In 1946, while recovering from a heart attack in a Los Alamos hospital ward, the mathematician Stanislaw Ulam passed the time playing solitaire. Watching the cards fall, he began to wonder: rather than solving a complex probability problem through brute-force mathematics, what if you simply ran the experiment hundreds of times and observed the outcomes? His colleague Nicholas Metropolis, with a flair for the theatrical, named the resulting method after the casino in Monaco where Ulam’s uncle was known to squander his wages at the roulette table. The Monte Carlo method, as it came to be known, is now a fixture of modern science, engineering, and — for our purposes — financial planning. When projecting retirement income, the instinctive approach is to plug in an average annual return and trace a smooth line forward. But retirement does not unfold in smooth lines. Retirees live through dot-com crashes, global financial crises, and the particular misfortune of a bad sequence of returns in the early years of a portfolio’s drawdown phase — precisely when it can do the most lasting damage. Monte Carlo simulation accounts for this disorder by running thousands of possible market futures and reporting back on how the plan holds up across all of them. Where the gambler hopes to profit from uncertainty, the retirement planner tries to measure it honestly. That measurement, however, carries a cost that is easy to miss, and easier still to ignore until it is too late to correct.
The Price of Sleeping Soundly
When building a retirement plan, advisors typically anchor their spending recommendations to conservative return percentiles such as 75th, 90th and sometimes 95th. These numbers carry a specific meaning: a plan built on the 90th percentile return assumption is one in which the portfolio performed at least as well as projected in 90 out of 100 modelled scenarios. More plainly, the plan was constructed to survive conditions that would have derailed nine out of ten less cautious approaches. The intuitive appeal is obvious. Retirement is not a time for recklessness. The consequences of running out of money late in life — when earning capacity is gone, when health is uncertain, when options narrow — are severe and largely irreversible. Conservatism, in that sense, is not timidity, but prudence. Prudence, however, has a price and is worth looking at honestly. A plan built on a conservative return assumption recommends lower spending than a plan built on a moderate or median assumption by design. The question that rarely gets asked in the first meeting, and sometimes never gets asked at all, is: what does that extra layer of protection actually cost, in real terms, over a thirty-year retirement?
Consider what that spending gap represents in practice. It is the trip to Portugal that gets pushed to next year. It is the kitchen renovation that stays on Pinterest. It is the cheque that does not get written to help a child with their first down payment, in the years when that help would have mattered most.
There is a concept in planning circles called the retirement smirk: the empirical observation that retirees naturally spend less as they age, with a modest uptick near the end of life when health care costs rise. The implication is that the years of greatest capacity for enjoyment — the active, mobile, cognitively sharp early years of retirement — are also the years when the conservative plan is most likely to hold spending back. The portfolio, meanwhile, continues to grow. The irony is not lost on anyone who has watched a client reach their eighties with a larger portfolio than they had at sixty-five, having spent fifteen years declining dinner invitations to protect it.
Figure to rebuild CHART: The Retirement Smirk: annual spending by age, 60 to 90. Original PDF, page 26.
The goal of a sound retirement plan, then, is not simply to avoid running out of money. It is to avoid running out of money while also making sure the money runs out at approximately the right time — which is to say, when you no longer need it. That is a far more difficult problem, and one that a single conservative assumption, locked in at age sixty-five and never revisited, is ill-equipped to solve on its own.
Meet Our Retirees
To make this concrete, consider three fictional couples, each of whom retired in 1996 at age sixtyfive with $3,000,000 in investable assets. They hold identical portfolios: 60% equities, represented by the S&P 500, and 40% fixed income, represented by the US Total Bond Market. These couples are, by any measure, identical at the starting line. No CPP, OAS, pensions, employment income, or taxes. Their spending needs will rise with inflation each year, and they have no intention of leaving an estate beyond funding their own retirement through 2026.
The only thing that separates them is the return assumption their financial advisor used when calculating how much they could safely spend.
Monty and Carla plan around the conservative 90th percentile assumption, with a projected annual return of 5.09%. Under this assumption, their advisor calculates they can spend $73,108 per year.
Stanis and Lana plan around the moderate 75th percentile assumption of 6.38% with recommended spending of $84,822 per year.
Alphy and Betty plan around the median 50th percentile assumption of 7.69% and spend $97,302 per year.
The difference between the most and least conservative couple is $24,194 per year. The more important story, however, begins in 2011, when each couple returns to their financial advisor for a mid-retirement review.
Figure to rebuild TABLE: Couple, Percentile, Return Assumption and Annual Spending (1996$) for Monty & Carla, Stanis & Lana, and Alphy & Betty. Original PDF, page 27.
The First Spin: 1996–2010
Between 1996 and 2011, all three couples lived through the same market history: the exuberance of the late nineties technology boom, its spectacular collapse, a partial recovery, and then the 2008 global financial crisis — arguably the most severe financial shock since the Great Depression. Their 60/40 portfolio returned an annualized 6.93% over that fifteen-year period: close enough to Stanis and Lana’s planning assumption that their plan performed roughly as expected, meaningfully better than Alphy and Betty’s optimistic scenario, and considerably better than Monty and Carla’s conservative one. When each couple arrives at their advisor’s office in 2011, they bring their portfolio statement with them. The numbers below are expressed in 1996 dollars — stripping out the distorting effect of fifteen years of inflation (which totalled 43.4%) to allow a direct comparison. A dollar in 2011 bought approximately 30% less than a dollar in 1996; expressed in real terms, the gap between the couples is even more instructive than the nominal figures suggest.
Figure to rebuild TABLE: All portfolio balances expressed in 1996 dollars (CPI-U adjustment, factor 1.434). Nominal 2011 balances: Monty & Carla $1,926,482; Stanis & Lana $1,533,017; Alphy & Betty $1,113,823. Original PDF, page 28.
Monty and Carla, who spent $24,194 less per year in real terms than Alphy and Betty, arrive at the halfway point of their retirement with a portfolio that is $566,847 larger in real purchasing power. In purely mechanical terms, the conservative plan worked exactly as advertised. The question is what it cost them to get there — and what it means for the road ahead.
No More Bets: The 2011 Review
A mid-retirement review is not simply an occasion to confirm that things are on track. It is an opportunity to recalibrate entirely, take what is now known about the first fifteen years of returns, observe the actual portfolio balance, and run the simulation fresh with the remaining time horizon.
In 2011, each couple sits down with their advisor and asks the most natural question available: given where we are, what can we afford to spend for the next fifteen years?
The following tables present the answer under three planning assumptions, alongside the “hindsight” figure — the spending level that each couple’s actual portfolio would support given the returns that would actually be delivered between 2011 and 2026, which came in at an annualized 9.06%. All figures are in 1996 real dollars.
Figure to rebuild TABLES: Monty & Carla, Stanis & Lana, and Alphy & Betty: portfolio in 2011 (1996$), with new annual spending and hindsight maximum under the 90th, 75th and 50th percentile assumptions and the hindsight (actual 9.06% return) case. Original PDF, page 29.
The divergence is striking. Monty and Carla, who spent fifteen years below their most comfortable level, emerge from the review with genuine options. Even adopting the conservative 90th percentile assumption for the second half of their retirement, they can now spend $104,082 per year in real terms — 42% more than they had been spending. They can afford to be generous with themselves. They have choices.
Alphy and Betty are in a materially different position. Their most optimistic sustainable spending level — applying the median assumption to what remains — is $70,724 per year in real terms. That is less, in real purchasing power, than Monty and Carla were spending from the very first day of retirement. While a projection can revise future behaviour with great precision, it cannot restore the depleted capital of a previously overzealous plan.
For Stanis and Lana, the outcome is neither triumphant nor dire. Their moderate assumption provided enough ballast to weather two of the worst financial crises in a century without severely constraining their options. In 2011, they find themselves with a reasonable range of spending levels that all fall around the consumption they’ve had since their original retirement date. That is the middle path: not as much later-in-life flexibility as the cautious couple, but far less constraint than what is being imposed on the optimistic one.
After the Wheel Stops
It is tempting, having seen these numbers, to declare Monty and Carla the winners and recommend that everyone adopt the 90th percentile assumption in perpetuity. That conclusion, however, would miss the point almost entirely.
Monty and Carla did not know, in 1996, that the next fifteen years would include two of the worst financial crises since the 1930s. Nobody did. Had history been kinder — had the late nineties technology boom simply continued, or had 2008 been a mild recession rather than a near-systemic collapse — the picture might have looked quite different; Alphy and Betty’s assumptions would have proved reasonable, and Monty and Carla could have arrived at 2011 having unnecessarily constrained their own retirement for fifteen years in order to protect against a catastrophe that never arrived.
After the wheel stops, the winning number looks obvious. Before it spins, every number remains possible.
What the data does illuminate, rather clearly, is that the choice made at the beginning of retirement has outsized consequences. That’s not because the initial assumption is necessarily right or wrong, but because it compounds across decades of actual lived experience. The gap between Monty & Carla’s spending and Alphy & Betty’s is $24,194 per year. That is a trip to Japan with the grandchildren; a year of private school tuition; the down payment assistance that can actually change the trajectory of a young person’s life, rather than appearing in a will decades later when the moment has long passed.
There is, of course, a genuine asymmetry in the risks. Running out of money in old age, when the capacity to recover is gone and dependence on others becomes real, is a different kind of suffering than having spent the early years of retirement somewhat more cautiously than necessary. Nobody would argue the risks are equivalent. The question is whether the asymmetry is so severe that it justifies treating the conservative scenario as the only scenario worth planning for, and treating every year of accumulated conservatism as an unambiguous victory rather than a cost.
The Case for Regular Reviews
The framing of this analysis is, by design, an extreme simplification. In practice, no well-managed retirement plan sits in a drawer for fifteen years while the world changes around it. Markets crash and recover. Portfolio balances diverge from projections in both directions. Retirees discover that they are spending less than they planned — because the grandchildren live across the country, or because the health that would have fuelled the travel itinerary has become less reliable, or because the pleasure of accumulation, after decades of building, turns out to be harder to unlearn than anticipated. They also discover, sometimes, that they are spending more than projected — that health care costs arrived earlier than the model assumed, or that the renovation turned into two renovations.
A regular review, whether annual or triggered by meaningful changes in circumstances, transforms the retirement plan from a one-time wager into an ongoing conversation. It is the mechanism by which overly conservative assumptions can be relaxed when the evidence supports doing so, and by which spending can be pulled back before damage becomes permanent. It is also the mechanism by which the cost of caution can be made visible, not as a criticism of the original plan, but as honest information about what has been preserved and what can now be safely enjoyed.
Monty and Carla’s situation in 2011 is not simply good news. It is also a question: did they need to spend fifteen years at $73,108 per year? Were there years — the early years of the technology boom, for instance, when markets delivered exceptional returns — when a review might have revealed room to spend more? Would a trip to Tuscany in 2003 have been financially supportable, even as the dotcom crash was still working through the system? These are not rhetorical questions, but the conversations that a good review process is designed to surface.
The goal is not to spend as little as possible until the evidence confirms that spending more is safe. The goal is to spend appropriately given what is actually known, and to update that estimate as knowledge accumulates.
Retirement Should Not Be a Roulette Wheel
Monte Carlo simulation borrows its name from a casino, but its purpose in financial planning is the opposite of gambling. It does not tell you where the wheel will stop, nor does it eliminate uncertainty, or pretend that the future can be known with confidence. Instead, it asks a more worthwhile question: across the full range of plausible futures, does this plan survive? And across that same range of futures, is this plan asking too much of you in exchange for protection you may not need?
Conservative assumptions earn their keep. The couples who retired in 1996 did, in fact, live through conditions severe enough to vindicate a cautious approach. Sequence of returns risk is real, and the early years of a retirement portfolio’s drawdown phase carry disproportionate weight. None of that is theoretical.
But the price of caution is also real and is not paid in the abstract. It is paid in the specific, irreplaceable currency of time: those years when your energy is high and your ambitions are intact; when the grandchildren are young enough to be shaped by the experiences you share with them; when the bucket list items are still physically possible and the window for generosity is still open.
The most defensible retirement plan is not always the most conservative one. It is the one that is reviewed regularly, updated honestly, and calibrated to what is actually known rather than what was feared at the outset. It treats security and enjoyment not as opposites to be traded against each other, but as two obligations that a well-managed retirement owes to the person who spent a lifetime building it.
Whether you are approaching retirement or already well into it, the most important conversation worth having is not whether your plan is safe. It is whether your plan is generous enough to fund the life you actually want to live, in the years when that life is still fully available to you.
The wheel is spinning. The odds change as life unfolds. The goal is not to leave too many chips sitting on the table.
Methodology Notes
The assumptions used in this analysis: a 60/40 portfolio (S&P 500 / US Total Bond Market), maximum CPP and OAS, no pensions or employment income, no taxes, spending indexed to inflation annually, and a thirty-year retirement horizon beginning in 1996. The annualized portfolio return between 1996 and 2011 was 6.93%; between 2011 and 2026 it was 9.06%. Portfolio balances and post-2011 spending figures have been restated in 1996 dollars (CPI-U: 156.9 in 1996, 224.9 in 2011; cumulative inflation 43.4%) to allow direct comparison across time periods. The couples and scenarios described are fictional and intended for illustrative purposes only. This article does not constitute financial advice.