SQUIRREL! ยท Ten Years Out
The Decommission Plan
I spend my working life sizing systems so they don't fail under load. Recently it occurred to me I'd never run that discipline on the one system I'm personally running out of runway on.
โ Back to SQUIRREL!I'm not retired. I'm close(ish). By the calendar I've got 12 years left in this career, maybe 10 if I'm lucky and the market behaves. That's exactly the window where the question stops being theoretical. Early in a career, retirement is a checkbox on an HR form and a number you don't look at too closely. Late in one, it turns into an actual number. The number, and numbers are the one thing I have never once been able to leave alone.
In my day job, nobody gets to say "we'll figure out capacity later." You size the box for the load it has to carry, you build in headroom for the load you didn't predict, and you stress-test the assumptions before the business bets on them. Somewhere around year fifteen of doing that professionally, it occurred to me that I had never run the same exercise on my own retirement account. I had a number in my head, like everyone does. I did not have a model behind it.
The 4% Rule Is a Rule of Thumb, Not a Law of Physics
The number most people reach for is the 4% rule. Withdraw 4% of the balance in year one, give yourself a raise for inflation every year after, and according to research going back to the '90s, thirty years later you probably haven't run the place empty. It's a solid starting point. It's also hiding the same thing every rule of thumb hides. Assumptions! Change the return, change the inflation, change how many years the money actually needs to survive, and 4% stops looking like a constant and starts looking like a slider.
Getting to Your Number
The simulator below assumes you already know your starting balance. If you're twenty years out, that number is still mostly fiction. Plug in anything and you're really just modeling numbers with extra steps. If you're close enough to count the years left on one hand plus a few fingers, you can actually project it. Not perfectly but, well enough to stop guessing.
So here's the buildout half of the exercise, before the drawdown half:
- Current balance.
- Years left.
- Salary/percent you're contributing.
- What your employer kicks in.
- Whatever's sitting in an old 401(k) or an IRA that isn't getting fresh money anymore.
- Run it forward and it hands you a projected balance at retirement.
- Drop that number straight into the simulator below to see how it actually holds up.
- Whew, I'm tired just reading it.
So I Built the Slider
Below is a full retirement drawdown simulator: starting balance, years in retirement, average return, average inflation, and a withdrawal rate you can drag anywhere from 0% to 10%. Run it and it'll show you the year the money runs out (if it does), what that balance is actually worth after inflation has been eating it for a few decades, and the highest withdrawal rate your specific assumptions can sustain without ever hitting zero. It isn't a plan. It's a sizing exercise. The same kind I'd run before telling a customer what they actually need to buy, except the workload here is the rest of somebody's life. If you already ran the buildout above, its handoff button drops your projected balance straight into the "Starting Balance" field below.
If You've Still Got a Decade of Paychecks Left
Here's the honest way to use this if you're where I am. Ten, twelve years out, still working, still contributing. Don't plug in today's balance and pretend you're retiring this afternoon. Run it forward. Plug in roughly what your balance will be when you actually plan to stop, and see how the withdrawal rate you're implicitly counting on holds up against the years you'll actually need it to last. Then go look at what you're contributing this year, because that's the only lever in this entire model you can still reach out and pull by hand.
I'm not a financial advisor, and none of this is financial advice. I'm an infrastructure guy with a spreadsheet habit and an occupational allergy to unstated assumptions. I've sat across the table from enough people who got surprised by systems they never stress-tested to know the pattern. Might as well run the model on the one balance sheet that's actually mine.