Why are long-term care costs so hard to predict?

There are a few reasons.  One, long-term care costs can vary from zero to well over a million dollars over the course of a client’s lifetime.  Two, long-term care costs are closely tied to the longevity of a client.  The longer a person lives, the more the chances of incurring a long-term care event.  A healthy person with a long expected lifespan will have more risk for long-term care costs than an unhealthy person with a short expected lifespan.  Third, a huge amount of actuarial calculations must take place to compute the probability distributions.  Fourth, an assessment of the longevity and propensity for long-term care of the client must be assessed. So a combination of the nature of long-term care costs and the work required to perform the computations is why these costs are hard to predict.

A long-term care policy from the client’s point of view

In my many years in insurance company actuarial departments, I have had the opportunity to participate in product development.  Always the concern was how profitable the products were as well as how much premium (and commission) will be generated (along with the questions of whether the product features will work on the administrative system and proposal system).  There were also questions of what interest rate and other guarantees were doable, as these were important pieces in the sales process.  Never was any talk given to how well the product performed from the point of view of the policyholder.  Did the product make the policyholder more well–off?

I have developed a program that gives vital information as to whether the purchase of long-term care insurance (either a policy or in a rider form)  is helpful to the client.  Let me explain the measures I use and how the results generally look.

I use a combination of three separate measures.  First, will the purchase of the product increase the chances that the client will meet goals, including the all-important goal of not outliving assets?  To answer this question, probability distributions of future long-term care costs are computed, customized to the client’s longevity and health profiles.  These distributions are then combined with many other aspects of the client’s financial situation (asset portfolio, investment/reinvestment/disinvestment assumptions, liabilities, expenses, tax and estate situations and many others) to compute the goal probability.  Then the process is repeated with long-term care insurance incorporated into the long-term care probability distributions.  The probabilities are compared.

The second measure is more detailed.  The measure addresses the questions: What are the chances that the client will spend less on long-term care costs over the client’s lifetime if the prospective insurance policy is purchased, as compared to not purchasing the insurance?  How much less?  For the chances of spending more, how much more?  The answer is displayed in a chart with the appropriate probabilities. 

So how do these comparisons usually work out?  For the first, the chances of success usually go down if the insurance is purchased.  Although unfortunate, the reduction is usually mild; say from 90% success chances without insurance to 86%( with insurance.  The second measure usually shows that about 65-85% of the time, the prospective insurance purchaser will pay less with insurance than without.  But the amount that they will save is usually on the small side.  On the other hand, if they do not purchase the insurance, 15-35% of the time they will pay more.  That  "extra" amount that they will have to pay can be huge.  Of course, this is the function of insurance, to protect against large losses.  The chart presents unbiased information about the policy, customized to the client, available nowhere else.  Often the client will see the big benefit of the insurance.


A third measure I recently developed deals with a major aspect of long-term care - how does the purchase of long-term care protection affect the amount of unpaid help to be provided by a spouse, by family and  by friends.  There are many articles as well as studies that talk about the great burden on the people close to the one who needs care.  My system actually quantifies this burden.  It produces a probability distribution of the amount of care that potentially could be provided by them, both with and without insurance (it assumes that care that would otherwise be provided by those close to the one who needs care will instead  be provided by professional care paid for by the insurance).  This analysis is only available through my program.  And just how well does the insurance reduce the burden for unpaid caregivers?  Contact me and I'll give you some interesting results!!

A few perils brought on by not incorporating the full range of your retiree client’s future health care costs into their retirement plans

Today’s retirement plans are often very sophisticated in many ways.  Estate plans, extensive asset analyses, what-if scenarios, tax analyses, spending broken down into much detail, percentage chances of meeting goals, and even a series of annual costs representing some of the health care costs your clients may incur are often parts of a modern retirement plan.  However, there is one critical component of costs that is mostly not fully recognized – the range of possible future health care costs, including the very variable long-term care costs.

Although all health care costs can significantly vary from year to year (hospital stays, the incurring of new chronic conditions and their associated prescription drug costs, and many others), the most variable is long-term care costs.  These long-term care costs can vary from zero to well over a million dollars over the course of your client’s lifetime.  It is impossible to pin down the costs to single amounts each year with any accuracy; the best that can be done is to produce probability distributions of future long-term care costs.

So, what can go wrong if the costs are depicted in the plan as something other than series of costs along with their associated probabilities of occurring?  Here are just a few:

1.       A long-term care event can easily exceed the amount assumed to be used for expenses in a given year.  The extra expenses can take a big bite out of the asset portfolio, leaving the client in a worse position than the plan’s projection.  The spending for all the previous years was calculated ignoring the long-term care event.   But the event may or may not occur, adding uncertainty that is not addressed in the retirement plan if probability distributions are not employed.

2.       The computation of the chances of success of the plan, including not outliving assets, is greatly influenced by long-term care costs.  My research shows that a computation of success without recognizing the full range of long-term care costs may be 90%; incorporating the full range of these costs can bring down the success rate down to under 50%!  The actual size of the asset portfolio is a key determinant of the impact of long-term care costs on success rates.  This is a complicated topic and I will have another blog on it.

3.       I can imagine the dissatisfaction of the clients who were expecting to not outlive assets but now find themselves in trouble because of long-term care costs (let alone their expected legacy!).  By recognizing these costs in the plan and adjusting spending, the chances of success can be computed more realistically. 

4.       The doozy is not to plan for the contingency that there will be someone who lives long with a long cognitive impairment. This often ruins the expected inheritance.  Incorporating probabilities of long-term care costs can give a more realistic picture of the size and chances of this contingency, as well as a more realistic picture of the client’s retirement.  Of course, to analyze this, the probabilities of living to various ages must be incorporated, preferably customized to the client’s health and longevity.

So what is the answer?  Prepare probability distributions of client’s health care costs, customizing them to the client’s longevity and health, and incorporating the distributions into the client’s retirement plans.