By Henry K. Hebeler

3/8/05

I retired sixteen years ago, and many of our friends are in their eighties and some in their nineties. Surprisingly few of those we have known many years have died, I suppose because our friends are still quite active both mentally and physically, have good eating habits, don’t smoke (anymore), drink modestly or not at all, and enjoy family life.

However, after twenty years or so of retirement, the majority of our friends have not fared as well financially as they have health-wise. Although they started retirement with good pensions and significant savings, they attribute their current financial situation to having spent too much too early in retirement.

Now this isn’t because they didn’t do any planning nor because they lived very extravagantly. They were brought up during the great depression and generally think of themselves as conservative spenders. For the most part, they had college educations and held professional jobs. They were not the kind of people that live day-to-day on whatever their income happens to be at the time. They are the kind of people that make a real attempt to plan ahead and never buy things on credit.

Their financial problems are due partly to forecasting overly optimistic returns from their investments because they did not account for costs, reverse dollar-cost-averaging, and migration to more conservative investments. But there are other important contributing factors as well--such as unanticipated financial burdens from elderly parents who were running out of money, or, in far too many cases, divorced daughters with children who had little other means for support than to turn to their retired parents.

An equally important planning shortcoming was that their early plans did not account for lives that could well stretch into their nineties. Those who were engineers generally did some kind of spreadsheet analysis assuming returns they read about in the media (which were universally optimistic for retirees). The combination of underestimating life spans with very optimistic returns was devastating later in retirement.

The less technical among our group of our friends relied on retirement analyses using tables from magazines. These not only suggested optimistic returns and short life spans, they almost also used a "gap" analysis that effectively assumed that pensions all had cost-of-living adjustments (COLAs) when in fact their pensions were fixed monthly payments.

But there is a greater problem that few people recognize even now. That’s a problem with planning methodology. This is equally as dangerous as making a plan without some reserves for emergencies, using highly optimistic returns, forecasting short life-expectancies, and assuming all pensions have COLAs. Let’s review this methodology.

Then, as still is the case, almost all professional planners used "exhaustion" analyses to develop projections of the future for their clients. An exhaustion analysis is one where the planner determines how long retirement savings will last at a specified spending level. This spending is inflation adjusted so that constant purchasing power is preserved until the investments are completely withdrawn.

Some professional planners don’t even recognize that they are using exhaustion analysis because they use Monte Carlo programs that are sold as state-of-the-art forecasting programs. They simply fill in the data and print the results. These professionals don’t think about the fact that even Monte Carlo analyses are the product of thousands of exhaustion simulations. Each of the simulations that make up a Monte Carlo analysis assumes constant inflation-adjusted spending.

Most financial planners are so confident in their Monte Carlo projections that they start expressing the chances of not using up their investments as a percentage, e.g., "If you spend $50,000 a year, adjusted for future inflation, you will have an 80% chance of not running out of money before you die." Of course this is absurd! This has two flaws that are devastating.

The first flaw is the assumption that the future will be the same as the past. This implies that, for example, the next twenty years or so will have the same statistics for both returns and inflation (and their correlation) as did the entire period of recorded investment history behind us. I remember Paul Samuelson, an economics Nobel Prize winner, telling our class that history is only one data point, and you can’t predict the future by extrapolating from one data point.

The second flaw is equally devastating, and it applies whether or not the professional is using Monte Carlo analysis. That flaw is the basic assumption of constant inflation-adjusted spending. This means that retirees would not adjust their affordable expense budgets either when their investment values go up or down enormously or when they finally get old enough to recognize that they are likely to live beyond the original period used in the exhaustion computation. This is equivalent to a retiree ignoring huge investment losses (as experienced by many who retired in 2000) while blithely continuing to increase spending each year thereby accelerating the exhaustion process.

Or consider how ridiculous it would be for an eighty year old person to rely on a planner’s analysis done at age sixty five. At sixty five, the planner may have assumed a twenty-one year life expectancy and believed dying at age eighty-six would be conservative. However, the eighty year old person now sees that she could well live more than six additional years and so must stretch the savings further or plan on being a pauper in six more years. This is not a theoretical point to many of my friends who have now advanced into their eighties and are now experiencing these consequences.

If, for no other reason than to get additional fees, planners would like to see their clients make a new projection every year. Isn’t there something inconsistent between basing a plan on not seeing the planner ever again vs. asking the client to return for a new plan every year? Why then don’t planners base their forecasts using math models that simply include a revised spending level for each year in the future? This is not difficult to do, yet virtually the entire planning community is ignoring this more realistic approach.

A mathematician would say that such a projection involves two dependent variables: spending and investments. Exhaustion analysis has only one dependent variable: investments. A client would see the difference easily because (1) investments would never be exhausted, and (2) there would be an additional graph that shows what happens to spending in each future year. The client’s spending will go up, down, and sideways depending on the return and inflation assumptions. Thus the spending will be "dynamic," that is, will not remain at a "static" value, using terms familiar to your high school physics teacher.

Even with the oversimplified assumption that returns and inflation will remain constant until eternity, annual spending will change every year in a dynamic analysis, and investments will never be exhausted. To illustrate this we have to make some life-expectancy assumptions. Remember the eighty year old now thinking that exhausting investments at age 86 was a bad assumption? Let’s begin by assuming that a professional planner would add five years to the IRS life expectancy thinking that would be conservative. After all, that is five years beyond the life span of the average person. That would give age dependent values as illustrated in Figure 1 which many planners would consider adequate.

Figure 1. You can never outlive your life-expectancy, but you could outlive your money.

In order to illustrate the misleading attributes of exhaustion analysis, let’s first consider a very simple situation using the following assumptions:

• Retire at age 65 with $1,000,000 investments for retirement expenses
• 5.5% constant return, 3% constant inflation, and no income tax.
• Add 5 years to the IRS life-expectancy for each age in the calculation.

Now, let’s compare pure exhaustion analysis with dynamic analysis in Figures 2 and 3 below for affordable expenses and investment history. In the exhaustion analysis, the planner assumes that spending will continue at a "real" (inflation adjusted) value of about $52,000 a year. In the dynamic analysis, the real value comes down each year as the perception of potential increased life span grows.

Figure 2. As retirees age, they continue to get closer to the point where exhaustion analysis will be devastating if they live longer than originally planned.

Figure 3. With exhaustion analysis, investments go to zero at age 91 as planned. With dynamic analysis, there is always some money left for potential longer life spans.

It is apparent from both Figures 2 and 3 that exhaustion analysis is not suited to determine a budget for those who may well live five years longer than the average person as represented by the term "life expectancy." So how much would we have to reduce initial spending to have relatively constant spending, even late in life?

The answer is a LOT! The amount depends on the actual return and inflation as well as the assumptions about return and inflation, even in this case where we assume the planner knows exactly what future returns and inflation will be. Don’t snicker. There are many planners who will do the analysis just as if they know future economic conditions. If they were more humble, they would use more conservative assumptions for the payment forecasting equations than in the compounding equations for investments.

Figures 4 and 5 illustrate what would happen in this highly simplified case if the planner used more conservative payment forecasting equations, thus reducing spending about 32% each year, or $36,000 in the first year instead of $52,000.

Figure 4. Dynamic analysis shows that, inevitably, retirees’ budgets will change annually even when economic conditions are constant and the forecasts are conservative.

Figure 5. Reduced spending early provides money for longer life spans.

A 32% spending reduction from conventional forecasts would correspond to reducing the return assumptions used in the forecasting equations by almost 3% or increasing the life expectancy to 47 years, i.e., age 112 for a person now age 65. This will sound like an enormous amount of conservatism to young people including young planners. However, my many friends in their eighties, who now recognize that they overspent early in retirement, would have appreciated more conservative spending guidance.

The situation can deteriorate even more using scenarios with time varying returns and inflation as in either actual historical scenarios or Monte Carlo analyses. One of the worst years to retire was 1965 because it was followed by several market failures and a period of high inflation. Figures 6 illustrates how much trouble conventional exhaustion analysis would have brought in a 1965 scenario based on an allocation of 40% S&P 500 stocks, 50% long-term corporate bonds, and 10% short-term treasuries with costs of 1.5%, 1.0% and 0.5% respectively. Figure 7 shows the importance of using both dynamic planning analysis AND conservative planning assumptions.

Figure 6. Those who retired in 1965 with the subsequent market failures and high inflation would have been decimated at age 82 by conventional exhaustion analysis.

Figure 7. Those who retired in 1965, and reduced expenses by 32% from conventional planning analysis, would have weathered the economics quite tolerably using dynamic analysis.

Many of my older friends have fixed pensions. Inflation’s destruction of fixed payments exacerbates their problems from spending down their investments too early. Numerous on-line and magazine publications still do not include the methodology to encourage fixed pension recipients to always save part of their payments to compensate for inflation later on. At the very least, fixed pension or annuity holders should be advised to spend no more than the after-tax value of their payments multiplied times their age divided by 100. For example, a 65 year old retiree would spend only 65/100 or 65% of the after-tax value of her fixed pension or annuity and would save 35% to help compensate for later inflation. A 90 year old person could spend 90% of the after-tax value, but would still have to save 10%. However, draws from the accumulated savings from previous years would more than compensate.

Many of my elderly friends now have financial problems because their plans failed to include some emergency reserves, used returns that failed to account for costs and reverse dollar-cost-averaging, underestimated their life spans, employed gap analysis that did not properly account for fixed pensions, and based the analysis exclusively on methodology which assumes constant inflation adjusted spending until investments are exhausted—presumably on the day they die.

Still today, the majority of computer retirement programs and many professional planners are misleading people with oversimplified exhaustion assumptions even in supposedly sophisticated Monte Carlo analyses. Instead, planners should use dynamic analyses which assume that people will adjust their spending upward when investments soar in value and reduce spending after major market declines. Such projections are harder to interpret because the results are not simply a plot showing investments disappearing on the day the retiree dies. Instead, there will always be some residual investment level at death and annual spending may go up and down with time in the forecast even if, by some magic, returns and inflation would be constant values throughout retirement.

We have shown that dynamic analysis may offer retirement spending budgets perhaps one-third less than exhaustion analysis. Even then, and even if Monte Carlo analyses are modified to use dynamic equations, planners should never tell their clients, nor should the media let their users believe, that retirees have a certain percentage chance of not exhausting their investments. The best they can say is that if the future economics are better than past economics, the projections of affordable spending and investments should turn out to be conservative. They should also say that preserving investments until death requires a new forecast every year, or even more frequently if investment values are declining rapidly.

Finally, those who do their planning, using exhaustion methods currently available from the media or even the more sophisticated Monte Carlo analyses, should probably reduce their retirement draws from investments by almost a third early in retirement to preserve funds that will be needed should they live into their nineties. Such advice almost twenty years ago would have preserved the funds needed for my elderly friends’ comfort and well being.

________________________

Note: All figures were derived from the Dynamic Financial Planning program from www.analyzenow.com

RETURN