Lowering sequence of withdrawals risk by living abroad and the craziness of Safe Withdrawal Rate analyses.
One widely known, but little discussed, aspect of the 4% rule is that is has a severe flaw around timing of retirement.
Adam and Bob identical twins, working identical jobs, and investing identically. Both 65 years old and both have $1,000,000 invested in an S&P 500 index fund. Adam decides to retire in January 2008 and the 4% rule tell him he can withdraw $40,000 a year for the rest of his life. Bob decides to work one more year and retires the following January. But now Bob’s portfolio has shrunk to $670,000 and the 4% rule tells him he can only withdraw $25,200 a year for the rest of his life.
But Adam and Bob are identical! Adam’s portfolio also shrunk…by the exact same amount as Bob’s. Why is he “allowed” to withdraw nearly 60% more for the rest of his life?
I explored this logical inconsistency almost a year ago…
The myopia of failure rates
Whenever you read about safe withdrawal rates, they usually lie to you about how often they fail. What if I told you…
…but today I saw a question about a real-world example of how this SWR-craziness manifests.
Let’s say you are going to retire early, at age 45, and have enough money to live off of 3.5% withdrawals. But you’d prefer a bit more of a buffer. So you’re going to go live in Thailand for 10 years, where you can live off of 1.5% withdrawals and then move home and live off of 3.5% of your original (10-years ago) portfolio value, safe in the knowledge that you’ve dodged the Sequence of Returns Risk bullet.
But have you?
It should be obvious that after 10 years, even with a very low 1.5% withdrawal rate, there’s no guarantee that your portfolio will above where you started.
How did we go from 100% success to 95% success by spending our first ten years in Thailand living on 1.5% withdrawals!?!
It is easy to check the historical results. What’s our portfolio value after 10 years of 1.5% withdrawals?
Of course, we really only care about the times when we ended up with less than what we started with. So let’s focus on those.
23 times out of 123, we ended up with less than we started with after 10 years. 18.6% of the time.
Now our time in Thailand is done. We want to move back home and start withdrawing $35,000 a year (3.5% of our original portfolio). But how much is that of our current portfolio?
It is a lot more than 3.5%. Let’s pretend this wasn’t 10 years into our retirement. Let’s pretend we had just retired now, at age 55, for the very first time. Would we think a 4% or 5% or 6% withdrawal rate is “safe” for a 55 year old? Clearly not.
Here’s a concrete example of what can go wrong. You retire in 1911 with $1,000,000 and move to Thailand for 10 years, living off of just $15,000 a year. But after a decade, your portfolio has fallen to $566,000 thanks to a series of financial panics. Would you really move back to the US and start withdrawing $35,000 from that portfolio? That’s a 6.1% withdrawal rate for a 55-year old.
What happened? When we view the entire retirement as an unbroken whole it looks “safe”. But when we start splitting into chunks and looking at those individual chunks, we see that some of them are unsafe by the very same definition.
We can see this more rigorous by composing the probabilities. Let’s go back to our 45-year old with $1,000,000 in an 80/20 portfolio using 3.5% withdrawals. What is the probability that it will last to age 100? That’s a typical SWR backtesting question, easily answered.
The answer: 3.5% never fails over a 55-year retirement. Of course, in part, that’s because we don’t have very much data about 55-year retirements. The last cohort begins in 1963.
But when split up the retirement into two segments, a 10-year retirement in Thailand followed by a 45-year retirement in the US, suddenly retirement becomes possible.
We have 138 total scenarios. In 115 of them we being our 45-year American retirement with a withdrawal rate at or below 3.5%, so we know that none of those fail. But what about the other 23 scenarios? 70% of them, nearly three-quarters, have a failure rate over 20%. 40% of them, two-fifths, have a failure rate over 40%.
We compose all 138 scenarios we have a success rate of 94.8%. How did we go from 100% success to 95% success by spending our first ten years in Thailand living on 1.5% withdrawals!?!
The way we’ve framed our scenario, it is obvious that we are “restarting” our retirement when we move back to the US and increase our withdrawal rate to 3.5%. But the reality is that, even when we don’t have a big, obvious change like that, we’re always “restarting” retirement.
Every January, we are “retiring again”. We are looking at our current portfolio value, our current withdrawals, our current life expectancy, and seeing if it all adds up. Safe Withdrawal Rate math ignores that reality. Or, perhaps more charitably, is implicitly sweeps it under the rug using “it’ll (probably) mean revert” as an excuse/explanation.
This is less of a problem for normal retirement lengths. But it is hard to ignore the problem for early retirees. Our intuition says that moving to Thailand and dramatically cutting expenses for a decade can only be a good thing when it comes to ensuring our retirement viability.
But the Safe Withdrawal Rate math says otherwise.