning in 1966 would have survived only 4% withdrawals. Those who retired in 1982 could split a pair of fives and probably still win.

My favorite example of super-intelligent people creating lovely, complex algorithms that failed miserably is the story of Long-Term Capital Management. This hedge fund, the subject of the book When Genius Failed by Roger Lowenstein, was initially highly successful under the leadership of a former Wall Street bond manager and two future Nobel laureates. Eventually, however, LTCM failed, bankrupted its founders and brought the global financial system to its knees.

(I highly recommend Roger Lowenstein's book, but Business Insider published a synopsis at the link below.)

As an aside, I often quote the William Bernstein dictum, "When you win the game, stop playing." Occasionally, someone will write me to say that's a bad idea (Note: Bernstein doesn't have a lot of bad ideas).

Buffett: …It’s…an interesting story…The whole story is really fascinating because if you take John Merriwether, and Eric Rosenfeld, Larry Hilibrand, Greg Hawkins, Victor Haghani, the two Nobel Prize winners, Merton and Scholes, if you take the 16 of them, they probably have as high an average IQ as any 16 people working together in one business in the country…an incredible amount of intellect in that room. Now you combine that with the fact that those 16 had had extensive experience in the field they were operating in…In aggregate, the 16 had probably had 350 or 400 years of experience doing exactly what they were doing. And then you throw in the third factor that most of them had virtually all of their very substantial net worths in the business. So they had their own money up. Hundreds and hundreds of millions of dollars of their own money up. Super high intellect, working in a field they knew. And essentially they went broke. And that to me is fascinating…

…But to make money they didn’t have and didn’t need, they risked what they did have and did need, and that’s foolish. That is just plain foolish."[5]

Sounds like Warren Buffett and William Bernstein agree on this point.

My point is this. Retirement finance is a probabilities game. There is no set of rules that guarantees a successful outcome like the set of rules for solving a Rubik's Cube, even though both are complex problems. Retirement finance is more of a Sun Tzu thing.

Had Sun Tzu been a retirement planner of the probabilist school, he might have said, "Know your enemy and know yourself and you need not fear about 95 out of a hundred battles."


[1] A Tiny Bit of Game Theory, The Retirement Cafe´.

[2] How to Solve a Rubik's Cube.

[3] Retirement Roulette, The Retirement Cafe´.

[4] What If Retirees Don’t Want To Run Out Of Money In 30 Years? by Wade Pfau.


[6] The Epic Story Of How A 'Genius' Hedge Fund Almost Caused A Global Financial Meltdown, Business Insider.


  1. love your blog, which i’ve been following for a long time. i tried to post a comment on your most recent post, but somehow that wouldn’t work, so i thought i’d just send an email which you could post if you think it’s worthwhile.

    how do you stop playing the game? i’m not referring to some kind of addiction to investing. i’m referring to the fact that there is no asset without risk. even if you try to hedge against what you think are the major risks, there is no portfolio without risk. furthermore, part of minimizing risk must be a constant monitoring of the threat landscape, with the possible need to move assets around as you change your view of the dangers you face.

    you can try to minimize risk, but you can’t eliminate it. and that means that, at least to some degree, you can’t really ever stop playing the game.

    [i leave aside the case of someone with so much wealth that if they deploy their assets conservatively, there is no plausible scenario in which they are fundamentally threatened. for most people, for me, that much wealth means they kept playing for a long time after "winning the game" in ordinary terms. or they just inherited.]



    1. Jeff, you are correct that there are no assets without risk, though I would argue that TIPS bonds come pretty close. Even if you funded retirement income with risk-free assets your expenses would still be risky. "Stop playing the game" doesn't mean avoiding retirement finance risk (you can't).

      Nor does it mean "stop investing in equities." That advice would be totally out of character for Bernstein or Buffett.

      The clue is in Buffett's comment, "…to make money they didn’t have and didn’t need, they risked what they did have and did need, and that’s foolish." To "stop playing the game" in the retirement context means to secure your standard of living before risking money that you have and need for the chance of improving that standard of living.

      Good question! Thanks for writing.

  2. Dirk,

    My wife turns 65 in August and will receive a 17,640-20,520 annual pension from the hospital she worked at for most of her career.

    We're both 64 and in good health and have family longevity in her parents and my Mom.

    Straight life pays 20,520
    50% survivor pays 18,960
    100% survivor pays 17,640

    It's the old "how long will we live" financial planning dilemma.

    If my personal discount rate is 3%, and we both lived to 100, the PV of the three would be:

    If you have the time to give me your thoughts and/or turn the problem into a blog post, I'd appreciate it.


    1. Good question. I don't consider "how long we will live" to be a dilemma for a reasonably healthy person. It's unknowable with any meaningful precision and somewhat meaningless to guess. A better approach is to assume that you will live a long time.

      The probability of both spouses living to 100 or more is tiny. Although it happens (for example), that is probably a bit too conservative an assumption.

      In general, though, I think you're asking the wrong question. Instead of asking how you can maximize total lifetime annuity payouts if you both live a really long time, I think you should be asking how to best mitigate longevity risk (the purpose of a pension).

      In other words, you should be asking how much money you will need if one or both of you lives a very long time. This will include three scenarios: only you live a very long time, only your spouse lives a very long time or, in the least likely case, you both live a very long time.

      Will the survivor(s) need $18,960 to maintain his or her standard of living in old age or $17,640? Will Social Security benefits and other assets meet the entire desired standard of living without any survivor benefits from this pension, at all?

      My recommendation would be to choose the option that, in conjunction with the rest of your retirement plan, meets or exceeds your desired standard of living in all three long-life scenarios, not necessarily the option with the highest net present value.

  3. Hi Dirk, I enjoyed this post. I also liked your response (Last comment - Buffet) to the first commentor. I struggle though with the idea of securing your standard of living before risking money that you have and need. My question is this. How does one determine his/her standard of living (sol) in the first place. My wife and I are currently living on 50K per year, but we anticipate our assets and guaranteed income will allow us to have income of about 130K per year in retirement (couple years). We currently have about 20-25% in equities. We could have much more in equities if we defined our standard of living as 50K (what we need) vs 130K which is what we currently project. To me the question then becomes how much are you willing to risk the 130K which is in part how I arrived at 20-25% in equities since this would mean the 130K might fluctuate 10% with that equity allocation (only about 70% of our income is not guaranteed). We could have more in equities (which would be nice as the market continues to advance), but it is not worth the chance of drawing down what is already a more than acceptable sol (130K) for us. Thanks for always thought provoking posts. Brad

    1. Let me make sure I understand your question.

      Your current standard of living is the amount you spend on living expenses excluding any retirement savings (retirement savings are the amount of standard of living you are forgoing today to improve your standard of living in future retirement). Sounds like your current standard of living is $50k per year.

      It also sounds like you expect be able to generate $130K of income per year in a couple of years. That is expected income, not standard of living. If you decide to spend all $130k annually in the future on living expenses, $130k will become your standard of living. In other words, your SOL would increase from $50k today to $130k in a couple of years.

      It also sounds like $50k is your minimum acceptable standard of living. If so, and you increase your standard of living to $130k in the future by spending all of that income on annual living expenses, then $50k might still be your minimum acceptable standard of living even then.

      How much of your standard of living you desire to secure with safe investments is, in my experience, highly personal. I have a client with a $75,000 current standard of living who thinks $55,000 is her minimum acceptable standard of living but she chooses to secure the entire $75,000 with fixed annuities. She does that because she would rather have that $75,000 of income with some certainty than to have only $55,000 of income with certainty and the upside potential of spending more but the downside risk of having less than $75,000 to spend.

      Others might choose to annuitize only the $55,000 minimum acceptable standard of living and invest more in hopes of increasing future spending. As I said, it's a personal decision.

      I'm not sure that answers your question because I'm not sure I understand what you're asking. If I answer your question "How does one determine his/her standard of living (sol) in the first place?" literally, the answer appears to be that your current standard of living is $50,000 but that you expect to have the potential to increase your standard of living to $130,000 in the near future.

      That begs the question, however, of why you don't smooth your income by borrowing some of that future $130,000 standard of living to increase the $50,000 and arrive at a smooth income somewhere in between (and I'm guessing close to $130,000)?

  4. Dirk, you have gotten it right. 50K is our current sol, and 130K is our expected retirement income. I think we fall under the scenario that you mention you have with your client. While 50K is our current sol and probably our minimum acceptable sol it is not what we would like to have secured with some degree of certainty in retirement. That figure is closer to 100K. We think with our expected retirement income currently at 130K we can have about 20-25% in equities and the rest in bonds to somewhat secure our 100K retirement income goal. So while we could have a much higher % of our assets in equities if we base our retirement needs on current sol (50K) we chose to base our equity % on our desired retirement income of 100K.

    The reason we don't smooth our income now is two-fold. First, one never knows what the future may hold in terms of surprises (long term care etc.), and second it is hard to switch from a life-long saver to a spender. I realize tomorrow is not promised so we may be foregoing some pleasures today that we may never realize, and that does bother me sometimes. But, to this point, future security has won out. Thanks for your response. Brad

    1. Have to chuckle about the "spenders" comment. A recent retiree called me a couple of weeks ago in a panic. She told me that her checking account was getting low and she won't get Social Security payments for another few months. "Where will I get the money to buy groceries and pay the rent until then?" she asked.

      Me: "You know that $2 million you saved for retirement?"

      Client: "Yes?"

      Me: "This is retirement."