Workshop on Poverty Analysis, Using Census Data1

1
Workshop on Poverty Analysis, Using Census
Data1

• Ben E. Aigbokhan,Ph.D
• Department of Economics
• Ambrose Alli University
• Ekpoma, Nigeria.

UNSD Capacity Building Workshop For LISGIS Staff
, Monrovia, Liberia 12-16 December 2011.
2
LECTURE ONE WHAT IS POVERTY AND WHY MEASURE IT?
1.1 What is Poverty? Objective of this season

• To acquaint participatants with working
  definition of poverty and the justification for
  measuring poverty.
• Poverty may be defined on the basis of monetary
  dimension or non-monetary dimension Either
  dimension views poverty as a manifestation of
  pronounced deprivation of well being.
• Well being is conventionally linked to command
  over consumption.
• The poor are therefore those who lack enough
  income or consumption to place them above some
  defined minimum level. This is the monetary
  dimension of poverty.
• The broader approach to well being focuses on the
  capabilities of individuals to function in
  society.
• Lack of capabilities is reflected in inadequate
  income, education, health, political freedom, or
  powerlessness.
• 1.2 Why Measure Poverty
• - To be able to identify the poor.
• - To be able to target appropriate interventions.
• - To be able to monitor and evaluate projects and
  policies targeted at the poor.
• - To be able to evaluate the effectiveness of
  institutions whose mandate is to help the poor.
•  

3
LECTURE TWO MEASURING POVERTY
2.1 Steps in Measuring Poverty

• Define an Indicator of Welfare.
• Most measures of welfare are based on household
  consumption expenditure or income. In developing
  countries preference is for current consumption
  rather than current income, for the following
  reasons
• In the short run it reflects more accurately the
  resources that households control. Income tends
  to be understated (table 1).
• Over the long run, it reveals information about
  incomes at other dates in the past and future.
• In poor countries, income is particularly
  difficult to measure accurately, especially in
  agriculture.
• Establish a minimum acceptable standard of the
  indicator to separate the poor from the non-poor.
• Generate a summary statistic or index to
  aggregate the information from the distribution
  of the welfare indicator relative to the poverty
  line.
• Comparisons across households at similar
  consumption level is meaningful when adjustments
  are made for differences in household size, age
  composition, the prices they face, the publicly
  provided goods to which they have access.
• When divided by the number of household members,
  this gives per capita measure of household
  consumption expenditure or income.
• When adjusted for age composition, and therefore
  household consumption needs, it gives adult
  equivalence scales.
• An adult equivalence scale typically measures the
  number of adult males to which that household is
  deemed to be equivalent. Each member of the
  household counts as some fraction of an adult
  male. Effectively, household size is the sum of
  these fractions. It is not measured in members
  but in numbers of adult equivalence.
• Both fall under the welfarist approach to
  measuring well being.
• A non-welfarist approach would focus on whether
  households have attained certain minimum levels
  of, say, nutrition or health. In this case,
  indicators would include caloric intake,
  proportion of income/expenditure devoted to food,
  housing conditions, child schooling, and infant
  mortality rate.

4
2.2 Sources of Data for Measuring Poverty
- Household Sample surveys- National population
census- Perception Assessment Interviews.Table
1 Income and Expenditure by per capita
Expenditure Quintiles Vietnam 1992/1993.
Lowest Lower-mid Middle Mid-upper Highest Overall
Income/capita 494 694 956 1,191 2,190 1,105
Expenditure/capita 518 756 984 1,338 2,540 1,227
Food spending /capita 378 526 643 807 1,382 747
Food as expenditure 73 70 65 60 54 61
Source Haughton and Khandker (2009)
P.23. Income tends to be understated for the
following reasons Recall problem. Intent to
limit tax burden. Reluctance to report income
from illegal sources. Some parts of income are
difficult to calculate accurately. The
understatement of both income and expenditure
means that poverty rates are overstated. It also
means that the estimates of income and
expenditure that are based on sample survey data
invariably fall short of the levels observed in
national accounts data.
5
LECTURE THREE POVERTY LINES

• Objectives of the session
• To acquaint participants with what a poverty line
  is and why is it needed.
• To enable participant develop the skill required
  for constructing poverty lines.
• 3.1 Defining a poverty line
• - The poverty line defines the level of
  consumption or income needed for a household to
  escape poverty.
• - The poverty line is obtained by specifying a
  consumption bundle considered adequate for basic
  consumption needs, then estimating the cost of
  those basic food and non food needs.
• - Another approach is to determine a food poverty
  line, based on some notion of the minimum
  caloric intake required to stay healthy and
  productive. The income required to purchase these
  basic needs food bundle is computed.
• - If the cost of basic nonfood needs is
  estimated, this added to the nonfood basic needs
  will equal the overall poverty line.
• 3.2 Approaches to Constructing a Poverty Line
  Cost of basic needs approach
• (i) One approach is to compute a poverty line for
  each household, adjusting it from household to
  household to take account of differences in the
  prices they face and their demographic
  composition. This gives different poverty line
  for each household.
• (ii) A second approach is to construct one per
  capita poverty line for all individuals, but to
  adjust per capita expenditure / income for
  differences in prices and composition. The
  adjusted per capita expenditure/income is then
  compared with the poverty line to determine
  whether the individual is living below the
  poverty line. This approach gives a single
  poverty line.
• (iii) A third approach is to construct poverty
  lines by region. Separate poverty line is
  constructed for each region, based on the prices
  prevailing in each region. Each household in a
  region is then assessed to be poor if its per
  capita expenditure / income is below the regional
  poverty line. (Table 2)
• - Over time, nominal poverty lines do change for
  a population
• (i) Due to change in price over time, and
• (ii) If the real poverty threshold is revised
  over time (Table 2).

6
Example Table 2 Summary of Poverty Lines for
Cambodia riels Per Person Per Day
1993/94SESC
1999CSES
Region Food Poverty Line Poverty Line Food Poverty Line Poverty Line
Phnom Penh 1,185 1,578 1,737 2,470
Other Urban 995 1,264 1,583 2,093
Rural 881 1,117 1,379 1,777
Source Haughton and Khandker (2009)
P.42 3.3 Relative and Absolute Poverty
Lines -Generally, the choice of poverty line
depends largely on the intended use of the
poverty rates. -Sometimes we are interested in
identifying the poorest segment of the population
for purposes of policy intervention. We may want
to identify the poorest one-fifth or two- fifths
or one- third. These are the relatively
poor. -Poverty line may in such cases be defined
as some percentage, say, two-thirds, of the mean
income / expenditure. -Were the objective is to
identify and target todays poor, a relative
poverty line is appropriate, if it is tailored to
the overall level of development. -Absolute
poverty line, on the other hand, is fixed in
terms of the standard indicator being used, and
fixed over the entire geographical space of the
poverty comparison. The poverty line is set so
that it represents the same purchasing power year
after year, but this line may differ from region
to region or country to country. -An absolute
poverty line is essential if one in concerned to
judge the effect of antipoverty policies over
time. -Legitimate comparisons of poverty rates
between one region and another, or between one
country and another can only be made if the same
absolute poverty line is used in both regions or
countries. -The World Bank/UN needs absolute
poverty lines to be able to compare poverty rates
across countries US1/day, revised in 2008 to
US1.25 a day is commonly used ( Table 3). The
poverty line refers to the poverty line used by
the 15 poorest countries in their sample,
converted to US dollars, using the most recent
measures of purchasing power parity.
7
Table 3 1/day Absolute Poverty Lines
Region of Population Living on less than 1/day in 1998 of population living on less than one-third of average national consumption for 1993 (in 1998)
East Asia Pacific 15.3 19.6
East Asia Pacifie Excluding China 11.3 24.6
Europe Central Asia 5.1 25.6
Latin America the Caribbean 15.6 51.4
Middle East of North Africa 1.9 10.8
South Asia 40.0 40.2
Sub-saharan African 46.3 50.5
Total 24.0 32.1
Total excluding china 26.2 32.0
Table 4 Absolute and Relative Poverty Rates
Country Year of population living on less than 1/day Country Year of population living on less than 1.25/day
Indonesia, rural 2005 24 China, rural 2005 26
Indonesia, urban 2005 19 China, urban 2005 2
Philipines 2006 23 India, rural 2004/05 44
Vietriam 2008 50 India, urban 2004/05 36
Vietriam 2006 22 Nigeria 2003 64
Source Haughton and Khandker (2009) P.49
8

• 3.4 Objective Poverty Lines
• The Cost of Basic Needs method
• This method requires composing a consumption
  bundle that is deemed to be adequate, with both
  food and nonfood components, and then estimate
  the cost of the bundle for each subgroup.
• The process is as follows
• Pick a nutritional requirement for good health.
  The standard value widely used, which has been
  proposed by the FAO, is 2,100 calories per person
  per day.
• Estimate the cost of meeting this food energy
  requirement, using a diet that reflects the
  habits of households close to the poverty line.
  Let this food component be Zf.
• Add a nonfood component, Znf.
• Sum up both to derive the basic needs poverty
  line, Zbn Zf Znf
• Table 5 Illustration of construction of cost of
  food component of poverty line for Nigeria.
• Table 5.1 Monthly per capital food poverty
  line-national urban average, 1996

Food items Monthly Consumption (kg) Calories Per kg Total calories Per month Unit required Prices N/kg 1996 Food Expenditure Poverty line
Cassava 2.64 3510 9,266.4 2.64 103.30 272.71
Beans 2.46 3,420 8,413.2 2.46 134.77 313.53
Rice 2.20 364.0 8,008.0 2.20 123.06 270.73
Maize 2.28 3.570 8,139.6 2.28 44.95 102.49
Millet 0.38 3,330 1,265.4 0.38 32.00 12.16
Yam 3.29 1,235 4,063.2 3.29 143.72 472.84
Meat 1.63 2,500 4,075.0 1.63 129.87 211.69
Fish(dried) 1.90 2,890 5,491.0 1.90 360.43 684.82
Eggs 0.44 1,400 616.0 0.44 42.111 18.53
Palm oil 0.52 8,750 4,550 0.52 186.15 96.80
Tomatoes 2.58 220 567.6 2.58 131.01 338.01
Pepper 0.87 940 8718 0.87 28.37 24.68
Vegetables 1.28 250 320 1.28 35.58 45.54
Fruits 0.34 430 146.2 0.34 24.8 8.43
Onions 0.57 410 233.7 0.57 47.26 26.94
Sugar 0.22 4,000 880.0 0.22 43.23 9.51
56,853.1 2,928.93
9
Table 5.2 Food poverty lines
1985/86 1992/93 1996/97
National 776.27 723.77 1,169.18
Urban 808.34 742.06 1,278.25
Rural 742.68 737.96 1,179.76
Gender Male-headed 770.52 833.93 1,237.59
Female-headed 760.18 711.82 1,154.23
RegionsNortheast 736.64 795.16 1,169.72
Northwest 827.64 1,181.95 1,169.92
North central 795.55 725.77 1,149.25
Southeast 800.93 807.85 1,379.39
Southwest 764.69 739.18 1,235.88
South south 781.31 824.29 1,156.61

• Source Aigbokhan (2000) p. 15.
• 3.5 Subjective Poverty Lines
• - This approach relies on self-rated or
  self-assessment poverty lines. It involves asking
  people simple questions such as
• (i) What income level do you personally consider
  to be absolutely minimal that is, below which
  you could not make ends meet?
• (ii) Do you consider your current consumption to
  be adequate to make ends meet?
• The answers provided, e.g. in (i) could be
  plotted, and a line fitted through the points to
  obtain a subjective poverty line, z.
• Example

10
LECTURE FOUR MEASURES OF POVERTY
Objectives of the session having obtained a
satisfactory measure of wellbeing and a poverty
line, this session is to enable participants
understand how to construct summary statistical
measures of the extent of poverty. There are a
number of such summary measures. 4.1 Headcount
Index - This is the most widely used measure of
poverty. It measures the proportion of the
population that is counted as poor P0 NP
(1) N E.g if 60
people out of a sample population of 300 are
poor, then P0 60/300 0.2 20. Equation (1)
is better written as P0 I I (yi lt
Z) . (2) N Yi is
household income / expenditure, Z is the poverty
line. If yi is less than Z, the household would
be counted as poor. The headcount index does not
capture the intersity of poverty, for example, if
a poor household were to give to a very poor
household, headcount index would remain unchanged
even though overall poverty would have
lessened. The index does not indicate how poor
the poor really are, and hence does not change if
people below the poverty line become
poorer. 4.2 Poverty Gap Index - This index
measures the extent to which households fall
below the poverty line (the poverty gaps) as a
proportion of the poverty line. - The sum of
these poverty gaps gives the minimum cost of
eliminating poverty, if transfers were perfectly
targeted. - More specifically, poverty gap, Gi,
is defined as the poverty line (Z) less actual
income (yi) for poor households Gi (Z-Yi) x
I (Yi lt Z) (3) The poverty gap index,
PI, may be written as PI I Gi .
(4) N Z The
measure is the mean proportionate poverty gap
index in the population, where nonpoor have zero
poverty gap. Example Calculating the Poverty Gap
Index Assume Z 125
11
Expenditure for each household Poverty gap
index (PI)
Expenditure in country 100 110 150 160
Poverty Gap 25 15 0 0
Gi/Z 0.20 0.12 0 0 0.08(0.32/4)
The index does not reflect changes in inequality
among the poor. 4.3 Poverty Severity (Squared
Poverty Gap) Index - The measure is a weighted
average of poverty gaps as a proportion of the
poverty gap. - The weights are the proportionate
poverty gaps themselves. That is, where the
poverty gap is, say 10, of the poverty line, it
is given a weight of 10 for a poverty gap of
50, it is given a weight of 50 - By squaring
the poverty gap index, the measure implicitly
puts more weight on observations that fall well
below the poverty line. Example Calculating
Squared Poverty Gap Index Assume Z
125 Square Poverty Gap Index (P2)
Expenditure in the country 100 110 150 160
Poverty Gap 25 15 0 0
G/Z 0.20 0.12 0 0
(G/Z)2 0.04 0.0144 0 0 0.0136(0.0544/4)
-The three indexes of poverty form what has been
termed a family of poverty measures proposed by
Foster, Greer, and Thorbecke (1984), which is
written as Pa I (Yi) a .
(6) N Z Where
a is a measure of the sensitivity of the index to
poverty when the parameter a 0, P0 is simply
the headcount index when a I, the index is
poverty gap, and when a 2, the index is poverty
severity index. 4.4 Checking for Robustness of
Poverty Lines - There are four main reasons why
measures of poverty may not be robust. (i) Samplin
g error, which is due to the fact that poverty
rate for the entire population of a country is
based on sample data which captures only a modest
number of household. (ii) Measurement error,
which is due say to underreporting. This tends to
overstate the degree of poverty. Unknown
differences in needs between households at
similar levels of income / consumption, even when
adult equivalence scale is used. Uncertainty and
imprecision due to choice of poverty line.
12
LECTURE FIVE POVERTY PROFILING

• Objective of the session
• To expose participants to what a poverty profile
  is and why it is useful, particularly in mapping
  the various dimensions of poverty.
• 5.1 What is a Poverty Profile?
• - A poverty profile sets out the major dimensions
  of poverty.
• - It shows how the pattern of poverty varies by
  demographic characteristics and geographical
  location of individuals or households.
• - Thus, a poverty profile is useful for
  comprehensive poverty comparison across sub-
  groups of a country.

Box 2 Key Questions to Guide in Preparing a Poverty Profile. -Does poverty vary widely between different areas in the country? -Are the most populated area also the areas where most of the poor live? -How is income poverty correlated with gender, age, urban and rural, racial or ethnic characteristics? -What are the main sources of income for the poor? -On what sectors do the poor depend for their livelihoods? -What products or services-tradables and montradables-do the poor sell? A tradable good is one that is, or easily might be, imported or exported. The prices of such goods are influenced by changes in the world price and the exchange rate. -To what extent are the rural poor engaged in agriculture? In off-farm employment? -How large a factor is unemployment? Underemployment? -What are the important goods in the consumption basket of the poor? How large are the shares of tradables and nontradable? -How is income poverty linked to malnutrition or educational outcomes? -What are the fertility characteristics of the poor? -To what public services do the poor have access? What is the quality of these services? -How important are private costs of education and health for the poor? -Can the poor access formal or informal credit market? -What assets-land, housing, and financial-do the poor own? Do properly rights over such assets exist? -How secure is their access to, and tenure over, natural resources?
13

• 5.2 Poverty Profile Presentation
• There are two main ways of presenting a poverty
  profile.
• By poverty status Poor versus nonpoor, or by
  expenditure per capita quintile. Then summarize
  the incidence of characteristics, e.g.
  educational level of each group.
• By some characteristics, e.g. age, educational
  level, occupation of household head, region of
  residence. Then report the poverty rate for each
  components.
• Examples on Nigeria and Indonesia
• 5.3 Poverty comparisons over Time
• To enable comparison of poverty oven time, based
  on more than one round of household survey,
  certain steps are necessary.
• Use data from highly comparable questionnaires
  with respect to sampling frame and definitions of
  income or consumption otherwise, adjust the
  definition of income por consumption aggregates
  over time to ensure that similar definitions are
  used.
• Adjust for price changes over time. Use temporal
  or regional price indexes to ensure similar
  definition of the poverty line over time and
  across regions to measure real income or
  expenditure over time.

Principal Sector of Employment Population Share, 1984 (percent) Contribution of sectoral change (x 100) Headcount Index Poverty gap index Poverty Severity Index (po)(percent) (p1) (percent) (p2), Contribution of sectoral change (x 100) Headcount Index Poverty gap index Poverty Severity Index (po)(percent) (p1) (percent) (p2), Contribution of sectoral change (x 100) Headcount Index Poverty gap index Poverty Severity Index (po)(percent) (p1) (percent) (p2),
Farming Self-employed Laborer 45.0 9.0 49.8 11.2 54.6 14.8 57.4 16.5
Industry Urban Rural 2.6 3.3 4.1 0.8 2.8 3.2 0.4 3.1 2.6 0.3 2.7 2.2
construction trade Urban Rural 5.4 6.6 3.8 2.2 7.2 3.6 1.6 5.6 2.7 1.4 4.7 2.2
Transport services Rural Urban Total sector effect (including omitted sectors) Contribution of population shifts Interaction effects 6.5 5.8 n.a. n.a. n.a. 1.0 2.9 89.3 13.2 -2.6 1.0 2.4 93.8 10.4 -4.3 0.9 2.0 95.1 9.4 -4.5
Total 100.0 100.0 100.0 100.0
14
Table 5.2 Sectoral Poverty Profile for
Indonesia, 1987
Principal Sector of Employment Population Share (percent) Headcount Index (po) (percent) Poverty gap index (p1) (percent) Poverty severity
Farming Self-employed Laborer Industry Urban Rural construction trade Urban Rural Transport services Rural Urban Other total 41.1 8.6 3.0 3.4 4.3 6.3 7.6 4.1 7.6 7.3 6.7 100.0 3.11 38.1 8.1 19.4 17.4 5.0 14.7 10.7 4.2 11.6 17. 21.7 6.42 7.62 1.26 3.00 2.92 0.71 2.42 1.53 0.61 1.84 3.55 4.22 1.97 2.21 0.32 0.76 0.80 0.17 0.61 0.34 0.14 0.49 1.03 1.24
Source Huppi and Ravallion, 1991  Table
5.3 Distribution of Poverty by Age and Gender
of Household Head in Cambodia, 1999
Head count Index (Po) Head count Index (Po) Head count Index (Po) Poverty gap Index (P1) Poverty gap Index (P1) Poverty gap Index (P1) Poverty severity c Index (P2), x100 Poverty severity c Index (P2), x100
Index (percent) Contribution To total (percent) Contribution To total (percent) index Contribution To total (percent) Share of Total Population (percent)
35.9 100.0 6.5 100.0 2.0 100.0 100.0
Age of head
18-30 years 36.7 10.7 5.6 9.1 1.4 7.5 10.5
31-35 years 35.4 10.9 5.4 9.2 1.6 8.8 11.1
36-40 years 43.6 21.2 8.0 21.6 2.7 23.3 17.5
41-45 years 40.3 15.7 7.3 15.8 2.2 15.3 14.0
46-50 years 36.5 14.4 7.7 16.9 2.4 16.9 14.2
51-60 years 28.3 15.8 5.3 16.3 1.7 16.8 20.0
61 and above 32.0 11.3 5.6 11.1 1.8 11.3 27.7
Male 36.4 84.4 6.6 84.2 2.1 85.1 83.3
Female 33.6 15.7 6.1 15.8 1.8 14.9 16.7
15
LECTURE SIX IDENTIFYING THE DETERMINANTS OF
POVERTY

• Objective of the session To enable participants
  identify main causes of poverty and understand
  how regression technique could be used to
  identify the immediate main causes of poverty,
  for purposes of identifying necessary policy
  interventions.
• 6.1 Causes of Poverty
• The main causes of poverty may be classified by
  certain characteristics
• Household and individual characteristics, the
  most important of these are
• Demographic characteristics, such as household
  size, age structure, gender of head, dependency
  ratio.
• Economic, such as employment status, occupation,
  hours worked, property owned.
• Social, such as health and nutritional status,
  shelter, education of head.
• Regional level characteristics, which include
  availability of infrastructure, proximity to
  markets, quality of governance, property rights
  and their enforcement, and vulnerability to
  weather conditions.
• 6.2 Analyzing the Determinants of Poverty
• - The most widely used technique to identify the
  contribution of different variables or indicators
  to poverty is regression analysis.
• - There are two main types of analysis
• (i) Attempts to explain the level of expenditure
  per capita (an indicator for living standard or
  welfare) as a function of a variety of variables
  or indicators, typically those discussed in 6.1
  above.
• (ii) Attempts to explain whether the household is
  poor, using a probit or logit regression. In this
  case, the independent variables are as in 6.1,
  but the dependent variable is binary, taking the
  value of 1 if the household or individual is poor
  and zero otherwise.
• - A regression estimate shows how closely each
  independent variable is related to the dependent
  variable, holding all other influences constant.
• A typical multiple regression equation as
  applied in poverty analysis is
• Log (Yi) a0a1Xi1a2Xi2a3Xi3..anXin
• Z

16

• Where z is the poverty line, yi is per capita
  expenditure or income, Xi are respective
  explanatory variables, and aj are the
  coefficients to be estimated.
• -Level of significance As a rule of thumb,
  t-statistic of less than 2, p-value of 0.05 and
  above, suggests that estimated coefficient is not
  statistically significant from zero.
• 6.3 Interpreting Regression Results on
  Determinants of Poverty
• Examples
• Table 6.1 Determinant of household welfare in
  Ivory Coast, 1993

Dept. variable lag(per capital) Urban Coeff t-statistic Rural Coeff t-statistic
Educ. Level of most Educated male Elementary Junior secondary Senior secondary University 0.38 5.3 0.62 8.6 0.80 9.6 0.93 9.4 0.08 0.9 0.05 0.4
Educ. Level of most Educated female Elementary Junior secondary Senior secondary University 0.11 1.7 0.24 3.1 0.34 4.1 0.52 4.1 0.07 1.0 0.27 2.2
Value of selected house Hold assets Home Business assets Savings 0.06 5.3 0.04 3.3 0.08 4.7 0.16 4.9
Hectares of agric. Land Cocoa trees Coffee trees 0.17 4.3 0.04 1.3
Distance to nearest paved road -0.04 -2.0
Distance to nearest market -0.09 -3.3
Unskilled wage 0.37 0.4
17
6.2 OLS Regression of determinant of economic
welfare in Nigeria, 1996 dependent variables
LnPCE
Constant Age of head Age2 Sex of head Edu. Of head Household size Household Size2 Adj. R2 Adj. R2
Middle Belt Middle Belt Middle Belt Middle Belt Middle Belt Middle Belt Middle Belt Middle Belt Middle Belt Middle Belt
Rural 8.73 (57.68) -0.002 (0.33) -0.00001 (0.19) -0.64 (13.29) 0.03 (1.46) -0.31 (22.23) 0.01 (13.18) 0.31  
Urban 7.49 (22.63) 0.02 (0.99) -0.0001 (0.89) -0.29 (3.15) 0.17 (6.11) -0.25 (13.49) 0.006 (7.63) 0.32  
North-East North-East North-East North-East North-East North-East North-East North-East North-East North-East
Rural 7.49 (42.63) 0.01 (1.59) -0.0001 (1.70) -0.12 (1.46) 0.09 (3.61) -0.32 (20.10) 0.01 0.29 0.29
Urban 7.26 (11.99) 0.01 (0.51) -0.0002 (0.69) 0.14 (1.14) 0.19 (4.44) -0.26 (6.64) 0.009 (3.75) 0.35 0.35
North-West North-West North-West North-West North-West North-West North-West North-West North-West North-West
Rural 7.60 (41.58) -0.02 (2.62) 0.0002 0.14 (1.18) 0.14 (4.10) -0.29 (16.65) 0.01 (9.40) 0.28 0.28
Urban 8.63 (12.82) -0.05 (1.77) 0.0004 (1.45) 0.12 (0.58) 0.05 (0.92) -0.20 (5.81) 0.004 (2.38) 0.38 0.38
South-East South-East South-East South-East South-East South-East South-East South-East South-East South-East
Rural 8.10 (36.35) 0.01 (0.63) -0.0001 (0.55) -0.16 (3.55) 0.12 (4.81) -0.35 (18.10) 0.01 (10.34) 0.31 0.31
Urban 7.92 (11.39) 0.03 (1.01) -0.0003 (1.11) 0.06 (0.24) 0.18 (2.62) -0.58 (7.46) 0.03 (4.76) 0.49 0.49
Source Aigbokhan (2000) Lecture 7 Poverty
Analysis based on Census data Objective of
session To acquaint participants with basic
methodologies useful in computing and analyzing
poverty estimates, based on national population
census data.
18

• 7.1 Poverty Mapping
• Governments are often interested in more detailed
  poverty rates for relatively small
  geographical units.
• Often in a given region there are typically wide
  divergence in standards of living
• Detailed poverty maps cannot be generated from
  household survey data alone.
• If one tries to use survey data to measure
  poverty in each district or region, such
  estimates would be based on just a few
  observations, and would as such be imprecise.
• One solution is to use population census data.
• Census data are more available for all households
  and can provide reliable estimates at highly
  disaggregated levels, such as small
  municipalities, towns and villages.
• Census do not contain the income or consumption
  information necessary to yield reliable
  indicators of welfare or standard of living, such
  as poverty and inequality measures.
• 7.2 Methodology
• One approach involves estimations of measures of
  poverty based on the access and quality of
  service. One such measure is the Living
  Conditions Index, (LCI). This index
• Summarizes access and quality of household source
  of water
• Source of drinking water
• Type of toilet facility
• The main method of garbage collection.
• The extent of crowding in the household (number
  of people divided by the number of bedrooms in
  the dwelling).
• This approach involves classifying responses into
  Level 1 ( high quality) through Level 5 ( low
  quality).
• Then assign a number of points to each level,
  say, 100,75,50,25, and 0 to the respective level.
• Let us specify h000y, to denote the relevant
  question number in the census questionnaire,
  where y denotes the code for the answer e.g
  water supply
  level 1if h171/h174 level 2 if h173/ h175
  level 3if
  h176/h177 level4 if h172 level 5 if
  h178/h179.
• Crowding Index level 1 one person to a bedroom,
  level 2 1 or 2 persons, Level 3 2 or 3
  persons, Level 4 3 or 4 persons, Level 5 4
  persons or more.

19
Table 1 A comparism of the ranking of rural and
urban areas based on the LCI and the EDMI (means
weighted by the number of households in the ED)
Poverty score Based on LC11 Poverty score based on EDMI
Rural 341 0.333
Urban 426 -0.782
Table 2 A comparism of the ranking of regions
based on the LCI and EDMI (means weighted by the
number of households in the ED)
Based on LCI Based on LCI Rank (poorest on top less poor at bottom) Based on EDMI Based on EDMI
Region 8 162 1 Region 1 2.125
Region9 184 2 Region 9 2.049
Region 1 207 3 Region 8 1.982
Region 7 259 4 Region 7 1.023
Region 2 278 5 Region 2 0.583
Region 3 352 6 Region 5 0.303
Region 5 355 7 Region 3 0.234
Region 10 364 8 Region 6 0.188
Region 6 373 9 Region 4 -0.137
20

• 7.3 Combining Household Survey and Census Data.
• An alternative approach is to combine survey data
  with census data to create more detailed poverty
  maps.
• The basic idea is to use more detailed survey
  data to construct household or individual level
  indicators of welfare as a function of variables
  that are common to both survey and census data..
  In(y?)X1? ßU?.(1)
• The second step is to use estimates from equation
  (1) along with census data on the Xi variables,
  to get predicted values of Yi for every household
  in the country.
• These predicted values can then be used to
  measure poverty to a more disaggregated level.
• Total poverty gap (TPG) in each district/ county
  can be estimated as TPGi P1i x Zi x Ni(2)
• Where P1i is poverty gap in location i, Zi is
  poverty line in location i, Ni is total
  population in location i.
• The share (Wi ) of the total poverty gap of
  location i among k locations within a county is
  calculated as
• ?TPG  i-ikWi TPGi...............................
  ..(3)
• 
  
• Inequality measures
• Useful to examine the relation between poverty
  and inequality.
• Correlation estimation between FTG poverty
  indices and inequality indices
• May be positive or negative in survey data.
• More likely to be negative in census data. This
  is because being a larger population than sample
  data, the larger the proportion of the population
  below the poverty line, the less scope there is
  for variance in the per capita consumption
  measure, and therefore the less inequality likely
  to be.
• 7.4 Exercises
• 1. Identify questions from National Census
  Questionnaire to construct the following
  variables
• EDMI VARIABLES

21

• -Proportion of dwellings built with
  mud/stick/reed/bamboo/thatch/mat/grass ( H04-H06)
• -Average number of household members per bedroom
  (H03)
• LCI VARIABLES
• -Number and quality of rooms the household
  occupy (H03-H06)
• -Main source of water supply L1 Pipe or pump
  indoorsL2 Pipe or pump outdoors/Public tap L3
  Close well or protected springL4 Open wellL5
  River/Lake/Spring/Vendors (H07)
• -Mai n source of fuel used for lighting/cooking
  L1 Electricity own generator, L2 Electricity
  power supply, L3 Gas/Kerosine, L4 Palm oil
  lamp/charcoal, L5 wood (H08-H09)
• -Type of human waste disposal system used by
  household L1Flush toilet for household alone, L2
  Flush toilet shared with other households, L3
  Covered pit latrine, L4 Open pit latrine, L5
  Bush/Beach (H10)
• -Distance from nearest facilities (H11-H13) L1
  Less than 20 mins, L2 20-39 mins, L3 40-59 mins,
  L4 60-79 mins, L5 80 mins and more (H11-H13)
• -Construction materials of housing unit
  (H04-H06)
• Outer wallsL1 cement blocks, L2 Mud blocks, L3
  zinc and iron, L4 mud and sticks, L5
  bamboo/reed/grass/mat
• Roof---L1 concrete/tiles, L2
  Asbestos, L3 Zinc, L4 Tapaulin, L5 Bamboo/thatch.
• Floor---L1 Tiles, L2 cement, L3 mud,
  L4 wood, L5 others.
•  

22
THANK YOU