Percentage vs percentage point: the difference that breaks financial news

A percentage is a ratio: 25% means 25/100 of something. A percentage point(often abbreviated “pp”) is an absolute increment on a percentage scale. The two get used interchangeably in news and casual conversation, which causes real arithmetic errors — especially around interest rates, vaccine coverage, election polls, and market shares.

The classic example: Fed rate changes

Headline: “Fed raises rates by 0.25%.”

What it actually means: the Fed raised rates from (say) 5.25% to 5.50%. That’s a 0.25 percentage pointincrease. Compared to the starting rate, it’s actually a 4.76% relative increase (0.25 / 5.25).

For a borrower with a 5.25% mortgage that moves to 5.50%:

• The rate increased by 0.25 percentage points (absolute).
• The rate increased by 4.76% (relative).
• Monthly payment increased by ~3%.
• Lifetime interest increased by ~5%.

Three different numbers, all correct, describing the same change. The headline “0.25% rate hike” is technically wrong but universally used.

When the distinction matters most

Interest rates and yields

Any time you’re talking about a rate that already is a percentage, the change should be in percentage points. “Yields rose 50 basis points” is the bond-market convention; 100 basis points = 1 percentage point. The unit removes the ambiguity.

Vaccine and treatment efficacy

Vaccine A reduces infection from 10% to 5%. The relative risk reduction is 50% (5/10 less infection). The absolute risk reduction is 5 percentage points. News coverage almost universally reports the relative number — which sounds dramatic for low-base-rate conditions where the absolute benefit is tiny.

A vaccine that takes infection from 0.1% to 0.05% has a 50% relative risk reduction (same headline number as above) and a 0.05 percentage-point absolute reduction (almost negligible).

Election polls and market shares

“Candidate A leads by 5 points” in polling usually means 5 percentage points of the vote share — e.g., 52% vs 47%. If reported as “5% lead” it could be misread as 5% of A’s support, which would be 2.6 percentage points.

Discounts and price changes

A 20% discount on a $100 item saves $20. A subsequent “additional 10% off” on the already-discounted price saves $8 (10% of $80). Total savings: $28, or 28 percentage points off the original price. Not 30%.

Even simpler: a 20% discount followed by a 25% markup does not return you to the original price. $100 → $80 (20% off) → $100 (25% markup of $80) — same as starting price. But $100 → 25% off ($75) → 20% markup ($90), which is below the start.

Quick test you can apply

If you’re reading about a change in something that is itself a percentage (interest rate, unemployment rate, survey share), ask: is this 0.25 absolute or 0.25 of the original?

• If it’s on a basis-point scale, it’s percentage points.
• If it’s “X rose Y%” in a news headline, it’s usually percentage points but written sloppily.
• If it’s “X grew Y%” or “X is Y% higher than”, it’s relative.

When ambiguous, ask the source. In financial reporting, getting this wrong by an order of magnitude is normal — a 0.5% interest-rate hike (relative) is dramatically less than a 0.5 percentage-point hike.

How to write it unambiguously

• Use “percentage points” or its abbreviation ppfor absolute changes on a percentage scale. “Rates rose by 0.5 pp.”
• Use “basis points” or its abbreviation bpsfor finer measurements. 100 bps = 1 pp. “The spread widened by 25 bps.”
• Use plain percentagesfor relative changes. “Rates rose by 4.8% in relative terms.”
• Show bothfor non-specialist audiences. “Rates moved from 5.25% to 5.50% — a 0.25 pp (or 4.8% relative) increase.”

The trap in absolute vs relative comparisons

Two ways to report the same vaccine result:

• “The vaccine reduced infection risk by 90%.” (Relative.)
• “The vaccine reduced infection risk by 0.45 percentage points (from 0.5% to 0.05%).” (Absolute.)

Same study, same data. The relative number is what marketers and headlines lead with; the absolute number is what you need to assess personal benefit. For common-disease vaccines, the absolute reduction is large; for rare-disease vaccines, less so. The 90% figure doesn’t distinguish.

The pragmatic bottom line

When you read or report a change in a percentage:

1. State whether the change is absolute (percentage points) or relative (percentage of the original).
2. For finance, basis points are the right unit. 100 bps = 1 pp.
3. For health and policy, both numbers should be shown — they answer different questions.
4. For marketing claims, distrust the absence of context.

See also our why percentage change isn’t percentage difference guidefor the related “asymmetric percentage” trap.

Worked example: unemployment going from 4% to 5%

Headline candidates for the same underlying fact:

• “Unemployment rose 1 percentage point.” (Absolute, precise.)
• “Unemployment rose 25%.” (Relative — 1 / 4 = 25%.)
• “Unemployment rose 100 basis points.” (Absolute, in bps.)
• “Unemployment rose 1%.” (Ambiguous; could mean either, frequently misread.)

All four describe the same change. Three are correct, one is journalistic shorthand that loses 25× of magnitude. Reading “unemployment rose 1%” as relative (4% → 4.04%) is a much milder claim than reading it as absolute (4% → 5%) — yet the headline doesn’t disambiguate. Always check the underlying data series before reacting to a percentage-of-a-percentage headline.

Common mistakes

• Compounding sequential percentages. A 20% gain followed by a 20% loss is not break-even. $100 → $120 → $96. Compounding asymmetry costs 4% on a round-trip; it costs more on volatile assets.
• Comparing relative changes across different base rates. A 100% increase in cancer risk sounds catastrophic. If the base rate is 1 in a million, the absolute risk is 2 in a million — still negligible. Relative ratios without base rates are noise.
• Treating percentage-of-GDP changes as percentages.“Government debt rose 5 percentage points of GDP” is absolute (from say 80% to 85%). “Government debt rose 5%” could mean the debt grew by 5% in dollars — a different quantity that depends on GDP growth too.
• Averaging percentages. Average of 10% and 20% returns is not 15% compound; (1.10 × 1.20)^0.5 − 1 = 14.89%. Geometric mean, not arithmetic, for chained rates. Our geometric mean glossary entry covers this.
• Misreading polling margin of error. A poll with ±3 pp margin and a 49% vs 47% result is statistically a tie, not a 2-point lead. The margin is percentage points; misreading it as percent of the result inflates apparent precision.

For the related “what counts as a meaningful discount” question, see what percent discount is actually good. For the deeper compounding-asymmetry story, the percentage change vs percentage difference guide covers it.

Sources: BIPM The International System of Units9th edition (2019) on percentage notation; Federal Reserve press-release style guide (2024); American Statistical Association “Statistical Significance and P-Values” on relative vs absolute risk communication; Spiegelhalter (2017) on risk communication.