With the aging of the baby boomers, retirement planning has received increased attention from financial planners, academics, and the public at large. On the practitioner side, the traditional wisdom is that retirement savings be invested aggressively when the saver is young, with a gradual shift towards more conservative investments over time. On the other hand, academics have focused primarily on how static allocations fare against each other. While more aggressive allocations have higher returns, they also have higher risk. The issue is whether time diversification reduces much of that risk for young investors.
In "A Dynamic-Programming Approach to Multiperiod Asset Allocation, " Professor Musumeci, together with Joe Musumeci, considers a variety of hypothetical investors with different attitudes toward risk. For each type of investor, he determines the optimal allocations for that investor when there is one period left until retirement. These allocations may differ not only for different investors, but even for the same investor as a function of his accumulated wealth. Once the optimal allocations are determined for the last year of saving, the dynamic programming procedure "rolls back" one year at a time to determine optimal allocations for each pervious year as well. This generates optimal allocations for a specified investor that are a function of the number of years remaining until retirement, the wealth accumulation to date, and the investor's tolerance for risk.
This paper was published in the Journal of Financial Services Research, Vol. 15, No. 1, in 1999.
