The Center for Retirement Research (CRR) at Boston College analyzed the traditional rules of thumb for withdrawal—including spending interest only, basing withdrawals on life expectancy or adopting the 4% rule—and compared them with the IRS’ RMD strategy (or a percent of assets that individuals are required to withdraw each year starting at age 70.5). In comparing the RMD with the 4% rule, the RMD strategy performed better. In dollar terms, a 65-year-old couple would need about $25,000 more (or 10%) of their $250,000 savings to be persuaded to use the 4% rule instead of the RMD strategy.
A spending interest only strategy can work for wealthy individuals, but has drawbacks for those who lack substantial retirement savings. One disadvantage is that when an individual dies, he will leave behind all of his initial wealth plus capital gains. In some cases, this unnecessarily restricts retirement consumption. Another drawback is that a retiree’s income and consumption are dictated by his asset allocation, running the risk of a portfolio allocation that does not minimize the risk for any given level of expected return on the portfolio. In other words, the retiree may overinvest in dividend-yielding stocks, losing the benefits of diversification.
A second strategy is to spend all financial assets over one’s life expectancy. The equation for calculating this is not simple for most people, and retirees face a high probability (a 50% chance) they will outlive their savings and be forced to rely solely on Social Security.
Spending a fixed percentage of one’s initial retirement savings is another popular strategy, commonly known as the 4% rule. The advantage is that the retiree has a low probability of running out of money; the drawback is that the rule does not permit retirees to periodically adjust consumption in response to investment returns. For instance, if returns are less than expected in a given year, the retiree should respond by reducing consumption to preserve the assets—a fixed 4% withdrawal does not allow this flexibility.