Why compartmentalizing money isn’t a cognitive bias — it’s a retirement superpower
There is a concept in behavioral finance called mental accounting. It describes the tendency people have to treat money differently depending on where it is mentally “kept” — to think of the twenty dollars in your wallet as different from the twenty dollars in your checking account, even though they are mathematically identical.
Most behavioral finance literature treats mental accounting as a bias to be corrected. A cognitive shortcut that leads to irrational decisions. Something a sophisticated investor should learn to overcome.
I spent forty years watching clients do the opposite — and I think the academics had it exactly backwards.
The Cookie Jar
My earliest clients — retirees in the 1980s who had grown up in households shaped by the Depression and the post-war economy — used to tell me about the cookie jars in their parents’ kitchens.
Not one cookie jar. Several.
There was a food jar. A utilities jar. A gifts jar. A go-to-the-movies jar. Cash, earmarked by category, visible on the counter. When the food jar was full, the family ate. When the movies jar ran low, they stayed home. The boundary between jars wasn’t a rule you had to remember or a budget line you had to track. It was a physical wall. You could see it. You could touch it. You could watch it go down. Nobody called it mental accounting. It was just how money worked when money was real and visible and finite.
Why It Still Works — For Everyone
Here is what surprised me across four decades of practice: the clients who never saw a cookie jar respond to the same concept just as powerfully.
A 62-year-old engineer who has spent his career managing spreadsheets and abstract financial projections lights up the same way when I show him that his income for the next ten years is sitting in a specific account — separate, named, funded — that the stock market cannot touch.
It is not nostalgia. It is not a generational quirk. It is something deeper than either.
Mental accounting isn’t a flaw in human cognition. It is, I would argue, an adaptation. The brain did not evolve to manage abstract pools of interchangeable capital. It evolved to manage discrete resources with discrete purposes. The food jar and the movies jar worked because they matched the architecture of human decision-making, not because they outsmarted it.
The segmented retirement income plan works for the same reason. When a client can see — concretely, in an actual account with an actual balance — that the money funding next month’s income check has no stock market exposure, the anxiety doesn’t just decrease intellectually. It decreases viscerally. The boundary is real. The fence is visible. The brain can relax in a way it simply cannot when income and growth are pooled in a single account managed to a probability percentage.
The Academic Objection
I have presented this concept at advisor conferences for decades, and there is always an objection — usually from the back of the room, usually from someone with an impressive academic background.
“It’s all smoke and mirrors,” they say. “It’s still one big portfolio. The labels don’t change the math.”
They are technically correct. The dollars in Segment 1 and the dollars in Segment 5 are, from a portfolio theory standpoint, part of the same pool of capital. The segmentation is, in a sense, a construct.
But here is what that objection misses entirely: the client is not a portfolio. The client is a person.
And for a person — a person who watches the news, who has a spouse, who has a mortgage paid off and a grandchild in college and a memory of 2008 — the construct matters enormously. The label changes the behavior. The behavior changes the outcome. And the outcome, ultimately, is what the plan is for.
The behavior gap documented by DALBAR — the consistent, decades-long gap between what the market returns and what the average investor actually earns — is not a math problem. It is a behavioral problem. Investors sell at the wrong times because they panic. They panic because they cannot see that their income is safe. They cannot see that their income is safe because everything is in one pool, managed to a probability, and when the pool drops 40% it feels like the income is dropping with it.
Because in a conventional plan, it is.
The Problem Monte Carlo Doesn’t Solve
Here is the specific failure I want to name, because it gets almost no attention in the industry.
Monte Carlo simulation is a sophisticated tool. I do not question its mathematical integrity. What I question is what it tells the client — and, more importantly, what it never tells them.
Monte Carlo takes a pool of assets, runs thousands of hypothetical market scenarios, and reports back a probability of success: the percentage of scenarios in which the money lasted as long as the plan required. If the probability is high enough, the plan is approved. If it isn’t, the solution is to reduce spending or increase assets.
What Monte Carlo never does is look at a client with $1,248,000 in assets who needs $1,000,000 to fund their income plan and say: “Do you realize you have $248,000 more than you need for income? Do you know what that money could do? Have we talked about what job you want it to do?”
That $248,000 is invisible in a probability framework. It raises the probability number. It improves the odds. But it has no name, no purpose, no jar.
In a structured, segmented plan, that $248,000 becomes the surplus segment. It gets a specific job — or a choice of jobs. It can serve as the early warning reserve, ready to backfill any income segment that underperforms its target. It can serve as the longevity reserve, a source of continued income if the client lives longer than the plan assumed. It can serve as the legacy pool, growing aggressively over the full horizon of the plan to maximize what passes to heirs.
The client has to choose. And that conversation — what do you want this money to do? — is one of the most important conversations in retirement income planning. It is a conversation that a probability framework never prompts, because in a probability framework, excess assets are just a buffer that raises the percentage. They have no cookie jar.
What the Cookie Jar Knew
The families who lined their kitchen counters with labeled jars were not unsophisticated. They were, in a specific and important sense, more sophisticated than the modern retirement income industry.
They understood that money without a purpose is money at risk — not market risk, but behavioral risk. The risk that it gets spent on the wrong thing at the wrong time. The risk that when times get hard, you reach into the wrong jar because you cannot see the boundary.
The segmented retirement income plan is the cookie jar, rebuilt for the complexity of a 25-year retirement in a volatile market. Each segment has a name. Each has a purpose. Each has a specific start date and a specific end date and a specific account that funds it.
The income for years one through five comes from here. The income for years six through ten comes from there. The money you will not need for twenty years is invested aggressively, in a jar you are not allowed to open yet — and that discipline, enforced by structure rather than willpower, is what lets the long-term money do what long-term money does best.
The academic in the back of the room is right that it is all one portfolio.
But the client at the kitchen table, watching the jars, already knew something the academic is still learning.
Structure is not smoke and mirrors. Structure is how human beings actually make good decisions under uncertainty.
Philip G. Lubinski, CFP®, is the co-founder of IncomeConductor and the author of Probabilities Are Interesting Until You Become the Statistic: A Case for Structure Over Probability.
© 2026 Philip G. Lubinski, CFP®. All rights reserved. IncomeConductor LLC.