Essays on individual risk-taking behavior and Social Security reform

Chapter 1 of this thesis presents an empirical study of individual investor behavior. Using confidential data on the daily trading activity of 78,000 clients of a large US discount brokerage house over six years, I find that individuals' tendency to hold on to losing investments (known as the disposition effect) is concentrated primarily in large-capitalization stocks. Trades in stocks at the bottom 40 percent of the market capitalization distribution exhibit a reverse disposition effect: investors keep their winners and realize their losers. This new evidence challenges the preference-based view of the disposition effect advocated by prospect theory.; Chapters 2, 3, and 4 present a series of studies on investment-based Social Security reform, co-authored with Martin Feldstein. Chapter 2 examines the investment-risk aspect of a defined-contribution Social Security plan where individuals deposit a fraction of wages in private retirement accounts (PRAs) invested in a 60:40 equity—debt index and annuitized after age 67. We perform Monte Carlo simulations of the 80-year experience of the cohort which is 21 years old in 1998. We find that with savings of 6 percent of wages, the probability that their variable annuity payout will be less than the benefit promised under current law (the “benchmark”) is 17 percent. Raising the savings rate substantially reduces that risk.; Chapter 3 analyzes the cost of introducing bequests in the plan described in Chapter 2. We find that pre-retirement bequests reduce the funds available for post-retirement annuities by about 16 percent. Among the variety of post-retirement bequests that we consider, the least costly, a ten-year-certain bequest, reduces annuities by 6 percent while, the most costly, a bequest equal to the remaining actuarial value of the PRA annuity at the time of death, reduces the annuities by 23 percent.; Chapter 4 proposes a market-based approach to managing the investment risk of PRAs. In our model, retirees purchase a put option which guarantees payouts above the benchmark. They finance it by selling a call option on the same payouts, effectively locking into a multi-period collar. We use risk-neutral valuation techniques to price the collar and find that saving 3 percent of wages in the PRA allows the retiree to purchase a collar which guarantees income between 100 and 145 percent of the benchmark when two-thirds of the benchmark comes from the pay-as-you-go program.