Discussant: John Karl Scholz, University of Wisconsin
Tom Steinmeier summarized the findings to date from the Health and Retirement Study (HRS). He said that only half of the people could correctly identify their pension type. Only roughly half come anywhere close to knowing the income stream they will receive from their DB pension. For DC pensions, only about 30% come anywhere close to knowing their account balance, 30% don’t know at all, and 30% systematically underreport their DC balances. He noted that some of the problems are: (1) The overall distributions of respondent reported versus administrative values is close, but there are large differences at the individual level; (2) Imputed values are noisy but not excessively so; and (3) Calculating pension incentives for research studies from respondent reports is exceedingly noisy.
John Karl Scholz said that as the paper evolves, it will be important to clarify the following: (1) What drives the discrepancies? (2) What questions suit current measures?; and (3) What are the policy consequences?
