I recently read Steve Vernon’s new book Retirement Game-Changers. I hadn’t heard of it before seeing it on some “end-of-year favourite finance books” list. Overall I found the book average — despite the name it doesn’t offer any game-changers and instead is pretty run-of-the-mill modern retirement advice. Defer taking Social Security, consider buying an income annuity, find something to “retire to” and not just “retire from”, and so on.
Since this blog focuses most on retirement withdrawal strategies, let’s zoom in on the strategy mentioned in the book, and see why I’m lukewarm about it.
The book calls its strategy “Spend Safely” and is the same kind of “floor & upside” strategy that virtually all advisors today recommend. First secure a safe level of income via some combination of Social Security, bond ladders, annuities, and pension. Then invest the rest in an equity-heavy portfolio.
The question is: how much to spend from the equity-heavy portfolio? Vernon recommends to just use the IRS RMD tables. This idea was first put forth by Sun & Webb in their 2012 paper, “Can Retirees Base Wealth Withdrawals on the IRS Required Minimum Distributions?”
The short answer is, yes, it seems to work okay. It has a few other potential benefits:
- RMDs are a known thing. They come from a trusted entity. It isn’t Joe Random Blogger telling you to withdraw 3.7%. 10-years from now, you don’t have to remember what random website you read the numbers on.
- Many financial institutions can send the RMD to you automatically. That means you don’t even need to do any work; money just shows up in your checking account. This is especially important when cognitive decline starts setting in.
However, the approach comes with some downsides as well and I felt like Vernon didn’t give them their due.
- There is really no such thing as an RMD before age 70. Vernon suggests using 3.5% withdrawals before you are 70. But this nullifies most of the above points. It is just a random number from a random person. Financial institutions won’t be able to do anything automatically for you. That means 5+ years of managing things on your own.
- It ignores taxable accounts. If you have any taxable savings (maybe you downsized the family home when you were 68?) then you’ll be left doing stuff manually anyway.
- Starting out with 3.5% withdrawals for the first 5 years of retirement is actually a substantial spending cut relative to most other retirement strategies. That cut comes when you are likely to be most active.
On top of that, a better method isn’t that much more complicated. Let’s look deeper at how the IRS RMD approach works compared to some other strategies.
First let’s take a look at our two high-level metrics for assessing retirement withdrawal strategies, WER and HREFF.
WER is “Withdrawal Efficiency Rate”. It tells us how much money the strategy leaves on the table — whether it is overly conservative. The closer the score to 1, the less money left on the table.
ConstantDollar is the traditional “4% withdrawals” rule. We already knows that this has a tendency to leave lots of money on the table. You often die with a portfolio of millions. We can see that IRS RMDs do reasonably well. They are definitely a lot better than 4% withdrawals, though they are also visibly a notch below VPW. VPW has a WER of 71% and IRS RMDs have a WER of 66%.
It is a small difference but we’ll see later how small differences can still result in noticeable improvements in the quality of our retired life.
Next let’s look at HREFF.
HREFF is a measure that tells us how often — and how far — our retirement spending drops below our “floor”. All of us have some kind of spending floor during our retirement. If a strategy doesn’t provide enough spending to meet that floor then we need to make cuts. The bigger the shortfall, the deeper the cuts, the more painful it is.
We can see that the 4% rule does terrible on this metric. This is because the 4% rule pulls out 4% a year…and then when the portfolio runs out we have to cut our spending to $0. This is a very deep cut, so it is heavily penalised in this metric, even though it is quite rare. In general, we’d rather cut spending $1,000 here and there rather than $40,000 in extreme scenarios.
We also see that the IRS RMD strategy does quite poorly here. The primary reason is that withdrawals start out quite low — just 3.5% of the portfolio compared to 4–5% for other strategies.
Does the lower initial withdrawal amount ever pay off? The intuition is “withdraw less now, so there is more later”. Does it really work that way, though?
At first glance, we might be tempted to say “yes, RMD wins in the long run”. But let’s dig a bit deeper. RMD quickly falls behind, due to the smaller starting withdrawals. It we track the total amount of spending over a lifetime, RMD never catches back up. What’s more, let’s remind ourselves:
- Money early is better than money later, simply because of the time value of money. You’d rather have $100 today than $100 a year from now.
- Money early is better than money later, because there is a chance you are dead later. By year 18 there is already a 50% chance you are dead.
- Money early is better than money later, because we all tend to become less active when we reach advanced ages. Our spending goes down because we’re not taking as many vacations or going to as many restaurants.
Instead of just eye-balling the chart, we can look at the “Certainty Equivalent Wealth”, which is a way of boiling variable withdrawals down to a single number. For the 1966 retiree, the CEW using the IRS RMD method was $29,857. That means that sequence of withdrawals was “like” just withdrawing a constant $29,857 a year instead. The CEW of the VPW method was $31,580. That’s a 5–6% improvement.
A 5–6% improvement isn’t massive but think about how you’d feel if your boss gave you a 5% raise (on top of inflation)? Pretty nice, right?
The IRS RMD method has one key insight: as we get older we should increase the percent of our portfolio that we withdraw, simply because we are closer to being dead, so we don’t need it to last as long. So we need to take that insight and apply it to our withdrawals. Whatever strategy we choose should do that.
If we are withdrawing 5% of our portfolio when we are 65 years old, there’s little reason to still be withdrawing 5% of our portfolio when we are 85 years old.
The major flaw of the IRS RMD method is simply that it is too conservative. When you are 75 years old, you’ll only be withdrawing 4.3% of your portfolio. We always need to balance portfolio longevity versus current spending but all available evidence suggests that the IRS RMD tables are overly conservative. Which isn’t a huge surprise — they weren’t designed for retirement needs. They were designed for tax reasons. They just happen to work pretty okay if you also use them for retirement spending.