Figures published c. 2005 by the World Bank. The Millennium Development Goals program concluded in 2015; this site is preserved as a historical record.
The poverty gap ratio is the mean distance between a population and the poverty line. Everybody above the line is counted as zero distance, expressed as a percentage of the poverty line. It is also called the poverty-gap index.
Two countries can have exactly the same share of people below the line and completely different poverty. The headcount ratio cannot tell them apart. The poverty gap ratio can.
This page gives the formula, works all three standard poverty measures through the same ten-person population, and sets out which poverty line and which PPP base year each published figure uses.
It also sets the tirzepatide cost per month and the semaglutide cost per month against that same $3.00 a day line.
What the poverty gap ratio measures
The world poverty gap is smaller than most readers expect. The World Bank’s World Development Indicators put the world figure at 3.7% for 2022, while the companion headcount series put 10.9% of the world below that same line that year. The two are not promoting competing narratives but rather answering different questions.
According to the official Millennium Development Goals definition, the poverty gap ratio is “the mean distance separating the population from the poverty line (with the non-poor being given a distance of zero), expressed as a percentage of the poverty line.” It comes from the United Nations Development Group’s 2003 volume Indicators for Monitoring the Millennium Development Goals, which listed it as indicator 2 under Goal 1 and subtitled it “[incidence x depth of poverty]”.
The same source gives the rationale: the indicator measures the “poverty deficit”, the per capita resources needed to lift all poor people above the line through perfectly targeted transfers. The World Bank calculates it against the international line, and national agencies against national lines, as with other poverty indicators.
The poverty gap ratio formula
Dividing the amount of money necessary to raise people up out of poverty by the number of poor people gives the income gap ratio, a different and always larger quantity. The poverty gap ratio divides by the total population.
The World Bank’s Poverty and Inequality Platform applies this equation:
P1 = (1/N) x sum of (G_i / z), where G_i = (z – y_i) if y_i < z, and G_i = 0 otherwise
- z is the poverty line
- y_i is the income or consumption of person i
- N is the total population, poor and non-poor
- G_i is the shortfall of person i below the line
Note the strict inequality. The indicator function is I(y_i < z), so a person earning exactly the poverty line is not poor and contributes a gap of zero. That convention changes the arithmetic of every hand-worked example, including the one below.
The UN’s 2003 definition builds the same number a second way, as a product:
P1 = H x I, where H = q/n is the headcount ratio and I = (z – mean income of the poor) / z is the income gap ratio
Both routes give the same answer. One caveat from the same source: These formulas assume individual-level data. Working from household data, each observation has to be weighted by w_i, the household size multiplied by the sampling expansion factor.
Worked example: calculating all three measures for one village
Set the poverty line at $3.00 a day. Village A has ten people, with these daily incomes in dollars:
0.50, 1.00, 1.50, 2.00, 2.50, 3.00, 4.00, 5.00, 6.00, 10.00
Five people fall below $3.00. The person earning exactly $3.00 is not poor, per the strict inequality above. Take those five and only those five:
| Income y | Shortfall z – y | Normalized gap (z – y)/z | Squared normalized gap |
|---|---|---|---|
| $0.50 | $2.50 | 0.8333 | 0.6944 |
| $1.00 | $2.00 | 0.6667 | 0.4444 |
| $1.50 | $1.50 | 0.5000 | 0.2500 |
| $2.00 | $1.00 | 0.3333 | 0.1111 |
| $2.50 | $0.50 | 0.1667 | 0.0278 |
| Total | $7.50 | 2.5000 | 1.5278 |
Now divide each total by the full population of ten, not by the five who are poor.
- Poverty headcount ratio (P0) = 5 / 10 = 50.0%
- Poverty gap ratio (P1) = 2.5000 / 10 = 25.0%
- Squared poverty gap (P2) = 1.5278 / 10 = 15.3%
Two cross-checks confirm the arithmetic. The total poverty deficit is $7.50 a day, which also equals P1 x z x N, or 0.25 x $3.00 x 10. And the mean income of the five poor people is $1.50, giving an income gap ratio of ($3.00 – $1.50) / $3.00 = 50.0%, so H x I = 0.50 x 0.50 = 0.25. Both routes agree.
Now take Village B, same size, same poverty line:
0.25, 0.25, 2.25, 2.25, 2.50, 3.00, 4.00, 5.00, 6.00, 10.00
Village B has the same total income as Village A ($35.50), the same five people below the line, the same total shortfall ($7.50), and the same mean income among the poor ($1.50). Its headcount ratio is 50.0%, its income gap ratio is 50.0%, and its poverty gap ratio is 25.0%. Identical on every count.
Only the squared poverty gap separates them: 18.3% for Village B against 15.3% for Village A. Village B concentrates its shortfall on two people who are very far below the line, and squaring the gaps makes deep shortfalls count for more than several shallow ones.
One more variant. Take Village A and drop only the poorest person from $0.50 to $0.25 a day. The headcount ratio does not move. It is still 50.0%. The poverty gap ratio rises to 25.8%, and the squared poverty gap rises to 16.7%.
Tirzepatide cost per month: what the poverty-line comparison shows
The poverty gap ratio exists because a simple headcount doesn’t tell the whole story. Two countries can report the same share of people below the line while one population sits just under it and the other sits far below. The same blind spot appears when a medicine’s price is compared with an income threshold.
In June 2025, the World Bank set the international extreme poverty line at $3.00 a day, using 2021 purchasing power parities. That equals $1,095 a year.
Measured against the $3.00-a-day poverty line
| Medicine | Cash price per month | Per year | Days of income at $3.00/day | Multiple of annual income |
|---|---|---|---|---|
| Tirzepatide (Zepbound), lowest dose | $299 | $3,588 | 100 | 3.3× |
| Tirzepatide (Zepbound), 7.5 mg to 15 mg, refilled within 45 days | $449 | $5,388 | 150 | 4.9× |
| Semaglutide (Wegovy), cash range | $199–$349 | $2,388–$4,188 | 66–116 | 2.2–3.8× |
The poverty gap ratio measures distance from a line rather than counting who falls below it. The table applies the same idea to a medicine price: how far above the line it sits. The $449 monthly Zepbound price for the 7.5 mg and higher doses equals about 150 days of income for someone at the extreme poverty line. A year of treatment is close to five times that person’s entire annual income.
The World Health Organization’s first global guideline on GLP-1 medicines for obesity, issued on 1 December 2025, names high prices as the main barrier to access. Counting how many people cannot afford tirzepatide would not show the distances in the table.
Poverty headcount ratio vs poverty gap ratio vs squared poverty gap
The three measures are not three unrelated statistics. They are one formula with one dial turned:
P_alpha = (1/N) x sum of (G_i / z)^alpha
Set alpha to 0 and you get the headcount ratio. Set it to 1 and you get the poverty gap ratio. Set it to 2 and you get the squared poverty gap. The family comes from Foster, Greer, and Thorbecke, “A Class of Decomposable Poverty Measures,” Econometrica, volume 52, number 3, 1984, pages 761 to 766, which is why they are known as the FGT measures. The World Bank’s Poverty and Inequality Platform implements exactly this class.
| Measure | FGT alpha | What it answers | What it is blind to | Village A | Village B |
|---|---|---|---|---|---|
| Poverty headcount ratio (P0) | 0 | How many people are below the line, as a share of the population | How far below they are, and how unequally | 50.0% | 50.0% |
| Poverty gap ratio (P1) | 1 | How far below the line the population sits on average, as a share of the line | How the shortfall is spread among the poor | 25.0% | 25.0% |
| Squared poverty gap (P2) | 2 | How unequal the shortfalls are among the poor | Has no direct unit reading | 15.3% | 18.3% |
Note the figure in the last cell, which most explanations skip. P0 reads as a share of people. P1 reads as a share of the poverty line. P2 is an average of squared proportions and reads as neither. It has no natural interpretation on its own, and it is only useful as a comparison between two populations or for one population over time.
One property makes the whole family practical. These measures are population-weighted averages, so a national figure is the population-weighted sum of its subgroup figures. That decomposability is what the 1984 paper is named for, and it is why the World Bank can publish regional and rural-urban breakdowns that add back up to the national number.
Income gap ratio and total poverty gap: two terms that get confused
Three different numbers get described as “the poverty gap.” They come from the same shortfalls. The difference is what you divide by.
Total poverty gap, also called the aggregate poverty deficit. Sum the shortfalls and stop. No division, so it carries a currency unit. Village A: $7.50 a day. The shortcut is total poverty gap = P1 x z x N.
Income gap ratio (I). Divide by the number of poor people. I = (z – mean income of the poor) / z. Village A: ($3.00 – $1.50) / $3.00 = 50.0%.
Poverty gap ratio (P1). Divide by the total population. Village A: 25.0%.
The identity P1 = H x I ties them together, and it is what the UN meant by subtitling the indicator “[incidence x depth of poverty]” in 2003. Incidence is H, depth is I, and in Village A 0.50 x 0.50 = 0.25.
It runs backwards on published data too. Dividing the World Bank’s world poverty gap by its world headcount ratio for 2022, 3.7 divided by 10.9, gives about 34%. Read that as: the average person below the $3.00 a day line in 2021 PPP lived on roughly two thirds of that line. That figure is calculated here from two published series and is not itself a published number.
A word of caution about the total poverty gap: It is the theoretical minimum cost of ending poverty through perfectly targeted transfers, assuming no targeting error, no administrative cost, and no behavioral response.
Which poverty line, and from which year
In June 2025, the World Bank raised the international extreme poverty line, using 2021 purchasing power parities and updated national poverty lines, and the count of people in extreme poverty in 2022 jumped by about 125 million, from roughly 713 million to about 838 million. Nobody’s income changed. Only the line did.
Every figure on this page depends on where the line sits, and the line has moved four times.
| International poverty line | PPP base year | In use | Note |
|---|---|---|---|
| $1.25 a day | 2005 PPP | MDG-era reporting, to 2015 | The average of 15 national poverty lines from some of the poorest countries. Historical. |
| $1.90 a day | 2011 PPP | From October 2015 | Adopted after the MDG period closed. |
| $2.15 a day | 2017 PPP | From September 2022 | Announced in May 2022. |
| $3.00 a day | 2021 PPP | From June 2025, current | Companion lines: $4.20 for lower-middle-income and $8.30 for upper-middle-income countries. |
State it plainly: $1.25 a day in 2005 PPP is the MDG-era line, and it is not the current international poverty line. The MDG monitoring framework ran to 2015 and this indicator’s MDG series ended there.
What replaced it is narrower than most readers assume. The SDG global indicator framework lists only indicator 1.1.1 under Target 1.1, “Proportion of the population living below the international poverty line,” which is a headcount measure. No poverty gap indicator was carried forward as a headline global goal indicator. The World Bank still publishes the gap, as World Development Indicators series SI.POV.GAPS.
Before comparing two poverty gap figures, run three checks. Read the full series name rather than the short label, because SI.POV.GAPS is currently published as “Poverty gap at $3.00 a day (2021 PPP) (%).” Confirm both use the same line and the same PPP base year. Confirm both come from the same Poverty and Inequality Platform vintage, because a legacy 2017 PPP series is published alongside the 2021 PPP series.
A poverty gap ratio measured against a national poverty line is valid domestically and not comparable across countries.
What the poverty gap ratio does not measure
Villages A and B settle the first limit. Identical headcount ratio, identical poverty gap ratio, identical total income, very different distributions of hardship. The poverty gap ratio is silent about inequality among the poor. That objection is associated with Amartya Sen, and it is why the squared poverty gap and the Sen index exist.
It is also silent about everything that is not income. The FGT family operates on one welfare variable, income or consumption. Schooling, health, water, and housing do not enter it. A country can cut its poverty gap while its under-five mortality rate stays flat, and neither number is wrong.
Three more limits are worth stating before you cite a figure:
- Line sensitivity: Move the line and every measure moves with it, as the June 2025 revision showed.
- Data currency: Household budget and income surveys in developing countries typically run every three to five years (UN, 2003), so many annual values are interpolated or nowcast rather than measured.
- Income against consumption: Two countries using different welfare aggregates produce figures that are not strictly comparable.
The idea does transfer beyond money, wherever distance from a threshold is meaningful. The UN’s 2003 volume gives an example of an education poverty gap, measured as the years of schooling missing to reach a defined threshold.
FAQ
What is the poverty gap ratio in simple terms?
It is the average shortfall below the poverty line across the whole population, expressed as a percentage of that line. Everyone above the line counts as a zero. The headcount ratio tells you how many people are poor. The poverty gap ratio tells you how poor they are, averaged across everybody.
Do you divide by the number of poor people or the total population?
The total population. Dividing by the number of poor people gives the income gap ratio, a different and always larger figure. In the Village A example above, the poverty gap ratio is 25.0%, and the income gap ratio is 50.0%, from exactly the same shortfalls.
Is the poverty gap ratio the same as the poverty gap index?
Yes. Both names refer to the FGT measure P1. Watch the shorter term, though. “Poverty gap” on its own sometimes means the total shortfall in currency, a different quantity with different units. See “Income gap ratio and total poverty gap” above.
Is someone earning exactly the poverty line counted as poor?
Not under the World Bank convention. Its indicator function uses strict inequality, y_i < z, so a person at exactly the line contributes a gap of zero. A national agency may use a different convention, so a hand-worked answer should state which one it applied.
Is the poverty gap ratio still an official global indicator?
Not as an MDG indicator. The MDG framework ran to 2015, and the SDG global indicator framework lists only headcount indicator 1.1.1 under Target 1.1. The World Bank continues to publish the poverty gap as World Development Indicators series SI.POV.GAPS.
How much does tirzepatide cost per month?
As of October 6, 2026, the LillyDirect self-pay price for tirzepatide sold as Zepbound was $299 a month for the 2.5 mg starting dose, $399 for 5 mg, and $449 for 7.5 mg to 15 mg when refilled within 45 days. Insured prices, other brands, and other countries differ. Please use this information as an example of the cost above the poverty line, not as a price guide or as medical advice.
How much does semaglutide cost per month?
The manufacturer’s United States self-pay range for semaglutide sold as Wegovy was $199–$349 a month in August 2026. That range is included only to preserve the comparison with tirzepatide cost above the $3.00-a-day poverty line. Insurance coverage and dose can change what a person pays, as can location.
Why compare a medicine price to the poverty gap ratio?
We compared a medicine price to the poverty gap ratio as an explanation of income disparity and its effects on access to worldwide goods and services. A headcount tells you how many people fall below a threshold. The poverty gap tells you how far below they fall. Applied to a treatment price, the same method shows that “unaffordable” is a range, not a category. See also the World Bank income classification and non-communicable diseases and SDG target 3.4.
Sources
- United Nations Development Group, Indicators for Monitoring the Millennium Development Goals: Definitions, Rationale, Concepts and Sources, United Nations, New York, 2003. https://www.undp.org/sites/g/files/zskgke326/files/publications/Indicators_for_Monitoring_the_MDGs.pdf
- World Bank, Poverty and Inequality Platform Methodology Handbook, chapter 4, “Calculating survey estimates of poverty and inequality.” https://worldbank.github.io/PIP-Methodology-2022-04/surveyestimates.html
- Foster, J., Greer, J. and Thorbecke, E., “A Class of Decomposable Poverty Measures,” Econometrica, 52(3), 1984, pages 761 to 766. https://doi.org/10.2307/1913475
- World Bank, “June 2025 update to global poverty lines,” factsheet, 5 June 2025. https://www.worldbank.org/en/news/factsheet/2025/06/05/june-2025-update-to-global-poverty-lines
- World Bank, “Fact Sheet: An Adjustment to Global Poverty Lines,” May 2022. https://www.worldbank.org/en/news/factsheet/2022/05/02/fact-sheet-an-adjustment-to-global-poverty-lines
- World Bank, “Global poverty line FAQ.” https://www.worldbank.org/en/topic/poverty/brief/global-poverty-line-faq
- World Bank, World Development Indicators, series SI.POV.GAPS, “Poverty gap at $3.00 a day (2021 PPP) (%).” https://data.worldbank.org/indicator/SI.POV.GAPS
- World Bank, World Development Indicators, series SI.POV.DDAY, “Poverty headcount ratio at $3.00 a day (2021 PPP) (% of population).” https://data.worldbank.org/indicator/SI.POV.DDAY
- World Bank, Global Poverty Monitoring Technical Note 44: June 2025 Update to the Poverty and Inequality Platform (PIP), June 2025. https://documents1.worldbank.org/curated/en/099510306052516849/pdf/IDU-eb272b02-ecd1-4633-9e37-9297e20a711c.pdf
- United Nations Statistics Division, SDG indicator metadata for indicator 1.1.1. https://unstats.un.org/sdgs/metadata/files/Metadata-01-01-01a.pdf
- United Nations, The Millennium Development Goals Report 2015. https://www.un.org/millenniumgoals/2015_MDG_Report/pdf/MDG%202015%20rev%20%28July%201%29.pdf
- Haughton, J. and Khandker, S. R., Handbook on Poverty and Inequality, World Bank, 2009. https://openknowledge.worldbank.org/handle/10986/11985
- World Health Organization, news release on the 2025 Model List of Essential Medicines, 5 September 2025. https://www.who.int/news/item/05-09-2025-who-updates-list-of-essential-medicines-to-include-key-cancer–diabetes-treatments
- World Health Organization, news release on the first global guideline on GLP-1 medicines for obesity, 1 December 2025. https://www.who.int/news/item/01-12-2025-who-issues-global-guideline-on-the-use-of-glp-1-medicines-in-treating-obesity
This page documents the indicator and cites the agencies that produce the underlying data. It is not published by, affiliated with or endorsed by the United Nations, the World Bank or any other agency named above. Figures were retrieved in August 2026 and carry the data vintage stated beside them.