Investment portfolio torture chamber
The success of a client’s retirement portfolio can be defined as follows: remains solvent while providing the needed annual income for the rest of the client’s life. Toward that end, every client needs to do their part to facilitate that success — such as not withdrawing too much money and utilizing an asset allocation strategy that gives their portfolio a chance to do its job.
The natural challenge in meeting the stated objective is the fluctuating performance of the underlying investments in the portfolio, combined with the periodic withdrawals made by the client.
In short, this is an investment portfolio torture chamber. Advisors need to help clients understand and meet these challenges, and the following analysis may help in that process.
The retirement portfolio asset allocation we use is based on four primary asset classes: large-cap U.S. stock, small-cap U.S. stock, U.S. aggregate bonds and U.S. cash. Large stock is represented by the S&P 500, while small stock is represented by the Ibbotson Small Stock Index from 1961 to 1978, and the Russell 2000 from 1979 to 2018. U.S. bonds are represented by the SBBI U.S. Intermediate Government Bond Index from 1961 to 1975, and the Barclays Aggregate Bond Index from 1976 to 2018. U.S. cash is represented by 90-day Treasury bills.
These are the only four primary asset classes for which performance data are available back to 1961. In practice today, you will likely use more than these four asset classes in your clients’ portfolios to achieve optimal diversity.
The annual returns of these indexes were reduced by 100 bps to reflect the cost of actual investment products and an advisory fee.
The portfolio asset allocation is 40% large U.S. stock, 20% small U.S. stock, 30% bonds and 10% cash (with annual rebalancing). For context, the average 25-year annualized return over 34 rolling 25-year periods for this 40/20/30/10 portfolio was 9.5% from 1961 to 2018.
Each rolling period represents a 25-year retirement window with a unique sequence of returns and a unique return. Each year’s withdrawal took place at the end of the year and was based on 4% of the portfolio’s balance at the end of the prior year — somewhat like RMD rules. This analysis did not impose a COLA on the annual withdrawals. The starting balance in the retirement portfolio was assumed to be $1 million. The portfolio was rebalanced annually.
Let’s start with an examination of all 34 rolling 25-year periods from 1961 to 2018. Picture this as a simulation of 34 different clients, each one starting into retirement one year later than the previous client.
Shown in “Many paths” are the annual withdrawals for all 34 clients. Each colored line represents a client and their withdrawals each year for 25 years. The highest gray-colored line has dots where each represents the portfolio withdrawal each year. (Dots on the other lines were removed for clarity of the lines.)
Clearly, each client had a unique experience based on the variation of the lines in the graph. Some were more fortunate (the lines near the top) and some less so (lines near the bottom of the graph). Why? Two reasons: The absolute performance of the portfolio over each 25-year period and the specific sequence of returns over each 25-year period.
As shown in “Best and least best,” the ideal 25-year period was from 1975 to 1999. This period started after the equity declines of 1973 and 1974, and finished before the equity market losses that started in 2000. This was a perfect 25-year window that produced an average annualized return of 13.43%.
As a result, the client withdrew a total of $3.75 million pre-tax dollars (or an annual pre-tax average withdrawal of $150,028) and had $9.49 million remaining in the portfolio after 25 years.
Amazingly, the withdrawal in the 25th year was over $350,000.
By contrast, the most recent 25-year period, from 1994 to 2018, was the least desirable for a client. Over the past 25 years the portfolio produced an average annualized return of 6.51%, which meant the client withdrew a total of $1.54 million pre-tax dollars (or an annual pre-tax average of $61,440) and had an ending portfolio balance of $1,848,430.
It’s important to stop right here and examine how good the worst-case scenario actually is. Over the past 25 years the client pulled out $1.54 million pre-tax dollars and finished with $848,430 more than they started with.
Perhaps the worst-case scenario is more accurately described as not-so-best — because the outcome was still rather impressive. Would that we all had such poor luck.
The average annual withdrawal for all 34 clients, assuming a 4% withdrawal rate, was $88,826 pre-tax (the average of all the lines in “Many paths”) and the average ending portfolio balance was $4,079,724. This is a remarkably encouraging result for clients.
Bottom line: Retirees who withdraw 4% of the prior year’s portfolio balance have a high likelihood of seeing their portfolio grow — maybe significantly — throughout their retirement years. Perhaps this awareness could reduce the fear of running out of money that haunts so many clients. However, this assumes a portfolio allocation that is at least 60% equity, and overall portfolio expenses that don’t exceed 100 bps.
Now we need to address the issue of sequence-of-returns, illustrated in the graph “Sequence of returns.”
We are using these two different 25-year periods (1983 to 2007 and 1969 to 1993) because the overall return, assuming a lump sum investment, was essentially the same (9.66% for the first period and 9.48% for the second). However, the annual withdrawals were considerably different because of the sequence of returns in the portfolio.
The 25-year window starting in 1983 was clearly superior to the window that began in 1969. The returns in the first six years for the favorable period were 17.25%, 5.58%, 25.30%, 12.80%, 0.76% and 13.70%.
The returns from 1969 to 1974 tell a very different story. They were -8.97%, 2.80%, 11.08%, 9.45%,
-10.95% and -13.06%. Three losses within the first six years made a powerful difference, even though the portfolio recovered in later years and finished the 25-year period with nearly the same overall return as during the 1983 to 2007 period.
The favorable 1983 to 2007 period allowed the client to withdraw a total of $2.57 million pre-tax (for an average annual pre-tax withdrawal of $102,828) and have a remaining balance of $3.94 million after 25 years.
The client with the less-favorable sequence of returns withdrew a pre-tax total of $1.67 million (an annual average of $66,644 pre-tax dollars) and had $3.78 million remaining after 25 years.
Because the two different time periods had nearly identical 25-year performances, the ending portfolio balances were nearly the same. But because of the differences in the sequence of returns, there was a difference of approximately $900,000 in total money withdrawn — in favor of the period that had better returns in the first six years.
So what do we learn from all this?
- A growth-oriented retirement portfolio that annually cost the client 100 bps performed admirably under the stress of a 4% withdrawal system. In every case, the portfolio survived intact for 25 years. In fact, the average ending portfolio balance after 25 years was $4,079,724 — or four times larger than the starting balance. The median ending balance was $3.8 million. The best-case outcome allowed the client to withdraw over $3.7 million pre-tax dollars and have nearly $9.5 million remaining after 25 years. But we can’t plan using best-case assumptions. Therefore, using our least-best-case experience, we see the retiree was able to withdraw over $1.5 million pre-tax dollars and have $1.84 million remaining after 25 years. Retirees with a prudently diversified growth-oriented portfolio can live their retirement years with much less fear if they are simply withdrawing 4% of their portfolio balance each year.
- The sequence of returns is not manageable within the portfolio (assuming the portfolio allocation is consistent through time). The portfolio simply does what it does based on the performance of the underlying mutual funds, ETFs and/or individual securities. A non-favorable initial sequence of returns in a retirement portfolio must be managed by liquid assets outside the portfolio — such as a cash account. This is why some portion of retirement assets should be in cash or a cash equivalent. A client who has 12 to 24 months of living expenses in a savings account would be an example of this approach.
In summary, we need to remember that, for many people, retirement is a long-run investing experience. The portfolio will have ups and downs, but a well-diversified portfolio will likely generate a return in the range of 6% to 13%, averaging around 9%. That level of return is clearly adequate to handle a 4% withdrawal rate.