Life After Full-time Work Blog

Perhaps after you read this, you’ll really understand it

If I’ve heard it once, I’ve heard it a hundred times: a mention, often an endorsement, of the famous “4 per cent rule” for decumulation, without any context or any reference to where it comes from. And that annoys me, because it implies that the rule applies to everyone at all times – which of course is nonsense. So let’s explore what the rule actually says, where it applies, and how it was derived – all of which will be highly educational. And that’s because Bill Bengen, who came up with the rule, is a most thoughtful person, and his initial derivation of the rule, and subsequent commentaries on it (including a new book, A Richer Retirement: supercharging the 4% rule to spend more and enjoy more) are excellent.

***

First, what is the actual rule? It’s essentially as follows.

How much money do you have accumulated at your retirement? Whatever it is, withdraw 4% of that amount in the first year of your retirement, and then, in each subsequent year, withdraw that same amount increased by inflation. So if, for example, you have accumulated $1,000,000, withdraw $40,000 in the first year; if inflation is 3%, withdraw 3% more (that is, $41,200) in the second year; if the next year’s inflation is 10%, withdraw 10% more (that is, 110% of $41,200, or $45,320) in the third year; and so on.

That’s the rule.

***

Wait a minute! What’s the goal, the problem, for which this rule is the solution? Does it apply to everyone, regardless of their age and state of health? Does it matter what asset allocation they use? What risk of failure is tolerable? How can it possibly be independent of those crucial factors? In fact, I often ask the person who cites the rule: “The 4% rule is the answer to a question. Can you tell me what that question is?” Typically, no.

So let’s dig deeper into it.

Let me first cite the question for which the 4% rule is the answer. You can phrase it many ways; for me, here’s the simplest.

“Let’s look at historical US data from 1926 through 1992. Let’s assume an asset allocation of 50% in a US common stock index fund and 50% in intermediate term US Treasury notes [never mind the specifics]. Rebalance the portfolio to this allocation after each year’s withdrawal. Withdraw x% of the portfolio at the end of the first year, and increase the dollar amount of withdrawal each year by the amount of inflation. What is the highest value of x for which the withdrawals would have lasted for at least 30 years?”

Answer: x = 4.

***

Why would you ever dream up a question like that? You wouldn’t. And Bengen didn’t.

He did a lot of experimenting, and found something really interesting that came out of his experiments.

He looked at lots of withdrawal time periods: 50 years, 40 years, 30 years, …, 5 years. He tested lots of asset allocations: not just 50-50, but 0-100, 25-75, 50-50, 75-25 and 100-0. He tried lots of initial withdrawal rates: 1, 2, 3, …., 7, 8 per cent. How to compare the results? He said simply: “I have quantified portfolio performance in terms of ‘portfolio longevity’: how long the portfolio will last before all its investments have been exhausted by withdrawals.” He added: “This is an intuitive approach that is easy to explain to my clients, whose primary goal is making it through retirement without exhausting their funds, and whose secondary goal is accumulating wealth for their heirs.”

And he found that “assuming a minimum requirement of 30 years of portfolio longevity, a first year withdrawal of 4 percent, followed by inflation-adjusted withdrawals in subsequent years, should be safe.”

And that’s how the 4% rule originated.

***

He could have done many things differently. (One minor thing I always wondered about was that his withdrawals started at the end of the first year: what would the retiree live on through the year itself? But that’s trivial, and doesn’t affect anything.) He could have shown similar results for many time periods, given that the longevity of future retirement depends very much on one’s current age – and therefore he could have enabled people to adjust their withdrawals as their expected future years changed. Indeed, he could have allowed his hypothetical test subject to adjust withdrawals modestly within pre-specified bounds, depending on whether investment returns had been good or bad, rather than sticking rigidly to inflation-adjusted withdrawals, and then proclaiming failure if even one thirty-year period failed to reach completion by as little as $1, while success was proclaimed simply through a “yes” answer for reaching the end, even if there was over $1,000,000 unspent at that stage. (In fact, he found that a 75% stock exposure produced almost as good a worst case outcome, as well as far superior outcomes in most years, so his actual advice to clients, based on his work, was to consider a stock exposure as close to 75% as possible – and of course, as you can imagine, his explanation of his work would have been far more complete than what he wrote up.)

But all of these objections would be totally unfair, because we think of them in retrospect, with hindsight, knowing what we know today. Remember, when Bengen published this intriguing set of experimental results in 1994, nobody had ever done any work like it, and that’s why it was so interesting – so interesting, as it turns out, that today the 4% rule is cited by people who haven’t a clue where it originated and apply it in circumstances that bear no match to Bengen’s work.

(There’s one other detail in the original piece that I think calls for minor criticism. With 1992 as the final year of market and inflation data available, there weren’t actually many 30-year periods available, the first being of course 1926-1955, and the last being 1963-1992. To construct more data points, for 30-year periods starting in 1964 (or 50-year periods starting in 1944), Bengen used actual returns and inflation until 1992, and then he added the requisite number of additional years, in each of which he assumed the average market returns and inflation. So: he assumed the actual average return, but didn’t show the impact of the associated volatility. But you know what? I don’t care.)

***

Once it became so popular, Bengen followed up with a number of papers. Adding additional asset classes to his initial simple analysis, he raised his recommendation to a first-year withdrawal rate of 4.5%. (But in popular literature it’s still the “4% rule” that has been enshrined.)

Any investor who retired in 1969 was his original “worst case” scenario (and the only one not to make it through 30 years with a 4.5% initial withdrawal rate). In 2012 Bengen considered the first twelve years of retirement for an investor who retired in 2000, so that the impact of two stock market declines (in the S&P 500 index) of more than 50% in the same decade could be considered. Surprisingly, the 2000 retiree was better off, at the end of the first 12 years, than the 1969 retiree? Why? Because the 1969 retiree encountered huge inflation, and had to increase annual withdrawals proportionately, whereas the 2000 retiree encountered substantially lower inflation. Yes, everything interacts: it’s not the nominal investment returns that matter to these retirees, it’s the real (i.e., after inflation) returns.

All of that led Bengen, in his 2012 paper, to loosen his rigid (because deliberately simple, for the 1994 study) requirement that retirees must annually withdraw the initial formula dollar amount and then inflation-adjusted amounts after that. And he had a bit to say about two relevant approaches in bad times.

One approach is to reduce spending. Here he came up with this rule of thumb. Calculate your inflation-formula withdrawal amount, in a subsequent year, as a percentage of your then available assets. If that percentage is more than a quarter higher than your initial percentage (e.g., if you start at 4% and in the year under consideration you are due to withdraw more than 5% of the remaining assets), take some pre-emptive action. (He doesn’t specify what action, but you might, as an example, not allow that recalculated withdrawal rate to exceed 125% of your initial withdrawal rate.)

The other approach is to find a way to try to increase your income, and he gives a few examples (which he says he finds interesting, though far from an exhaustive list). Invest in immediate fixed or inflation-indexed annuities. Use a reverse mortgage. Depart from a buy-and-hold investment approach. [My observation: Potentially dangerous, if you’re trying market timing. I note that Bengen did not say what might replace buy-and-hold investing. And actually he never advocated buy-and-hold in the first place, because rebalancing annually is not buy-and-hold.]

***

So there you have Bengen up to 2012.

Now there’s his new book – which had not yet been released when I wrote this. But I intend to read it, of course.

***

Takeaway

The 4 per cent rule was meant to apply to one set of clearly-defined circumstances, and even after deriving it Bill Bengen suggested using 4.5% and a more aggressive allocation to his clients. Most people who cite the rule have no idea when it applies.

2 Comments