ROI Explained: Return on Investment, Annualised Return, and Its Limits

Return on investment (ROI) is the profit from an investment divided by its cost, expressed as a percent. It is the most quoted performance measure in finance and the most misused, because it answers only one question — how much, relative to what was put in — and ignores the two questions that matter equally: over how long, and at what risk. This guide covers the formula, the annualised form that restores the time dimension, the break-even and loss cases, and the comparisons where ROI misleads and where it is the right tool. The time dimension’s full treatment, with the compound annual growth rate, is in the companion article on rate of return linked below.

The formula: one subtraction, one division

ROI = (final value − initial cost) ÷ initial cost, times 100 for a percent. The numerator is the net profit: everything received, minus everything paid, including purchase price, fees, and any additional investment. The denominator is the cost: the money put in. Invest $10,000, receive $15,000 three years later, with no other flows: profit $5,000, ROI 50%. The formula cares about nothing else — not the years, not the risk, not the timing of any intermediate flows. That narrowness is its strength for a quick comparison and its weakness for a decision.

Costs belong in both halves. A property bought for $200,000 and sold for $240,000 with $15,000 of buying and selling costs and $10,000 of renovation has a profit of $15,000, not $40,000, and an ROI of 7.5%, not 20%. A stock bought for $10,000 and sold for $12,000 with $200 of commissions and $300 of dividends received has a profit of $2,100 and an ROI of 21%. The ROI Calculator takes the gain and the cost as entered, and the accuracy of its answer is the accuracy of the inputs: omit a cost and the ROI is overstated by exactly that cost’s share.

Multiple flows need care. An investment that pays dividends or rent during its life has a profit that includes those flows, not only the final sale. An investment with additional contributions has a cost that includes them, and the simple ROI formula — final minus initial, over initial — understates the cost and overstates the return when contributions arrive late. For irregular flows, the internal rate of return (IRR) is the right measure, because it weights each flow by its timing; ROI is the right measure only when there is one inflow and one outflow, or when the flows are few and their timing is immaterial.

Annualising: restoring the time dimension

Annualised ROI — the compound annual growth rate, CAGR — is (final ÷ initial)^(1 ÷ years) − 1. The $10,000 → $15,000 investment over 3 years: (1.5)^(1/3) − 1 = 14.47% a year. The same 50% over 10 years: (1.5)^(1/10) − 1 = 4.14% a year. The total ROI is the same; the annualised return differs by more than threefold, and the annualised figure is the one that compares with a bank rate, a bond yield, or another investment’s return. Any ROI quoted without its period should be treated as incomplete, and any comparison of two ROIs over different periods should be annualised first.

Annualising assumes the growth compounds smoothly, which real investments do not. A 50% total return earned as +80% then −17% then +5% annualises to the same 14.47% as a steady 14.47% a year, but the experience — and the risk — was different. Annualised ROI is a summary, not a history, and it should be read alongside the path: the volatility, the worst year, and the drawdown. The companion article on rate of return covers the annualised form, the arithmetic mean versus the geometric mean, and the time-weighted versus money-weighted distinction in full.

Simple annual return — total ROI divided by years — is the other annualisation, and it overstates the compound figure: 50% ÷ 3 = 16.67%, against the true 14.47%. The difference is the compounding the simple form ignores. Use the compound form for any comparison that matters, and treat a simple annual return as an approximation that flatters the investment by the gap between the two.

Losses, break-even, and the limits

Losses are asymmetric. A 20% loss needs a 25% gain to break even, not 20%: $100 falls to $80, and $80 needs $20 — 25% of $80 — to return to $100. A 50% loss needs a 100% gain. The asymmetry grows as the loss deepens, and it is the arithmetic behind the advice to avoid large losses rather than to chase large gains: the recovery from a deep loss is the hardest return in investing. ROI states the loss as −20%, which is correct, and the recovery as +25%, which is also correct, and the two do not cancel in the reader’s ear unless the asymmetry is understood.

ROI is meaningless when the cost is zero or negative. A leveraged investment — $10,000 of own money controlling $100,000 of property — has an ROI on the own money that is ten times the ROI on the property, which is correct leverage arithmetic and a misleading headline if the debt is not mentioned. A subsidised investment with zero net cost has an infinite ROI, which is arithmetic nonsense, not infinite value. In both cases the measure to use is the return on the capital at risk, with the leverage and the debt stated, not the bare ROI. And ROI ignores risk entirely: a 50% return earned by risking the whole capital is not the same as a 50% return earned by risking a tenth of it, and no ROI figure distinguishes them. Risk-adjusted measures — the Sharpe ratio, the maximum drawdown, the volatility — exist for that reason.

ROI’s proper uses are narrow and real: comparing similar investments over the same period, with the same risk, where the flows are simple; screening projects against a hurdle rate; and communicating a result to a non-specialist audience. Outside those uses, it should be paired with its period, its risk, and its flows. A 50% ROI over 3 years at low risk is an excellent result; a 50% ROI over 10 years at high risk is ordinary. The number is the same; the judgment is not, and the judgment is the work.

The hurdle rate: the return the investment must beat

Every investment should be compared with the return available without it — the hurdle rate. For a US investor, the natural hurdle is the yield on a government bond of matching maturity, or the rate on a high-yield savings account for money that must stay liquid: the return that requires no risk, no work, and no lock-up. An investment whose ROI does not beat the hurdle, after fees and taxes, has not earned its risk, however positive its ROI looks in isolation. A 6% ROI when the hurdle is 5% has earned 1 point for its risk; a 6% ROI when the hurdle is 7% has lost 1 point against doing nothing, and the loss is real even though the ROI is positive.

The hurdle rises with the risk and the commitment. Money locked away for ten years must beat the ten-year bond yield, not the savings rate, because the alternative for ten-year money is the ten-year bond. Money in an illiquid investment — a property, a private business — must beat the liquid alternative by a premium for the illiquidity, because the holder cannot exit when the thesis changes. And money whose return is uncertain must beat the hurdle by a margin for the uncertainty, because the expected return, not the hoped-for return, is what the hurdle compares. A hurdle that ignores risk, term, and liquidity is not a hurdle; it is a number chosen to make the investment pass.

The hurdle also disciplines the comparison between unlike investments. A property returning 8% and a stock portfolio returning 8% have not performed equally if the property required leverage, management, and illiquidity while the portfolio required none of them: the property’s hurdle was higher, and its 8% cleared a higher bar by less. The ROI figure alone cannot carry that comparison, and the hurdle is the adjustment that lets it. State the hurdle with every ROI that is presented as a success, and the success becomes checkable: did it beat the alternative, by how much, and at what risk.

Property ROI: cash-on-cash versus total return

Property has two ROIs, and they answer different questions. Cash-on-cash return is the annual cash flow divided by the cash invested: the worked rental example gives $4,415.16 of cash flow on $100,000 of down payment, a cash-on-cash return of 4.42%. Total return adds the other components: appreciation, the principal portion of the mortgage payments (which builds equity), and the tax effects, divided by the same cash invested. If the $200,000 property appreciates 3% in the year, that is $6,000 of gain, and the total return before tax and before counting principal paydown is ($4,415.16 + $6,000) ÷ $100,000 = 10.42% — more than double the cash-on-cash figure, on the same property, in the same year.

The gap between the two is leverage and appreciation, and both are double-edged. Leverage multiplies the cash-on-cash return when the property’s return exceeds the loan rate, and multiplies the loss when it does not: a property that falls 3% loses $6,000 on $100,000 of cash, a −6% return before cash flow, and the cash flow must cover it. Appreciation is unrealised until sale, and it is taxed on sale, with depreciation recapture — so the total return quoted before sale and before tax is a paper figure, and the realised, after-tax total return is the one that matters. Quote both, label both, and never present the paper total as the cash return.

The property ROI that should be compared with other investments is the realised, after-tax, risk-adjusted total return over the holding period, annualised. It is the hardest figure to compute and the only one that compares fairly, because every other figure omits something — the cash-on-cash omits appreciation, the total omits tax, the pre-tax omits the sale costs, and the unannualised omits time. The ROI calculator’s total and annualised figures are the starting point; the property’s full return needs the sale, the tax, and the years, and the honest presentation says which of those are in the figure and which are not.

Frequently asked questions