The income replacement rate (IRR)—income immediately following retirement divided by income immediately preceding retirement—is commonly used as a measure of economic preparation for retirement. Various online calculators, for example, suggest that retirement income replacement rates should exceed 70% of pre-retirement income.

In “Measuring Economic Preparation for Retirement: Income Versus Consumption,” a paper by Michael D. Hurd and Susann Rohwedder, two senior economists at RAND Corporation, the researchers contend that the widely accepted concept omits a number of relevant issues. They consider two variations of IRRs to address some of the complexities of contemporary labor and investment markets to workers nearing retirement.

Their finding’s imply that the income replacement rate is of little use for assessing economic preparation for retirement. The chances that someone with a low income replacement rate is well prepared are not much different from the chances that someone with a high income replacement rate is well prepared, they contend.

The standard calculations may have made sense in an era when preretirement income almost exclusively comprised earnings and when post-retirement income comprised Social Security and defined benefit (DB) pensions. The formula’s simplicity and transparency of the concept have contributed to its use.

For one thing, defined contribution (DC) plans have supplanted most DB plans—and DC values depend not just on the plan participant’s contributions, but also on movements in the markets in which those plans are invested.

Some people, for instance, make nontraditional transitions from full employment to full retirement. Members of a married couple transition to retirement at different times. Many people have other forms of financial wealth that are not always thought of as sources of income but that could serve as such.

In its simplest form the IRR has not been redefined to accommodate these other potential income sources. As a result, the IRR understates the degree to which workers have adequately prepared economically for retirement relative to enhanced forms that do account for those income sources.

Post-retirement income options have also diversified to the extent that one household may have two earners with retirement ages that differ, either because the two differ substantially in age, or because one prefers to retire at a different age. Given the complexity added by a second earner, determining the timing of a couple’s retirement and quantifying pre- and post-retirement incomes is a difficult planning problem that cannot be solved by a simple ratio with a few values.

To address individual retirement accounts (IRAs) and other sources of wealth that could produce post-retirement income, the researchers assumed an annual 4% drawdown of financial assets and of IRAs—a rate considered prudent by financial advisers. This allowed Hurd and Rohwedder to account for possession of financial wealth, which can result in considerable improvement to economic preparation for retirement.

Annuitization of wealth resulted in even higher levels of preparation for retirement, the paper said, but also noted that annuitization of financial assets is rare—which made the 4% drawdown more relevant. They emphasized a 4% drawdown in their comparisons.

The researchers also compared economic preparation for retirement using the various IRRs and economic preparation factoring in consumption-based measures. The consumption-based measure is theoretically preferable, the paper says, because consumption is more likely to translate directly into well-being than income, which has to be consumed. The paper’s estimated consumption-based measure indicates retirement preparation at 59% for single people, well over the quantity derived from IRRs. For couples, the consumption-based retirement preparation rate is a much higher 81%.

The lack of a sensible or anticipated relationship between singles’ IRRs and those of couples argues against putting much stock in these ratios as indicative of retirement preparation adequacy. The consumption-based measure does differ in the expected way across singles and couples. Moreover, there is little relationship between the income replacement measures and the consumption-based measures. Adequacy of preparation, as measured by the consumption-based measure, does not increase substantially with increasing income replacement ratio, regardless of how it may be modified for improvement. The researchers conclude that neither the IRR nor the modifications they considered is a good guide to economic preparation for retirement of a household: either one may be quite misleading, they say, concluding that other consumption-based metrics should be put to wider use.

“Measuring Economic Preparation for Retirement: Income Versus Consumption” can be accessed online, free of charge.