Monte Carlo Simulation for Retirement Planning: A Guide for Serious Investors

Retirement planning has traditionally relied on deterministic projections: take an assumed portfolio return of 5% per year, apply it every year for 30 years, and read off the ending balance. The problem is that this approach is unrealistic in a fundamental way: actual investment returns do not arrive uniformly at 5% per year. They vary substantially from year to year, and — critically for retirees — the sequence in which those returns arrive matters enormously.

Monte Carlo simulation addresses this by modelling the uncertainty in future returns explicitly. Instead of a single deterministic projection, it runs thousands of simulations of possible future return sequences, producing a distribution of outcomes and a probability of success for any given retirement income strategy. This guide explains how Monte Carlo simulation works, what its outputs mean, and how to use it in the context of UK retirement planning.

What Monte Carlo Simulation Is

A Monte Carlo simulation models the random nature of investment returns by repeatedly sampling from a distribution of possible return outcomes. In a typical retirement planning implementation:

1. The planner specifies assumptions: expected long-run returns and volatility for each asset class (equities, bonds, cash), correlation between asset classes, and inflation assumptions.

2. The simulation randomly generates a sequence of annual returns for each asset class over the planning horizon (say, 30 years), drawing from the specified distributions while respecting the correlation structure.

3. The simulation applies the client's specific circumstances to each scenario: starting portfolio value, planned withdrawals (adjusted for inflation), any additional income sources (pension, rental income, part-time work), and any large expenditures.

4. Steps 2 and 3 are repeated many times — typically 1,000 to 10,000 iterations — to produce a distribution of possible ending portfolio values.

5. The outputs include a "probability of success" — the percentage of simulations in which the portfolio survives the full planning horizon without being depleted — and percentile charts showing the range of possible outcomes from poor (bottom 10%) to good (top 10%).

The Probability of Success Metric

The central output of a retirement Monte Carlo analysis is the probability of success: the proportion of simulated scenarios in which the portfolio does not run out of money before the end of the planning period (typically set to life expectancy plus a safety margin, often age 90 or 95).

A probability of success of 90% means that in 90% of the 10,000 simulated return sequences, the portfolio sustains the planned withdrawals for the full period. In 10% of sequences, the portfolio is depleted before the period ends.

What probability of success should investors target? This depends on:

Risk tolerance and flexibility: An investor with minimal spending flexibility — who cannot reduce withdrawals or generate additional income if the portfolio underperforms — should target a higher probability of success (85–95%). An investor with meaningful flexibility (could reduce discretionary spending, receive additional income, or downsize a property) can accept a lower probability (70–80%).

The consequences of failure: If portfolio depletion in a bad scenario would mean genuine hardship, plan conservatively. If it would mean adjusting lifestyle modestly, some probability of shortfall is acceptable.

Interaction with guaranteed income: State pension, defined benefit pension income, and annuity payments act as a floor of guaranteed income. A client with significant guaranteed income relative to their total spending requirement needs a lower probability of success from the investment portfolio because the consequences of portfolio underperformance are partially buffered.

Sequence of Returns Risk

One of the most important concepts that Monte Carlo simulation illuminates is sequence of returns risk — the risk that poor investment returns occur early in retirement rather than later, with a disproportionate impact on portfolio longevity.

Consider two retirees, each with a £1,000,000 portfolio, each withdrawing £50,000 per year, both experiencing the same average annual return of 7% over 30 years. If Retiree A experiences poor returns in the early years (say, -20% in year 1 and -15% in year 2) and good returns later, and Retiree B experiences the reverse (good returns early, poor returns later), their outcomes will be dramatically different despite identical averages. Retiree A, depleting capital at a depressed value in years 1 and 2, runs out of money before Retiree B.

This asymmetry — which deterministic projections completely miss — is sequence of returns risk. Monte Carlo simulation captures it by modelling actual year-by-year sequences rather than a single average. The spread of outcomes in a Monte Carlo analysis reflects, in large part, the range of possible sequences.

Safe Withdrawal Rates: The Kitces Research

The question of how much a retiree can safely withdraw each year without depleting the portfolio is addressed by safe withdrawal rate research. The foundational work is William Bengen's 1994 study "Determining Withdrawal Rates Using Historical Data", which found that a 4% initial withdrawal rate (of the starting portfolio, adjusted for inflation annually) had historically survived all 30-year historical periods in US data — now known as the "4% rule".

Michael Kitces and Wade Pfau have subsequently refined this analysis. Key findings:

• The 4% rule is calibrated to 30-year planning horizons in US historical data. For longer horizons (35–40 years, relevant for early retirees) or for portfolios with lower equity allocations, the safe withdrawal rate is lower — perhaps 3.3–3.5%.

• Valuations matter: the safe withdrawal rate is lower when equities are highly valued at the start of retirement (high Shiller CAPE) and higher when valuations are depressed.

• Flexible withdrawal strategies — reducing withdrawals in years when the portfolio has underperformed and increasing them when it has outperformed — are substantially more efficient than rigid fixed withdrawals, allowing higher average income over the retirement period while maintaining portfolio survival probability.

• The 4% rule has a US equity bias. For international investors using global equity allocations, the findings broadly hold but with somewhat higher variability.

UK-Specific Application: The State Pension as Annuity Floor

UK retirement planning has an important structural element that US-centric research does not fully capture: the State Pension. For 2026/27, the full new State Pension is £241.30 per week, or approximately £12,548 per year (subject to change). For a couple, both entitled to the full State Pension, this represents approximately £25,000 per year of guaranteed, inflation-linked income — an implicit annuity with substantial value.

In a Monte Carlo framework, the State Pension is modelled as guaranteed income that reduces the net drawdown from the investment portfolio. For a couple needing £80,000 per year total income, the State Pension contribution of around £25,000 means only about £55,000 needs to be drawn from investments — substantially improving the probability of portfolio survival. The effective safe withdrawal rate as a percentage of portfolio is higher when a significant income floor exists.

Private defined benefit (final salary) pension income and annuities purchased with pension funds provide similar buffering effects and should always be explicitly modelled in retirement planning.

Adjustable vs Fixed Withdrawal Strategies

Research consistently finds that flexible or adjustable withdrawal strategies substantially improve retirement outcomes compared to rigid fixed withdrawals. Several practical approaches:

Guardrails approach (Guyton-Klinger rules): Reduce withdrawals by a defined percentage if the portfolio falls below a specified threshold; increase them if the portfolio significantly exceeds plan. This adapts spending to portfolio performance while maintaining a planned range of income.

Percentage-of-portfolio withdrawals: Take a fixed percentage of the portfolio value each year (say, 4.5%). In good years, income rises; in poor years, it falls. Guarantees portfolio survival at the cost of income certainty.

Rising equity glidepath: Some research (Kitces and Pfau) suggests that starting retirement with a lower equity allocation and gradually increasing it reduces sequence of returns risk — counterintuitively, maintaining more upside exposure later in retirement when the portfolio is smaller reduces the long-run risk of depletion.

Limitations of Monte Carlo Simulation

Monte Carlo simulation is a powerful tool but has real limitations that users should understand:

Garbage in, garbage out: The probability of success is only as reliable as the input assumptions (expected returns, volatility, correlations). If future returns are materially lower than historical returns — as many analysts project, given current valuations — probability of success estimates using historical parameters will be optimistic.

Normal distribution assumptions: Most Monte Carlo implementations model returns as normally distributed. Actual equity returns exhibit fat tails — the probability of extreme negative events is higher than normal distributions suggest. This means that the "failure" scenarios in a Monte Carlo analysis may be more severe than the simulation implies.

Static assumptions: Life changes. Spending needs change. Health changes. Income sources change. A static Monte Carlo plan must be revisited regularly — ideally annually — to incorporate changes in circumstances.

The single-number trap: A probability of success of 87% is not a hard guarantee of 87% success. It is an estimate of probabilities conditional on a set of assumptions, all of which are uncertain. Treat the output as one useful data point among several, not a definitive verdict.

Software Tools and Professional Implementation

Professional financial planning in the UK increasingly uses Monte Carlo-capable cashflow modelling platforms. Timeline (by Timeline App), Voyant, and CashCalc are among the most widely used tools in UK financial planning practices. Private banks and wealth managers typically use proprietary systems with embedded Monte Carlo capability.

For investors who want to engage meaningfully with their adviser's planning analysis, asking to see the Monte Carlo output — the distribution of outcomes, the probability of success, and the sensitivity of the result to key assumptions — is entirely reasonable. A good adviser should be able to explain the assumptions driving the analysis and the actions that would improve the probability of success if circumstances change.

Compliance and Regulatory Note

Monte Carlo simulation is a planning tool that provides probabilistic estimates of possible future outcomes; it does not predict or guarantee future investment performance. All investment returns are uncertain and investments can fall as well as rise. Tax treatment, State Pension entitlement, and pension rules may change over the planning horizon. A probability of success figure is contingent on stated assumptions and is not a reliable guarantee of outcome. This article is for information only and does not constitute personal financial advice. Seek qualified professional advice before making retirement planning decisions.

How Global Investments Can Help

Retirement income planning for internationally mobile, high-net-worth individuals involves complexities beyond standard UK cashflow modelling: multiple currencies, cross-border pension entitlements, property assets, offshore vehicles, and international tax considerations. At Global Investments, we work with clients to build retirement income strategies that account for the full balance sheet, use evidence-based withdrawal frameworks, and are stress-tested against historical drawdown scenarios. If you would like to discuss your retirement income strategy — including a Monte Carlo analysis of your current position — please contact our advisory team.