When should someone retire? The answer may be fraught with danger if the retirement portfolio is overly weighted to stocks or other risky investments. There is one small window of time surrounding the retirement date in which sharp declines in stock values can ruin retirement.
Retirement investing may span 60-80 years: 30-40 years of preretirement investing, then 30 or so years of withdrawals during retirement. Substantial risk, maybe 70% to 90+% in stocks, may well be advisable during most of that 60-80 years, but in a roughly six-year period surrounding retirement, risk means danger.
A typical retirement story underpins today’s analysis: people save and invest many years for retirement, then withdraw a little each year to support themselves during retirement.
If a portfolio tanks near retirement, either just before or after, a retiree may become much poorer than planned. By continuing to withdraw inflation-adjusted annual living allowances, that poverty may worsen and last the rest of retirement. The good news is that people can manage and solve this problem.
The analysis here uses a model of retirement investing in which stocks and bonds move up and down year-to-year according to actual returns from two mutual funds: Vanguard Total Stock Market Fund Investor Shares (VTS), and Vanguard Total Bond Market Fund Investor Shares (VTB). The time period under study is from the beginning of 2000 through 2013. Given a beginning portfolio value ($100), the computer grows or shrinks the initial $100 at the rates experienced by the Vanguard funds.
At the end of the first year, the retiree withdraws 4% of the initial value, which is $4. to help support herself during the next year, and the remaining stocks and/or bonds stay invested. Then the portfolio is rebalanced to the original proportions of stocks and bonds. The computer again grows or shrinks the investments at the rates experienced by the Vanguard funds for that second year. Again the retiree withdraws the predetermined amount, rebalances her portfolio, and leaves what’s left invested for a third year. The computer carries this process along for 14 years and reports an end value for the portfolio.
How do different portfolios and withdrawals affect the end value? Clearly, if stocks and bonds increase by large amounts, that growth will overwhelm the withdrawals and the retiree will be better off at the end than at the beginning. Similarly, but harder to see, if the investments decline early in the period, that decline combined with annual withdrawals can send the portfolio rapidly toward zero. In other words, people can outlive their investments and end up in real poverty.
With those possibilities in mind, the model can help show what portfolio management strategies will best protect a retiree’s ability to stay solvent for her full retirement.
The two Vanguard funds selected as model investments are high-quality funds that many investors use as the core of their portfolios.
A table at the end of the post shows the returns published by Vanguard. The selected time period (2000-2013) is the most recent, and it started with three years of sizable drops in stock values. Therefore it is a good period to see if early losses endanger retirement.
In all cases reported below, a retiree had $100 at the beginning of retirement on January 1, 2000. Using $100 allows readers to easily convert results into percentages and apply them to any portfolio value. A nearby chart shows the results for five portfolios.
The main results are:
- Portfolio A is 100% stock (VTS). A retiree is assumed to increase the $4 annual withdrawal by an average inflation rate of 2.5% each year. In this example, the portfolio value in January 2014 is $57, which is dangerously low. In fact, if the next 14 years were just like 2000 through 2013, this portfolio would be exhausted by 2021. The first three years of negative stock returns combined with an inflation-adjusted withdrawal overwhelms stock growth in later years.
- Portfolio B is also 100% stock (VTS), and it has the same inflation adjustment as A. Portfolio B, however, is a hypothetical portfolio designed to test the impact of the three beginning years of negative returns. In B, the first three years of negative returns were moved to the last three years of the period (2011, 2012, 2013), and the remaining years were moved back in time by three years. That means the actual returns for 2003, 2004, through 2013 were moved to 2000, 2001, through 2010. (See the middle column of returns in the table at the end.) Moving the negative returns to the end of the period has a dramatic effect on portfolio performance: at the end of the period in January 2014, the portfolio has $111, which is 95% greater than the $57 of the previous example. The timing of negative investment returns has a dramatic effect on the portfolio’s performance in retirement.
Again, Portfolio B is hypothetical. If a person actually did retire at the beginning of 2000, she would have experienced three years of negative returns at the beginning. There are only two remedies for such a situation: she can withdraw less money, and she can diversify her portfolio with other asset classes.
- Portfolio C is 100% stocks (VTS), but the withdrawal is held constant at 4% of the initial value ($4). That is, there is no inflation adjustment. The value at the beginning of 2014 is $76, still dangerously low, but better than Portfolio A ($57).
- Portfolio D is 50% stocks (VTS) and 50% bonds (VTB), and the withdrawal increases at 2.5% annually to adjust for inflation. Here the end value is $102, which is about the same as at the beginning. This retiree is more financially secure because she is nearer the end of life (her portfolio need not last another 30 years), but she is less secure because her withdrawals are a higher share of the portfolio value. With 14 years of inflation adjustments, she withdrew $5.65 at the beginning of 2014, which exceeds 5% of the portfolio’s ‘before-withdrawal’ value of $108.
- Portfolio E, like D, is 50/50 stocks (VTS) and bonds (VTB), but the withdrawal is held constant at $4 (no increases for inflation). Here the ending value is $119, which is the best result. The portfolio was $123 before the 2014 withdrawal. Because the portfolio grew but the withdrawal did not, the $4 is now only 3.3% of the $123. She could safely increase her withdrawal by a small amount, perhaps resetting it next year to 4% of the portfolio value at the end of 2014.
The results show that although stocks (with greater variability) are great for the long run, they can be disastrous near the time of retirement. Portfolio’s A and B highlight that.
During that small window of maybe six years or so, the results also show that diversification to bonds and modest reductions in annual withdrawals compensate for the declining stock values early in retirement. That is important.
These lessons will remain important in the future. The period used here—2000 through 2013—witnessed good, stable bond returns (see table). Many commentators expect interest rates will rise and bond values decline overt the next several years. If that happens, and if stocks are as volatile as they have been in the past 14 years, then retirees may need to rely more on moderating their withdrawals. And they will want to consider more asset classes like REITs, international stocks, and perhaps high dividend stocks.
One general point from the analysis is that retirees should not put their retirement portfolio on autopilot: choose basic rules at age 65 and ignore things for the next 30 years. Regardless of withdrawals or diversification, retirees must stay vigilant and adaptable to ensure a successful retirement.
A second point is that at other times, in the long period leading to three or four years before retirement and in the later years of retirement, stocks (risk) present less danger and offer better long-term returns. We’ll look more closely at this point in a future post.