Savings and retirement in the new millennium

The typical pension plan has changed dramatically in the past twenty-five years. There has been a large increase in the coverage of defined contribution (DC) plans, in which the pension is generally paid in unannuitized form, and where the amount of the pension depends on investment returns and the amount contributed. There has been a corresponding decline in the coverage of defined benefit (DB) plans in which the pension is some function of salary and years of service.; DC plans place greater responsibility on the individual to make adequate provision for his retirement. They also lack many of the age related incentives of DB plans that have been shown to influence retirement.; The first chapter, “The Effect of 401(k) Eligibility on Household Asset Holdings: New Evidence from the Health and Retirement Study” examines the question of whether contributions to 401(k) plans, the most common type of DC plan, have merely displaced other non pension savings, or whether they represent additional saving that would not otherwise have been made. A comparison of the savings of participants with those of non-participants would likely produce biased estimates, as participants probably have a stronger unobserved taste for savings than non-participants. I therefore follow previous practice in comparing eligible with ineligible households. Eligibility is, however correlated with DB plan membership, which may also affect savings behavior. I therefore restrict my analysis to households that are also members of DB plans, a relatively homogenous group. Previous researchers use asset balances imputed from donor households of a different pension type, leading to further potential bias. I therefore re-impute missing data, including eligibility among my covariates.; Using the Health and Retirement Study, a panel dataset, I find no evidence of any displacement of other savings in either financial assets or home equity by 401(k) contributions. If anything, eligible households save more in other financial assets, a finding that I attribute to the financial education often associated with the introduction of DC plans.; In my second chapter, co-authored with Professor Friedberg, I use data from the same dataset to show that retirement patterns have changed as DC plans have spread. We estimate that the financial incentives in DB plans lead people to retire almost two years earlier on average, and that the decline in DB plan coverage over the period 1983 to 1995 has increased average retirement age by two to four months.; In my third chapter, I examine whether the observed failure of the elderly to dissave is a result of concerns about end-of-life expenditure. Using data from the Asset and Health Dynamics among the Oldest Old (AHEAD) dataset, I show that married couples save more if they expect to enter long-term care, but that single women do not. I attribute the difference in behavior between the two groups to a desire on the part of couples to make adequate provision for the surviving spouse.