Percentage Change vs Percentage Points — Why 10% to 15% Isn't a 5% Rise

Published 2026-04-07 · By the CalculatorHive editorial team

Two headlines describe the same central bank decision. One says the rate rose by half a percent. The other says it rose by fifty percent. Both can be literally true, and the gap between them is the single most common numerical error in public writing. This post gives you the two formulas, shows exactly when each one is the honest choice, and covers the asymmetry that trips up investors, discount shoppers, and anyone reading a poll.

What is the difference between percentage points and percent?

Use percentage points when you are comparing two quantities that are already percentages. Use percentage change when you are describing how much one quantity grew or shrank relative to where it started.

• Percentage-point difference = new percentage − old percentage. A savings rate that goes from 10% to 15% rose 5 percentage points (usually abbreviated pp).
• Percentage change = (new − old) ÷ old × 100. The same move is (15 − 10) ÷ 10 × 100 = 50%.

Neither number is wrong. They answer different questions. The percentage-point figure tells you how much of the underlying population, budget, or balance changed hands. The relative figure tells you how big the change was compared with the starting point. Saying "the rate rose 5%" when you mean 5 percentage points is what's wrong, because a reader will divide by the base and get a completely different picture.

The trap is that the same percentage-point move means wildly different things depending on where you start:

A one-point move off a 2% base is a huge story. The same one-point move off a 40% base is barely news. That is why the choice of phrasing is an editorial decision as much as a mathematical one — and why you should be suspicious when a press release picks the flattering framing without saying which one it used. Our percentage change calculator reports both figures side by side so you can't accidentally quote the wrong one.

Why does this matter for interest rates, polls, and unemployment?

Three domains where the confusion causes real misreadings:

Interest rates. A mortgage rate going from 4% to 5% is a 1 pp increase and a 25% relative increase. Neither number tells you what happens to your payment, because payments do not scale with the rate — on a 30-year loan, that 1 pp change raises the monthly principal-and-interest payment by roughly 12%, not 25%. Rates are quoted in basis points precisely to avoid ambiguity: 1 bp = 0.01 pp, so a 25 bp move is unmistakably 0.25 pp.

Polling. A candidate moving from 42% to 45% gained 3 percentage points. Reporting that as "a 7% gain" is technically the relative change but useless to a reader trying to judge whether the lead is real. Crucially, a survey's margin of error is stated in percentage points — plus or minus 3 points means the true value plausibly sits anywhere in a 6-point band, and a 3-point shift between two polls each carrying that margin is not evidence of movement.

Labour statistics. Unemployment moving from 3.6% to 4.2% is a 0.6 pp rise, which is a 16.7% relative rise — the second number sounds alarming and is rarely how economists talk. Agencies such as the Bureau of Labor Statistics report the level and the percentage-point change; a "16.7% jump in unemployment" headline usually signals someone converting the figure for effect. The same applies to inflation: if annual inflation goes from 3% to 6%, prices are rising twice as fast, but "inflation doubled" does not mean prices doubled.

Why does a 50% loss need a 100% gain to recover?

Percentage changes are not symmetric, because the base changes underneath you. Lose 50% of 100 and you have 50. To get back to 100 you must add 50 to a base of 50 — a 100% gain. The general formula for the gain needed to recover a fall of f (as a decimal) is:

Required gain = 1 ÷ (1 − f) − 1

The practical consequence: equal-and-opposite percentage moves do not cancel. Gain 10% then lose 10% and you have 1.10 × 0.90 = 0.99 of what you started with — a 1% net loss. Repeat that pair ten times and you are down about 9.6%. Volatility alone erodes a value even when the average percentage move is zero, which is why investment returns over multiple periods are compounded, never averaged.

Percentage of versus percentage off

These are two different operations that share a word. "30% of 80" means 0.30 × 80 = 24. "30% off 80" means 80 − 24 = 56, which is the same as 70% of 80. Multiplying by the retained fraction (0.70) is faster and less error-prone than computing the discount and subtracting.

That shortcut also settles the stacked-discount question. A "30% off, then an extra 20% off at checkout" promotion is not 50% off. You keep 0.70 × 0.80 = 0.56 of the price, so the true discount is 44%. On an 80-unit item you pay 44.80, not 40.00. The order of the two discounts doesn't matter — multiplication commutes — but the total never equals the sum. Our percent off calculator handles stacked discounts and sales tax; the percentage calculator covers the plain "what is X% of Y" and "X is what percent of Y" forms.

One more relative in this family: percent error, used in labs and engineering, is |measured − accepted| ÷ accepted × 100. It is a percentage change with a sign stripped and a known-true value in the denominator. Measure gravitational acceleration as 9.72 m/s² against an accepted 9.81 m/s² and the error is 0.09 ÷ 9.81 = 0.92%. Note it is a percent, not percentage points — the percent error calculator makes the denominator explicit so you don't accidentally divide by the measurement instead of the reference.

The misreadings that come up most

• "Sales rose 200%." That means they tripled, not doubled. A 100% rise is a doubling. If something merely doubled, say it doubled.
• "A 150% decrease." Impossible for a quantity that can't go negative. A decrease caps out at 100%.
• Averaging percentages of different-sized groups. If a 90%-pass group of 10 people and a 50%-pass group of 90 people are combined, the overall rate is 54%, not the 70% you get by averaging 90 and 50. Weight by group size.
• Percentage of a percentage. "Our 5% market share grew 20%" gives 6%, not 25%.

Common questions

When is it acceptable to write "percent" for a change between two percentages?

Only when you genuinely mean the relative change and the context makes the base obvious — for example, "conversion improved 50%, from 2.0% to 3.0%," where both endpoints are stated. If you are writing a single number with no endpoints, use percentage points. Style guides for economics and science writing (and most newsroom guides) treat "pp" as the default for differences between rates.

What exactly is a basis point, and why do finance people use them?

One basis point is 0.01 percentage points, so 100 bp = 1 pp. The unit exists purely to remove ambiguity: "the spread widened 25 basis points" cannot be misread as a relative change, whereas "the spread widened 0.25%" can. You'll see basis points in bond yields, central bank decisions, loan margins, and fund expense ratios.

How do I reverse a percentage change to find the original value?

Divide, don't subtract. If a price is 120 after a 20% increase, the original was 120 ÷ 1.20 = 100, not 120 − 20% = 96. The same applies to tax-inclusive prices: to strip a 20% VAT from a 60.00 total, divide by 1.20 to get 50.00. Subtracting the percentage from the final figure is the single most common arithmetic mistake in this area.

Can percentage change be calculated when the starting value is zero or negative?

Not meaningfully. Dividing by zero is undefined, so a jump from 0 to 5 has no percentage change — report the absolute change instead. With a negative starting value (a company moving from a loss to a profit), the formula produces a number with a misleading sign, which is why analysts write "n/m" (not meaningful) rather than a percentage in those cells.