The Mathematics of Compounding and Sequencing of Returns Risk
The compound-interest identity FV = PV × (1 + r)n describes the long-run trajectory of capital, but it conceals a brittle assumption: that r is constant across each period. In real markets, periodic returns are realised as a stochastic sequence — a path — and the order in which positive and negative returns arrive can dramatically alter terminal wealth even when the arithmetic mean is identical. This phenomenon, known as sequence-of-returns risk, is the single most underappreciated source of dispersion in retirement and accumulation modelling, and it is the reason any projection that quotes a single “expected balance” is at best the midpoint of a wide distribution.
Geometric vs. arithmetic mean returns
A portfolio that gains 50 % and then loses 50 % has an arithmetic mean return of zero — but a geometric mean return of approximately −13.4 % per period, because the multiplicative chain 1.5 × 0.5 = 0.75 leaves the investor with seventy-five cents on the dollar. The gap between arithmetic and geometric means widens as volatility increases, an effect quantified by the volatility-drag approximation μg ≈ μa − σ²/2. Higher variance is not free; it is taxed silently by the geometry of compounding, which is why two strategies with identical average returns can produce terminal wealth that differs by an order of magnitude.
Why sequence risk dominates decumulation
During the withdrawal phase, identical average returns produce wildly different outcomes depending on whether the worst years are clustered at the beginning or the end of retirement. A retiree who suffers a 30 % drawdown in year one while withdrawing 4 % annually may exhaust capital in roughly twenty-two years; the same retiree experiencing an identical drawdown in year twenty-five may leave a substantial bequest. Our projection engine treats path dependency as a first-class citizen, sampling thousands of return sequences rather than relying on point estimates, and reports outcomes as a percentile distribution rather than a single deterministic curve.
Why Fixed Savings Fail to Beat Real Asset Volatility
Fixed-rate savings vehicles — high-yield savings accounts, money-market funds, short-duration Treasury bills, and most certificates of deposit — offer the comforting illusion of stability at a brutal long-run cost. Their nominal yield, however attractive in a high-rate regime, is consumed by two relentless forces: realised inflation and the opportunity cost of capital that could be earning the equity risk premium. The result is that “safe” instruments are often the largest source of unforced wealth destruction in a multi-decade plan.
The real-return trap
A 5 % nominal savings yield against 3.5 % CPI growth produces a real yield of just 1.5 % — and that figure ignores taxation, which in most jurisdictions taxes nominal interest at marginal income rates, frequently pushing the after-tax real return below zero. Over a thirty-year accumulation horizon, the gap between cash-like instruments and a diversified equity sleeve compounds into a multiple of terminal wealth, not a percentage point. Investors who anchor on the headline yield routinely underweight the base-rate effect of how small differences in real return compound across decades.
Volatility is a premium, not a penalty
Modern Portfolio Theory frames volatility as risk, but for long-horizon capital that does not need to liquidate during drawdowns, equity volatility is largely a risk premium harvested over time. The historical equity risk premium of roughly four to six per cent above short-term Treasuries is, in practical terms, the compensation paid to investors willing to tolerate temporary mark-to-market dislocation. Refusing that premium in exchange for liquid certainty is a defensible choice — but it is a choice, and our calculator makes the present-value cost of that choice explicit so it can be evaluated rather than assumed.
How Our Multi-Variable Projection Engine Models Alpha Overlays
Beneath the clean interface, the projection engine resolves a system of stochastic equations approximated through discrete-time Monte Carlo simulation. Each scenario evolves a portfolio along a sampled return path drawn from a parameterised distribution, with optional alpha overlays representing active-management edge, factor tilts, or systematic-strategy returns layered on top of the passive beta benchmark. Inputs flow through the engine as random variables, not point estimates, and outputs are reported as percentile bands rather than a single false-precision line.
Distributional assumptions and fat tails
We do not assume a pure log-normal distribution. Real equity returns exhibit excess kurtosis and negative skew — fat left tails and a higher central peak — which a Gaussian model badly underestimates. The engine optionally substitutes a Student’s t-distribution with calibrated degrees of freedom to reflect the empirical frequency of three-sigma and four-sigma events observed in historical S&P 500, MSCI World, and emerging-market index data. The practical consequence is that worst-case percentile outcomes are reported honestly, rather than understated by an assumption of normality the data refuses to support.
Alpha as an additive, decaying overlay
Skill-based excess return is modelled as an additive component with explicit decay: a two-hundred-basis-point gross alpha assumption is haircut for transaction costs, capacity constraints, and the empirically observed half-life of factor premia. The engine never assumes alpha is permanent; users can specify a decay schedule so that long-horizon projections do not extrapolate edge into perpetuity. This produces conservative terminal-wealth distributions that remain robust to the well-documented decline in active-manager outperformance once a strategy crowds, and it prevents the most common modelling error in retail projection tools — treating discovered alpha as a perpetuity.