Skip to content

Model Limitations

monteplan is an educational and planning tool. Understanding its limitations is essential to interpreting results responsibly.

General Limitations

This is not financial advice. Results are simulations based on simplified models and assumptions. Always consult a qualified financial advisor for real planning decisions.

Simplified Tax Model

  • Only US federal taxes are modeled. State and local taxes are not included.
  • The model assumes a single filing status for the entire simulation horizon.
  • Tax-loss harvesting, estate taxes, and alternative minimum tax (AMT) are not modeled.
  • Capital gains are simplified -- the model does not track individual lot-level cost basis.
  • Tax bracket tables are static (2024 values); they do not adjust for future legislation.

No Behavioral Dynamics

  • The model assumes perfect adherence to the chosen spending policy (no panic selling, no lifestyle inflation beyond the plan).
  • Contribution rates are fixed; real-world saving behavior varies with life events.
  • Healthcare costs, long-term care, and other age-dependent expenses are not explicitly modeled.

Return Model Assumptions

  • MVN: Returns are independently and identically distributed (i.i.d.) across time. Real markets exhibit momentum, mean reversion, and volatility clustering.
  • Student-t: Captures fat tails but not skewness or time-varying volatility.
  • Bootstrap: Limited to the range of observed historical data. Cannot generate worse-than-observed crashes. Sensitive to the choice of historical sample period.
  • Regime switching: Regime definitions are user-specified, not estimated from data. The number of regimes and transition probabilities are assumptions, not inferences.

Inflation Simplifications

  • Inflation is modeled as a single aggregate rate, not component-based (healthcare inflation often outpaces general CPI).
  • The OU process assumes inflation is stationary and mean-reverting. Structural shifts (like the 1970s) require regime switching.

Portfolio Simplifications

  • The model assumes continuous rebalancing is possible at zero cost (no transaction costs, bid-ask spreads, or market impact).
  • Asset classes are represented as broad categories; individual security selection is not modeled.
  • Real estate, private equity, and other illiquid assets cannot be properly modeled in a mark-to-market framework.

Social Security / Guaranteed Income

  • Benefit amounts are assumed to be known and fixed (in real dollars before COLA).
  • The model does not account for Social Security claiming strategy optimization (when to claim).
  • Pension and annuity default risk is not modeled.

Longevity Risk

  • The simulation runs to a fixed end_age. There is no stochastic mortality model.
  • Users must choose a conservative end_age to account for the possibility of living longer than expected.

What This Means for Interpretation

  1. Success probability is approximate. A result of "80% success" does not mean there is exactly an 80% chance of a plan working. It means that under the assumed model, 80% of simulated paths survived.

  2. Sensitivity analysis is your friend. Because all results depend on assumptions, test how sensitive your plan is to changes in key parameters. A plan that is robust across a range of assumptions is more reliable than one that depends on specific values.

  3. Use multiple return models. Run the same plan under MVN, Student-t, and regime switching. If the plan succeeds under all three, it is more robust.

  4. Stress test explicitly. Overlay crash and high-inflation scenarios to see how the plan handles adverse conditions.

  5. Update regularly. Re-run simulations as your financial situation changes, and update market assumptions to reflect current conditions.