The most important investing decision you'll make.
Studies suggest your asset allocation explains roughly 90% of your portfolio's behavior. Try a few mixes, see how the expected return, volatility, and outcome range shift.
A 60/40 stock/bond mix on $50,000 over 20 years. Roughly 9 in 10 historical outcomes would land between $86,846 and $385,272.
More stocks usually mean a higher expected return, and bigger drops along the way. The right mix depends on your time horizon and how much volatility you can sit through without selling.
Blend the two assets at your 60/40 mix to get the portfolio's expected annual return.
Portfolio return = (0.60 * 8.5%) + (0.40 * 4.0%) = 6.70%
Combine the two assets' swings into one portfolio volatility, assuming they move together only loosely (correlation 0.1).
Portfolio volatility = 10.1% a year
Compound your $50,000 at the expected return for 20 years to get the median ending balance.
Median = $50,000 * (1 + 6.70%) ^ 20 = $182,919
Spread that median into a 90% range using a lognormal model over the full horizon.
Pessimistic = $86,846 (5th percentile) Optimistic = $385,272 (95th percentile)
The 1.645 multiplier is the z-score for a 90% range under a normal log-return distribution. Real return distributions have fatter tails, so extremes happen more often than this implies.
Educational only. These are simplified long-run assumptions, not predictions or recommendations. Real returns are not guaranteed and will be reduced by taxes, fees, and inflation.
The 90% range is a guidepost, not a guarantee.
The "pessimistic" and "optimistic" numbers are a 90% range, meaning historically, roughly 9 in 10 outcomes would land between them. They aren't worst-case or best-case, just educational guideposts. Real markets occasionally do worse than the pessimistic case (1929, 2008) and occasionally better than the optimistic case.
More on this in the asset allocation lesson, which walks through how the split shifts as your time horizon and risk tolerance change.
Assumptions
- Stocks: 8.5% nominal expected return, 16% annual volatility. Educational defaults consistent with long-run U.S. equity averages before inflation; not a prediction.
- Bonds: 4.0% nominal expected return, 6.0% annual volatility. Educational defaults consistent with long-run intermediate-term bond averages; not a prediction.
- Stock-bond correlation: 0.1. Real correlation moves around (was near zero for decades, briefly went positive during the 2022 inflation shock).
- Returns are lognormally distributed. The 90% range is symmetric in log space (the optimistic side reaches further in dollars than the pessimistic side falls).
- No contributions or withdrawals during the horizon. This is starting-amount-and-time only.
- No taxes, no fees, no inflation.
Limitations
- The 90% range is not a worst-case or best-case. Roughly one in twenty real outcomes falls outside it on each side, including the rare big losses (1929, 2008) and rare big gains.
- Lognormal models underestimate the frequency of extreme events. Real return distributions have fatter tails.
- Only stocks and bonds are modeled. Cash, real estate, alternatives, and foreign-currency exposures behave differently.
- The two-asset variance formula assumes the correlation is constant. Real correlations move in stress periods, often the wrong way.
- It is not a prediction of any specific portfolio's actual outcome.
- It is not a recommendation to pick a specific stock-bond split. The right split depends on age, risk tolerance, account type, and other holdings.
- It is not after-tax or after-inflation. Both reduce the real value of the dollar figures shown.
- It is not personalized advice. A CFP can help with the household-specific call.