Friday, February 2, 2018

What's a Floor?

After my last post, The Retirement Café: Unraveling Retirement Strategies: Floor-and-Upside (An Update), I received several comments and emails regarding floor portfolios that made me realize that the definition of a floor isn’t universally applied and that I need to communicate the definition that I use more clearly.

In "How retirement savers construct an income floor"[1], Stan Haithcock suggests the following:
"You need a solid income base to build on and to hopefully add to those guaranteed amounts. These income sources can include your Social Security, pensions (if so fortunate), income-producing real estate, dividends from stocks, bonds, and contractual annuity payments."
While I find that the column provides generally sound advice, I don't agree that all the assets in this list should be used to build a floor. Real estate income depends on real estate market performance, stocks have market risk so their dividends do, as well. Bond income has bond market risk unless laddered and held to maturity.

I received other comments from readers that considered potential floor assets to include a rolling 10-year TIPs ladder. Deplete your portfolio and how will you buy future rungs? There was even a suggestion that RMDs are floor income, although they totally depend on portfolio performance.

Some seem to define a floor portfolio as an income portfolio complementing the upside stock portfolio and believe that any investment that yields income is suitable for a floor. I don't view floors that way and I don't believe that Bodie, Merton and Samuelson[2] had that in mind when they envisioned lifecycle finance.

I found the following an excellent explanation from a Bogleheads thread[3]. "bobcat2" explains:
"The life-cycle approach ("floor" plus upside approach) is the general economics approach to financial planning including retirement planning. The older approach (called mean/variance or "probabilistic") is based on risk-return tradeoffs along the efficient frontier and is a special case of the life-cycle approach. In that special case, the floor goal is either non-existent or very low, and the aspirational goal is soft. ("I would like to have this much or more, but perhaps not realizing that the “or more” reduces the chances of meeting the goal.")

There is no pure life-cycle approach. You pick two goals. One goal is what you want [upside]. The other goal is a lower conservative goal that typically you want to hit with very high probability [the floor]. You are serious when you set or reset the goals and you employ investment strategies that are explicitly targeted to meet the goals. If you want to hit the lower goal with near certainty, you are going to have to hedge or insure, not diversify, the risk of reaching that goal. That means you need a matching strategy to reach that conservative goal both before and during retirement." 
We "insure" the "near certain" floor with annuities, Social Security benefits, pensions, and possibly a very long and expensive TIPs bond ladder or a shorter, non-rolling ladder supplemented with a deferred income annuity at its end[4] that's less expensive.

Dividends, bonds or bond funds other than laddered TIPs held to maturity, RMDs, real estate income and rolling ladders are not "near certain" and, therefore, not predictable.

What's a floor, anyway?
[Tweet this]

I view a floor portfolio as a safety net of near-certain income in case factors beyond my control should leave me with nothing else.

My grandparents suffered the effects of hyperinflation that destroyed their finances. People used to refer to their predicament as “living on a fixed income” but what they meant was that they no longer received pay increases from an employer to offset inflation. The purchasing power of their pensions and savings accounts eroded quickly.

When I think of a floor portfolio as a safety net, I imagine that my client’s other assets are depleted and that they are forced to live only on income from the floor. I want to make sure that their floor income can withstand some pretty serious inflation, so I consider inflation protection a critical component of a floor portfolio.

A dear friend lost his entire $4M retirement portfolio during the Tech Crash just a few years before his planned retirement. When I think of a floor portfolio as a safety net, I imagine that my client’s investment portfolio is depleted and he is forced to live only on income from the floor portfolio. I consider mitigation of market risk critical in a floor portfolio. Market risk (any market) belongs in the upside portfolio.

For a decade now, retirees have been hurt by historically low interest rates that have left safe income sources like CDs and money market returns barely worth the effort, so I consider the mitigation of all capital market risk to be critical in a floor portfolio.

Growing up in a rural community, I had several relatives and friends of relatives who were trying to get by on Social Security benefits alone. It wasn’t pretty.[5]

They were mostly widows whose husbands, often deceased for a decade or two, likely hadn't been able to afford to delay Social Security benefits or didn't understand the value of delaying, which greatly reduced their spouse’s survivors benefits. When I think of a floor portfolio as a safety net, I imagine that my clients don’t want to end up living off Social Security alone in old age. I recommend they delay Social Security benefits as long as possible. I consider mitigating longevity risk to be a crucial component of a floor portfolio.

It is also worth considering building a floor with as many judgment-proof and bankruptcy-proof assets as you can.

Those three goals, mitigating inflation risk, mitigating capital market risk, and mitigating longevity risk are, in my personal view, essential to a floor portfolio’s design. Combined, they provide a pretty strong safety net of near-certain, real lifetime income.

Of course, not all risk can be mitigated. Spending shocks, for example, can destroy our finances even with an adequate floor. The floor defines the amount of safe income available but it has no sway over costs.

One last but critical consideration is the amount of floor income you target. Don't overdo it. Even small floors can be expensive. Design the rest of your plan such that having to live off floor income alone is very unlikely.

Sleeping on the floor is a more tolerable consideration if the chances of ending up there are small enough. At the other extreme, retirees who plan to spend 5% of their investment portfolio for thirty years in retirement probably want a cushier floor. If I had a 10% chance of losing my bed, I'd keep an air mattress nearby.

I don't get to define what constitutes a floor but it is important to understand the definition I have in mind when I use the term in posts.

It's fine if you or your retirement planner use a different definition as long as you agree and understand how it differs from the lifecycle economics definition. Just know that, with a different definition, some of your floor might not be there in certain scenarios when you need it.

It's challenging to build a perfect floor for several reasons. The cost of retirement is unpredictable and changes over time[6]. You may not claim Social Security the year you retire and filling the safe income gap until you do can be problematic. Pensions provide lifetime income but most aren't inflation-protected. An inflation-adjusted immediate annuity is in many ways the best answer but many retirees won't buy one. Flooring is very expensive in today's economy.

For these reasons, you are unlikely to find a perfect answer and will need to make concessions. But, it's important to begin the process with some basic goals in mind and to imagine the future scenarios in which you might have to live off floor income.

I begin with the goal of a floor portfolio that provides near-certain safety-net income and I try to fill it with assets that in combination mitigate inflation risk, capital market risk, and longevity risk to guarantee that I can survive improbable but worst-case outcomes. Because floors are expensive, I build mine as low as I think I could tolerate and then I structure the rest of my retirement plan to minimize my chances of rolling off the bed.


[1] How retirement savers construct an income floor, MarketWatch.

[2] Videos - Robert C. Merton Finance Class at MIT

[3] Wade Pfau: Lifecycle Finance - Page 3 -

[4] The TIPS plus DIA strategy is discussed in this column by Wade Pfau. It contains links to the original research. Safe Retirement Income with TIPS and a deferred annuity, Wade Pfau.

[5]9 Ways to Retire on Social Security Alone, AARP.

[6] Estimating the True Cost of Retirement, David Blanchett. >


  1. i don't see how to inflation-proof my floor. i've priced step-up annuities [increased by 2 or 3% per year] and imo they are way overpriced. inflation indexing requires going to a lower-rated insurer, and those are also incredibly overpriced.

    even in that case you are assuming that cpi-u is somehow equal to the cost increases that YOU will experience. i think that is an overly optimistic as well as unlikely assumption,

    tips have the same problem of being tied to cpi-u. social security is tied to the same index. it's something, but imo it's inadequate inflation protection.

    otoh fixed spia's act more like DEflation insurance. in combination with one's inadequately indexed social security, they provide a deflation-protected floor.

    an inflation-targeted variable portfolio [in addition to or -more likely- instead of, your "upside" investments] can hedge inflation more effectively.

    1. You can get a CPU-adjusted annuity from a strong carrier. Don't know if you should.

      "Overpriced" is subjective. If you live a long time and experience high inflation then the "overpriced" option will more than pay for itself. If that happens, was it overpriced? If your goal is to protect yourself from that scenario and it does, was the option overpriced? Are vaccines overpriced? Not if you get the flu. Inflation-adjusted annuities inoculate you against inflation but are a waste of money if high inflation doesn't develop.

      If you feel they are overpriced because you think insurers should accept lower margins, I think you're tilting at windmills. (I think Teslas are overpriced. I wish I could buy one cheaper. No one seems to care.)

      Jeffry, I think you're perhaps focussing too much on indexing. An index is an average and of course it is unlikely to mirror your individual situation. Your market returns are unlikely to match the S&P 500 index returns.

      CPI-U is an average of inflation in urban areas. If you don't live in a city, it probably won't match your inflation. In fact, it's an average for many cities and, further, an average of the averages within each of those cities. Until they come out with CPI-J (for Jeffry), you're going to have this issue. The problem isn't that the index doesn't work because of your age but because you aren't the average guy in the average city among all the cities in the index.

      Lastly, equity portfolios don't hedge inflation but they have historically outperformed it. They usually don't perform well during periods of high inflation. If you deplete your upside portfolio, your floor portfolio should save you from catastrophe. How will your upside portfolio have protected you from inflation then?

      You can't "inflation-proof" a retirement plan. I don't believe I have suggested that anywhere. At best you can mitigate inflation risk.

      As I mentioned, there is no perfect answer to retirement planning. If there were, I wouldn't be writing a blog.

      Regards. . .

  2. Bobcat2 and other proponents of liability matching seem to confuse the mean (the portfolio) with the end (the income). Where a floor is needed is for the income. This does NOT mean that one needs to partition their means (portfolio and other sources of income) accordingly. Bucketing strategies are plainly flawed, partitioning things which are NOT additive, or leading to strategies artificially altering long-term income for the sake of mental accounting. The concept of portfolio floor just doesn't make much sense. What is more sensible is to discuss strategies leading to an income floor with fairly solid (and yet not necessarily absolute) guarantees.

    1. So, your argument is that Paul Samuelson, Robert Merton, Zvi Bodie and William Bernstein are confused?

      Let me consider than and get back to you.

      Thanks for writing! :-)

  3. i think the indexed annuities are overpriced because the insurers have difficulty estimating, let alone hedging, future inflation. thus they build in a hefty margin for the uncertainty. that may be proper pricing in a sense, but it's VERY expensive for the annuitant.

    the main problem with cpi-u is health care, which is severely under-represented in the index. and among a retiree population it is all the more relevant. similarly, shelter is under-represented. the bls constructed an index, cpi-e, which attempts to capture the differences, and cpi-e runs above cpi-u, and of course that difference compounds.

    when i talk of an inflation hedging portfolio i am NOT talking especially about equities. for myself, i am following as a superstructure the "golden butterfly" portfolio at portfoliocharts dot com. the allocation to equities is 40%. from that allocation i subtract investments i have in private real-estate partnerships. then i manage the equity component using a taa strategy i assembled at allocatesmartly dot com. at the moment that strategy is 80% invested in equities+commodities. at the moment that has me, overall, 16% in equities, 2/3 of which are foreign [eem,ieur,ewj,vxus]. the superstructure calls for 40% split among fixed income, of which i have about half in a 9 year tips ladder starting at 2020, about 1/4 in currencies [fxe,dsum,fxy,cny], a bit in edv, some in cash. then i have about 20% in phys [closed end canadian-based fund that holds audited physical gold and is not subject to all the questions about gld].

    i am not recommending this to anyone else, i'm just saying that i'm attempting to hedge inflation with a very diversified rules-based portfolio.

    but you're right that the big problem is indexing. i can't accept cpi-u as a worthwhile index for retirees.

  4. Good article.

    I provided a link below to article that Merton wrote (credit BobK) where he outlines that Tier 2 floor income can be constructed with a mix of short and long term TIPS funds where the duration matches the duration of your liability (ie, your expenses in retirement).

    What are your thoughts on TIPS funds when used to duration matched as opposed to building a long TIPS ladder? The cost should be roughly the same, perhaps even a bit more due to the funds’ expenses, but it may be more straight forward for those who do not work with an advisor.,PDF.pdf

  5. Good questions, Matthew.

    TIPs aren't perfect because they don't completely hedge longevity, ladders are difficult to build, they're expensive because there are no mortality credits and there aren't TIPs with maturity dates matching several years. SPIAs aren't perfect because inflation protection is expensive. So, we have to build a floor with imperfect assets or, more likely, a combination of imperfect assets. A long TIPS ladder is difficult.

    I don't have an answer to the TIPs bonds versus funds question, though I'm searching for one. You can't easily match funds to liabilities because you need to match both duration and convexity, which basically means that you would have to build the fund from the same bonds as the ladder. You have no control over when the fund sells, of course.

    I used to think you couldn't match liabilities with funds but I now question that for two reasons. I believe it might be possible to match the funds closely to a ladder but I suspect that would require constantly updating your fund holdings. I'm trying to get a firm answer on that. It may be achievable but too much trouble.

    Second, I'm not sure how much difference it makes if you only approximate the ladder with funds. I don't know the answer to that one, yet, either.

    I approximate a ladder with funds but I have a large margin of error, so that's another issue. If you're wealthy enough, it probably doesn't matter.

    When I get the answers, I'll share.

    Thanks for the question.

  6. Thanks for the thoughtful and honest response. I also found the following article written by BobK to be particularly useful on this topic. Enjoy!

  7. Thanks for all your writing! Thanks in this particular case for the footnote link to the videos of Merton’s finance class lectures. What a resource!