In a previous article, the concept of buffered withdrawals was introduced, but some details were omitted. This article will explore that idea with a little more depth and describe how such a strategy could be implemented in practice.
Before moving on, I would like to give credit for the idea of this strategy. I first encountered the idea of buffered withdrawals on the bogleheads.org site’s wiki page for withdrawal methods. This page alludes to the Galeno Strategy but doesn’t describe what it is. Intrigued, I set out to find The Galeno Strategy and discovered that it is just a way to buffer the withdrawals. In Galeno’s case, he uses a 5-year buffer with the buffer invested in safe money-market funds and the rest of his portfolio invested in stocks.
I wanted to take this a step further and look at different size buffers and different stock allocation strategies to see how this affected the math.
The basic idea of the spending buffer is to split your portfolio (
P) into the
V) and withdrawal buffer (
P = V + U
Each year, you take a percentage of your portfolio (your withdrawal rate) and add
it to your buffer. Then you divide that buffer by the size (in years) and withdraw
that amount as your income for the coming year. If your buffer size is
this means that normally you will have
k-1 years worth of spending sitting in the buffer.
Your buffer will only temporarily have
k years of spending in the short time
between when you add annual withdrawal to the buffer and then withdraw your
spending funds from the buffer.
Why do this? The main reason to do this is to buffer the ups and downs of the
stock market. The bigger your buffer is, the smaller the variation will be
on your withdrawal income. Statistically, if you sample data that has a
standard deviation of
σ, then taking
N samples of that data will have
a standard deviation of
σ/sqrt(N)1. Applying that principal to a
withdrawal buffer, if we sample a portfolio that has a 20% standard deviation
per year, a 4-year buffer of that data would have a 10% standard deviation.
In short, a withdrawal buffer reduces the volatility of your income compared to
directly withdrawing a fixed percentage from your portfolio.
Let’s define a single years’ worth of the buffer to be
P = V + (k-1)W
Typically the invesment piece (
V) would be made up of some combination of stocks (
and bonds (
V = S + B
And the withdrawal buffer would be invested conservatively in cash-equivalents since that is the money will support your living expenses in the near future.
Galeno’s version of this strategy invests everything in stocks which means that
the percentage of the overall portfolio in stocks fluctuates.
I would prefer to keep the stock percentage at some fixed value so I will
diverge from the Galeno Strategy on this point.
If the stock percentage is
S = sP
The reason I like to keep the stock percentage fixed is to that it enables you to take gains when the market is up and move more money back into stocks when the market is down, simply by rebalancing to that fixed percentage. This is in keeping with the asset allocation strategies that we all know and love.
If we define the withdrawal rate (
w) as a percentage of the overall portfolio,
W = wP
The portfolio equation now looks like this,
P = sP + B + (k-1)wP
The bond piece is the part of the portfolio that fluctuates to accommodate swings in the stock market and can be written as,
B = (1-s)P - (k-1)wP
Notice that for the degenerate case of no buffer (k=1), the bond percentage is
(1-s). So for a 60% stock target, the bonds would be 40% as expected.
Keeping Up with Inflation
If you want your portfolio to keep pace with inflation, then you could require that the portfolio that is left over after a withdrawal must increase enough to cover the money lost to the withdrawal as well as covering the effects of inflation,
(P - kW)(1 + r) + (k-1)W > P(1 + i)
W=wP and solving for
w < (r - i) / (1 + kr)
Interestingly, even in the degenerate case where we don’t have a buffer (k=1),
then the maximum withdrawal rate that keeps up with inflation is not simply
but is reduced by
1/(1+r). For example, if your portfolio returns 7% and inflation
is 3%, then the maximum withdrawal would not be 4% but rather it would be,
w < (.07-.03)/(1.07) = 3.74%
This is because your portfolio remaining after the withdrawal has to overcome both inflation and the effects of withdrawing part of the portfolio.
Also notice that your withdrawal rate decreases with bigger buffer sizes. This is because it is assumed that the buffer has zero gains. You could mitigate this by investing a part of your buffer in bonds but you will be increasing your income volatility slightly if you do so.
The buffering concept plays really well with a Roth conversion ladder. A Roth conversion ladder is a way to convert traditional IRA funds to Roth IRA funds in an attempt to avoid early withdrawal penalties. The IRS allows you to convert a traditional IRA to a Roth IRA and may use the converted funds without penalty after a 5-year waiting period.
So if you pair a buffered withdrawal strategy using a buffer of 6 years, with a Roth ladder, your buffer could be entirely made up of Roth IRA funds that are converted annually from traditional IRA funds. Then you simply withdraw 1/6th of that buffer each year for living expenses.
The reason for choosing a 6-year buffer instead of a 5-year buffer is to avoid the possibility of withdrawing slightly more than 1/5 of your buffer due to market fluctuations. But a 6-year buffer is a pretty large amount to not be invested so it makes sense to invest 1-3 years of that buffer in something relatively conservative like bonds. One safe way to do that would be to simply buy bonds with a 3-5 year maturity using the withdrawn funds.
Implementing the Buffered Strategy
Here are the basic steps for using a buffered withdrawal. In these steps, by way of example, let’s assume a buffer size of 6 years and a withdrawal rate of 3%.
Determine your initial buffer size. Simply multiply your buffer size by your withdrawal rate (6*3% = 18%).
Set aside your initial buffer size somewhere in an account with safe funds. For example, if your portfolio is $1,000,000 then set aside 18% or $180k as your spending buffer.
Divide your buffer by the buffer size (in years) to get this year’s spending. For example, if your buffer has $180k and you have a 6-year buffer, remove $30k of spending for this year. There is now $150k left in your buffer that you will not touch until next year.
Rebalance your stock allocation. For example, if you use a 60% stock allocation, rebalance your remaining portfolio so that you have 60% in stocks (e.g. 60% of $970k = $582k).
Invest the remainder in bonds (e.g. 970-582-150=$240k).
Enjoy life for 1 year.
After 1 year, recompute the size of your portfolio, including the buffer. For example, if your stocks went up to $640k, and your bonds went up to $250k and your buffer is $150k, then your total portfolio now has $1,040,000.
Multiply your new portfolio value by your withdrawal rate to get your buffer increment. For example, $1,040,000*3% = $31,200.
Move the buffer increment amount into your existing buffer. You might want to consider your allocation before moving funds so as to avoid moving funds twice.
Go back to step 3 and repeat.
Implementing a buffered withdrawal strategy is one way to reduce the volatility of spending in early retirement by spreading portfolio gains and losses between multiple years. It also can be used in conjunction with a Roth conversion ladder for retirees needing to convert the tax-deferred IRA funds to Roth IRA funds. Implementing such a strategy is fairly easy if you follow the steps outlined above.