Start here
What should readers check before trusting a withdrawal-study headline?
Check what was tested: the spending rule, portfolio, time horizon, data, currency, fees, taxes and definition of success. Historical tests replay observed sequences; simulations generate paths from assumptions. Neither converts a reported success percentage into a personal guarantee. A portfolio can remain positive while spending falls below essential needs. Comparing studies requires matching their spending promises and examining shortfall severity, not simply choosing the largest starting withdrawal percentage.
What you’ll learn
- Historical survival and simulated success are different types of evidence.
- Positive ending assets do not necessarily mean that essential spending was maintained.
- Record the complete scenario, costs and currency before using a research percentage.
Read the question before reading the percentage
A withdrawal study turns assumptions into outcomes. To interpret its headline, identify the spending rule being tested, the portfolio, the retirement horizon and the event counted as success. A large percentage is not inherently more useful than a small one if it funds a different spending promise.
One paper may test constant purchasing power for 30 years. Another may permit frequent reductions, assume later pension income or evaluate only whether a positive balance remains. Comparing their starting withdrawals without those distinctions can create a false ranking.
A strong reading habit is to translate each headline into a full sentence: this policy met this criterion, over this horizon, under these data and assumptions. The qualifications are part of the finding.
Historical tests replay observed sequences
Historical analysis applies a withdrawal policy to past returns and inflation in their observed order. Bengen's 1994 paper examined US stock and Treasury-bond histories with inflation-adjusted withdrawals and discussed assets held in tax-deferred accounts. It is a historical US framework, not a direct test of a Canadian household's spendable income. Bengen: Determining Withdrawal Rates Using Historical Data
Historical sequences preserve events that occurred together: inflation, market falls and recoveries. But the observed record is finite. It cannot establish that a future sequence will be no worse than the hardest one in the sample, nor that another country's investors faced the same outcomes.
Overlapping periods also share evidence. A hypothetical dataset containing 50 complete years permits 21 complete 30-year annual windows: years 1–30 through years 21–50. Adjacent windows share 29 years. Those are not 21 independent histories of the economy.
Check how a study handles recent retirement dates. A start date fewer than 30 years before the data endpoint cannot already have 30 observed subsequent years. Any extrapolation, truncation or other treatment should be visible in the method. Otherwise the reader may mistake partly assumed futures for fully observed outcomes.
Simulations generate paths from a model
A simulation constructs possible sequences using specified return, volatility, inflation and relationship assumptions. This allows more paths than the historical record directly supplies. It also makes the results dependent on how the model generates those paths.
Morningstar's December 2025 US-dollar research used 1,000 simulated paths per allocation. Its base case required inflation-adjusted withdrawals for 30 years, with a 90% success standard and positive ending assets; fees and taxes were excluded. Its 3.9% highest base-case rate occurred at 30%–50% equities, with bonds and cash making up the remainder. Morningstar: The State of Retirement Income, 2025
A modelled 90% success rate is conditional on the model and inputs. It is not a personally measured 90% probability for a specific Canadian household. Generating additional trials can reduce simulation noise while leaving a mistaken return assumption or missing expense unchanged.
A simulation is therefore useful evidence about a specified system. The count of trials alone does not establish that the system describes the household's future.
Ask what must survive: assets, spending or both
A portfolio can remain positive while its withdrawals become too small to fund necessities. This can occur with a fixed percentage of the current balance: the payment shrinks with the portfolio. A model that celebrates avoiding literal depletion may therefore miss a substantial loss of living standards.
Conversely, a plan can deliberately spend down capital and still fund all scheduled expenses. Requiring the original principal to remain intact would define a different objective. An inheritance target needs its own amount, date and inflation convention.
These are distinct measures:
- Whether the investment balance stays positive.
- Whether every planned payment is funded.
- Whether essential spending is maintained after any allowed cuts.
- Whether a specified legacy remains at the end.
None can substitute silently for another. A flexible strategy's reported success should be read alongside the size, duration and timing of its permitted cuts.
Fictional CAD example: one table, three success rates
Consider a teaching dataset of 1,000 invented 30-year outcomes. It is not a simulation we ran or an empirical estimate. Each outcome concerns a flexible policy funding Canadian-dollar spending; all amounts are in initial-year purchasing power, after assumed fees and taxes. The essential portfolio-funded spending requirement is $30,000 annually. A separate legacy goal is $100,000 of real capital at year 30.
Suppose the invented outcomes fall into these mutually exclusive categories:
| Outcome | Number of cases |
|---|---|
| Essentials funded every year; at least $100,000 remains | 760 |
| Essentials funded every year; positive balance below $100,000 remains | 140 |
| Positive balance remains, but allowed cuts breached essential spending | 60 |
| Portfolio depleted and essential spending was not fully funded | 40 |
| Total | 1,000 |
On ending-balance survival, the success rate is 96%: 960 cases retain assets. On funding essentials throughout the period, it is 90%: 900 cases meet that test. On funding essentials and the stated legacy, it is 76%.
All three calculations use the same cases. Selecting the largest percentage does not improve any outcome; it changes the question being answered.
Change the success criterion and inspect the failures
Keep the invented cases unchanged but make the essential-spending test the headline instead of portfolio survival. The reported rate falls by six percentage points, from 96% to 90%. The 60 reclassified cases reveal a household problem previously hidden by the positive-balance criterion.
Even a single failure category can conceal different consequences. A $1,000 shortfall in year 29 differs from repeated $20,000 shortfalls beginning in year eight. A percentage alone does not show their severity, duration or the outside resources available to respond.
For research that reports median outcomes, ask about the less favourable cases as well. A comfortable median does not describe the experience of the household encountering an early market decline and persistent inflation. The distribution matters because retirement is lived along one path, not at the average of all paths.
Match costs, taxes and currency before applying findings
Read whether portfolio returns include investment expenses. The Ontario Securities Commission's investor-education site explains that mutual-fund fees reduce returns and distinguishes ongoing expenses from other charges. A model using fund returns already net of a particular expense should not deduct that same expense again. Separate fees may still remain. OSC: Mutual fund fees
Similarly, a gross withdrawal is not necessarily a net household payment. Tax depends on the account, income and applicable rules. USD market histories cannot become Canadian spending histories simply by replacing the currency symbol; exchange rates, inflation and the portfolio's actual exposures matter.
The review sheet should also record annual versus monthly withdrawals, beginning versus end-of-period timing, rebalancing, benefit start dates and the horizon. Small implementation differences can compound, especially when balances are falling.
Build a reproducible reading note
For a learning exercise, return to the fictional table and count cases that both fund essentials and retain some assets: 760 + 140 = 900, or 90%. Then ask whether that same result proves a 90% chance of success for a real household. It does not; the cases were invented and no predictive model was established.
For a real paper, save its date, exact scenario, data source and stated limitations beside any percentage copied into a planning worksheet. If an assumption is missing, record it as unknown. The result becomes useful when its boundaries are clear enough to compare with the actual question being studied.
Evidence behind the explanation
Sources & checks
Linked sources were consulted on October 28, 2026. The examples are original educational calculations. Source consultation and arithmetic checks do not constitute independent professional review.
- Determining Withdrawal Rates Using Historical Data (October 1994)William P. Bengen; Journal of Financial Planning · Accessed 2026-10-28
Original historical US investment and inflation framework, and explicit tax-deferred-account assumption.
- The State of Retirement Income: 2025 (December 3, 2025)Morningstar · Accessed 2026-10-28
Simulation methodology, 1,000 trials, 90% success, 30-year base horizon, 3.9% with 30%–50% equities, costs and taxes excluded.
- Mutual fund feesOntario Securities Commission; GetSmarterAboutMoney · Accessed 2026-10-28
Ongoing and other mutual-fund charges and their effect on investment returns.
When this article needs another review
- A cited research publication is corrected or its methodology clarified.
- A new primary study is added to the comparison.
- The example success criteria change, requiring the classification arithmetic to be checked.
Check the linked official guidance before applying a rule that may have changed.
About this educational article
Published by North of Numbers and created with AI assistance using the sources listed above. This article has not been reviewed by a qualified financial professional. Sources, assumptions and hypothetical examples are shown so readers can examine the reasoning.
This article explains general concepts; it does not assess individual suitability or provide personal financial, investment, tax, insurance, mortgage or legal advice.