From the August 2008 issue of Research Magazine • Subscribe!

August 1, 2008

Investor Spending: Risk

A generic retirement planning process should first build a floor under the retiree's income and spending risks. In a second stage, it should create upside consistent with preserving the floor. This article looks at the retirement planning process from the perspective of what investors do with their retirement income, their spending patterns and their spending risk.

The purpose of a retirement plan is to improve the investor's ability to adapt productively to change. The purpose of a retirement plan is to create better outcomes than would be available in the absence of a plan. One such outcome is the provision of a floor, a reliable level of retirement income. However, there seems to be many factors that limit the effectiveness of such purpose and intent. Such factors include risks, uncertainties and unknowns.

Risks refer to situations where the outcome cannot be known with certainty but where the probability of any given outcome can be known. Uncertainties refer to situations where neither the outcome nor the probability of any given outcome can be established with certainty. Unknowns speak for themselves. Clearly there are risks, uncertainties and unknowns that limit our ability to predict and control retirement planning outcomes, including the retiree's spending risk.

The real world is rarely a simple story. The number of variables that go into a retirement plan is most likely greater than the number of variables that go into an investment plan. The number of combination of the variables that go into a retirement plan can rapidly become too great for computation. For contrast, consider that an investment plan for an accumulation portfolio of 100 risky assets requires individual estimates of more than 5,000 variables (e.g., expected returns, standard deviations and covariances) before computation starts. How much larger can this number of required estimates grow for a retirement plan that not only takes financial capital into consideration but also includes the retirees' human capital and social capital?

Furthermore, the degree of uncertainty associated with some of the variables and estimates (e.g., future tax rates) may not allow for the computation of an optimal solution. Some variables may be estimated to the 20th decimal (e.g., probability densities). Others may only be identified at the level of the first integer (e.g., expected longevity). Some may even be imponderables (e.g., the relevant rate of inflation, 10 years from now).

Consider the following: In games like chess where all relevant information is available to all observers, computing an optimal solution can only be done in a very small number of all possible positions. Given the risks, uncertainties and unknowns that go into a retirement plan, we can expect that such a plan is considerably more complicated than a game of chess. We can also expect that we will not be able to compute an optimal solution. So, what can we do, and in particular, what can we do about the investor's spending risk? We need to form an opinion. We need to learn to form an opinion based on the best available data and the most plausible and professional reasoning.

Varying Over Time The Bureau of Economic Analysis' NIPA tables make it possible to show spending on discretionary items as a percentage of all personal consumption expenditures (PCE). For instance Figure 1.2, on page 5 of Pamela Danziger's book Why People Buy Things They Don't Need shows, for each decade, the evolving share of consumer expenditures from 1930 to 2000.

o Durable items such as cars and furniture increased from a 10.3 percent share of PCE in 1930 to a 12.2 percent share in 2000.

o Non-durables such as food, clothing, gas and oil decreased as a share of PCE from 48.4 percent to 29.6 percent.

o Services such as transportation, recreation and medical care increased from 41.3 percent to 56.2 percent.

This means that we spend slightly more on cars and furniture as a percent of our total budget than our grandparents did. On the other hand, we spend much less on food and clothing as a percent of our total budget. However, heath care takes a vastly greater percentage of our budget than theirs. So, how will our collective spending patterns evolve over the rest of our time in retirement? What fashion, ideas, technologies will change the relative weights of the retiree's budget categories and expenditures?

Varying With Income Stephen J. Rose in the 2007 edition of Social Stratification in the United States shows a poverty line threshold of $17,055 for a family of three people. He also identifies a low-budget threshold ($29,847), a medium-budget ($46,902) and a high-budget ($72,485). Spending patterns vary from budget to budget. For instance,

o The low-budget family most likely rents housing without air conditioning, performs its maintenance and repairs, eats less meat than at the higher budget levels but eats more potatoes, drinks less wine and liquor but drinks more beer.

o The medium-budget family probably owns its own home but has many years left on the mortgage, buys used cars and keeps cars for a long time.

o The high-budget family owns its own home, buys new cars, has a complete set of appliances, pays for household services rather than performing maintenance and repairs, and takes vacations abroad.

Such budget levels when mapped against the number of households create heavily skewed, non-normal curves of the types discussed in my May column. Many people have the lower-level budget, while few people have the higher-level budget. As a result, how many categories of retirement budgets should we create?

Varying Over LifecyclesHarry S. Dent Jr. in The Roaring 2000s developed charts showing average annual family spending. For example, Figure 1.5 on page 37 shows average annual family spending peaking just before age 50 at $46,500 and declining steadily thereafter to reach a low of just about $10,000 at age 80. Not only will spending patterns reflect the changing fashion and technologies of the unfolding decades, not only will specific categories of retirement budgets drive certain types of expenditures, but budget levels will also change as investors move through the rest of their lifecycle.

Spending patterns can also be mapped into various psychological categories. According to Pamela Danziger, 18 percent of consumers are "Impulsives," 17 percent are "Conflicted," 15 percent are "Bargain Hunters," 28 percent are "Self-Expressives" and 23 percent are "Careful Indulgers." This brings the discussion back to behavioral finance. Not only is the world full of risks, uncertainties and unknowns but we are not all that good at perceiving these correctly. On the other hand, we are very good at seeing what we expect to see as opposed to what is really happening.

This is one of the areas where advisors can help. Steve Mahalec, an investment representative with Fidelity Investments, makes an interesting observation: "Spenders are spenders, savers are savers. Those who spend in accumulation spend in retirement. Those who save in accumulation save in retirement. I have to create two different types of asset allocations for these two types of clients."

Retirement management professionals are increasingly asked to help investors with the "soft" aspects of retirement planning. This includes helping investors put a ceiling on their spending risk. Investors and advisors who can perceive and integrate spending risks, uncertainties and unknowns better than others will have a great comparative advantage. Adapting productively to change is even more valuable when others do not. Ask yourself, how does your advice increase your client's ability to adapt productively to spending changes? How does it create better spending outcomes than would be available in the absence of a retirement plan?


Francois Gadenne is chairman and executive director of the Retirement Income Industry Association in Boston; see

Reprints Discuss this story
This is where the comments go.