Life After Full-time Work Blog

Learn about preparing for life after full-time work through posts from Don's upcoming book.

#173 An Easy Way To Create A Lifetime Income Stream

Of the five ways to create a sustainable lifetime income, one is particularly simple to implement and has many desirable features.


I was honored to feature in the Defined Contributions (DC) West conference organized by Pensions & Investments (P&I) in Carlsbad, CA, in October. My session was called “Creating income streams post-retirement,” and it was constructed as a fireside chat in which I was interviewed by Nikki Pirrello, the Chief Operating Officer of P&I. My remarks were aimed at DC plan sponsors, to show them how to create a default option for the decumulation (that is, pay-out) phase of their DC plans. But they apply just as much to individuals creating their own income streams outside DC plans.

Nikki got to the point immediately. “Nobel Prize winner Bill Sharpe, whom you’ve worked with, calls this ‘the nastiest, hardest problem in finance.’ And you have the solution for it?”

No, I don’t have the solution for it. But Bill means the optimal solution. And if he struggles with it, there’s no chance on earth I could solve it. But I’m not searching for the optimal solution. I’ll take my cue from another Nobel Prize winner, Herbert Simon. He revived the word “satisfice”, which he said is how we tend to solve problems in practice. The word means what it sounds like, a combination of “satisfy” and “suffice”. The solution has to meet a high standard – otherwise it doesn’t satisfy. And when we find that sort of solution, it suffices – it’s enough. We adopt it. We move on to another problem. It’s not worth saying, “Wait, let’s not adopt it, let’s put it off and keep searching until we find the optimal solution” – because we may never find what’s optimal, and even if we do, we may not recognize that it’s optimal, or it may be too tough to implement.

So that’s what I’m suggesting. A solution that meets the high standard of working for everyone, that (if adopted) improves their situation, and that is very easy to implement. In fact, those are the conditions a default option should satisfy. A retiree can elect not to use it, but for the vast majority it solves the problem and improves their situation.

Nikki, direct again: “OK, if you can do that, it’s worth considering very seriously. Tell us how you get there. Even more precisely, tell us what the problem looks like, what makes it so difficult, and what you suggest.”

The problem is to create a roughly level income stream that lasts throughout your life. And what makes it so difficult is that there are two big unknowns, your investment return and how long you’ll live. The one original contribution I’ve made to this whole subject is to compare the effects of the two unknowns. Since then Bob Collie explained it much better, but I told the audience about the simple thought experiment I conducted, about 15 years ago.

Imagine two highly unusual planets. On Planet A, longevity is certain: you know exactly how long you’re going to live. But investment returns are uncertain, so there’s uncertainty about how much capital you need to finance your desired lifestyle. On Planet B, investment returns are certain. But longevity is uncertain, so again there’s uncertainty about how much capital you need. And my question is: on which planet is there more uncertainty?

The answer depends on your age. For a male aged 60 (or a female aged 65), longevity uncertainty has a smaller financial impact than being invested in fixed income or in a balanced stock/bond portfolio. That’s not a problem. As age increases, the uncertainty curves cross over. By male 75 (female 80), longevity uncertainty has a larger financial impact than being 100% invested in equities. Wow, that’s way riskier than we can stand! So if at that age you’re uncomfortable being 100% in equities, you should hedge or reduce your exposure to longevity uncertainty. If you don’t, you’re effectively more than 100% in equities.

In other words, for retirees, that longevity issue eventually becomes the one you simply must deal with. And people aren’t good at dealing with it, to their own detriment. Let me give you two indications of how true that is.

The first is that surveys consistently show that a big fear is: Will I outlive my assets?

And the second is from a country where their version of Social Security is an individual accumulation plan: Australia. People are so afraid of running out of money that they prefer not to touch it at all, and try to live off the income their savings generate. The published evidence is that, when they die, they leave on average more than 90% of what their savings amounted to when they retired. Imagine how much better they could have lived if they hadn’t been too scared to touch their capital!

It’s a big, big issue, and most people would benefit enormously from knowing they can draw down their capital at a pace that won’t exhaust it.

And my proposed solution?

Well, there are five possible ways to do it.

The first is to buy a guaranteed lifetime income (also called an annuity) from an insurance company. People don’t like it, because they lose access to their capital, and they lose it if they die early. This approach generates a high level of income, but those are psychologically big negatives.

The second is a longevity pool, involving the same pooling of capital but with no guarantee about the size of the payments. Which many people view as an advantage, because it permits investment in growth-seeking assets. But this solution isn’t available in the States yet, though it exists in Australia and Canada.

With both these solutions you can also get a return of unused capital if you die early, but it dramatically reduces the periodic payments.

The third approach is to split your prospective future years into two parts, like up to age 85, and beyond. And you only buy the guaranteed lifetime payments beyond age 85. That’s called a deferred annuity. (In the States, in a DC plan, it’s often called a QLAC – never mind why.) Because it doesn’t start until beyond your life expectancy, it only costs roughly 10-15% of what the full lifetime income annuity costs. With the remaining 85-90% you invest the money yourself any way you want, you own it, and you have a fixed period to make withdrawals, so you can calculate how much capital you can withdraw each year to make it last up to age 85.

In most countries a deferred annuity is simply not available, although it is in the States.

The fourth way is to estimate a future lifetime horizon, somewhat longer than your expected future survival years to build in a margin of longevity safety, and in that way you self-insure your longevity risk. And then you have 100% investment freedom and asset ownership, and you draw down capital at a rate that will last to the end of your planning horizon.

One thing you have to do with this approach is to recognize that your expected survival age gradually increases over time, because the longer you survive, the more select a group you belong to, and your expected survival age increases.

This approach is actually an expensive way to generate lifetime income, because self-insurance requires you to provide for a much longer lifespan than most people encounter. It’s like setting aside enough money to replace your home, just in case of fire, instead of buying fire insurance.

Those are the ways in which longevity is explicitly taken into account.

The fifth way isn’t really a planning approach, it’s built by the US tax authorities (the IRS) for income tax purposes. It’s commonly called RMD, or Required Minimum Distribution. Each year you withdraw an amount equal to your remaining capital divided by your expected survival years. (And those expected survival years automatically consider your gradually increasing expected survival age.) You can take your spouse’s age into account too, so it lasts until the second death if you want. You can invest the money any way you want, you always own the assets, and the formula ensures that you never withdraw all your capital.

Nikki slowed me down, and called for clarification here. Am I suggesting that any of those five approaches can be used as a default?

No, I think there’s one that’s superior to the others. And you may be surprised to learn that I think it’s the RMD approach. It’s very simple to administer, and apart from the fact that (as with any uncertain investment return) there’s no guaranteed annual withdrawal, it has all the other desirable features: ownership, investment flexibility, it allows you to withdraw capital, and it never runs out. As the evidence shows, for most people this is a huge advantage.

Ever the practical one, Nikki asked if it really works in practice, or just in theory.

Well, I was very pleased with myself, I have to say, when I thought of this notion. So I did what I always did, as a consultant: I researched it, before taking it public. And I found, to my delight, that I wasn’t the first to think of it! I found that Steve Vernon, who was then a research scholar at Stanford University, and someone I know and respect enormously, had not only already reached the same conclusion, but he had also tested it and published his research. (Also see his book, Don’t Go Broke in Retirement.) And yes, it works.

As Nikki pointed out, this would work anywhere in the world, not just in the US. (In fact, Canada has a similar approach, with its Registered Retirement Income Fund minimum distribution rules, but it’s not explicitly linked to future expected survival years.) And if you don’t have to follow US tax rules explicitly, you can be more flexible with the formula in other countries.

But let’s stick with the US, because Steve Vernon goes further. He suggests using your assets first as a bridge to delay claiming Social Security to the latest age possible, and then using the rest with the RMD approach. But that may be tougher to implement as a default option for every member of the plan.

My own addition is that, if the plan member can identify the annual size of spending required for the essentials of life, then a life inflation-linked annuity for any excess of this amount over Social Security would lock it in, and the balance of the capital can use the RMD approach for the nice-to-have aspects of life. But sadly, inflation-linked annuities are available today only in Australia and the UK. This is a huge gap in the US market – but that’s a topic for another day.

Anyway, the takeaway from all of this is that the RMD approach is easy to implement as a default option, and it substantially improves the spending power for most people.

Let me make something else clear. People with lots of money don’t actually need this solution if, together with their government “first pillar” benefit (like US Social Security, Canada or Quebec Pension Plan, UK State Pension, etc) and defined benefit pension plan income, they have sufficient DC and other savings to support their desired lifestyle to an advanced age, though the discipline of the RMD approach could benefit them. My proposed default is really aimed at those who don’t have that sort of adequacy, and for them it would help enormously.

Nikki knows her audience, and wanted to make it very clear for them. Many plan sponsors are considering developing decumulation solutions to implement “in plan” for their retirees. So she asked me explicitly: is the RMD approach one that could be successfully implemented by plan sponsors, or are plan participants and retirees better served by accessing a solution like this outside of the plan? In other words, could it actually become a default decumulation option “in plan” at some point?

In a word: yes! No ifs, ands or buts – just simply “yes.” This could be an “in plan” decumulation default. Anyone could elect out of it and either use it as a partial solution, or not use it at all, outside the plan. But as a plan default, it improves the situation for most retirees, and it’s easy to administer. You only need to know two things, which you already know: the retiree’s date of birth and asset balance. Divide the asset balance by the expected future survival years (a.k.a. life expectancy) corresponding to that date of birth, which in the US is specified by the IRS, and pay it out over the following calendar year. The sponsor has to decide on an asset allocation, mid to low risk, but that’s a separate issue we didn’t have time for in our session.

Nikki noted that it’s a simple solution, but it’s still a complex problem. Did I have any advice for how best to communicate with retirees or those nearing retirement about how to start thinking about the decumulation process and the possible solutions that might best serve them? We’ve come such a long way on the savings/accumulation side. Are there learnings that apply here to help solve the next piece of this conundrum for plan participants?

To which I responded: Again, yes. But this is, as Nikki said, complex. And we’re back to Bill Sharpe saying that it’s the nastiest, hardest problem in finance. Explaining the solution is, in a way, even tougher than solving it. Participant education is essential, and it needs to start before retirement, maybe five years before. And that education will only come about gradually. But if I had to try to put together one paragraph to sum up its theme, it would be something like:

Everyone fears outliving their money. Because of that, they actually try not to touch their retirement assets, and live only off the annual income generated by the assets. This reduces their standard of living by a lot, unnecessarily. The default option finds a way for you to gradually draw down your assets, without the fear of it running out.


In my next post I’ll mention some of the reactions I’ve got, and my Mt Everest analogy for explaining how much this solution does for the average participant.



It’s a huge uplift in a retiree’s lifestyle if the retiree can spend not only the income generated by accumulated savings, but also the capital itself – without the fear of running out of capital by living too long. A simple way to achieve this is through a formula known in the US by the acronym RMD: the minimum distribution required by the IRS for US income tax purposes. 


I have written about retirement planning before and some of that material also relates to topics or issues that are being discussed here. Where relevant I draw on material from three sources: The Retirement Plan Solution (co-authored with Bob Collie and Matt Smith, published by John Wiley & Sons, Inc., 2009), my foreword to Someday Rich (by Timothy Noonan and Matt Smith, also published by Wiley, 2012), and my occasional column The Art of Investment in the FT Money supplement of The Financial Times, published in the UK. I am grateful to the other authors and to The Financial Times for permission to use the material here.

8 Responses to “#173 An Easy Way To Create A Lifetime Income Stream”

  1. Michael Dodd says:

    Don congrats on another brilliant insight! So glad to see Decumulation for DC Plans continuing to get thought leadership ideas.
    Thank you!

  2. Thomas Philips says:

    The RMD strategy is fundamentally sound, and in fact, is surprisingly close to optimal! The earliest reference I have seen to it is Sun, Wei and Anthony Webb. 2012. “Should Households Base Asset Decumulation Strategies on Required Minimum Distribution Tables?” Working Paper 12-10. Chestnut Hill, MA: Center for Retirement Research at Boston College.

    They had a less techical version in “Can Retirees Base Wealth Withdrawals on the IRS’ Required Minimum Distributions?”, also published by the Center for Retirement Research.

    • Don Ezra says:

      Thanks very much, I wasn’t aware of those papers, and now I’ll read them eagerly, as I hope my interested readers also will.

  3. Barbara Warren says:

    Don, thank your for this terrific analysis! One question: you say that “In most countries a deferred annuity is simply not available, although it is in the States.” What about in Canada, specifically Ontario? I thought there was something called an “Advanced Life Annuity” which if I understand correctly sounds like the deferred annuity you mentioned. Many thanks!

    • Don Ezra says:

      Glad you like the analysis. I haven’t been able to obtain an actual quote for an Advanced Life Deferred Annuity in Canada, though in principle such a contract is legally feasible. You may know more than I do about this, in which case do bring it to the attention of our readers.

  4. Cindy Deere says:

    Thanks Don! I’m a huge fan of decumulation strategies in DC plans. Just wonder how long it is going to take for plan sponsors to actually act in this space. It seems like these ideas have a long gestation time! Anything you (or maybe we) can do to encourage the thinking is a step in the right direction!

    • Don Ezra says:

      Thanks, Cindy. I’ve no idea what to do to advance the cause — I simply hoped that bringing it to the attention of those at the conference might help. I’m happy to help in any way that I can.

Leave your question or comment