What’s wrong with replacement rates?

I've written several papers on the use of replacement rates in retirement planning and, more specifically, in judging the adequacy of retirement income. This paper written with Glenn Springstead of SSA looked at different ways of defining replacement rates, and pointed out that private sector financial planner define replacement rates differently than the SSA does. (In the private sector the replacement rate is generally retirement income/income immediately preceding retirement; for SSA, it's retirement income/the wage-indexed average of lifetime earnings). So statements like,

While Social Security replaces about 40 percent of the average worker's pre-retirement earnings, most financial advisors say that you will need 70 percent or more of pre-retirement earnings to live comfortably.

just don't work, as they're calculated using different denominators.

This second paper looks at how adjusting replacement rates for household size and the presence of children affects how we measure retirement income adequacy. While these adjustments can shift replacement rates up or down, depending on circumstances, the paper found that the typical Social Security replacement rate rose from 49 to 63 percent when adjusted for household composition, while the typical Social Security-plus-pension replacement rate rose from 75 to 92 percent. (These figures for members of the 1940 birth cohort.) The takeaway is that a) household circumstances matter a lot, and b) current retirees may not be as badly off as we sometimes hear.

All of this is a prelude to an interview with University of Wisconsin economist John Karl Scholtz, who with his co-authors has looked closely at how we measure retirement income adequacy and how replacement rates measure up as a financial planning tool. Excerpts of the interview, in the current ProManage newsletter, follow:

Chuck Miller: What has been the replacement ratio "rule of thumb" and is the rule grounded in real research?

John Karl Scholz: As mentioned, a common financial planning rule of thumb is that households need to replace between 70 to 85 percent of preretirement income to be financially secure.

The replacement rate rule of thumb has some apparent logic to it. Target replacement rates are thought to be less than 100 percent for three main reasons. First, upon retirement, households typically face lower taxes than they face during their working years, if for no other reason than Social Security is more lightly taxed than wages and salaries. Second, households typically save less in retirement than they do during their working years, so saving is a smaller claim on available income. Third, work-related expenses generally fall in retirement.** Accounting for these diminished income needs, conventional financial planning advice suggests that people need to replace at least 70 percent of pre-retirement income in retirement to maintain accustomed living standards.

In fact, when we calculate the median replacement rate that arises from our lifecycle model, applied to a sample of households born before 1954, it is 0.68 – very close to the 70 percent rule-of-thumb. The problem, however, is that the range of "optimal" replacement rates is much larger than the 70 to 85 percent range commonly discussed. Some households should be saving more. Many can comfortably save less.

CM: Where does the rule come up short? What are some of the factors the rule fails to consider?

JKS: A large number of factors will affect optimal target replacement rates. Optimal rates will be larger for couples than for singles. The evolution of average tax rates will have a substantial effect on optimal replacement rates. The reduction in average tax rates over the period we study, particularly for affluent households, implies that replacement rates for high-income households are lower than they otherwise would be absent the tax changes. Of course, if taxes increase in the future, replacement rates will need to reflect tax increases that will be borne by high-income households. Earnings shocks, particularly those incurred after children have left the household will also have substantial effects on optimal target replacement rates. Shocks to earnings are common and persistent, which makes durable rules of thumb difficult to formulate.

Perhaps the biggest and most straightforward-to-understand limitation arises because of children. Financial planning rules of thumb do not vary with the number of children in a family. But the resources needed to equate the discounted marginal utility of consumption in retirement for parents (assuming an intact married couple) is smaller if household resources during the pre-retirement period were devoted, in part, to raising four children than if the couple was childless.

Put differently, an otherwise equivalent household with many children will have a smaller optimal replacement rate than their childless counterpart. Conceptually, the ages when children are born, due to the interactions of credit constraints and optimal consumption profiles, and the timing of income realizations, will also affect target replacement rates.

Check out the whole interview. Several of Scholtz's papers are available on his website. Here's a link to a paper specifically on replacement rates.