If there is such a thing as safe withdrawal rates, it follows that there must also be unsafe rates. There would actually be a continuum starting at safe, getting less safe and crossing into unsafe. What does this range look like and what do we do with this information?

My most recent thinking about this issue was spurred by my coverage of a now published paper (May issue of the Journal of Financial Planning), "Safe Savings Rates: A New Approach To Retirement Planning Over The Lifecycle" in which Wade Pfau, Ph.D., looked at the accumulation phase and distribution phase of a saver's life asĀ  a single, long period. Much has been written about safe initial withdrawal rates but little connected that information with the savings behavior.

As a practicing financial planner, I cannot help but look at research from the perspective of how I might apply or not apply its findings in real-world client circumstances. One of the things that jumped out to me in Pfau's paper was the wide range of initial withdrawal rates that were sustainable over a 30-year retirement. The initial withdrawal rate is the amount of the first year's withdrawal divided by the portfolio balance at that time. A $50,000 withdrawal in year one from a $1 million portfolio equates to a 5% initial withdrawal rate.

Like Bill Bengen, Pfau's paper is a historical reenactment. He too found the lowest initial withdrawal rate to be about 4%. This is the root of the oft-called 4% rule. Other research of the last 20 years uses different parameters and put a reasonable initial withdrawal rate at higher levels; some purport that even a rate near 6% will be reasonable. The more able and willing the client is to adjust their withdrawals if markets do not cooperate, the more likely an initial rate higher than 4% will still be plausible.

Making an assessment of an appropriate initial withdrawal rate is just the beginning. Regardless of what the initial withdrawal rate may have been, the issue is quite quickly turned away from a safe initial withdrawal rate to whether or not a current withdrawal rate indicates danger of exhausting a family's funds. The current withdrawal rate is the amount of the withdrawal divided by the portfolio value at that time. It is easy for clients to assume that if a low initial withdrawal rate is "safe" then a higher current rate can be unsafe.

A closer look at the data in just about every study I have seen indicates that a high current rate is almost inevitable. In my opinion, one of the most underappreciated aspects of Bill Bengen's original paper from 1994 is his discussion of this very phenomenon. He describes "black hole" and "star" retirees. Black holes retired right before particularly bad market environments, such as one retiring in 1929, while stars were fortunate enough to retire at the beginning of boom periods for the markets. Bengen walks us through a couple of scenarios including how a "Star" retiree could find their current withdrawal rate as high as 8%.

In the lead up to writing this column, I contacted professor Pfau to see what he could tell me about the volatility of current withdrawal rates in his research, which went more than 50 years farther back in time than Bengen. He provided me with some interesting information about four different retirees: the 1918 retiree who experienced the worst overall scenario; the 1921 retiree who had the distinction of having accumulated the least amount of money at retirement; the 1966 retiree who had the lowest maximum withdrawal rate in history; and a 1980 retiree who is the most recent of the 30-year retirements he examined.

With the exception of the 1921 retiree, the current withdrawal rate spiked higher than the initial withdrawal rate within the first five years and retested or breached that higher level in the middle of retirement. The 1921 retiree would have qualified as one of Bengen's stars because very good markets drove down the current withdrawal rate well below the initial withdrawal rate.

The current withdrawal rate can go up because of either an increase in the withdrawal amount, a decrease in the portfolio value, or a combination of the two. With bear markets occurring on average every five years, planners would be wise to give their clients the expectation that the withdrawal rate will increase and that this increase is not necessarily a reason to take drastic action in of itself.

Conversely are the behavioral issues strong markets can produce. Bengen noted that a client retiring in 1958 would have experienced favorable markets and low inflation such that their current withdrawal rate would have dropped to a clearly safe 2.4% by 1967. Such a client would have felt quite confident increasing the withdrawals to make the current withdrawal rate 4%. At that level, by the end of 1974, her current withdrawal rate was 8% and her purchasing power had been cut in half. While we know 1975 helped quite a bit, it is reasonable to think any comfort level this retiree had in 1967 may have been severely damaged by the end of 1974.