The aggregate decline in defined benefit pension coverage raises concerns that more retirees will face severe economic hardship later in retirement for two reasons: (1) DB pensions are a form of forced savings, so their decline reduces the probability of arriving at retirement with sufficient assets, and (2) the decline in DB pensions reduces the amount of annuity income one has to insure against outliving one’s savings. We propose to use Internal Revenue Service (IRS) tax data to measure how the trajectory of Americans’ income from pre-retirement through death has been changing over the last two decades. We will examine how the composition of income sources, the mean level of income, and cross-sectional inequality in income have evolved across birth cohorts at each age. We will also measure how the fraction of households that rely exclusively on Social Security for retirement income has been changing.
Although previous research has explored the evolution of late-life income, our use of IRS data makes this study unique. By using administrative tax data rather than survey data, we will be able to conduct our analysis without relying upon self-reports that are subject to significant reporting error, and our large sample sizes will allow us to conduct extensive subgroup analyses and explore the entire distribution of outcomes instead of just the mean. We already have access to the IRS data and are currently analyzing it for a currently funded RRC study. The proposed project will extend this existing project’s analysis to understand the dynamics of late-life income paths across birth cohorts. Our results will be written up in a manuscript.