There is a raging debate in the think tank world and among some academics that has spilled over to the op-ed pages: the question whether there is a looming retirement crisis for American workers. Both sides of the debate use the same underlying data sets and similar methodological approaches. One side says that most American households are woefully underprepared for retirement and that aggressive government policy, like increasing Social Security benefits or creating new mandatory government-run retirement plans, is needed. The other side says that the assumption-based crisis threat is substantially exaggerated. It contends that not only are no new government programs needed but that the existing ones, like Social Security and government employee pension plans, are them- selves weak reeds that must be repaired and trimmed soon.
Before this debate heated up, a colleague and I addressed the issue comprehensively with a better method than most of the cited studies. We came to the following somewhat paradoxical conclusions. The analysis indeed indicates that many American working households are not on the path to a well-funded retirement, but the solutions to this ‘‘crisis’’ lie in evolving and changing private behavior — longer working lives, higher savings, longer planning horizons, and more efficient investment strategies before and during retirement.
The Retirement Crisis Crowd
Among several analysts, the retirement crisis viewpoint is best represented by researchers at the Center for Retirement Research at Boston College, in particular through the periodic National Retirement Risk Index (NRRI) developed there.1 The basic method used is the following: Project a replacement rate (that is, retirement income as a share of pre- retirement income) based on current assets and income for each household in a nationally representative data set; construct a target replacement rate that would allow each household to maintain its pre-retirement standard of living in retirement; and compare the projected and target replacement rates to find the percentage of households at risk of inadequate retirement preparedness. Naturally, the first two steps in this process are based both on data and on many assumptions.
In the NRRI, relevant household assets used to produce retirement income include 401(k) plans and other retirement account assets, other financial wealth, and housing equity, less debt, as reported in the most recent Survey of Consumer Finances (SCF) conducted by the Federal Reserve Board. These assets are projected to retirement based on a stable relationship between wealth-to-income ratios and age evident in SCF data collected over the years. Retirement income is based on the asset conversion rates implicit in estimated prices for inflation- indexed immediate life annuities. For housing, there are two sources of income: the rental value derived from living in the house, and the annuity income converted from a lump sum borrowed from housing equity through a reverse mortgage. Other sources of retirement income include defined benefit pension income and Social Security benefits. Earnings before retirement are calculated by creating a wage-indexed earnings history and averaging these indexed wages over the individual’s lifetime. Interest on debt is subtracted from income, and investment income is added.
In the NRRI, the target replacement rate is generally less than full pre-retirement income, like 75 percent, to account for lower spending on taxes, savings, and mortgages in older ages. These target replacement rates are estimated for different types of households by a few broad demographic and income categories, with consideration of average pension coverage and homeownership.