Advisor Perspectives welcomes guest contributions. The views presented here do not necessarily represent those of Advisor Perspectives.
Many advisors use the same end-of-life assumptions for all of their clients. However, there are numerous benefits to using the Actuarial Longevity Illustrator (ALI) in combination with other longevity calculators to develop personalized end-of-life assumptions and improved financial plans.
Dr. David Blanchett reaches several conclusions regarding selection of reasonable lifetime planning period assumptions for retired clients in his article, “How to Estimate ‘The End’ of Retirement”. These include:
- “Client-specific information can and should result in significantly different forecasted retirement periods.”
- “A better [planning] approach is to consider the probability of surviving to various ages and selecting some period longer than life expectancy to ensure a cushion will exist should the individual survive longer than average.”
- “Modeling periods longer than life expectancy is especially important for couples, since they typically must plan for the longer potential retirement period of either member, which introduces additional tail risk into planning.”
- “An example of an online tool, available for free, that provides insight into the probability distribution of survival for both an individual and a couple is the ‘Longevity Illustrator’ tool made available by the American Academy of Actuaries and the Society of Actuaries.”
For a hypothetical couple, both age 65, non-smokers with average health, Dr. Blanchett compared the 25% probability of survival with the 50% probability of survival from the “Planning Horizon” section of the Actuarial Longevity Illustrator (ALI). He concluded:
Adding five years to projected life expectancy for a single household and eight years to the longest life expectancy of either member of a joint household (or to each member if separate end ages are used), at retirement, is an appropriate retirement end age assumption if the outcome variable is the probability of success (using a Monte Carlo model) and only fixed periods can be modeled (mortality rates are not directly incorporated).
While I agree with the major conclusions presented in Dr. Blanchett’s article, I will suggest a few modifications.
For purposes of this article, I will refer to the longer periods based on lower probabilities of survival as lifetime planning periods (LPPs) versus periods based on 50% probabilities of survival (generally referred to as life expectancies). As noted by Dr. Blanchett, there can be a significant difference between LPPs and life expectancies. In addition, there can also be significant differences between life expectancies and LPPs based on gender, health status, or other factors. Good planning should reflect all these differences.
Like Dr. Blanchett, I believe selection of the 25% probability of survival from the planning horizon of the ALI produces more reasonable LPP assumptions than using the 50% probability of survival. However, some clients with different risk tolerances may wish to use lower or higher probabilities of survival assumptions in the model used for their planning. This decision will also be affected by the level of conservatism employed in the other assumptions used in the advisor’s model.
Actuarial Lifetime Illustrator (ALI)
The ALI is a free online tool developed jointly by the American Academy of Actuaries and the Society of Actuaries that can assist both advisors and their clients in better understanding longevity expectations. The ALI is found at http://www.longevityillustrator.org.
The projections in this tool are based upon mortality tables used by the Social Security Administration and are adjusted by Society of Actuaries members for expected future mortality improvement. Since it is based on Social Security mortality experience, it does not consider some significant factors that may affect client longevity. For this reason, I believe that the results of other available reliable longevity calculators should also be considered when developing reasonable LPP assumptions for clients.
Once you have entered preliminary data, the information used in the ALI is very easy to change. For example, if you want to see the impact of changing the health status of one of the members of the household from “Excellent Health” to “Average Health,” you can do this easily by changing the data in the box on the left of the screen and hitting the “update results” button.
The planning horizon section of the ALI produces numbers of years until expected demise for various probabilities of survival. Some retirement experts and financial advisors believe that probabilities of mortality or survival at every age are more sophisticated and should be incorporated into the process of determining how much households can afford to spend each year. However, I disagree. Pieces of us do not die each year. In any future year, we will either be 100% alive or we will be 100% dead during that year — we just don’t know in advance which year that is. Therefore, we must plan conservatively.
I propose a three-step process for developing reasonable longer-than-life-expectancy LPP assumptions for your clients.
Three-Step Process for Determining Reasonable LPP Assumptions
Step 1
Estimate client life expectancy (single or for each member of the couple) using the 50% probabilities of survival from the planning horizon of the ALI. Also estimate life expectancies from three or four other reliable life expectancy calculator tools available on the website that generally consider more factors than the ALI. Such tools include:
- Life Span Calculator (Northwestern Mutual)
- John Hancock Life Expectancy Calculator
- Blueprint Income
- Living to 100 Life Expectancy Calculator
- Big Life
- The Healthy Life Expectancy Calculator
One of the related benefits to your clients associated with this step may be to encourage them to engage in healthier lifestyles to increase their life expectancies.
Step 2
Compare the 50% probabilities of survival life expectancies from the planning horizon of the ALI with the average life expectancy results obtained from the internet calculators. Determine appropriate adjustments in your client’s current age for significant differences. This may look like adding years to the client’s current age if internet life expectancy results are generally shorter than ALI results, or by subtracting years from the client’s current age if internet results are generally longer.
For example, if the ALI indicates a life expectancy for a 65-year-old male in excellent health is 23 years but the average of internet calculators supports a 25-year life expectancy, then subtract 2 years from the male’s current age. Do the same thing for a spouse. I call these age adjustments “age setbacks” or “age set forwards.”
Step 3
Go back to the ALI and enter adjusted ages from Step 2 without changing the health status or gender. Update the results and take the adjusted results from the 25% probability of survival from the planning horizon section. This will give you one longer-than-life-expectancy LPP for a single client or four LPPs: “Person 1,” “Person 2,” “At least one Alive,” and “Both Alive” for a married couple.
Examples
Example 1
Harry and Sue have recently retired and are meeting with Betty, their financial advisor, to see how much they can afford to spend in retirement. Harry and Sue are 65 and 62, respectively. Both are non-smokers who believe they are in excellent health. They visit the ALI, which indicates that the 50% probability of survival for a 65-year-old non-smoking male in excellent health is 23 years (end age 88) and the 50% probability of survival for a 62-year-old non-smoking female in excellent health is 28 years (end age 90).
Several of the internet calculators, however, indicate that based on a number of other measures, including their healthy lifestyles, etc., Harry’s average life expectancy is around 91 and Sue’s is around 93 — both three years longer than developed by the ALI. Therefore, Betty will use a three-year age setback for both of them. She goes back to the ALI and changes Harry’s current age to 62 (still excellent health), Sue’s current age to 59 (still excellent health) and hits the “Update Results” button.
The resulting 25% probability LPPs from the ALI planning horizon are:
Harry: 31 years (age 96); Sue: 37 years (age 99). At least one alive: 38 years; both alive: 28 years.
Example 2
Betty now meets with Bert and Edith, who are the same ages as Harry and Sue, but both believe they are in poor health (but not smokers). Betty takes them to the ALI which indicates that the 50% probability of survival for a 65-year-old non-smoking make in poor health is 17 years (age 82) and the 50% probability of survival for a 62-year-old non-smoking female in poor health is 25 years (age 87).
The internet calculators, however, indicate that Bert and Edith’s health and other factors are actually worse than they thought and they both require an age set forward of three years to make their life expectancies consistent with the ALI generated life expectancies. Betty goes back to the ALI and changes Bert’s age from 65 to 68 and Edith’s age from 62 to 65 (still both in poor health) and hits the update results button.
The 25% probability LPPs for the ALI planning horizon for this couple are:
Bert: 20 years (age 85); Edith: 26 years (age 88). At least one alive: 27 years; both alive: 17 years
Financial Impact of Using LPP Assumptions Developed From 3-Step Process
All things being equal, clients with longer LPPs will require more funding to provide the same level of annual recurring expenses in retirement — or the same level of funding for a client with longer LPPs may generate lower amounts of annual recurring spending.
For the couples in the above examples, the following table shows the present values of providing $50,000 of real annual recurring spending based on a 5% investment return assumption, a 3% rate of inflation, and unreduced spending lasting as long as at least one member of the household remains alive (also known as “joint and 100% to the last survivor”). Also shown, for comparison, are present values assuming payment over other fixed periods of years.
This table illustrates that there can be a significant difference in the expected present values of recurring expenses even for same-age couples. Using the same fixed period of years for all of an advisor’s couples may not be sufficient to cover expected expenses for some couples and may be too high for others.
It should be noted, however, that the present value of expected recurring expenses is not a particularly good measure of the financial impact of assuming to live longer or shorter, as it does not reflect the expected impact of assuming longer or shorter LPPs on the couples’ present value of assets. A more meaningful measure would be a comparison of the expected impact of longer or shorter LPPs on the couples’ funded status (the ratio of the present value of assets to the present value of spending liabilities).
Reduction of Expenses Upon the First Death
LPPs from the ALI show that there may be a significant number of years between the first death within a couple (when they cease to be both alive) and upon the second death within a couple (when either of them ceases to be alive). The ALI indicates that this period of time for both of the hypothetical couples described above is 10 years at the 25% probability of survival level.
Instead of assuming that recurring expenses will remain the same after one partner dies, it is reasonable to discuss with clients how much they expect their recurring expenses will decrease upon the first death. By providing four LPPs, the ALI provides sufficient information to enable advisors to estimate the financial effect of assuming that there will be a decrease in real household expenses of a certain amount upon the first death within the couple.
For example, the advisor and clients may agree that recurring household spending will decrease by 33% upon the first death within the couple. Therefore, the present value of $50,000 in annual recurring expenses based on the assumptions above for Harry & Sue is reduced from $1,360,990 to $1,293,985. For Bert & Edith, it is reduced from $1,063,208 to $980,417.
These calculations can be confirmed using the Actuarial Financial Planner for Retired Couples. Calculations based on the ALI LPPs are sometimes off somewhat due to rounding of each of the LPPs to the nearest whole year.
Other Financial Implications of Planning to Live longer Than Life Expectancy
Prudent retirement planning requires planning to live longer than one’s life expectancy (whether that life expectancy is based on poor, average, or excellent health). Financial planning in or near retirement involves comparing household assets with household spending liabilities. I encourage advisors to annually calculate their client’s funded status, which is the present value of household assets divided by the present value of household spending liabilities.
If the client expects to live longer than their life expectancy and their health is not poor, the advisor should encourage the client to consider investing in assets that reward longevity to fund the present value of their essential expenses (and better match their spending liabilities). These assets include pensions, deferring commencement of Social Security, and life annuities. Purchasing or investing in these types of assets will frequently increase the client’s funded status by increasing the present value of their assets relative to the present value of their spending liabilities.
I frequently run across break-even analyses for deferral of Social Security or other analyses involving lifetime income products. Such analyses generally ignore the liability side of the planning equation and the need to plan to live longer than one’s life expectancy. Reasonable analysis of the financial viability of various types of retirement assets should be based on the client’s LPPs — not their life expectancies.
Periodic assessment of risks that assumptions used to determine client spending may not be realized in the future should also be part of prudent retirement planning. Therefore, an advisor should periodically stress test LPP assumptions by measuring the impact on the client’s funded status of assuming longer or shorter LPPs so that results may be discussed with the client and actions taken to manage this risk if necessary.
Selecting reasonable LPP assumptions should be part of an advisor’s annual review process. Mortality assumptions used in the ALI change from time to time, and the client’s health statuses may also change over time. In addition, the older a client becomes, the later their expected date of demise will generally be.
Summary
In order to develop prudent spending plans for clients that are consistent with their spending goals, advisors should employ reasonable client-specific LPP assumptions. The three-step process and tools highlighted in this article can help advisors get started. Furthermore, for married couples, advisors are encouraged to explore and develop reasonable assumptions regarding decreases in household expenses upon the first death so that the client’s spending needs in retirement are not unintentionally overstated.
Ken Steiner is a retired actuary with a website titled, "How Much Can I Afford to Spend in Retirement?"
