Monday, March 24, 2014

The Thing That's Missing

After recently writing about four major retirement spending strategies, Life Annuities, Time-Segmentation, Floor-and-Upside and Systematic Withdrawals, I received a number of comments at this blog and at Wade Pfau's Retirement Research Blog that began with, "The thing that's missing from this discussion is. . ."

Some mentioned that the topic of Required Minimum Distributions was missing. Others mentioned the risk of high long term care expenses late in life wasn't discussed. Truthfully, many risks are omitted from the discussion of retirement spending strategies.

There's a risk of very high but uncovered medical costs. A reader recently told me that he needs medication that needs to be prescribed off-label and will costs five figures a month that will not be covered by insurance. There is a risk of financially-debilitating divorce. There is liability risk. There is a risk that your children or grandchildren will need financial help. And, yes, there is a chance that you will have high long term care costs late in life. (The risk may not be as great as you have been led to believe, but that's another topic.)

(Wade Pfau just posted two excellent pieces on retirement risk. In my opinion, we spend way too much time talking about income and not enough about risk.)

The reason these risks weren't discussed is that the spending strategies I reviewed are intended primarily to mitigate only longevity risk, or the risk that you will run out of money before you die. A retirement plan should address all identifiable financial risks of retirement, but these four strategies aren't retirement plans, they're potential pieces of a comprehensive retirement plan.

I think this gets a little confused by systematic withdrawal strategies that lump all future liabilities into one pile and attempt to mitigate all risks by creating a large enough pile of money to address whatever pops up. That may be the most efficient way to create a lot of wealth, but it might not be the most effective way to prioritize, manage and meet liabilities. Many economists prefer to match resources with future liabilities so each can be addressed in he best way.

Let's take the LTC cost liability, for example, and assume that you want to save money to offset those potential costs. (I'm not sure that's a great approach for most people, but let's go with it for the moment.)

LTC costs, as I showed in a post about Long Term Care Insurance a while back, will be near zero for about 40% of retirees, manageable out of pocket for more, and catastrophic for a few. The spinal injury Christopher Reeve suffered, however, is estimated to cost about a million dollars for the first year of the injury and nearly $2M annually for subsequent years, according to the Christopher and Dana Reeve Foundation. That's an enormous range, from zero to $2M a year. And your LTC costs, if you experience them at all, may be decades in the future.

Hardly anyone can set aside millions of dollars "just in case".

In contrast, we can estimate annual "normal" retirement income needs relatively accurately and we know for certain that we will have those costs. The only real question is for how long. That is a very different animal than LTC cost risk and I prefer to address each with its own portfolio.

"Why do you recommend a separate portfolio?" one of my readers responded. "Why not just change your asset allocation?" another wondered.

Glad you asked.

First, because you need to decide what portion of your financial resources you are willing to dedicate to each goal. Most people can't secure their retirement income needs and save another several hundred thousand dollars to address a large liability they might never have. Building two portfolios forces you to decide how to allocate your investments. If forced to choose, I suspect most people will value sustainable retirement income more than avoiding spending the last of their wealth at the end of life.

Second, you need to decide on a portfolio asset allocation. If you want to minimize longevity risk, allocate something like 30% to 50% to stocks. That will dampen portfolio volatility and reduce the chances of depleting your portfolio, according to William Bengen in Conserving Client Portfolios During Retirement.

To have the best chance of growing a very large amount of money to cover large LTC costs 20 to 30 years in the future, you would want a much higher stock allocation, maybe 70% to 80% stocks. You can't do both with a single portfolio. A compromise might not meet either goal.

Third, we sell stocks and spend from a retirement income portfolio, but presumably not within an LTC savings portfolio, at least not for many years. A single portfolio would unnecessarily introduce sequence of returns risk into our LTC investments.

As you can see, the goals and risk and reward characteristics of a retirement income portfolio are different in many ways from those of a portfolio intended to save for a large, uncertain, future liability. And, that's why I would recommend separate portfolios.

Ultimately, what's important is how much money you have and not how many portfolios or accounts you divide it into. Your net worth will be a function of the weighted allocation of all your portfolios.

Instead of having two portfolios, one with 90% of your total assets invested in 60% stocks and 40% bonds and a second with 10% of your assets allocated 80% to stocks and 20% to bonds, for instance, you could have one portfolio with 62% stocks and 38% bonds (the weighted average allocation). Your expected returns would be the same in either case, as would your net worth's volatility.

But the same argument could be made for budgeting: why not just divide everything into "income" or "expense"? The reason is that you can't manage your finances very well that way.

Like putting all of your bills in a single pile and considering only their sum, by creating a single portfolio to address all liabilities, you would have difficulty understanding where you stood at any point in time in meeting your individual liabilities and which might be at risk. That's an important thing to know unless you value all of your liabilities equally.

Also, any actions you took within your single portfolio would have an equal effect on all of your goals, whether or not that is what you intended. It's a personal financial management issue.

So, why didn't I include LTC cost and other risks in the discussion of spending strategies? Because those risks are different than the longevity risk we try to mitigate with spending strategies. Spending strategies are about maximizing the amount of money you can spend each year and minimizing the risk that you will run out of savings prematurely. An LTC saving strategy would be about paying for a potentially very large, uncertain expense near the end of life.

LTC cost risk is just one example of a retirement risk that isn't addressed by a spending strategy, so I guess "the thing that's missing" from those strategies would be the rest of the plan.


  1. Many thanks, Dirk! I'm glad I posed some of those questions. You gave a thorough answer. You also got a new topic to write about. :-)

    Best regards.

  2. A few topics, actually. Thanks for commenting and I hope it helped!

  3. Thanks Dirk. A very lucid explanation. I have decided that ,for me , having a separate "bucket" for LTC ( choosing to self fund) is the most helpful way to go. I hope it can reduce some behavioral problems that I might otherwise be at risk of with just looking at "one big bucket".
    I have my LTC bucket at 80% stocks and hopefully will not need to touch it for 20 years (ideally never and then it becomes a legacy). If stocks tumble I will not be tempted to sell because know it has a very high chance of recovering before I need it. I have not considered more than 20% safe assets because ideally I would like this bucket to grow to provide even more LTC. If things do not "pan out" (such as needing LTC much earlier than expected, and just after a stock crash) then its not the end of the world: Medicaid is always there as a safety net. It may not be great but it is way better than nothing.
    For basic spending needs that need to last potentially a long life: these are in much safer investments: an annuity, eventually TIPS, CDs and Stable Value funds and only 30% stocks.
    There is a danger of this type of approach of "mental accounting". But for me the advantage is the peace of mind of having a safe spending bucket for basic needs, then a riskier bucket that will hopefully grow, but if it doesnt: "oh well". I think as long as one is aware of the dangers of "mental accounting" hopefully they can be avoided. Liability matching has a lot going for it as an approach. The "one big bucket" approach, I believe, is more likely to lead to misguided portfolio decisions.

  4. Thanks for the comment, Derek. I don't think of this as "mental accounting". That comes into play when you make a purchase or investment decision based on irrational criteria. I'm not suggesting that you can't spend LTC savings dollars for other expenses or spending-portfolio dollars for LTC expenses (that would be mental accounting); quite the opposite.

    I'm suggesting that by using different portfolios for outcomes that you value differently (lifetime spending vs. preserving end-of-life wealth in this case) you can better understand which liabilities are covered and which are not. If you decide late in retirement that you value additional spending more than you value a terminal portfolio, then you can start spending LTC money. Or, vice versa.

    Conceptually, it's no different than a floor-and-upside strategy where you invest capital for non-discretionary spending in safer assets. Some retirees might want to make certain their basic spending needs were covered before saving for possible LTC expenses.

    I'm mulling over a few potential future posts on this topic.