How well-off is a household? The answer underlies poverty rates, targeting, and most welfare policy — and it is built from a chain of measurement choices, each contestable, that can swing the resulting numbers substantially. This module covers measuring household welfare: why consumption beats income, the craft of the consumption aggregate, and the poverty lines and equivalence scales that turn it into a poverty rate.
Consumption versus income
Why consumption is preferred in developing countries
To measure material welfare you could use either INCOME or CONSUMPTION, and in developing-country analysis CONSUMPTION is generally preferred, for several reasons: • Income is hard to measure where most people are self-employed, farmers, or informal — there's no payslip; farm and business income is irregular, hard to recall, and entangled with the household enterprise; and income is systematically UNDER-REPORTED (people understate earnings). • Income is VOLATILE (seasonal for farmers, irregular for the informal sector), so a snapshot of income is a poor measure of a household's standard of living, while consumption is SMOOTHER (households smooth consumption across income fluctuations through savings, borrowing, and reciprocity — the behavioural-economics theme), so it better reflects permanent/typical welfare. • Consumption is more directly tied to WELFARE (what you actually consume is closer to your living standard than what you earn). So most developing-country poverty measurement is consumption-based. (In rich countries with formal employment and good income records, income is often used.) This choice itself matters — income-based and consumption-based poverty can differ substantially, and knowing which a number uses is essential to interpreting it.
The consumption aggregate
A craft full of judgment calls
Building the CONSUMPTION AGGREGATE — the single number summarising a household's total consumption — is a craft full of consequential judgment calls (Deaton-Zaidi), not a mechanical sum. The components and their difficulties: • Food — including food the household PRODUCED ITSELF and consumed (subsistence farming — a huge share of poor households' consumption), which must be VALUED at appropriate prices; plus food bought, gifts received, and meals eaten out. The valuation of own-produced food and the treatment of seasonality are major choices. • Non-food — clothing, transport, fuel, etc., with appropriate recall periods (frequent small items vs lumpy purchases). • Durables — you consume the SERVICE FLOW of a durable (a bicycle, a phone) over years, not its purchase price in the year bought, so you must impute a USE VALUE — a tricky adjustment. • Housing — owner-occupiers consume housing services they don't pay rent for, so you must impute a RENTAL VALUE — another judgment-laden estimate. Each choice (what to include, how to value own-production, how to handle durables and housing, recall periods, regional price adjustments for comparability) affects the aggregate, and inconsistent choices make households or surveys non-comparable. The consumption aggregate looks like an objective number but embeds dozens of contestable decisions — which is why poverty comparisons across surveys or countries that used different methods can be misleading.
Poverty lines
A poverty line is the consumption threshold below which a household is 'poor'. Types: ABSOLUTE poverty lines are anchored to a fixed standard of needs — typically the cost of a basket meeting basic CALORIE/nutritional requirements (the food poverty line) plus an allowance for essential non-food (the cost-of-basic-needs approach); the international extreme-poverty line (currently $2.15/day in 2017 PPP) is an absolute line used for global comparison, derived from the national lines of the poorest countries. RELATIVE poverty lines are set relative to the population's standard of living (e.g., 60% of median income — common in rich countries, measuring inequality-flavoured poverty). The choice of line — its level, whether absolute or relative, the calorie threshold, the non-food allowance, the PPP conversion — directly determines the measured poverty RATE, and reasonable people disagree about all of these. A poverty rate is meaningful only relative to its (contestable) line.
Equivalence scales and poverty measures
Household size and the depth of poverty
Two final choices. EQUIVALENCE SCALES: a household of 5 does NOT need 5 times the consumption of a household of 1 to be equally well-off — there are ECONOMIES OF SCALE in households (shared housing, utilities, durables) and children may need less than adults. Equivalence scales adjust household consumption for size and composition (per 'adult equivalent', with economies of scale), rather than naive per-capita division. The choice of scale (how much economies of scale, how to count children) significantly affects WHO counts as poor (per-capita measures make large households look poorer than equivalence-scaled measures do) — and there is no single 'correct' scale. POVERTY MEASURES: beyond the HEADCOUNT (the % below the line), the Foster-Greer-Thorbecke (FGT) measures capture more — the poverty GAP (how FAR below the line the poor are, i.e., the depth of poverty) and poverty SEVERITY (weighting the poorest more) — which matter because two countries with the same headcount can have very different depths of poverty, and a policy can reduce the gap without moving the headcount. The headcount is the famous number but it ignores depth; good poverty analysis reports the gap and severity too. All these choices — line, equivalence scale, poverty measure — mean the poverty 'number' is the end of a long chain of contestable decisions, each of which can move it.
Exercise
A country reports a national poverty headcount of 30%, and a critic argues the true figure could be quite different depending on measurement choices. (1) Explain why consumption rather than income was likely used and the advantage. (2) Identify three judgment calls in building the consumption aggregate that could move the 30%. (3) Explain how the poverty-line and equivalence-scale choices affect the number. (4) Explain why the headcount alone is an incomplete picture and what to add.