Tweaking withdrawals for lifelong income

This course is no longer eligible for CE credits. Go to cecorner.ca to find eligible courses.

Methodology

In most retirement plans, forecasts are based on “average” growth rates and inflation, i.e. they follow the Law of Averages. In contrast, we use Murphy’s Law – anything that can go wrong, will go wrong – to ensure extreme risks are covered.

We use actual market history, which we call “aftcasting” (as opposed to forecasting). It uses the market history starting in 1900 and ending at the end of 2011. We do not use Monte Carlo simulators.

Aftcasting displays, on the same chart, the outcome of all historical asset values of all portfolios since 1900, as if a scenario starts in each one of the years between 1900 and 2000. It gives a bird’s-eye view of all outcomes. It provides the success and failure statistics with exact historical accuracy, as opposed to simulation models because it includes the actual historical equity performance, inflation and interest rates, as well as the historical sequencing of all these data sets.

Part 1: What are the Chances of Recovering from a Loss?

Example: Bob is 65 and has $500,000 invested — 40% in S&P500 index and 60% in fixed income. He is worried about losses in his portfolio during retirement. Let’s look at four different scenarios:

Scenario A: No withdrawals

Bob has plenty of other income, so he doesn’t need it from this portfolio. Thus, his initial withdrawal rate (IWR) from this portfolio is 0%. You, as his advisor, tell Bob, “Don’t worry about losses. Over the long term, markets always come back.”

The chart in Figure 1 shows the chances of recovering from a loss using actual historical data since 1900. The vertical scale indicates the probability of a lower portfolio value. The horizontal axis indicates age.

Figure 1: Chances of recovery from a loss, no withdrawals

The chart shows that one year later, there was a 27% chance that Bob’s portfolio would be lower than what he started with. However, the portfolio inevitably recovered from the loss by age 73 in the worst case. In other words, if there are no withdrawals from the portfolio with this 40%/60% asset mix, historically it recovered from the worst loss after about 8 years.

Scenario B: 3% Initial Withdrawal Rate

In this scenario, Bob needs to withdraw $15,000 annually, indexed to CPI, starting at age 65, for the rest of his life. This is an initial withdrawal rate of 3% — well below the sustainable rate. As his advisor, can you still say, “Don’t worry about losses, markets always come back”?

The chart in Figure 2 shows the chances of recovering from a loss. One year later, there was about 37% chance that Bob’s portfolio would be lower than what he started with. What’s worse, even with a low withdrawal rate of 3%, there was a 20% chance that his portfolio would never fully recover from a loss. I am not saying he will run out of money; but I am suggesting you not say to the client, “Don’t worry about losses because markets always come back.” In this case markets do come back, but the portfolio might suffer a permanent loss.

Figure 2: Chances of recovery from a loss, 3% initial withdrawal rate

Scenario C: 5% Initial Withdrawal Rate

Let’s increase Bob’s withdrawals to $25,000 annually (5% IWR), indexed to CPI, starting at age 65, for the rest of his life.

The chart in Figure 3 shows the chances of recovering from a loss. The odds are Bob will never see the original amount of his investment ($500,000) after a loss.

Figure 3: Chances of recovery from a loss, 5% initial withdrawal rate

Scenario D: 8% Initial Withdrawal Rate

Now we increase Bob’s withdrawals to $40,000 annually (8% IWR), indexed to CPI, starting at age 65, for the rest of his life.

The chart in Figure 4 shows that after 10 years, there is 100% certainty that Bob will never see his portfolio’s initial value of $500,000 after a loss.

Figure 4: Chances of recovery from a loss, 8% initial withdrawal rate

It is important to understand that the concept of “investing for the long-term” applies only to accumulation portfolios. As soon as a portfolio is switched from accumulation to distribution, the concept of “long-term” no longer applies and the “luck factor” takes control of the outcome — even if withdrawals are below sustainable.

If withdrawals are below sustainable, then luck determines how much money is left to estate.

If withdrawals exceed sustainable, then luck determines how soon the portfolio depletes. If you don’t want to rely on luck for lifelong income, then you need to pool the risk.

Part 2: How much of a pay cut is necessary to ensure lifelong income?

Bob decides to withdraw $22,500 annually, indexed to CPI, starting at age 65, for the rest of his life. This means his initial withdrawal rate is 4.5%, which William Bengen, the inventor of the now famous “4% rule,” said he was comfortable with (See William P. Barrett, “The Retirement Spending Solution,” Forbes.com.).

Bob’s primary concern is sustainability of income for life, and he is worried about outliving his money. He wants his assets to last until age 95 and does not want the probability of depleting his portfolio to exceed 10%.

Figure 5 shows the aftcast. This is our base case. There is one thin, black aftcast line starting at the left vertical axis for each year since 1900. We define the bottom decile (bottom 10%) of all outcomes as the “unlucky” outcome, the top decile (top 10%) as the “lucky” outcome. The blue line indicates the median outcome, where half of the scenarios are better and half are worse.

In this example, the probability of depletion by age 95 is 39% — well beyond the 10% limit Bob has set.

“How can this happen?” you might ask. “The initial withdrawal rate is only 4.5%, as Bengen suggests, and the plan is still inadequate?” The answer is yes, and it’s because the 4.5% withdrawal rate ignores portfolio costs. We don’t. Portfolio management costs, advisor fees, trading costs and all other expenses make the famous “4% rule” too optimistic.

Figure 5: The aftcast of $22,500 annual withdrawals, indexed to inflation, from an investment portfolio, starting capital of $500,000

Let’s modify our base case. Bob is willing to have a 15% pay-cut whenever his portfolio grows less than 0% in the preceding calendar year. We call this 0% the “growth threshold.”

Figure 6 depicts the aftcast. The probability of depletion by age 95 is reduced from 39% to 33%. When we compare these two charts (Figures 5 and 6), we can hardly notice the difference — they both look bad.

Figure 6: The aftcast of $22,500 annual withdrawals, indexed to inflation, from an investment portfolio, starting capital of $500,000, withdrawals reduced by 15% whenever portfolio growth is less than 0%

Varying the Growth Threshold

Can we solve the problem by varying the growth threshold? What happens if we increase it from 0% to 6%?

For example, Bob takes a 15% pay cut if his portfolio grows by less than 6% in the preceding calendar year. We find that the probability of depletion by age 95 is 26% — somewhat better. But now, the pay-cuts come more often.

The table below shows the effect of varying the growth-threshold on the frequency of pay-cuts:

Growth-threshold	Frequency of Pay-cuts (historical average)
0%	21%	About 1 in 5 years
3%	31%	About 1 in 3 years
6%	48%	About 1 in 2 years

Be careful with the meaning of “historical average.” Just because it says “1 in 5 years” it does not mean the pay-cut happens once every fifth year. You might never see a pay-cut, or you might see three pay-cuts in five years. It’s just an average across all 30-year rolling periods over the entire twentieth century.

Here are two basic points you might want to remember:

• The larger the growth-threshold, the more often you get pay-cuts.
• The larger the growth threshold, the smaller are the pay-cut amounts for the same initial withdrawal rate and time horizon.

Finally, we want to know how much of a pay-cut Bob needs to ensure that the probability of depletion is 10% or less by age 95. Figure 7 illustrates the answer.

Figure 7: Pay-cut calculation to bring the probability of depletion down to 10% for a 30-year withdrawal time horizon

Example: Bob’s initial withdrawal rate is 4.5%. He does not want the probability of depletion to exceed 10% by age 95. He is willing to have a pay-cut in the year following a bad year, which he defines as portfolio growth of 0% or less.

Question 1: How much is the pay-cut in the year following a bad year?

Answer: Look up 4.5% initial withdrawal rate on the horizontal scale. Follow the vertical red dashed line up until it intersects the blue line (0% threshold). From that point, draw a horizontal line until it meets the left axis. Read the pay-cut: about 48%.

Question 2: On the average, how often did a pay-cut happen historically?

Answer: Read the description on the blue line: 1 in 5 years

The Pay-Cut Zones

We divided the chart into 4 zones:

• If required pay-cut is 10% or less, we call it “tweak zone”
• If required pay-cut is between 10% and 20%, we call it “austerity zone”
• If required pay-cut is between 20% and 30%, we call it “dog food zone”
• If required pay-cut is over 30%, we call it “destitute zone”

Guidelines

Never forget that the pay-cut a client can tolerate depends on what this money is needed for. It might be for basic necessities such as housing, living, transportation and health care. Or, it might be for discretionary expenses such as travelling, donations, gifts, helping children, etc.

• Basic Necessities: If these withdrawals are for basic necessities, and if the pay-cut exceeds 10% (outside the tweak zone), it is probably not an acceptable outcome. In these situations, you might suggest other remedies such as reducing expenses, delaying retirement, creating income from other assets (renting part of the home, cottage, summer home), working part-time during retirement, downsizing home, selling home outright and moving to rental, etc.

Also, the frequency of pay-cuts might be another concern. If the pay-cuts are required too often, or are too deep, you might want to look at annuitization strategies.

• Discretionary Expenses: If this income is earmarked for discretionary expenses, then tweaking income can work. You might also want to follow withdrawal strategies based on the percentage of the portfolio value or portfolio growth (see Part 3).

Some overly optimistic academic studies define this type of income adjustment strategy as “tweaking.” The ramifications of pay-cuts can be a lot more severe than the word “tweaking” implies. We think it is far better to disclose potential pay cuts and their frequency to clients up front and quantitatively, as in the examples given above.

There is one more thing to be careful of: Most Monte Carlo simulators are inadequate to model the sequence of returns and sequence of inflation, resulting in conclusions that are significantly distorted on the optimistic side.

The material in this article gives you sufficient background on “tweaking” withdrawals and enables you discuss the pros and cons of this strategy comfortably with your clients. Now, let’s look at other withdrawal strategies based on the portfolio’s value or growth.

Part 3: Withdrawal Strategies Based on Portfolio Value or Portfolio Growth

If income is required for discretionary expenses you can implement other withdrawal strategies. Here, we cover two methods:

• Withdrawal rate based on the growth of the portfolio: This strategy may be suitable for purchasing non-urgent, big ticket items like cars.
• Withdrawal rate based on the portfolio value: This strategy is more suitable for ongoing discretionary expenses, such as travelling, helping children or donations.

Withdrawal Rate is Based on the Growth of the Portfolio:

Example: Bob, 65, sets aside one of his portfolios — he calls this his “car bucket” — to finance his car purchases throughout retirement. The market value of this car bucket is $75,000, with an asset mix of 40% equity and 60% fixed income. He is flexible about the timing of the purchase.

Strategy: Each year he withdraws double the growth of the portfolio, up to $7,000. He saves this money in a high-interest bank account. When it reaches $25,000 he buys his next car. He is not planning to buy a car after age 85, so withdrawals from the car bucket stop at that age. The aftcast is shown in Figure 8.

Historically, in the worst case, he would have to wait 10 years to buy his next car (after the 1929 crash). In the best case, he would have waited 4 years. On the average, he accumulates $25,000 in 5 years and 2 months.

As for remaining assets at age 85, he would have about $26,400 (worst case) or $57,000 (median) left in the car bucket, which he can use for anything he wants.

Figure 8: The aftcast of Bob’s “car bucket”

Withdrawal Rate is Based on the Portfolio Value

Example: Bob, 65, wants to travel during retirement. He sets aside one of his portfolios as his “travel bucket” to finance these expenses. This portfolio is currently worth $100,000 and has an asset mix of 40% equity and 60% fixed income. He is flexible about his travel budget and does not think he will travel as much after age 85.

Strategy: Each year he withdraws 10% of the current portfolio value, not exceeding $9,000, for travelling expenses. The aftcast is shown in Figure 9.

Historically, in the worst case (1929), he would receive about $84,000 over the next 20 years and have about $16,500 left in the portfolio at age 85. His average annual travel budget would be about $4,200.

As for the median scenario, he would receive about $140,000 over 20 years, and have $44,000 left in the portfolio at age 85. His average annual travel budget would be about $7,000.

Figure 9: The aftcast of Bob’s “travel bucket”

Now that you’ve finished reading, complete the exam to receive your CE credits. If your score is 85% or higher, send an e-mail to jim@retirementoptimizer.com with your name and proof of your score (a screen shot will do) to get a free retirement calculator based on aftcasting, and a read-only, pdf copy of Jim Otar’s 525-page book, Unveiling the Retirement Myth.