REPÚBLICA DE MOÇAMBIQUE MINISTÉRIO DA ECONOMIA E FINANÇAS DIRECÇÃO NACIONAL DE ESTUDOS E ANÁLISE DE POLÍTICAS ANTECENDENTES E CONCEITOS BASICOS Curso sobre Análise de Pobreza Maputo, 6-10 Julho 2015
Outline Introduction Basic conceptual and practical considerations Poverty measurement Issues and choices The poverty line Alternative poverty measures Decomposition Empirical observations: who are the poor (correlates)? The functional impact of poverty Conclusion (including policies for poverty alleviation)
Introduction Why focus on poverty? Ethical (normative) considerations (basic human right) Functional (positive) considerations Coming to grips with the nature of the development process: No guarantee for trickle-down to poor groups is growth pro-poor? The growth-poverty-inequality triangle Global poverty headcount 1990: HC > 1 billion individuals and it continues high even if considerable global progress can be noted over the last two decades (the MDGs) Expectations versus reality (at global and at national levels) In the economic definition of poverty there is typically an emphasis on income or consumption, but
The experience of poverty is multidimensional When one is poor, she has no say in public, she feels inferior. She has no food, so there is famine in her house; no clothing, and no progress in her family. WOMAN IN UGANDA When food was in abundance relatives used to share it. These days of hunger, however, not even relatives would help you by giving you some food - YOUNG MAN IN ZAMBIA When my husband died, my in-laws told me to get out. So I came to town and slept on the pavement. - MIDDLE-AGED WIDOW IN KENYA Source: World Bank: Voices of the poor 2002
So poverty in reality also refers to Housing Personal security Vulnerability to external shocks Absence of morbidity Opportunities to learn Access to social life, markets and politics Freedom from discrimination and dependence Etc. Poverty is in reality inherently subjective in nature: Still, a core problem(at least in poorer countries) is lack of income and assets which can give access but note access is broader!
A variety of measures just to illustrate Dimension Component Indicators Economic Consumption and income Head count index (poverty lines) Poverty gap Social Nutrition Calorie intake to requirement and anthropometric measures Enabling environment Sanitation and water Health Education Means of production Natural endowments Geographic infrastructure Access to sanitation and potable water Access to primary health care, family planning, and immunisation Primary school enrolment rate and literacy rates Access to land, access to employment Agro-climatic conditions Access to service providers and to markets
Conceptual and practical issues surrounding poverty measurement Poverty line at the core Income-/expenditure-data vs. consumption-data Absolute vs. relative poverty Temporary vs. chronic poverty Household vs. individual poverty National and international poverty lines an example These points to be discussed one by one in what follows
The poverty line: the foundation for poverty measurement Definition of poverty line Represents a minimum level of acceptable economic participation in a given society at a given point in time The global poverty line of 1 US$ (one dollar a day) (revised) Definition of poverty Income/expenditure below poverty line (0-1) A standard measurement method Nutrition-based minimum expenditure need (to achieve 2,150 calories/day) -> the cost of basic human needs (CBN) bundle (food and non-food) Remember: Just above or just below may be equally bad. Global number of poor more than doubles if poverty line is raised to 2 US$ (revised) Differences across countries - the PPP adjustment Differences within countries (urban versus rural and over time)
Income/expenditure-data vs. consumption-data Is a person poor when the actual, observed consumption basket falls below a threshold or when the total income/expenditure falls below the minimum required to achieve this threshold? Not the same: an example the ascetic rich, and nutrition levels may not rise with rising income Total consumption-data Willingness to consume Total income/expenditure-data Capacity to consume
Absolute vs. relative poverty Absolute (fundamental) needs Food/nutrition (calories, vitamins, etc.) (inherently relative to many things) (even if there exists biological imperatives, but ) And clothing? And shelter? Acceptable level of participation not so clear Relative (fundamental) needs Radio/television Bicycle Poverty line Determined relative to socioeconomic standards (food-nonfood) Yet, remains absolute (income/expenditure) measure: items must fulfill some absolute notion of the ability to function in society) How not to define the poverty line: poverty = % with less than half of average income (confuses poverty and inequality)
Temporary vs. chronic poverty Income-fluctuations among the near-poor Weather-dependent agriculture Income may temporarily drop below (income-based) poverty line Access to (informal) credit market consumptionsmoothing, but It matters for policy: Temporary poverty Short-term access to goods & services Income-based poverty measures reflect temporary poverty Chronic poverty Long-term access to goods & services Consumption-based poverty measures seems to reflect chronic poverty better
12 Household vs. individual poverty Expenditure/consumption data only available at household level Intra-household allocation Skewed against females/elderly (why?) Need for complementary micro-studies Adult equivalence scales Low consumption needs of children Fixed costs/irts in running family Poor families with many children overstatement of poverty
International vs. National Poverty lines? A Warning China: percentage of the population below the poverty line 28% 32% 29% 17% 17% 11% 12% 9% 7% 6% 1987 1990 1993 1996 1999 $1/day poverty line National poverty line
Measurement is difficult Measurement error Poverty line: focus on extreme poor or poor? Which measure Aggregation and substitution, and how to go from nominal to real? Robustness Political objectives and norms
Note! Poverty lines are always approximations to a threshold that is fuzzy Especially when we recall that poverty effects are often felt at a later stage Quantitative estimates are not to be memorized and taken as nothing but approximations: don t do 5 decimals!!!! And remember confidence intervals.
Poverty measures 16 Headcount Ratio (HCR) HC = i=1,n 1 (yi <p) HCR = HC/n (n = population size) Focus: Relative incidence of poverty no attention to depth of poverty or inequality among the poor Poverty Gap Ratio (PGR) PG = i=1,n (p-y i )1 (yi <p) PGR = PG/(m*n) (where m = mean income) Focus: Relative amount of resources required to eradicate poverty (may look small in unequal societies) Income Gap Ratio (IGR) IGR = PG/(HC*p) Focus: relative amount of resources to eradicate poverty (where relative is measured relative to total income needed to make poverty go away) -> depth of poverty Discuss the policy dilemma of targetting HCR only (policy bias) What about minimizing the PGR or IGR?
Examples Example 1: Poverty line 1000 Mt/month -> 200 people are poor Poor group 1: 100 people with 500 Mt/month Poor group 2: 100 people with 900 Mt/month Budget available: 20,000 Who should be targeted if you target the poor and forget poverty line? What if President what HC to go down as much as possible Example 2: Minimize the PGR or the IGR Not susceptible to policy distortion as in example 1 but only capture per capita intensity of poverty Show both measures (and they both ignore relative deprivation among the poor =inequality among the poor) Example 3: 200 poor the 100 poorest now transfer 50 to less poor -> 450 and 950 HCR and PGR the same
Visualising the poverty gap Annual income Annual income Poverty line Poverty line 50 Percentage of population 100 50 Percentage of population 100
Analysing poverty measures Amartya Sen proposed an axiomatic approach to defining a poverty measure A poverty measure should have the following desirable properties: 1. If the number of the poor increases, the measure should rise 2. If the poor get poorer, the measure should go up 3. If the distribution of income among the poor becomes more unequal, the measure should go up The Head Count ratio rates satisfy the first, but not the other two incidence of poverty The Poverty Gap Ratio satisfies the second, but need not satisfy the other two depth of poverty The Foster-Greer-Thorbecke (1984) FGT measure (more advanced) (poverty head count, poverty gap, squared poverty gap) an overview plus show the Squared Poverty Gap - severity
FGT General Math Form Onde: n é a população, y é o consumo per capita, z é a linha de pobreza, e é um parâmetro não negativo. 0, 1 1 z y z y n P
FGT notes By far the most common poverty measure employed. Typically use for = 0, 1, 2 =0: Headcount (incidence) =1: Poverty gap (depth) =2: Squared poverty gap (severity) Note that the headcount is implicitly a part of all three measures
Z = 3 Another example A= (1, 2, 3, 4) -> HCR = 0.75 = 3/4 PGR = 0.25 = ((3-1) + (3-2) + (3-3))/4*3 B= (2, 2, 2, 4) -> HCR = 0.75 =3/4 PGR = 3(3-2)/3*4 = 0.25 Same HC and PGR but in A the poorest only get half of what the poorest in B gets - > need for squared poverty gap SPG (A) = 0.14 = (5/9)/4 and SPG (B) = 0.08 = (3/9)/4
Decomposing poverty a poverty profile Poverty profiles are decompositions of an aggregate poverty measure by sub groups (for example regions) Consider two regions A og B with population shares n A and n B The aggregated poverty measure is P n P n P A A B B
Poverty empirical observations Recall there are large numbers of poor Characteristics of the poor (who are the poor?) Demographic features Rural and urban poverty Assets Nutrition Etc etc and now just a couple of examples
Demographic features Poor households tend to be: Large High dependency ratios (a cause and an effect of poverty) But exercise care in assessing the correlation using per capita figures Children may consume less Economies of scale
The rural urban divide Region and Country % of Urban Population under Nat. Poverty Line % of Rural Population under Nat. Poverty Line Sub-Saharan Africa Lesotho (1993) 27.8 53.9 Nigeria (1993) 30.4 36.4 Zambia (1991) 46 88 Asia India (1994) 30.5 36.7 Philippines (1991) 39 71 Vietnam (1993) 25.9 57.2 Latin America Brazil (1990) 13.1 32.6 Nicaragua (1993) 31.9 76.1 Peru (1991) 50.3 68
The Glewwe warning (30% defined as poor definitions matter)
The functional impact of poverty: poverty, credit and insurance Poverty affects access to markets Credit Lack of collateral Incentives for poor to repay may be limited (utility loss of repayment) see next slide (market fails due to different incentives) 28
Functional impact of poverty credit failure 29
Poverty, nutrition and the poverty trap The human energy balance Energy input Resting metabolism Energy required for work Storage and borrowing Nutrition affects the body s capacity to perform work so poverty bad nutrition low work capacity low earnings poverty (a poverty trap) The work capacity curve -> tends to promote unequal allocations within the household (the lifeboat ethic unequal allocations create greater household work capacity) 30
The work capacity curve 31
Poverty and the household the unequal sharing of poverty (a two person family, first in Y* and then in Y) 32
Poverty is not shared equally within household gender inequality Education and empowerment of women Ratio women/men Developed countries 105 India 93 Saudi Arabia 82 Discrimination or lifeboat ethic? 33
Conclusions Poverty: intrinsic and functional aspects Poverty is multidimensional Standard poverty measurement The poverty line Complementary measures Poverty correlates different indicators do not identify the same people as poor Policy objectives, indicators and policy formulation (it is easy to get it wrong type 1 and type 2 statistical errors) Poverty, under nutrition and work capacity The poverty trap The intra household allocation problem and discrimination Discussion about policy and policy priorities 34