Empirical Validation of HRS Pension Wealth Measures

On average, pension wealth amounts to a major fraction of household wealth with considerable variance at the individual level. Empirical analysis of household behavior therefore ought to account for it, but often does not, because of lack of appropriate data. The Health and Retirement Study (HRS) allows computing two measures of pension wealth: one derived from the self-reports and one that uses information on pension plan rules obtained from employer-provided documents. This study assesses the reliability of these two pension wealth measures on the basis of their explanatory power in three examples of economic frameworks. The analysis demonstrates that in straightforward analysis both measures may be thought of as reasonably powerful covariates. Yet, taking into account implications from the way the measures are constructed leads to quite different conclusions which favor the self-reports in empirical analysis.