Life Timing: What Did Lynn Hopewell Teach Us?

HRE PR Pic 2013

Harold Evensky CFP® , AIF® Chairman

You’re not average, so don’t plan the quality of the rest of your life based on averages.

I was sitting front row center in a big conference room at our national planning symposium; I’d been looking forward to this talk for a while. The speaker, financial planner Lynn Hopewell, was a good friend and one of the most thoughtful practitioners I knew. My partner, Deena, and I had been responsible for planning this program and we invited Hopewell to speak because he told us he had a few major concepts he wanted to share with his peers. Here’s what he shared that day:

And End Not So Near

Welcome, everyone. I have few stories to tell that I hope will be a wake-up call for the financial planning profession. The first is about my planning for an engineering client, Ms. Jane. She is sixty-three, a very successful and accomplished civil engineer, and president of a major structural engineering firm. She hired me to work with her in developing a comprehensive retirement plan. Well, since I too am an old engineer, I know how they think—detail, detail! So I worked very hard to provide Ms. Jane a plan that would resonate with her. Finally, I was sitting down with her, ready to blow her socks off, and after going through my complete analysis, I thought I had.

“Mr. Hopewell,” she said, “I’m very impressed with the thoroughness and depth of your plan. I have only one small question.”

Well, needless to say, I was beaming at the compliment and looked forward to answering her “one small question.”

She went on, “I understand that selecting a mortality age—the age the plan assumes I die and will no longer need income—is a critical element in the planning process.

“Obviously,” I said, “if we arbitrarily use a very old age, such as one hundred, we’re likely to have to tell you to reduce your spending so that your nest egg will last to that age. Of course, if you die before one hundred, you’ll be leaving a lot of money on the table that you could have enjoyed spending while you were alive. If we assume a much younger age and you’re long-lived, the consequences could be even worse because you’d run out of money before you ran out of time. As a consequence, we work hard to select a reasonable planning age.”

“That makes sense to me,” she said, “and I understand that the age you selected for the plan is based on the projected age of my death from a national mortality table.”

“Correct! And not just any mortality table. We spent quite a bit of time consulting with actuaries to determine which table reflected the most current actuarial data.”

“I understood that. What I’m still a little confused about is the meaning of that age. As I understand it, if the table says my mortality age is eighty-eight, that means half the people will have died by eighty-eight and half will still be alive.”

“Correct.”

“Well, doesn’t that mean if I plan to age eighty-eight, I’ll have a 50 percent chance of outliving my plan?”

That question hit me like a Mack truck. Ms. Jane was correct. Even worse, in thinking about it, I realized that anyone with the resources to need the advice of a financial planner was likely to have had better health care and nutrition than the average of the universe of individuals making up the mortality table. That means Ms. Jane had better than a 50 percent chance of outliving my plan. This was a major wake-up call for me and should be for any practitioner relying on a traditional mortality table. Lynn said, “After acknowledging Ms. Jane’s point and scheduling a follow-up visit to give me time to consider the ramifications of her simple question, I hunkered down in my office to consider how I might resolve this problem.”

So, I went back to my own office and did the same. After additional conversations with my actuary friends, I concluded that a reasonable solution would be to use more customized actuarial tables—those that allowed me to factor in whether the client is a smoker, nonsmoker, her current health, and whether the lifespan of her immediate family is long, average, or short. Then, using the appropriate customized table, we would select an age that represented only a 30 percent chance of her outliving the age indicated in the table.

Here’s an example that shows how big a range the mortality age can be depending on these factors:

Life Timing Chapter Image file - mortality age range

Obviously, there is no guarantee that the age selected will coincide with the client’s mortality; however, following this process is likely to provide a much more realistic estimate.

Well, Lynn was right. That was a major wake-up call, because I’d been using a standard actuarial table and mortality age for my planning assumption. That was about to change.

Even if Lynn had stopped there, this information would have justified all of the time and cost of attending the three-day symposium, but there was more. Lynn’s next story was about the ah-ha moment he had one day when developing a college funding recommendation.

College Calculations

Not long ago I was preparing a simple college funding recommendation for a client. You know how that goes. It’s a simple time-value calculation that requires input on how many years until college, how many years of college the client wants to pay for, the annual cost, and the college tuition inflation rate. My input looks something like this:

Life Timing Chapter Image file - college calculations

A financial calculation would result in a recommendation that the client set aside about $145,000 to fund this expense. When I presented this to the client, he asked how confident I was about my number. When I thought about his question, I realized the answer was not very. My estimate was what we refer to as a “point estimate.” This means that unless every assumption I made was exactly right, my recommendation would either over- or underfund the college tuition bill.

As a former engineer, I remembered that when trying to estimate the probability of uncertain events, we used a technique known as a “Monte Carlo simulation.” Developed at Los Alamos National Laboratory during the Second World War for the design of nuclear weapons, Monte Carlo is really a simple concept. Rather than making a single guess regarding a possible outcome, we make guesses about the likely ranges of the outcomes. We then simulate thousands of possible futures with different combinations of those possible outcomes.

Let’s expand the table I showed a minute ago to more realistically reflect the uncertainty in our estimates.

What we know with some certainty:

  • Years to college 4
  • What we’re making an educated guess about
  • Tuition somewhere between
  • Annual cost $30,000 to $50,000
  • College costs inflation 5 to 7 percent
  • Investment return 6 to 10 percent

With these ranges, there are many thousands of possible outcomes, for example:

Life Timing Chapter Image file - college calculations no. 2

The Monte Carlo simulation calculates for each of these examples how much money that investors would need to set aside today if they want to fully fund four years of education. If the analysis ran a thousand examples, the results, listed in order of decreasing savings, might look something like this:

Life Timing Chapter Image file - college calculations no. 3

In this case, the question was how much should you put away now if you want an 80 percent probability of meeting your goal? The answer would be $167,000, because 80 percent (800/1,000) of the simulations would have succeeded with that amount of savings or less.

Well, this was another major wake-up call for me. In hindsight, it seemed obvious that a point estimate was inappropriate and that a Monte Carlo simulation could provide a more meaningful answer. In wrapping up his discussion, Lynn reminded us that expanding the input matrix meant making more guesses. Despite the mathematical rigor of a Monte Carlo simulation, adding more guesses does not justify adding two more decimal places to the answer. His point was that we should use Monte Carlo as an educational tool and not suggest it is a mathematically accurate answer.

The Takeaways

When planning retirement, don’t assume average mortality—you’re not average.

When attempting to quantify an uncertain future, don’t default to a single estimate. Use a Monte Carlo simulation to develop an understanding of the likelihood of possible outcomes, but don’t take the results as gospel.

This blog is a chapter from Harold Evensky’s “Hello Harold: A Veteran Financial Advisor Shares Stories to Help Make You Be a Better Investor”. Available for purchase on Amazon.