The Methuselah Effect: The Pernicious Effects of Unreported Deaths on Old-Age Mortality Estimates

• Activity Number: 68 - Government Health Statistics
• Type: Contributed
• Date/Time: Sunday, July 30, 2017 : 4:00 PM to 5:50 PM
• Sponsor: Government Statistics Section
• Author(s): Lynne Schofield* and Dan Black and Seth Sanders and Lowell Taylor and Yu-Chieh Hsu
• Companies: Swarthmore College and University of Chicago and Duke University and Carnegie Mellon University and NORC
• Keywords: mortality rates ; Hispanic paradox ; black-white crossover ; measurement error

Abstract

We examine the use of matching survey members to death records to make inference about old age mortality. We show that even relatively small rates of failure to match respondents may lead to substantial bias in estimates of old-age mortality rates. This measurement error has important consequences for three strands of the literature on older age mortality: 1) whether there is a deceleration in mortality rates at old ages, 2) the black-white mortality crossover and 3) the low rate of old age mortality of Hispanics (often called the "Hispanic Paradox"). Using the National Longitudinal Study of Older Men matched to death records in the U.S. Vital Statistics system and the Social Security Death Index, we demonstrate that even small rates of missing mortality data can 1) lead to mortality deceleration when none exists and 2) suggest a black-white mortality crossover when none exists if the rate of death undercount is higher for blacks than for whites. We confirm these findings using data from the National Health Interview Study matched to the U.S. Vital Statistics system. Moreover, we show that the Hispanic Paradox is also likely explained by a similar undercount.