A retirement projection that assumes 7% annual returns every year for thirty years isn't a plan. It's a best-case scenario dressed up as analysis. Real markets don't deliver smooth, average returns. They deliver sequences: some years up 20%, some years down 30%, with no predictable pattern and enormous consequences depending on when the bad years arrive.
Monte Carlo simulation is the tool that replaces the single-path projection with a realistic picture of retirement plan outcomes. Instead of asking 'what happens if you earn 7% every year,' it asks 'what happens across thousands of possible return sequences, including the historically bad ones?' The answer is a success rate: the percentage of scenarios in which your plan doesn't run out of money before your planning horizon.
A success rate of 90% means the plan survived in 900 out of 1,000 randomly generated return sequences. A rate of 65% means it survived in 650. These numbers don't predict the future. They describe the resilience of your plan across the realistic range of what markets might deliver.
Understanding what Monte Carlo means, how to interpret a success rate, what drives it up or down, and how behavioral flexibility interacts with it is essential for anyone building or evaluating a retirement income plan. This paper covers all of it.
The financial planning industry spent most of the twentieth century building retirement projections on deterministic models: assume a constant rate of return, run it forward, and show the client a line chart that ends with a satisfying balance at age 90. These models are intuitive, easy to build, and wrong in a systematic way.
The problem is that deterministic projections are essentially optimistic by construction. A consistent 7% return is more favorable than the same 7% average delivered in a volatile sequence because volatility during the withdrawal phase creates sequence of returns risk. The deterministic model doesn't capture this. It shows you the average outcome. Monte Carlo shows you the distribution of outcomes including the ones that end badly.
With markets at historically elevated valuations as of the mid-2020s and interest rates having gone through a significant repricing cycle, the range of plausible future return sequences is unusually wide. This is precisely the environment where the difference between a deterministic plan and a stress-tested one is most consequential.
Monte Carlo simulation generates a large number of hypothetical return sequences using statistical parameters derived from historical market data or specified by the planner. The simulation used in the retirement calculator at plan.johnkoyle.com generates 500 scenarios, each representing a possible sequence of annual returns over the full planning horizon.
In each scenario, the portfolio starts at the same balance, the same withdrawals are made, the same income sources are applied, and the same inflation adjustment is applied to spending. The only thing that differs is the sequence of annual returns. Some scenarios deliver strong early returns. Some deliver poor early returns. Most fall somewhere in between.
At the end of each scenario, the simulation checks whether the portfolio survived to the end of the planning horizon with a positive balance. The success rate is the percentage of scenarios that passed this test.
Four variables dominate the Monte Carlo success rate: the withdrawal rate, the portfolio's assumed return and volatility, the planning horizon, and the level of guaranteed income relative to spending needs. The withdrawal rate is the most powerful lever. A 3% withdrawal rate on a balanced portfolio produces a near-certain success rate across essentially all historical sequences. A 6% withdrawal rate produces a success rate that varies dramatically with the return assumption and the planning horizon.
Guaranteed income sources, Social Security, pensions, and annuities, reduce the effective withdrawal rate from the portfolio. Every dollar of guaranteed income is a dollar that doesn't need to come from the portfolio in any scenario, good or bad. This is one reason maximizing Social Security directly improves Monte Carlo outcomes.
A common question is what success rate is 'good enough.' The answer depends on what you're willing to do if the plan runs into trouble and how much flexibility you have in your spending. A retiree with a mostly fixed spending requirement, large fixed expenses, no ability to work, and no discretionary budget to cut needs a higher success rate than one who has significant discretionary spending, some ability to return to part-time work, and flexibility to adjust their lifestyle in a poor market environment.
Most financial planners consider a success rate above 85% to be solid, 70-85% to be acceptable with active monitoring and willingness to adjust, and below 70% to warrant material changes to the plan. These aren't official standards, but they reflect the professional consensus on what represents an acceptable margin of resilience.
One of the most important limitations of Monte Carlo simulation is that it assumes you withdraw the same inflation-adjusted amount every year regardless of what markets do. Real retirees don't do that. They reduce discretionary spending in bad years, delay major purchases, and adjust their lifestyle in response to market conditions. This behavioral flexibility means that most retirees' real-world outcomes are better than their Monte Carlo success rate suggests.
Research by Wade Pfau and Jonathan Guyton on dynamic withdrawal strategies shows that retirees who reduce withdrawals modestly in years following significant portfolio losses can meaningfully increase the portfolio's longevity without dramatically reducing their standard of living. The Monte Carlo success rate is therefore a somewhat conservative measure of real-world plan resilience for retirees who are willing and able to adapt.
William Bengen's 1994 research that established the 4% rule used historical sequence testing rather than Monte Carlo simulation. He examined every historical thirty-year period starting from 1926 and found that a 4% initial withdrawal rate, adjusted annually for inflation, survived every historical sequence. This approach is sometimes called historical simulation and is complementary to Monte Carlo rather than competing with it.
The limitation of pure historical sequence testing is that it can only model the return sequences that actually occurred. Monte Carlo generates sequences based on statistical parameters, allowing it to model a wider range of possibilities including scenarios more extreme than any in the historical record. Modern retirement planning uses both approaches as cross-checks.
Wade Pfau has compared Monte Carlo and historical simulation extensively. His research shows that the two methods often produce similar success rates for standard portfolio allocations and withdrawal rates, but can diverge significantly in scenarios with unusual starting valuations. When current market valuations are high relative to historical averages, Monte Carlo simulations based on long-run historical return parameters may be optimistic because they don't explicitly model the lower forward returns that elevated valuations have historically predicted.
Pfau advocates for using both methods and being appropriately cautious when the two diverge significantly. If historical simulation gives a success rate of 85% but Monte Carlo, using long-run parameters, gives 91%, the 85% may be closer to the realistic forward-looking answer given current conditions.
Morningstar's retirement research team has published work on what success rates actually mean for retirees' quality of life. Their analysis shows that a success rate of 90% doesn't mean there is a 10% chance the retiree will run out of money and face destitution. In most failure scenarios, the portfolio lasts well into the retirement horizon before being depleted, leaving the retiree with Social Security and other income sources. The failure mode is 'portfolio depleted before end of plan' not 'no income at all.'
This nuance matters for interpreting success rates. A plan that fails in Monte Carlo simulation typically fails because the portfolio is exhausted, not because the retiree has zero income. Social Security and other guaranteed sources remain. The severity of failure scenarios is often less dire than the binary success/failure framing suggests.
Michael Kitces has written extensively on how dynamic withdrawal strategies, including the guardrails approach and flexible spending rules, improve Monte Carlo success rates without requiring higher initial portfolio balances. His research shows that allowing withdrawals to decline modestly in response to poor market performance, by 5 to 10% in years following a significant loss, can improve success rates by 10 to 15 percentage points compared to a rigid fixed-dollar withdrawal strategy.
The most fundamental mistake in retirement planning is building the entire plan around a single assumed annual return. A constant 7% annual return is not conservative planning. It's optimistic planning that happens to use a number below recent historical averages. It ignores volatility, sequence risk, and the realistic distribution of outcomes. Any retirement plan worth having should include Monte Carlo analysis or historical sequence testing as the primary analytical tool, not a supplementary one.
Monte Carlo results vary slightly between runs because the return sequences are randomly generated. A success rate of 82% and 85% on identical inputs in two different simulation runs are not meaningfully different. They reflect the same underlying plan resilience with statistical noise. Retirees who make significant planning changes in response to small variations in Monte Carlo output are optimizing noise rather than signal.
Monte Carlo provides a probability distribution of outcomes, but it's also valuable to run the plan under specific stress scenarios. What happens if the first five years of retirement deliver returns similar to 2000-2009? What if inflation averages 4% rather than 2.5%? What if healthcare costs spike? These scenario analyses complement Monte Carlo by testing the plan against named, historically grounded adverse conditions.
A 90% Monte Carlo success rate doesn't mean there's a 90% chance you'll be financially secure. The success rate is a narrow technical measure: the percentage of simulated scenarios in which the portfolio didn't deplete before the end of the planning horizon. It doesn't capture the quality of life in failure scenarios, the availability of other income sources, or the behavioral adjustments you could make in response to poor market performance. It's a useful tool, not a complete description of your financial situation.
Many retirees use 85 or even 80 as the planning horizon in their Monte Carlo analysis. For a 65-year-old couple, the probability that at least one spouse survives to 90 is over 60%, and to 95 is around 35%. A plan that succeeds 90% of the time to age 85 but fails in most scenarios beyond that isn't providing the security most retirees assume. The planning horizon should be set at 90 to 95 for most retirees, not at expected life expectancy.
The quality of a Monte Carlo analysis is only as good as its inputs. The key variables to review are the expected return assumption and standard deviation, the withdrawal rate, the planning horizon, the income sources that offset portfolio withdrawals, and the inflation assumption for spending. Conservative inputs, a slightly lower return assumption, a higher standard deviation, a longer horizon, produce a more realistic picture of plan resilience than optimistic ones.
The success rate is a summary statistic. The more informative picture is the distribution of outcomes across scenarios. Looking at the 10th percentile outcome, the 25th percentile, the median, and the 90th percentile gives a richer picture of how bad the bad outcomes are and how good the good outcomes are. The retirement calculator at plan.johnkoyle.com shows this distribution across the Monte Carlo scenarios, not just the summary success rate.
Monte Carlo is most useful not as a one-time snapshot but as an interactive tool for testing the impact of specific decisions. What does the success rate look like if you delay Social Security to 70? What if you reduce planned spending by $10,000 per year? What if you do $50,000 of Roth conversions annually for the next ten years? Running these scenarios and comparing success rates turns the Monte Carlo from a report card into a planning tool.
The answer should describe whether the simulation uses historical returns, parameterized distributions based on historical data, or some combination. If the advisor uses a return assumption that is meaningfully more optimistic than long-run historical averages, the success rate they're showing you is overstated.
The advisor should give you a specific number, explain what the failure scenarios look like in terms of when the portfolio depletes and what income remains, and explain what the success rate means for your planning confidence.
If the advisor is using 85 or 90 as the horizon, ask what happens at 95. The answer reveals whether the plan has meaningful tail risk that isn't visible in the standard analysis.
This question turns the simulation into a decision support tool. The answer should identify two or three specific, actionable changes, whether spending adjustments, Social Security timing, Roth conversion amounts, or other variables, and quantify their impact on the success rate.
This is the stress test question. Given current valuation levels, lower-than-historical forward returns are a plausible scenario. The advisor should be able to run the simulation with a return assumption 1 to 2 percentage points below the base case and show you the impact. If the plan fails this test, it's more fragile than the base case suggests.
The retirement planning calculator at plan.johnkoyle.com was built to model exactly the dynamics discussed in this paper. The Monte Carlo tab in the retirement calculator at plan.johnkoyle.com runs 500 simulated return sequences using your inputs and shows your plan's success rate across those scenarios. The simulation also displays the distribution of outcomes at the 10th, 25th, 50th, 75th, and 90th percentiles so you can see not just whether the plan succeeds but how it performs across the range of possibilities. The Action Plan tab identifies the specific changes that would most improve your success rate, ranked by impact. Enter your portfolio balances, your income sources, your planned spending, and your retirement age, and the calculator shows you where your plan stands across the realistic range of market outcomes.
John Koyle, AIF®, is the Co-Founder of Red Cedar Wealth Advisors, headquartered in Pocatello, Idaho. He holds the Accredited Investment Fiduciary (AIF®) designation, awarded by the Center for Fiduciary Studies (Fi360), which signifies completed coursework, a rigorous examination, and ongoing continuing education in fiduciary responsibility and prudent investment practices.
John serves individuals and families in or approaching retirement throughout eastern Idaho, the Pacific Northwest, and across the country via Zoom. Clients work directly with John, not a junior team. Red Cedar Wealth Advisors operates under Osaic Wealth, Inc. (Member FINRA/SIPC) and Osaic Advisory Services, LLC for investment advisory services.
Every client relationship is built on five integrated disciplines. The proportions shift with the client and the moment. The integration is constant. Most of the actual value sits in how these disciplines connect, not in any single one of them.
Retirement planning is a discipline that barely existed a generation ago. For most of modern history, people worked until they physically couldn't and then they died, usually fairly close together. The idea that an individual should spend decades saving, then spend decades drawing down those savings, with a plan that accounts for inflation, taxation, sequence risk, healthcare, longevity, and estate transfer, that's maybe forty years old.
Which means almost nobody your age grew up watching their parents do it properly. There's no cultural muscle memory. The advice industry backfilled that gap with rules of thumb that work sometimes and fail catastrophically the rest of the time. The 4% rule. 'You can take Social Security at 62.' 'Target-date funds will handle it.' These are not bad starting points. They are terrible ending points. Real planning is specific, personal, and built on principles that hold up across the full range of outcomes, not just the average one.
The tax code is a set of instructions Congress wrote to shape behavior. Aligning your financial life with those instructions is not aggressive planning. It is the planning. Roth conversion sequencing, capital gains compression on concentrated positions, charitable remainder trusts, Social Security timing, beneficiary coordination. None of these are obscure. They're all legitimate, all written into the code intentionally, and all under-used. Most CPAs handle compliance. The strategy work is a different discipline entirely.
Seventy percent of family wealth doesn't survive two generations. The reason isn't bad markets. The cause is failure to communicate, outdated documents, and no plan for preparing heirs. Beneficiary designations regularly override well-crafted wills. Trust structures created a decade ago no longer match current law or current family. The portfolio that built the wealth isn't the structure that transfers it. The work here is coordinating with your attorney to align documents, beneficiaries, gifting strategies, and trust funding with what you actually want to happen.
Every week, before any portfolio decision, John runs through four layers of market health: economic conditions, market internals, valuations, and sentiment. Each layer gets scored and those scores combine into a composite that maps directly to a portfolio posture, from aggressive on the positive end to defensive on the negative. The work is in the scoring. Once the scoring is done, the positioning follows. That removes one of the most dangerous things in investing: making it up as you go.
The protection gap quietly kills more plans than markets do. A significant net worth paired with inadequate liability coverage is a lawsuit away from a serious problem. Long-term care is more acute: a multi-year care event for one spouse can consume what was meant for the survivor. Most advisors relegate risk to a footnote because insurance conversations are uncomfortable. The math doesn't care. Coverage adequacy, umbrella sizing, long-term care planning, and life insurance structure sit alongside the portfolio in any complete plan.
These are the principles behind every plan John builds.
To begin a conversation, visit johnkoyle.com, use the retirement planning calculator at plan.johnkoyle.com, or reach John directly at john@redcedarwealth.com or (208) 915-8400. Initial consultations are complimentary and carry no obligation.