A CLOSE LOOK AT DISSAVING IN THE LONGITUDINAL RETIREMENT HISTORY SURVEY

Hurd (1987) provided evidence from the Longitudinal Retirement History Survey that the retired elderly dissave. This paper attempts to refin e and extend that analysis in several ways. I determine how sensitive his results are to his method for estimating missing data. I calculate median dissaving rates, a more robust measure of dissaving. I provide confidence intervals for those rates. Finally, I compute estimates ba sed on a method that weights all households equally. The results are m ixed. Hurd's dissaving finding is confirmed in many specifications, bu t questions are raised about dissaving among couples and the method of elderly dissaving.