The Official Poverty Measure (OPM) has existed since the 1960s, when it was developed by Mollie Orshansky, a researcher at the Social Security Administration at the time (Fisher, 2008). Though innovative in its time, the OPM (as it was enacted into federal statistical practices) has long been subjected to criticism on methodological terms. In response, the Census began releasing the Research Supplemental Poverty Measure (SPM) in 2011 (Short, 2011). The SPM makes a number of changes to the measurement of poverty, including: (a) altering the definition of the family to include not just those related by blood, marriage, or adoption but also cohabiting unmarried partners and their children; (b) basing poverty thresholds not on the cost of food in the 1960s but rather current consumer expenditures on food, clothing, shelter, and utilities (FCSU) plus a little more for other necessary expenses; (c) geographically adjusting poverty thresholds to account for variation in the cost of living across the United States; (d) including after-tax income and in-kind benefits in the counting of resources; and (e) subtracting so-called “non-discretionary” expenses like medical, work, and child care costs from income.

This last decision has been undoubtedly the most controversial, particularly when it comes to medical out-of-pocket expenses (MOOP). One notable finding under the SPM is that elderly poverty rates are much higher under the SPM than under the OPM, which is driven primarily by the decision to subtract MOOP expenses from income. For instance, Korenman and Remler (2013) show that, making all other changes under the SPM except subtracting MOOP would result in poverty rates that, if anything, would be below published OPM rates (8.6 percent vs. 9.0 percent in 2010). Meyer and Sullivan (2012) find that in the total population, those newly classified as poor under the SPM have substantially higher assets at the 75^th and 90^th percentiles than those who are poor under both measures and those who are newly classified as not poor under the SPM. This raises the question of whether the increases in elderly poverty under the SPM are “real,” or whether those elderly individuals classified as poor under the SPM really do not have sufficient resources to afford the levels of MOOP spending that they exhibit in the data. Ultimately this is an empirical question.

This paper utilizes data from the Health and Retirement Study (HRS), a nationally-representative survey of older Americans that collects detailed information on health, income, and wealth, to provide the first estimates that descriptively document the portfolios of various groups of elderly Americans classified as poor under different poverty measures. We first benchmark our SPM estimates against those found in the CPS, and then examine various assets among different groups of older Americans. As such, we provide critical information about the financial situations of the elderly in America, with implications for how to best measure poverty in the future.