Percentage Change Calculator

Understanding the Percentage Change Calculator

A percentage change calculator compares two values and reports the increase or decrease as a proportion of the original. The choice of denominator is what makes it meaningful, and using the wrong one is the most common error in the whole calculation.

How it actually works

Enter an original and a new value. The calculator takes the difference and divides by the original. From 80 to 100 is an increase of 25%.

The deeper context most people miss

Percentage change and percentage difference are distinct calculations answering different questions. Change implies a direction from one value to another, while difference treats two values symmetrically and therefore divides by their mean, giving a figure that is the same whichever value you name first.

Why increases and decreases are not symmetric

Going from 80 to 100 is a 25% increase, and going from 100 back to 80 is a 20% decrease, because each is computed against a different starting point. This asymmetry is inherent rather than a quirk of convention, and it has consequences wherever changes compound. Recovering from a percentage loss requires a larger percentage gain, specifically x/(100−x) percent to recover a loss of x percent, which grows steeply: a 20% loss needs 25%, a 50% loss needs 100%, and a 90% loss needs 900%. This is the arithmetic behind why volatility damages compounded returns and why the arithmetic mean of period returns overstates actual performance, with the geometric mean being the correct average for anything multiplicative. In retail, successive discounts multiply rather than add, so 20% off followed by a further 10% is 28% off rather than 30%. In pay and pricing, a percentage rise followed by an equal percentage cut leaves the value below where it started. The general principle is that percentage changes are multiplicative factors, so sequences of them multiply, and any statement combining percentage changes by addition is wrong unless every change is measured against the same fixed base. Log returns are used in finance partly because they do add, which makes them convenient for aggregation over periods.

A worked example: percentage points against percent

From 80 to 100 is a 25% increase, and if those figures were themselves percentages the language changes: moving from 80% to 100% is a rise of 20 percentage points and a 25% relative increase, and both statements are correct while conveying very different impressions. This ambiguity is exploited routinely. A treatment reducing risk from 2% to 1% halves relative risk, a 50% reduction, while reducing absolute risk by 1 percentage point, and reporting only the relative figure makes a small absolute benefit sound substantial. Medical and policy reporting does this frequently enough that number needed to treat, the count of people who must receive a treatment for one to benefit, has become a preferred statistic precisely because it cannot be inflated by framing. In that example it is 100. The reverse framing also misleads: a relative increase of 100% in a very rare outcome may be practically negligible, and reporting only the relative change makes it alarming. The discipline that resolves both is to give absolute and relative figures together with the baseline, since a percentage change without its base is uninterpretable and a reader cannot judge magnitude from a relative figure alone.

Deciding how to present a change honestly

Several practices reduce misleading, including yourself. State the base alongside any percentage change, since the same relative figure means different things at different scales. Use percentage points for differences between percentages and say which you mean. Give absolute counts where the base is small, since percentages from small numbers are unstable and one additional case can move them substantially, which is why percentages from samples under about twenty are best avoided or shown with counts. Watch for changes computed from a base that is itself changing, which is a common error in year-on-year comparisons where the denominator differs each period. Beware chained percentages, where applying successive changes and reporting a total by addition understates or overstates the actual result. Consider whether an index would communicate better for a series, setting an initial period to 100 and expressing subsequent values relative to it, which makes cumulative change readable at a glance. For anything comparing groups, check that the denominators are comparable, since normalisation choices including per capita, per unit output, or per exposure produce different rankings from the same data and the choice is substantive rather than neutral.

Compounding, and why small rates matter over time

Repeated percentage changes compound multiplicatively, and the divergence from linear intuition grows quickly. The rule of 72 gives a serviceable approximation for doubling time, dividing 72 by the percentage rate per period, so 6% growth doubles in about 12 periods and 2% inflation halves purchasing power in about 36 years. That tool makes several consequences visible. An annual fee of 1% on an investment compounds into a substantial share of the final value over decades, because it applies to a growing base each year, and the intuition that 1% is small fails badly over a working life. Small differences in growth rate produce large differences in outcome over long horizons, which is why long-run projections are so sensitive to assumed rates and why small changes in those assumptions swing results enormously. In finance, nominal and effective rates differ when compounding occurs more than annually, which is why disclosure rules require an annual equivalent figure so products can be compared, and comparing nominal rates alone misleads. Continuous compounding gives the limiting case using e. In the other direction, exponential decay describes half-lives, depreciation, and drug clearance, and the same failure of intuition applies in reverse, with people consistently underestimating how much remains after several half-lives.

Variations: related measures and index numbers

Percentage difference divides by the mean of two values and is symmetric, suiting comparisons where neither value is the reference. Percentage error divides by the accepted or true value and is used in measurement. Relative change and relative difference are the general terms. Log change, computed as the natural logarithm of the ratio, is additive over periods and approximately equals percentage change for small values, which is why it is used in finance and econometrics for aggregating returns. Basis points express hundredths of a percent and exist specifically to avoid the percentage point ambiguity in finance. Index numbers rebase a series to a reference period, with consumer price indices being the familiar example, and the choice of base period and weighting method affects the result, which is why index construction is a technical subject with competing formulas including Laspeyres and Paasche. Compound annual growth rate expresses multi-period growth as an equivalent constant rate and is a geometric mean. Percentage of a percentage requires care, since taking 10% of a 20% share gives 2% of the whole rather than 30% or 10%.

Calculating and reporting percentage change

Divide by the original value for percentage change, and by the mean of the two for a symmetric percentage difference, since they answer different questions. Use percentage points for differences between percentages and say explicitly which measure you are quoting. State the base alongside any percentage, since the figure is uninterpretable without it. Report absolute numbers alongside relative ones where the base is small, since a percentage from a handful of observations moves substantially with one case. Multiply rather than add successive changes, since 20% off then 10% off is 28% off. Remember that recovering a loss of x percent needs a gain of x/(100−x) percent, so a 50% fall requires a 100% rise. Use the geometric mean rather than the arithmetic mean for averaging growth rates or returns. Use the rule of 72 to check compounding intuitions. Give both absolute and relative change when reporting risk or effect sizes, since relative figures alone systematically mislead about magnitude. And check that denominators are comparable when comparing across groups.

What people get wrong

• Dividing by the new value rather than the original, which computes a different quantity and makes an increase and its reversal appear equal when they are not.
• Adding successive percentage changes, when they multiply, so two discounts of 20% and 10% total 28% rather than 30%.
• Reporting a relative change without the baseline, when halving a 2% risk and halving a 40% risk are both 50% reductions with entirely different practical significance.
• Averaging growth rates arithmetically, when compounding requires the geometric mean and the arithmetic average systematically overstates actual performance.

Where the math comes from

Percentage Change = (New − Original) / Original × 100, using the original value as the denominator. Percentage difference between two values without a reference divides by their mean instead, making it symmetric. Changes compound multiplicatively, so successive changes multiply their factors rather than adding, and recovering a loss of x percent requires a gain of x/(100−x) percent.

Questions and answers

Related calculators

Golden Ratio · Set Operations · Percentage · Proportion Solver · Decimal to Fraction