Historical inputs
When a visitor enters stock or ETF tickers, the model fetches historical monthly data for each holding. The lookback period can be changed in the simulator. Longer periods may include more market regimes, while shorter periods can reflect newer funds or securities with limited history.
The model estimates two different return series. For reinvested-dividend assumptions, it uses adjusted prices because adjusted prices include the effect of distributions. For no-DRIP assumptions, it uses regular closing prices so the price return is not treated as if dividends were reinvested.
- Expected return is estimated from the historical compounded annual growth rate.
- Volatility is estimated from historical monthly return variation.
- Covariance is estimated from overlapping monthly returns between the selected holdings.
- Dividend yield and dividend growth are estimated from available distribution history.
Portfolio path generation
Each simulation path applies stochastic returns across the holdings. The covariance matrix helps the portfolio behave like a group of related assets rather than a list of independent tickers. This matters because diversification depends on how assets move together during both normal markets and stress periods.
At each step, the simulator applies the user's assumptions for contributions or withdrawals, dividend reinvestment, recurring tax drag, expense drag, and rebalancing. The model tracks market value, estimated tax basis, and final-year dividend income so the results can show more than just a single ending balance.
Percentiles and distributions
The final output is summarized by percentiles. A P10 outcome means 10% of simulated paths finished below that value. A P90 outcome means 90% finished below that value and 10% finished above it. The percentile range is often more useful than the average because it shows how much the future can spread out over time.
Portfolio Modeler also estimates maximum drawdown for each path. A drawdown is the decline from a prior peak to a later low. This gives the user a rough sense of how painful a portfolio path could become before reaching the final year.
What the model does not know
No historical simulator can know the future. Tax law, inflation, interest rates, fund policies, investor behavior, business fundamentals, and valuation regimes can all change. The model is best used as a planning tool for comparing assumptions, not as a forecast or recommendation.