- Advanced
- About 16 min
- 5 practice questions
Measuring poverty: headcount, poverty gap and squared gap
Compute the headcount ratio, poverty gap index, squared poverty gap and Sen index from raw incomes, and see what each measure ignores.
If something looks wrong, please tell us.
Advanced questions on poverty rarely stop at 'what percentage of people are poor'. They give a small set of incomes and a poverty line, then ask for two or three different indices, or ask which index changes when money moves between poor people. This lesson builds the full family of poverty measures from first principles so that every such question becomes a short calculation.
The poverty line and the three FGT measures
Every measure starts with a poverty line z: the income or consumption below which a person is counted poor. Let N be the total population, q the number of poor, and y the income of a poor person. The Foster-Greer-Thorbecke (FGT) family gives three standard measures. P0, the headcount ratio, is q / N: the share of people below the line. P1, the poverty gap index, is the average over the whole population of the relative shortfall (z − y) / z, counting zero for the non-poor. P2, the squared poverty gap, averages the square of that shortfall, so a person far below the line weighs much more than one just below it.
| Measure | Formula (sum over the poor only) | What it captures |
|---|---|---|
| P0 headcount ratio | q / N | How many are poor |
| P1 poverty gap index | (1/N) Σ (z − y)/z | How far below the line, on average |
| P2 squared poverty gap | (1/N) Σ [(z − y)/z]² | Severity: inequality among the poor |
A useful identity links the first two. Define the income gap ratio I as the average shortfall among the poor only, Σ(z − y)/(q z). Then P1 = H × I, where H = P0. Examiners often give two of H, I and P1 and ask for the third.
Worked example 1: all three measures
Five people have monthly incomes 40, 60, 80, 120 and 150, and the poverty line is 100. The poor are the first three, so H = 3/5 = 0.6. Their relative gaps are 60/100 = 0.6, 40/100 = 0.4 and 20/100 = 0.2, which add to 1.2. P1 = 1.2 / 5 = 0.24. The income gap ratio is I = 1.2 / 3 = 0.4, and indeed H × I = 0.6 × 0.4 = 0.24. For P2, square the gaps: 0.36 + 0.16 + 0.04 = 0.56, so P2 = 0.56 / 5 = 0.112.
Worked example 2: what each index ignores
Take the same five people and move 10 from the poorest person (40) to the third person (80). Incomes of the poor become 30, 60 and 90. Nobody crossed the line, so H stays 0.6. The gaps are 0.7, 0.4 and 0.1, still adding to 1.2, so P1 stays 0.24. But P2 = (0.49 + 0.16 + 0.01) / 5 = 0.66 / 5 = 0.132, up from 0.112. Only P2 notices that the poorest person got poorer. This is the transfer axiom: a measure should rise when income moves from a poorer to a less poor person. P0 and P1 fail it; P2 satisfies it.
Now suppose instead that a government with a small budget gives 25 to the person at 80. That person reaches 105 and leaves poverty, so H falls from 0.6 to 0.4. A planner judged only on the headcount ratio is therefore tempted to help those just below the line rather than the poorest, which is the classic criticism of P0 as a target.
The Sen index
Amartya Sen combined the three ideas into one number: S = H × [I + (1 − I) × Gp], where Gp is the Gini coefficient of income among the poor. If every poor person has the same income, Gp = 0 and S reduces to H × I, the poverty gap index. Inequality among the poor pushes S above P1. For example, with H = 0.5, I = 0.3 and Gp = 0.2, S = 0.5 × (0.3 + 0.7 × 0.2) = 0.5 × 0.44 = 0.22.
P0 counts the poor, P1 measures how far below the line they are, P2 weighs the poorest most. Remember P1 = H × I, and only P2 rises when income moves from a poorer to a less poor poor person.
Key takeaways: count the poor strictly below the line; divide P1 and P2 by the whole population, not by the number of poor; I is the average gap among the poor only; the headcount ratio ignores depth and distribution, so cheap transfers to the near-poor can lower it while leaving the poorest untouched.
Check your understanding
-
1. The poverty line is 50. Six people have incomes 20, 30, 45, 60, 80 and 100. What is the headcount ratio?
- 0.33
- 0.5
- 0.6
- 0.45
Show answer
Answer: B. Incomes below 50 are 20, 30 and 45: three people out of six. H = 3/6 = 0.5.
-
2. For the same six people (line 50; incomes 20, 30, 45, 60, 80, 100), what is the poverty gap index P1?
- 0.367
- 0.55
- 0.183
- 0.110
Show answer
Answer: C. Gaps: 30/50 = 0.6, 20/50 = 0.4, 5/50 = 0.1; total 1.1. P1 = 1.1 / 6 = 0.183. Dividing by 3 instead (0.367) gives the income gap ratio I, not P1.
-
3. Money is moved from the poorest person to a less poor person, and both remain below the line. Which measure rises?
- Headcount ratio
- Poverty gap index
- Squared poverty gap index
- None of them
Show answer
Answer: C. No one crosses the line, so P0 is unchanged; the gaps change by equal and opposite amounts, so P1 is unchanged; squaring gives more weight to the larger gap, so P2 rises.
-
4. If the headcount ratio is 0.25 and the income gap ratio is 0.4, what is the poverty gap index?
- 0.65
- 0.16
- 0.625
- 0.10
Show answer
Answer: D. P1 = H × I = 0.25 × 0.4 = 0.10.
-
5. A government is judged only on the headcount ratio and has a small budget. Which people would it be tempted to help first?
- The poorest people first
- Those just below the poverty line
- Everyone equally
- Those just above the line
Show answer
Answer: B. Lifting someone just below the line out of poverty costs the least per person removed from the count, so the headcount ratio falls fastest that way, even though the poorest gain nothing.
Sign in to keep track of the lessons you finish.