What measures are useful in retirement planning besides the probability of goal success?

When analyzing whether a financial plan meets the retirement goals of the client, there are two measurements, other than the probability of whether the client meets goals, that I find useful.  They are both related to what I would call the "magnitude" of success.

One is the number of years the client is alive and without assets.  This I call the number of years "in ruin".  A plan with a 95% success rate, but with an average number of years of ruin of 7 for the 5% of the time the plan fails, might be not as good as a plan with a 90% success rate, but with an average number of years of ruin of 2 for the 10% of the time the plan fails (the client can decide which plan is better).

The other measure is, for the scenarios that are successful, the amount (either an average or a probability distribution) of assets remaining at the time of death.  This can be considered a potential "legacy" for the heirs of the client, or as a "margin of error" for the various scenarios.

These two additional measures gives a much fuller picture of the possible outcomes for the client.

Safety First Retirement Plans and Health Care Costs – Are these two items compatible?

Two major approaches for retiree financial planning are the creation of (1) Probability Based Retirement Plans and (2) Safety First Retirement Plans

Very briefly, what are these two approaches?

Probability Based Retirement Plans calculate (usually using Monte Carlo techniques) the probability that the client will meet his various financial goals, including the goal of not outliving his assets.  This technique takes into (or should take into) account all of the financial aspects of the retiree’s plan.  The goals can be examined together or prioritized and examined separately.

Safety First Retirement Plans involve segregating the goals of the retiree into needs and wants.   The plan first makes sure that the needs are met.  Assets are devoted to the satisfaction of the needs.  Often guaranteed annuity/insurance products are employed to meet these goals.  Inherent in these strategies is the assumption that the amount of assets to meet goals can be determined or at least estimated within a small range.  Once the needs are met, then the wants can be examined.  The wants could be or not be satisfied, depending on many factors, including the size of the client’s asset portfolio.

The trouble with safety first strategies is that the requirement that the amount of assets needed to meet future health costs be determined or  at least estimated within a small range is not met.  Health care costs can vary a great deal over the course of a retiree’s lifetime, from a small amount to well over a million dollars.  That makes it almost impossible to rationally allocate the amount of assets to that need (and health care costs are a need and not a want).

Is it possible to employ insurance or annuities to reduce the variability of costs associated with health care? 

The retiree can buy Medicare Part D or a Medicare Advantage Plan, as well as Medicare Supplement Part F.  In some cases, this would reduce some of the variability associated with some health care costs, but not fully.

However, it is rarely possible to reduce the variability of the costs associated with long-term care.  The only possible way I see is to purchase long-term care insurance with an unlimited benefit period and with a 5% inflation rider.  Unpaid help is not a reasonable approach for the most difficult and costly long-term care needed.

The troubles with this insurance approach are:

More than half of retirees are simply not insurable for long-term care insurance.

Unlimited benefit periods are rarely available now, partly because the insurance companies who have sold this have lost money on unlimited benefit policies.

The premiums are very high and not suitable for all but the very wealthy.  Also, there is the real risk that the insurance company may raise the premiums after purchase.

Certain long-term care events are not covered by most long-term care policies

In summary, safety first strategies run into problems, often insurmountable, because of the variable and potentially extremely high cost of long-term care.

Therefore, I strongly recommend retirement plans constructed using safety first strategies be tested using probability tests, due to health care costs! 

Only PDRP Plus has the capability to correctly probability test plans, since it is the only system to correctly incorporate health care costs into its Monte Carlo testing.

Is there an intuitive way to understand the effect of health care costs on safe withdrawal rates? a/k/a "not convinced that health care costs greatly affect safe withdrawal rates?!"

Think of this topic this way: Fidelity Investments publishes annually an amount of assets that has to be set aside from a portfolio to pay for future health care costs.  Although their calculation does not include long-term care costs, and is assumed to be for a couple with a fixed set of conditions (fixed life expectancy, no chronic conditions and others), it is a good number to use for this blog.  The asset amount has been above $200,000, but I’ll use exactly that number to make the example easy.

First, assume a couple aged 65 has a portfolio of $600,000.  $200,000 is earmarked for health care costs, leaving $400,000 for spending.  Assume the 4% rule is applicable to that amount.   4% of $400,000 is $16,000, which is the amount of spending allowed.  But $16,000 is 2.67% of $600,000.  So a safe withdrawal rate here is 2.67% and not 4%.

Now assume that couple has a portfolio of $1,200,000.  The amount left after $200,000 is set aside for health care costs is $1,000,000.  4% of $1,000,000 is $40,000.  So the safe withdrawal rate percentage of the portfolio is $40,000 divided by $1,200,000, or 3.33%.

This is very rough, but is a good example of how health care costs can affect safe withdrawal rates downward, and how the calculation is specific to the individual retiree (much more so than is illustrated here!).

What is wrong is using a single date of death assumption in your analysis of your retiree client's goals?

The main advantage of using a single age at death assumption is that it is makes the calculations much simpler and faster.

There are some serious drawbacks to the use of a single age at death assumption, however.

In general, there are many financial aspects of retirement that are not analyzed if the fact that the client can die at any age is ignored.  The client needs to have an analysis that is consistent with this fact of life.  There is no substitute for planning that recognizes the client’s life expectancy and the associated probabilities of living to various ages.

Different pitfalls arise depending on how the single age at death chosen compares to the true life expectancy (and the probabilities of living to various ages) of the client.  I’ll examine a few of these pitfalls (there are many more I’m not covering in this brief blog) in the different possible cases – (1) when the client’s true life expectancy is greater than the chosen age, (2) when the client’s true age is less than the chosen age, and (3) even when the client’s true life expectancy is equal to the chosen age.  Without loss of generality, I’ll use a current client age of 65 and a true life expectancy of 20 years (bringing the client’s age at death to 85):

 IF THE SINGLE CHOSEN AGE AT DEATH IS LOWER THEN THE TRUE LIFE EXPECTANCY OF THE CLIENT (AGE 80 IS USED):

The annual expense budget will be overstated.  The spending may be designed to last to age 80, but the client will live on average to age 85.  The client’s asset portfolio could become depleted, even though spending recommendations were followed.

Health care costs will be understated (assuming they are even addressed).  The issue of needing long-term care after age 80 will be ignored.  The client may very well live past age 80 and may incur high health care costs.  This could greatly increase the chance of financial ruin.

The analysis of whether immediate or deferred income annuities are helpful will be incorrect assuming the client will die at age 80 in all instances – the analysis would show that annuities would be harmful to the client when they may actually be helpful.

IF THE SINGLE CHOSEN AGE AT DEATH IS HIGHER THEN THE TRUE LIFE EXPECTANCY OF THE CLIENT (AGE 95 IS USED):

The annual expense budget will be understated.  This would cause an unnecessary cutback in the client’s lifestyle.  Sometimes an adviser will set a very high age “so that the client can sleep at night”.  But this gives the client wrong information.  

Health care costs will be overstated (assuming they are addressed, which of course they absolutely should be).  There will be an overstatement of the long-term care and other health care needs.

There will be too good a result shown for immediate and deferred income annuities.  They will be shown to pay out more in benefits than should be shown.

IF THE SINGLE CHOSEN AGE AT DEATH IS EQUAL TO THE TRUE LIFE EXPECTANCY OF THE CLIENT (AGE 85 IS USED):

Health care costs will probably (if the life expectancy is not extremely high) be understated (again, assuming they are addressed).  Health care costs are greatest at the high ages, and assuming the client dies just at the life expectancy will probably understate these costs greatly.  

The analysis of whether immediate or deferred income annuities are helpful will probably be incorrect, even though the correct life expectancy is used.  They will generally illustrate worse than they should, and it is unclear whether the analysis will correctly show if the annuity would be helpful or not.  The actual extent that they would be helpful or harmful will probably be incorrect, no matter whether the direction the analysis gives is correct.

In summary, these drawbacks are serious.  Clients are not getting the information they need to properly plan unless the probability of living to different ages, based on the life expectancy of the client, is incorporated into the planning!

Budgeting for a retiree, after solving for the safe withdrawal percentage incorporating health care costs- what is different?

Some retirees use a safe withdrawal rate to determine how much they can withdrawal from their portfolio each year.  This is done so that the retiree will not outlive his/her assets.  A budget for the retiree can be done in accordance with the safe withdrawal to give guidance to the client.

How does budgeting differ for a retiree when health care costs are separately calculated (as they are in PDRP Plus)?

Budgeting for retirees has traditionally involved projecting the expenses for the retiree for the upcoming period of time, usually one calendar year.  Typical expenses would include food (at home and at restaurants), utilities, car expenses, vacations, real estate taxes,gifts to charities and many other items. This budgeting generally includes health care costs.  Of course, that is a problem with budgets, since the healthcare costs can vary a lot (due to the uncertain risks of long-term care, hospital and doctor visits and drug costs), and the budget will not always predict the costs accurately.

BUT:

Because PDRP Plus includes health care costs separately, there is no need to budget for these unpredictable costs.  By living within the safe withdrawal rate computed by PDRP Plus, there is much more certainty that the goal of not outliving assets will be met than by budgeting some fixed amount of healthcare expenses and hoping for the best.

TO REALLY BE TECHNICAL:

The following expenses are “built in” to the safe withdrawal rate computed by PDRP Plus.  The retiree does not have to budget the following items:

Health care costs, other than dental and eye care

Federal and state income taxes

Federal estate tax

State estate and/or inheritance tax

This simplifies the budgeting process and reduces guesswork!