Feed the Future Ethiopia 
Zone of Influence End-line Assessment Report 2018
August 2019
Prepared for the United States Agency for International Development, USAID Grant Number 
AID-663-I0-13-00001
Recommended Citation:
Bachewe, Fantu, Guush Berhane, Kalle Hirvonen, John Hoddinott, Bethelehem Koru, Fanaye 
Tadesse, Alemayehu Seyoum Taffesse, Abdulazize Wolle, Abenezer Wondwossen, Ibrahim 
Worku, and Yisehac Yohannes, 2019. Feed the Future Ethiopia Zone of Influence End-line Assessment 
Report 2018. International Food Policy Research Institute.
USAID/Ethiopia Contact:
Semachew Kassahun (skassahun@usaid.gov) 
Alemayehu Seyoum Taffesse/International Food Policy Research Institute (IFPRI) 
IFPRI C/O ILRI, Bole Sub-City, Kebele no 13, Box 5689, Addis Ababa, Ethiopia
Email: A.SeyoumTaffesse@cgiar.org
USAID Bureau for Resilience and Food Security Contact:
rfs.ald@usaid.gov
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 3 
Table of Contents
List of Tables.......................................................................................................................... 5
List of Figures......................................................................................................................... 8
List of Acronyms................................................................................................................... 9
Executive Summary............................................................................................................. 10
Background ..................................................................................................... 10
ZOI Survey 2018 Phase One End-line Assessment Indicators............ 10
ZOI Survey 2018 Assessment Data Sources.......................................... 11
Summary of Key Findings............................................................................. 12
1. Background ..................................................................................................... 15
1.1 Feed the Future Overview ............................................................ 15
1.2 Feed the Future P1-ZOI Profile.................................................... 15
1.2.1 Rationale for P1-ZOI Selection.................................... 16
1.2.2 Demography of the P1-ZOI.......................................... 17
1.2.3 Agriculture in the P1-ZOI ............................................. 20
1.3 Purpose of this Assessment........................................................... 21
2. Methodologies for Obtaining 2018 Values for Feed the Future 
Indicators ......................................................................................................... 22
2.1 Data Sources..................................................................................... 22
2.1.1 Primary Data: The Feed the Future Ethiopia 
ZOI Survey 2018 ............................................................. 22
2.1.2 Comparability of Data Sources Used for the 
Feed the Future Phase One Baseline and End￾line ZOI Survey Assessment......................................... 27
2.2 Measures and Reporting Conventions Used 
Throughout This Report................................................................ 29
2.2.1 Standard Disaggregates.................................................. 29
2.2.2 Reporting Conventions.................................................. 31
3. ZOI Survey 2018 P1-ZOI Population ....................................................... 32
3.1 Demographics................................................................................... 32
3.2 Living Conditions............................................................................. 37
3.3 Education ........................................................................................... 40
4. Household Economic Status....................................................................... 43
4.1 Daily Per Capita Consumption Expenditures............................ 43
4.2 Prevalence and Depth of Poverty in the P1-ZOI ..................... 46
4.2.1 The $1.25 Poverty Threshold....................................... 47
4.2.3 The National Poverty Threshold ................................. 51
5. Hunger and Dietary Intake.......................................................................... 53
5.1 Household Hunger.......................................................................... 53
5.2 Dietary Intake................................................................................... 55
5.2.1 Dietary Diversity among Women Age 15-49 
Years................................................................................... 55
5.2.2 Infant and Young Child Feeding.................................... 62
6. Nutritional Status of Women and Children............................................ 68
6.1 Body Mass Index of Women Age 15-49 Years......................... 68
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 4 
6.2 Stunting, Wasting, Underweight among Children 
Under 5 Years.................................................................................. 72
6.2.1 Stunting (Low Height-for-Age)..................................... 72
6.2.2 Wasting (Low Weight-for-Height).............................. 73
6.2.3 Underweight (Low Weight-for-Age) .......................... 76
7. Women’s Empowerment in Agriculture.................................................. 80
7.1 Overview ........................................................................................... 80
7.2 Production ......................................................................................... 88
7.3 Resources.......................................................................................... 90
7.4 Income....................................................................................................... 94
7.5 Leadership ................................................................................................ 94
7.6 Time.................................................................................................... 96
8. Summary and Conclusions.......................................................................... 99
References 101
Appendix 1. Supplementary Data and Figures........................................................... 103
Appendix 2. Methodology .............................................................................................. 105
A2.1 Sampling and Weighting .............................................................. 105
Determining sample size ............................................................. 106
A2.2 Poverty Prevalence and Consumption Expenditure 
Methods .......................................................................................... 111
Measuring poverty and comparing poverty across 
groups ............................................................................. 111
A2.3 Criteria for Achieving Adequacy for Women’s 
Empowerment in Agriculture Indicators................................. 114
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 5 
List of Tables
Table ES1: Feed the Future PI-ZOI indicator estimates, by key disaggregates: Ethiopia 
2013-2018 ............................................................................................................................ 13
Table 1.2. Number of households, by category, in the P1-ZOI, Ethiopia 2018 ...................... 20
Table 2.2. Comparison of results of household and individual interviews in the PI￾ZOI, in total, Feed the Future phase one baseline and end-line ZOI 
Surveys.................................................................................................................................. 27
Table 3.1a Household demographic characteristics in the P1-ZOI ............................................ 32
Table 3.1b Comparison of household demographic characteristics between the Feed 
the Future phase one baseline and end-line ZOI Surveys......................................... 33
Table 3.2a.: Comparison of characteristics of primary adult female decision-makers in 
the PI-ZOI, Feed the Future phase one baseline and end-line ZOI Surveys ........ 35
Table 3.2b. Comparison of characteristics of primary adult male decision-makers in 
the PI-ZOI, Feed the Future phase one baseline and end-line ZOI Surveys ........ 36
Table 3.3a Comparison of household dwelling characteristics between the Feed the 
Future phase one baseline and end-line ZOI Surveys................................................ 38
Table 3.3b. Comparison of household dwelling characteristics, by gendered 
household type, between the Feed the Future phase one baseline and 
end-line ZOI Surveys......................................................................................................... 39
Table 3.4a. School attendance, educational attainment and literacy in the P1-ZOI................. 41
Table 3.4b. School attendance, educational attainment and literacy in the P1-ZOI, by 
age and sex........................................................................................................................... 41
Table 3.5. Comparison of household member educational attainment in the PI-ZOI, 
in total and by sex, Feed the Future phase one baseline and end-line ZOI 
Surveys.................................................................................................................................. 42
Table 4.1a Daily per capita consumption expenditures by household characteristic 
(in constant 2010 USD1
) in the P1-ZOI ........................................................................ 44
Table 4.1b Comparison of daily per capita consumption expenditures in constant 
2010 USD at 2005 PPP between the Feed the Future phase one baseline 
and end-line ZOI Surveys1
................................................................................................ 45
Table 4.1c Comparison of mean daily per capita consumption expenditures in 
Constant 2010 USD at 2005 PPP in the PI-ZOI, in total and by selected 
household characteristics, Feed the Future phase one baseline and end￾line ZOI Surveys................................................................................................................. 46
Table 4.2. Poverty in the ZOI at the international poverty line of $1.25 (2005 PPP) - 
2018 ....................................................................................................................................... 48
Table 4.3a: Comparison of prevalence of poverty at the $1.25 (2005 PPP) per person 
per day threshold in the PI-ZOI, in total and by selected household 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 6 
characteristics, Feed the Future phase one baseline and end-line ZOI 
Surveys.................................................................................................................................. 49
Table 4.3b: Comparison of depth of poverty at the $1.25 (2005 PPP) per person per 
day threshold in the PI-ZOI, in total and by selected household 
characteristics, Feed the Future phase one baseline and end-line ZOI 
Surveys.................................................................................................................................. 50
Table 4.4 Total (absolute) poverty line in Birr (average price).................................................. 51
Table 4.5: Comparison of prevalence of poverty at the national threshold of 13.02 
Birr (baseline) 18.66 Birr (end-line) in the PI-ZOI, in total and by selected 
household characteristics, Feed the Future phase one baseline and end￾line ZOI Surveys................................................................................................................. 52
Table 5.1a. Comparison of household hunger in the PI-ZOI, by severity, in total and 
by selected household characteristics, Feed the Future phase one baseline 
and end-line ZOI Surveys................................................................................................. 54
Table 5.1b. Comparison of household hunger between the Feed the Future phase 
one baseline and end-line ZOI Surveys......................................................................... 55
Table 5.2. Women’s mean and median dietary diversity scores in the P1-ZOI...................... 57
Table 5.3. Percentage of women of reproductive age achieving minimum dietary 
diversity in the P1-ZOI...................................................................................................... 59
Table 5.4. Percentage of women who consumed foods in the P1-ZOI, by 
achievement of minimum dietary diversity status....................................................... 60
Table 5.5. Comparison of women’s dietary diversity between the Feed the Future 
phase one baseline and end-line ZOI Surveys............................................................. 61
Table 5.6. Prevalence of exclusive breastfeeding among children 0-5 months in the 
P1-ZOI .................................................................................................................................. 62
Table 5.7. Prevalence of children 6-23 months who receive a minimum acceptable 
diet in the P1-ZOI.............................................................................................................. 64
Table 5.8.: Comparison of percent of children 6-23 months of age in the PI-ZOI 
achieving minimum feeding frequency, dietary diversity, and consuming 
specified foods, in total and by breastfeeding status and age, Feed the 
Future phase one baseline and end-line ZOI Surveys................................................ 65
Table 5.9. Comparison of children’s dietary intake between the phase one baseline 
and end-line ZOI Surveys................................................................................................. 67
Table 6.1. Mean BMI and prevalence of underweight, normal weight, overweight, 
and obese women in the P1-ZOI ................................................................................... 69
Table 6.2 Comparison of the nutritional status of women between the Feed the 
Future phase one baseline and end-line ZOI Surveys................................................ 70
Table 6.3. Prevalence of stunting and mean height-for-age z-scores among children 
under 5 years in the P1-ZOI............................................................................................ 73
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 7 
Table 6.4. Prevalence of wasting and mean weight-for-height z-scores among 
children under 5 years in the P1-ZOI............................................................................ 75
Table 6.5. Prevalence of underweight and mean weight-for-age z-scores among 
children under 5 years in the P1-ZOI............................................................................ 77
Table 6.6 Comparison of the nutritional status of children between the Feed the 
Future phase one baseline and end-line ZOI Surveys................................................ 78
Table 7.1 WEAI domains, indicators and definitions of adequacy............................................. 81
Table 7.2a: Comparison of average percent of primary adult decision-makers in the 
PI-ZOI achieving adequacy in the 10 WEAI indicators, using censored 
headcounts, by sex and age, Feed the Future phase one baseline and end￾line ZOI Surveys................................................................................................................. 82
Table 7.2b: Comparison of adequate achievement in each WEAI indicator in PI-ZOI 
using censored headcount ratios, by sex and age, Feed the Future phase 
one baseline and end-line ZOI Surveys......................................................................... 84
Table 7.2c. Comparison of WEAI, 5DE and GPI scores, and average empowerment 
gap in the PI-ZOI, Feed the Future phase one baseline and end-line ZOI 
Surveys.................................................................................................................................. 86
Table 7.3. Participation in economic activities and input into activity decision-making 
among female and male primary adult decision-makers in the P1-ZOI 
end-line Survey.................................................................................................................... 89
Table 7.4a. Household and female and male primary adult decision-maker ownership 
of productive resources in the P1-ZOI end-line Survey ........................................... 91
Table 7.4b. Household and female and male primary adult decision-maker control 
over productive resources in the P1-ZOI end-line Survey....................................... 92
Table 7.5a. Access to credit among female primary adult decision-makers in the P￾ZOI end-line Survey........................................................................................................... 93
Table 7.5b. Access to credit among male primary adult decision-makers in the P1-
ZOI end-line Survey........................................................................................................... 93
Table 7.6. Input into decision-making on use of income among male and female 
primary adult decision-makers in the P1-ZOI end-line Survey................................ 94
Table 7.7a. Group membership among female and male primary adult decision￾makers in the P1-ZOI end-line Survey .......................................................................... 95
Table 7.7b. Comfort with speaking in public among female and male primary adult 
decision-makers in the P1-ZOI end-line Survey.......................................................... 96
Table 7.8a. Time allocation among female primary adult decision-makers in the P1-
ZOI end-line Survey........................................................................................................... 97
Table 7.8b. Time allocation among male primary adult decision-makers in the P1-ZOI 
end-line Survey.................................................................................................................... 98
Table A1.1. ZOI Survey 2018 Feed the Future indicator estimates...........................................103
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 8 
Table A1.2. Women’s Empowerment in Agriculture Index results for indicators that 
compose the 5 domains of empowerment, using uncensored headcount 
ratios ...................................................................................................................................104
Table A2.1: Household composition of the EA sample ................................................................. 109
Table A2.2: Feed the Future sample woredas, grouped by Feed the Future program...........109
Table A2.3: Feed the Future sample households, planned by region..........................................110
List of Figures
Figure 1.1. Map of Ethiopia: Feed the Future P1-ZOI .................................................................... 16
Figure 4.1. Share of consumption per quintile: Feed the Future P1-ZOI................................... 45
Figure 7.1: Contribution of each indicator to the disempowerment of women and 
men in the P1-ZOI end-line Survey................................................................................ 87
Figure 7.2a: Contribution of each of the five domains to the disempowerment of 
women in the P1-ZOI end-line Survey.......................................................................... 88
Figure 7.2b: Contribution of each of the five domains to the disempowerment of men 
in the P1-ZOI end-line Survey......................................................................................... 88
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 9 
List of Acronyms
5DE Five Domains of Empowerment
A-WEAI Abbreviated Women’s Empowerment in Agriculture Index
BFS Bureau for Food Security
BMI body mass index 
CI confidence interval
CPI Consumer Price Index
DEFF design effect 
DHS Demographic and Health Survey
EA Enumeration area
FTFMS Feed the Future Monitoring System
GPI Gender Parity Index
HHS Household Hunger Scale
IFPRI International Food Policy Research Institute
LCU local currency unit
LSMS Living Standards Measurement Survey
M&E monitoring and evaluation
MAD minimum acceptable diet
MDD-W Women’s Minimum Dietary Diversity
MDG Millennium Development Goal
PBS population-based survey
P1-ZOI Phase One Zone of Influence
PPP purchasing power parity
PPS probability proportional to size
SD standard deviation
SDG Sustainable Development Goals
UNDP United Nations Development Program
UNICEF United Nations Children’s Emergency Fund
USAID United States Agency for International Development
USD United States Dollar
WDDS Women’s Dietary Diversity Score
WEAI Women’s Empowerment in Agriculture Index
WHO World Health Organization
ZOI Zone of Influence
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 10
Executive Summary
Background
Feed the Future seeks to reduce poverty, hunger and undernutrition among women and children, 
and to increase income, resilience, women’s empowerment, dietary diversity and appropriate 
feeding practices, and hygienic environments. Program efforts are designed to impact the 
population in zones of influence (ZOI) in Feed the Future target countries. Progress in achieving 
Feed the Future’s objectives is tracked using population-based performance indicators collected 
at baseline then periodically thereafter.
This assessment reports Feed the Future’s population-based phase one indicators from the Feed 
the Future Ethiopia Zone of Influence (ZOI) Survey 2018. The 2018 survey serves as the end-line 
survey for phase one of Feed the Future. The assessment compares indicator estimates and 
selected demographic and household characteristics from the 2018 ZOI Survey with those 
obtained from the baseline survey conducted in 2013 to assess trends and to test for statistically 
significant differences between the two time periods in the phase one ZOI (P1-ZOI) in Ethiopia. 
The assessment includes only Feed the Future phase one indicators that were reported at baseline
and for which comparable baseline and end-line survey data are available.
The Feed the Future P1-ZOI in Ethiopia includes rural areas in Amhara, Oromia, Tigray, Somali 
and Southern Nations, Nationalities, and People’s (SNNP) regions (see Figure 1.1). 
This end-line assessment will provide the United States Government interagency partners, USAID
Bureau for Food Security (BFS), USAID Missions, host country governments and development 
partners with information about long-term progress of the Feed the Future population-based PI￾ZOI indicators in Ethiopia. The assessment is designed for use as a monitoring tool, and as such 
provides indicator point estimates that were generated to have an acceptable level of statistical 
precision. The assessment is designed to also measure changes in indicator estimates and select 
demographic and household characteristics from the Feed the Future baseline ZOI Survey, which 
was conducted in 2013. Feed the Future ZOI Survey sample calculations, however, are not 
designed to support conclusions of causality or program attribution.
The Feed the Future Ethiopia ZOI Survey 2018 End-line Assessment is a product of the 
International Food Policy Research Institute (IFPRI), which is responsible for specific elements of 
performance monitoring and impact evaluation supporting the Feed the Future initiative.
ZOI Survey 2018 Phase One End-line Assessment Indicators
The Feed the Future phase one population-based ZOI indicators included in this assessment are: 
1. Daily per capita expenditures (2010 USD)
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 11
2. Prevalence of poverty: Percent of people living on less than $1.25/day 2005 PPP 
3. Depth of poverty: Mean percent shortfall of the poor relative to the $1.25/day 2005 PPP 
poverty line 
4. Prevalence of moderate and severe hunger 
5. Women dietary diversity: Mean number of food groups consumed by women of 
reproductive age
6. Prevalence of underweight women of reproductive age 
7. Prevalence of underweight children under 5 years of age
8. Prevalence of stunted children under 5 years of age
9. Prevalence of wasted children under 5 years of age
10. Prevalence of children 6-23 months receiving a minimum acceptable diet 
11. Prevalence of exclusive breastfeeding of children under 6 months of age 
12. Women’s Empowerment in Agriculture Index (WEAI)
Data for measuring the prevalence of anemia among women of reproductive age and the 
prevalence of anemia among children 6-59 months and consumption of targeted nutrient-rich 
value chain commodities were not collected for the Feed the Future Ethiopia ZOI Survey 2018 
and therefore are not included in this assessment. 
ZOI Survey 2018 Assessment Data Sources
Data for the ZOI indicators presented in this assessment are drawn from two sources: baseline 
and end-line household surveys conducted in June/July in 2013 and September/October 2018, 
respectively. 
The 2018 Ethiopia ZOI Survey was conducted by International Food Policy Research Institute 
(IFPRI) in conjunction with its data collection partner, BST Consulting PLC. (BST for short), with 
some support (enumerator recruitment, logistics and survey equipment) from the Central 
Statistics Agency (CSA). Fieldwork for the ZOI Survey took place between September and 
October 2018. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 12
Summary of Key Findings
Household Economic Status
The daily per capita expenditures in the Feed the Future ZOI in 2018 was $2.19 (in 2010 USD) 
– a level that is statistically significantly higher than the level at the baseline ($1.77). The 
population that fell below the $1.25/day (2005 PPP) poverty line was 32.3 percent in 2018 and 
lower than the 39.9 percent recorded in 2013. This decline in the prevalence of poverty is 
statistically significant. The depth of poverty is estimated at 12.2 percent for 2018 and 13.2 
percent for 2013. However, this difference is not statistically significant.
Hunger and Dietary Intake
Prevalence of households with moderate or severe hunger is 3.3 percent in the end-line, and this 
is statistically significantly lower relative to the prevalence in the baseline, 4.9 percent. The dietary 
diversity in the ZOI has been and remains low. Nevertheless, there has been considerable 
improvements during the 2013-2018 period. Women of reproductive age consume from an 
average of 2.32 food groups in a day in 2018. This is nearly a 50 percent, statistically significant, 
increase relative to the corresponding figure in 2013 (1.59 food groups). Only 1.84 percent of 
women achieve the minimum dietary diversity or consume from at least five food groups in the 
baseline while this number increased to 4.64 percent in the end-line, or by over 150 percent 
during the 2013-2018 period (Table 5.5). Only 3.4 percent of children 6-23 months old received 
a minimum acceptable diet in 2013 while the proportion more than doubled to 8.4 percent in 
2018. Both of these changes are statistically significant. Prevalence of exclusive breastfeeding 
among children under 6 months of age is stayed high in Ethiopia at about 74 percent in 2018 with 
little change during the period. 
Nutritional Status of Women and Children
Nearly 22 percent of women of reproductive age are underweight in 2018, and this is statistically 
significantly lower than the 26 percent in 2013. The prevalence of severely underweight women 
declined even more considerably, by nearly 40 percent, from 3.8 percent in 2013 to 2.3 percent 
in 2018. Prevalence of stunting among children under 5 years of age was about 39 percent in 
2018, which, relative to the 50.3 percent in 2013, was 23 percent lower. The prevalence of 
wasting among children under 5 years of age declined by 31 percent from over 14 percent in 
2013 to 9.7 percent in 2018. Similarly, the proportion of underweight children declined from 31.5 
percent in 2013 to 22.5 percent in 2018. The changes reported for all three child anthropometric 
measures are statistically significant. 
Women’s Empowerment in Agriculture 
Women’s empowerment is measured through the full Women Empowerment in Agricultural 
Index (WEAI). In 2018, the WEAI for the ZOI in Ethiopia is estimated as 0.75 and did not change 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 13
significantly from 2013. Access to and input into decisions on credit is a major constraint to 
women’s empowerment. Lack of leisure time, group membership and speaking in public are also 
among important contributors to women’s disempowerment. 
Baseline and end-line estimates of indicator values in the ZOI are shown in Table ES1 below. The 
main takeaway from the table is that most indicators are improving and some considerably. 
Further analysis is in subsequent chapters.
Table ES1: Feed the Future PI-ZOI indicator estimates, by key disaggregates: Ethiopia 
2013-2018
 Baseline (2013) End-line (2018)
Diff.
p￾valueb Sig.c Indicator Est. 95% CI na Est. 95% CI na
Mean daily per capita consumption expenditure in constant 2010 USD (2005 PPP) 
All households 1.77 1.70-1.84 4,618 2.19 2.10-2.27 3,876 0.42 0.000 ***
Gendered household type 
Male and female adults 1.73 1.65-1.81 3,534 2.21 2.12-2.31 3,089 0.48 0.000 ***
Female adults only 1.89 1.74-1.99 855 1.95 1.79-2.11 665 0.08 0.383 n/s
Male adults only 2.35 2.06-2.63 224 2.05 1.53-2.56 115 -0.35 0.250 n/s
Children only ^ ^ 2 ^ ^ 7 ^ ^ ^ 
Prevalence of poverty: Percentage of people living on less than $1.25/day (2005 PPP)
All households 39.9 38.5-41.4 4618 32.3 30.8-33.8 3,876 -7.6 0.000 ***
Gendered household type 
Male and female adults 40.4 38.8-42.0 3,534 31.3 29.6-32.9 3,089 -9.1 0.000 ***
Female adults only 38.9 35.6-42.2 855 40.8 37.1-44.6 665 1.9 0.504 n/s
Male adults only 31 24.9-37.1 224 36.3 27.3-45.1 115 5.3 0.445 n/s
Children only ^ ^ 2 ^ ^ 7 ^ ^ ^ 
Depth of poverty: Mean percent shortfall relative to the $1.25/day poverty line (2005 PPP)
All households 13.2 12.6-13.8 4,618 12.2 11.4-12.8 3,876 -1.0 0.054 n/s
Gendered household type 
Male and female adults 13.3 12.6-14.0 3,534 11.7 10.9-12.5 3,089 -1.6 0.010 * 
Female adults only 13.2 11.7-14.6 855 15.8 13.9-17.6 665 2.6 0.058 n/s
Male adults only 9.7 7.3-12.1 224 12.5 8.5-16.4 115 2.8 0.279 n/s
Children only ^ ^ 2 ^ ^ 7 ^ ^ ^ 
Women’s Empowerment in Agriculture Indexd
All women 0.702 0.69-0.72 3265 0.750 0.75-0.76 3047 0.052 0.000 *** 
Woman’s age 
18-29 years 0.705 0.68-0.71 1194 0.720 0.72-0.73 612 0.015 0.000 ***
30 years or older 0.690 0.69-0.72 2071 0.760 0.76-0.77 2435 0.066 0.000 ***
Prevalence of moderate and severe hunger 
All households 4.9 3.8-6.2 4,224 3.3 2.6-4.2 3,888 -1.6 0.014 * 
Gendered household type 
Male and female adults 3.7 2.8-5.0 3,235 2.5 1.9-3.3 3,105 -1.2 * * 
Female adults only 9.2 7.0-12.1 776 6.8 5.0-9.3 661 -2.4 n/s n/s
Male adults only 7.0 4.0-11.8 209 6.6 3.1-13.4 116 -0.4 n/s n/s
Children only ^ ^ 4 ^ ^ 6 ^ ^ ^ 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 14
Women’s dietary diversity: Mean number of food groups consumed by women of reproductive agee
All women 15-49 years 
of age 1.59 1.5-1.7 3,978 2.32 2.2-2.4 3,956 0.8 0.000 ***
Prevalence of exclusive breastfeeding among children under 6 months of agee
All children 72.2 65.2-78.3 204 74.5 63.5-83.0 111 2,3 0.709 n/s
Child’s sex
Male 71.5 61.5-79.7 96 73.9 58.9-84.8 57 2.4 0.778 n/s
Female 72.9 62.2-81.4 108 75.1 62.0-84.8 54 2.2 0.769 n/s
Prevalence of children 6-23 months of age receiving a minimum acceptable diete
All children 3.4 2.0 - 5.7 633 8.4 6.4-11.0 639 5.0 0.003 **
Child’s sex 
Male 3.5 1.6 -7.6 328 7.5 5.0-11.1 324 4.0 0.081 n/s
Female 3.3 1.7 -6.4 305 9.3 6.4-13.1 315 6.0 0.007 **
Prevalence of underweight women of reproductive agee
All non-pregnant women 
15-49 years of age 26.0 23.8-28.3 3,517 21.9 19.9-24.0 3,604 -4.1 0.002 **
Prevalence of stunted children under 5 years of agee
All children 50.3 47.2-53.5 1,553 38.9 36.1-41.8 1,963 -11.4 0.000 ***
Child’s sex
Male 52.9 48.7-57.1 778 38.8 34.9-42.8 976 -14.1 0.000 ***
Female 47.7 43.6-51.8 775 39 35.8-42.3 987 -8.7 0.001 **
Prevalence of wasted children under 5 years of agee
All children 14.1 11.9-16.6 1,471 9.7 7.8-12.0 1,916 -4.4 0.010 **
Child’s sex
Male 14.3 11.8-17.3 733 11.0 8.7-13.9 950 -3.3 0.101 n/s
Female 13.9 11.1-17.2 738 8.4 6.3-11.2 966 -5.5 0.007 **
Prevalence of underweight children under 5 years of agee
All children 31.5 28.6-34.6 2,102 22.5 20.5-24.6 2,153 -9 0.000 ***
Child’s sex
Male 32.2 29.1-35.3 1,050 23.9 21.2-26.7 1,071 -8.3 0.000 ***
Female 30.9 26.9-32.3 1,052 21.1 18.7-23.7 1,082 -9.8 0.000 ***
Est.=estimate, n/a=not available, CI=confidence interval
^ Results not statistically reliable, n<30
a Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample size 
reflects this loss in observations; therefore, disaggregate sample sizes may not total to the aggregated sample size. b Significance tests were performed to determine whether a difference exists between the baseline and end-line estimates at the overall 
indicator level. . 
c Differences found to be statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001.
d The full Women’s Empowerment in Agriculture Index (WEAI) was calculated at baseline and end-line. e Estimates are based on de facto household members. 
Notes: 
Estimates are sample-weighted; numbers of observations are unweighted. 
Estimates are based on de jure household members, except where noted.
Sources: Feed the Future Ethiopia ZOI Survey 2013; Feed the Future Ethiopia ZOI Survey 2018.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 15
1. Background 
This chapter provides background information on Feed the Future in Ethiopia, including a 
description of the program and the Feed the Future phase one Zone of Influence (P1-ZOI), 
demographic information on the P1-ZOI population and a summary of the agriculture situation 
in the P1-ZOI.
1.1 Feed the Future Overview
Feed the Future is a major United States Government program that aims to address the root 
causes of global hunger by sustainably increasing agricultural productivity to meet the demand for 
food, supporting and facilitating access to markets, and increasing incomes for the rural poor so 
they can meet their food and other needs, including reducing malnutrition. Ethiopia has been 
designated a priority country for the Feed the Future Initiative. It is within this context that 
USAID Mission Ethiopia has developed an approved Feed the Future strategy, which is being 
implemented by a host of implementing partners.
The Feed the Future program requires each USAID Feed the Future Mission to focus and 
concentrate its efforts in a defined area of coverage – the Zone of Influence (ZOI) – in order to 
measure impact. The ZOI for USAID Mission Ethiopia comprises the 149 woredas where the 
Feed the Future projects are implemented over the 5 years period of 2013-2017 (see Figure 1.1 
below). As part of its Feed the Future strategy, USAID has made a strong commitment to timely 
and high-quality assessments aimed at providing information and analysis that prevents mistakes 
from being repeated; and increases the possibility that future investments will yield even more 
benefits than previous investments. As part of this commitment, each USAID Mission is required 
to conduct a Population Based Survey (PBS) across its ZOI. 
1.2 Feed the Future P1-ZOI Profile 
The geographic focus of the Feed the Future Ethiopia ZOI Survey 2018 data collection for the 
phase one end-line analysis is the P1-ZOI—the geographic area where the Feed the Future 
program is expected to have an impact on hunger, poverty and nutrition during phase one of the 
program. The P1-ZOI geographic area has not changed since the baseline survey in 2013. 
However, the Feed the Future Ethiopia ZOI Survey 2018 could not be fielded in the five woredas 
of Somali region due to the instability in the region during the period leading up to and including 
the end-line survey. 
The ZOI in Ethiopia consists only rural households in five regions: Amhara, Oromia, Tigray, 
Somali and Southern Nations, Nationalities, and People’s region. The total population in the Feed 
the Future ZOI is 17.5 million individuals, residing in 4.1 million households in 2018 (not including 
the Somali sub-population due to the difficulty of conducting the PBS in that region during the 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 16
end-line). In terms of the agro-ecological zones, Feed the Future in Ethiopia operates in “drought 
prone,” “moisture reliable” and “pastoral” areas (Figure 1.1). 
A map of the Feed the Future P1-ZOI in Ethiopia is provided in Figure 1.1. The five program 
areas identified on the map together constitute the P1-ZOI.
Figure 1.1. Map of Ethiopia: Feed the Future P1-ZOI 
Note: AGP = Agricultural Growth Program, GRAD = Graduation with Resilience to Achieve 
Sustainable Development, PRIME = Pastoralist Areas Resilience Improvement through Market 
Expansion 
1.2.1 Rationale for P1-ZOI Selection
To define the ZOI, Feed the Future Ethiopia selected woredas that are conducive for 
agricultural growth (e.g. AGP activities), where investments create a “pull” factor, or those 
characterized by high levels of chronic food insecurity and/or pastoralist areas, where market 
components create a “push” factor.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 17
1.2.2 Demography of the P1-ZOI
Tables 1.1 and 1.2 present individual and household population estimates, respectively, for the 
P1-ZOI in 2018 based on the panel PBS conducted in 2018 by International Food Policy Research 
Institute (IFPRI) in conjunction with BST Consulting PLC., with some support from the Central 
Statistics Agency. Estimates of the total population as well as sub-populations of the ZOI are 
presented. The sub-population categories correspond to the various sub-populations for the Feed 
the Future indicators and disaggregates. The ZOI estimates for the total population of individuals 
as well as households are also disaggregated by gendered household type.1
 
First, a brief note on how the 2018 population estimates and disaggregates were obtained. The 
baseline population and sub-populations were estimated using the sample survey and the sample 
weights derived from the population census and the complete listing of households in each 
surveyed enumeration area (EA). In short, the sample levels were inflated by the inverse of the 
selection probabilities for each EA as inflation factors (or weights). Given a panel sample, the 
same procedure cannot be followed for the end-line for three reasons – attrition, demographic 
changes in sample households (marriages, divorces, age profile, etc.), and population growth. 
These changes make the sample less representative of the population given the passage of 5 years 
between the baseline and the end-line. It is thus necessary to use a different procedure to 
estimate the ZOI’s population in 2018.
Recall that population growth by gendered household types or other demographic categories of 
Tables 1.1-1.2 are not available from the Census. The simplest and most feasible approach is thus 
followed. The common national population growth rates for the 2014-2018 period provide the 
basic ingredient for the estimation. The growth rates were computed from CSA’s medium variant 
of projections (2008-2037) based on its Intercensal Population Survey of 2012 (2.46 percent, 2.41 
percent, 2.36 percent, 2.31 percent, and 2.28 percent per year, respectively, for 2014-2018). 
These growth rates were used to go from the 2013 population to the 2018 population with the 
assumption that the average household size remains the same during the period. The same 
procedure is used for estimation of the population of households. The resulting estimates are 
reported in the Tables 1.1 and 1.2 below.
In this regard, it is important to keep in mind the limits to the representativeness of the sample 
noted above. This is particularly critical in making direct comparisons across baseline and end￾line using data in Tables1.1 and 1.2. 
In 2018, the total ZOI population (excluding the Somali Feed the Future woredas) is estimated 
to be nearly 17.5 million individuals in 4.1 million households, all residing in rural areas of Ethiopia. 
More than 88 percent of the people in ZOI reside in households that contain both male and 
1
 See Section 2.2.1 Standard Disaggregates for the definition of gendered household type.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 18
female adults. Moreover, the ZOI consists about 1.7 million children who are less than 5 years 
of age and nearly 3.2 million women of reproductive age.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 19
Table 1.1: Population of Individuals in the PI-ZOI, by Category, Ethiopia 2013-2018 
Category of individuals Baseline (2013) End-line (2018)
Na Percent Na Percent
Total number of individuals 15,563,872 100 17,489,900 100
Total number of individuals, by key sub-population 
Children 0-5 months 162,689 1.0 86,651 0.5 
Children 0-23 months 672,018 4.3 588,691 3.4
Children 6-23 months 509,329 3.3 502,040 2.9
Children 0-59 months 1,993,095 12.8 1,714,885 9.8
Youth 15-29 years 3,990,272 25.6 3,964,116 22.7
Women of reproductive age 
 
3,285,838 21.1 3,151,115 18.0
Primary adult female 
 
2,679,107 17.2 2,518,136 14.4
Total number of individuals, by residenceb 
Urban n/a n/a n/a n/a
Rural 15,563,872 100 17,489,900 100
Total number of individuals, by gendered household type
Male and female adults 13,198,163 84.8 15,496,051 88.6
Female adults only 1,914,356 12.3 1,703,516 9.7
Male adults only 435,788 2.8 265,846 1.5
Children only ^ ^ ^ ^ 
Children 0-5 months, by sex 
Male 76,045 0.5 43,492 0.2 
Female 86,644 0.6 43,159 0.2 
Children 0-23 months, by sex 
Male 336,971 2.2 299,045 1.7 
Female 335,047 2.2 289,646 1.7 
Children 6-23 months, by sex 
Male 260,926 1.7 255,553 1.5 
Female 248,403 1.6 246,487 1.4 
Children 0-59 months, by sex 
Male 982,706 6.3 859,163 4.9
Female 1,010,390 6.5 855,722 4.9
Youth 15-29 years, by sex 
Male 1,988,576 12.8 2,127,707 12.2
Female 2,001,695 12.9 1,836,407 10.5
Women of reproductive age, by pregnancy statusc 
Pregnant 306,339 2.0 267,634 1.5
Non-pregnant 2,809,145 18.0 2,828,049 16.2
Primary adult female decision-makers, by age 
18-29 years 965,494 6.2 502,276 2.9
30 years or older 1,722,013 11.1 2,032,593 11.6
Source: Feed the Future Base line Survey of Ethiopia, 2013; Feed the Future End-line Survey of Ethiopia, 2018.
^ Results not statistically reliable, n<30. a Number of individuals in the population
b The urban/rural disaggregate uses the Ethiopia-specific definition of urban and rural reflected in the sampling frame at the time the sample was 
drawn.
C The number of pregnant and non-pregnant women of reproductive age do not sum up to the total number because 0.1 percent did not know 
their pregnancy status and 1 percent did not want to reply in 2013. The respective numbers were 0.1 and 0.3 percent in 2018. 
Note: Population growth rates from CSA’s intercensal projections.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 20
Table 1.2. Number of households, by category, in the P1-ZOI, Ethiopia 2018 
Category of households End-line 
(2018)
 Percenta N 
Total number of households 100 4,099,092
Male and female adults 79.8 3,272,715
Female adults only 17.1 702,584
Male adults only 2.8 116,414
Children only ^ ^ 
Urban - - 
Rural 100 4,099,092
 
Tigray 6.2 253,734
Amhara 32.9 1,346,552
Oromiya 40.4 1,654,394
SNNP 20.6 844,413
Source: Feed the Future End-line Survey of Ethiopia, 2018. 
^ Results not statistically reliable, n<30. Note: Regional population growth rates were from CSA’s intercensal projections.
1.2.3 Agriculture in the P1-ZOI
The ZOI in Ethiopia is large and agro-climatic conditions vary enormously. The four main regions, 
Amhara, Oromia, Tigray and SNNP, are mostly located in the highlands (with relatively large 
fractions of Oromiya and SNNP in the lowlands). In contrast, the Somali region is in the lowlands. 
Agriculture in the ZOI is largely rainfed. While most areas rely on only one rainy season (meher), 
some areas in the country cultivate in two rainy seasons (meher and belg). 
An average household in the ZOI cultivated about a hectare of land. This average masks the
considerable heterogeneity in landholding sizes across the ZOI. About 96 percent of the 
households in the ZOI practice crop-agriculture. Crop-agriculture in Ethiopia is heavily focused 
on cereal production with the five most important cereals being teff, barley, wheat, maize and 
sorghum. Coffee is the most important non-teff cash crop. The southern part of the country also 
cultivates enset that forms an important part of the diet in that area. 
About 88 percent of all households in the ZOI own livestock. The average household owns about 
4.7 Tropical Livestock Units (TLU). The major risks and constraints that households face in 
livestock production are water shortages, livestock diseases and lack of grazing land. 
With regard to marketing of crop output, livestock and livestock products in the Feed the Future
ZOI, more than half of the sample households in 2018 sold part of their produced crop output 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 21
with notable differences across various crop categories. The percentage of households who sold 
livestock and livestock products is around 34 percent and 23 percent, respectively. 
1.3 Purpose of this Assessment 
The purpose of this assessment is to provide the United States Government interagency partners, 
USAID/BFS, USAID Missions, host country government and development partners with 
information to monitor long-term progress of the Feed the Future phase one population-based 
ZOI indicators. The assessment is designed for use as a monitoring tool, and as such provides 
indicator point estimates that were generated to have an acceptable level of statistical precision. 
The assessment is designed to also measure changes in indicator estimates and select 
demographic and household characteristics from the Feed the Future baseline ZOI Survey, which 
was conducted in 2013. Feed the Future ZOI Survey sample calculations, however, are not 
designed to support conclusions of causality or program attribution.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 22
2. Methodologies for Obtaining 2018 Values for 
Feed the Future Indicators
This chapter describes the methodology used to obtain the Feed the Future population-based 
ZOI indicators. It provides information on the data sources and describes measures and reporting 
conventions used throughout the report.
2.1 Data Sources
Data for all the ZOI indicators presented in this assessment are drawn from two sources: Feed 
the Future baseline and end-line household surveys. As noted elsewhere in the report, these 
surveys took place June/July 2013 and September/October 2018, respectively. 
2.1.1 Primary Data: The Feed the Future Ethiopia ZOI Survey 2018
This section describes the Feed the Future Ethiopia ZOI Survey 2018, including discussion of the 
sample design and sample size, questionnaire customization, fieldwork, response rates and 
limitations of the survey. Appendix 2.1 provides additional details on the sampling and weighting 
methodology.
Survey Sample Design 
The Feed the Future Ethiopia ZOI Survey 2018 included a panel of households that were initially 
selected in the baseline survey of 2013. In other words, the ZOI end-line survey was a panel 
survey that covered the same households selected at baseline (in 2013) via a multi-stage cluster 
sampling design.2 Though it started in 2013 as a representative, random sample of the entire 
population living in the P1-ZOI, attrition and demographic changes since then made it less 
representative. Nevertheless, the design ensured that the total sample size included the necessary 
number of households in the P1-ZOI to assess changes in phase one indicators from the 2013 
ZOI Survey to the 2018 ZOI Survey. 
At baseline, woredas (or districts) were randomly selected in the first sampling stage from among 
woredas in the ZOI using proportions derived from population size and Feed the Future-funded 
project coverage. More specifically, the woreda composition of the sample was made to reflect 
the distribution of Feed the Future-supported projects by using the percentage of Feed the 
Future’s ZOI (or woredas) each major Feed the Future project covered as a basis of its share in 
the sample. Three enumeration areas (EAs) were then selected from each woreda using 
probability proportional to size (PPS) methodology. Finally, households were randomly selected 
in each EA from a complete list of households prepared specifically for the survey through a 
2
 The FTF baseline survey report (Bachewe et al., 2014) describes the sample design and selection in detail. Additional 
information is also available in section A2.1 below.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 23
complete household listing exercise. The Feed the Future baseline survey report (Bachewe et al., 
2014) discusses the considerations and assumptions in the determination of sample size for the 
Feed the Future baseline.
The Feed the Future Ethiopia ZOI end-line survey conducted in 2018 is a panel survey that 
followed the same sample households covered by the Feed the Future baseline survey in 2013. 
Upon USAID’s decision that the end-line survey should be restricted to households residing in 
the Feed the Future ZOI woredas, IFPRI prepared to conduct the Feed the Future end-line survey 
in those 56 woredas selected at the baseline. However, due to considerable insecurity in the 
Somali region during the period leading up to the survey, the five woredas in that region could 
not be covered by the end-line survey. Consequently, the survey covered 51 of the 56 Feed the 
Future ZI woredas. The issue this raises is discussed in section A2.1 in the annex. 
Questionnaire Design
Two questionnaires were developed for the Feed the Future end-line survey. These are the 
household questionnaire and the community questionnaire.3
 
The core of the household survey instrument consists of the modules found in the standardized 
survey instrument developed by the Analytics and Learning Division in USAID’s Bureau of 
Resilience and Food Security. These were modified to suit the circumstances in Ethiopia. 
Additional modules were also included for two primary reasons. First, it is necessary to obtain 
detailed information on the livelihood characteristics and options of the households and 
communities in order to capture the context within which Feed the Future investments occur. 
Second, such data are also required towards answering some of the questions in Feed the Future’s 
Learning Agenda.4 The household questionnaire’s modules are:
a. Household identification cover sheet 
b. Household roster and demographics
c. Dwelling characteristics
d. Household consumption expenditure
e. Household hunger scale
f. Women’s empowerment in agriculture index
g. Women’s anthropometry and dietary diversity
h. Child anthropometry and infant and young child feeding
3 A copy of the household questionnaire is available as part of the Feed the Future Ethiopia Zone of Influence End￾line Survey data asset in USAID’s Development Data Library at https://data.usaid.gov/. 4 Consequently, data collected in some of the additional modules of the household questionnaire as well as data 
collected via the community survey are not used in this report, which is focused on reporting on Feed the Future’s 
standard indicators.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 24
i. Agricultural productivity and input use
j. Crop utilization
k. Agricultural extension, technology and information networks
l. Livestock ownership and income from livestock and livestock products
m. Shocks and resilience 
n. Off-farm income and business activities – own business activities
o. Trust, control and agency
p. Household assets 
q. Transfers/gifts 
The community questionnaire provides information on community- or Kebele-level resources 
that will affect take-up of Feed the Future interventions. Modules in the community 
questionnaire include:
a. Site identification 
b. Location and access
c. Water and electricity
d. Services (general)
e. Education and health services 
f. Production, marketing and extension
g. Local wages and input costs
h. Current and last year food prices 
i. Public works 
j. Migration
k. Perceived changes in production and input uses
l. Government of Ethiopia and/or Feed the Future programs/projects operating in the 
locality or Kebele (e.g., the PSNP, AGP)
The IFPRI team in collaboration with the BST taskforce worked on the preparation of survey 
instruments based on the generic Feed the Future household questionnaire. Before the actual 
field work, IFPRI research staff, 56 supervisors from BST and CSA trainers commented on the 
household questionnaire. 
In parallel, the Computer-aided personal interviewing (CAPI) version was developed as an CSPro 
application or program. The program was put through a series of rigorous tests and modified as 
necessary. This process continued until the end of the enumerators’ training process.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 25
Fieldwork 
i. Training
Prior to fieldwork, all field staff were trained on survey procedures, including preparing for 
fieldwork, questionnaire content, human subjects protection, fieldwork procedures, data 
management, reporting and communications. Training included hands-on training and practice 
sessions that covered use of all technical equipment that was used in the survey, as well as a 
tablet pretest and a survey pilot test. 
Training of trainers (all CSA staff participating on their individual capacity) and BST supervisors 
was conducted during Aug. 8-10, 2018, at the International Livestock Research Institute 
compound in Addis Ababa. The training was conducted by a team composed of IFPRI and CSA 
staff (again participating in their individual capacity). This training of trainers (ToT) helped to 
refine the survey instrument further.
The main training (or training of enumerators) was conducted at the Ethiopian Civil Service 
University during Aug. 20-Sept. 4, 2018. The enumerators selected for fieldwork at the relevant 
CSA branches traveled to Addis Ababa during Aug. 17-19, 2018. The training involved going 
through the substantive aspects of the questionnaires module by module and was based on the 
CAPI versions of the questionnaires. 
Sessions dedicated to anthropometric data collection and data transfer protocols have been part 
of the training program. This was particularly true of the household questionnaire, which was 
implemented in digital form. A mock interview at the end of the training helped in reinforcing 
what was learned during training as well as finding any remaining bugs in the program. All parts 
of the training were implemented successfully as planned. 
The training was completed in 15 days as planned (with a 1-day break for the Arafa Holiday). 
ii. Data collection
Data were collected using the two questionnaires described earlier – household and community 
(with price modules) questionnaires. As noted above, BST had the responsibilty for fielding the 
survey with IFPRI’s leadership. The IFPRI team also traveled to sample sites to assess 
implementation and help solve unanticipated problems in the field. IFPRI’s supervision of survey 
implementation proved crucial as in previous surveys, particularly for data saving and transfer. 
A major task embedded with data collection was data transfer. The digital household data 
collected using CAPI questionnaire had to be regularly transferred during the data collection 
period to IFPRI. Three objectives were to be achieved by doing so: to detect and correct 
collection errors as quickly as possible; to reduce the likelihood of data loss; and to maintain the 
integrity of the collected data. A purposely designed transfer protocol was adopted. For the most 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 26
part data transfer from the field was conducted as planned. However, poor internet connection 
and remoteness of survey areas resulted in cases in which the data were transferred irregularly. 
The planned length of the data collection was 25 days during Sept. 7-Oct. 2, 2018 (with a 1-day 
break for the Ethiopian New Year). This was achieved in many areas. However, two additional 
days were required in some woredas due to longer travel time and the use of paper questionnaire. 
Consequently, the data collection was completed on Oct. 4, 2018. 
As planned, all interviewed households were administered the CAPI version of the questionnaire 
with the exception of Chilga woreda, Amhara region. A number of households in two EAs of that 
woreda were interviewed using the paper version of the questionnaire. The CAPI version of the 
questionnaire could not be administered because enumerators were unable to recharge their 
netbook batteries due to lack of access to electricity. The enumerators latter transferred the 
paper-based data collected from the households onto the CAPI. 
The household data collected using a CAPI-based questionnaire was delivered by BST for IFPRI 
in the two weeks after the completion of the survey. In contrast, BST completed the entry and
basic cleaning of the community data and transferred them to IFPRI on Nov. 23, 2018.
ZOI Survey Response Rates for 2013 and 2018
Table 2.2 presents the response rates for the Feed the Future Ethiopia Survey 2018 in the P1-
ZOI. The components and the response rates for the sampled households, women of 
reproductive age (15-49 years), primary adult male and female decision-makers, and children 
under 5 years are also presented. The numbers indicate an attrition rate of 8.5 percent or only 
8.5 percent of the households selected in 2013 were not surveyed in 2018.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 27
Table 2.2. Comparison of results of household and individual interviews in the PI-ZOI, 
in total, Feed the Future phase one baseline and end-line ZOI Surveys
Response rate Baseline (2013) End-line
(2018)
Households 
Number of households selected 4,704 4,251
Number of households occupied 4,635 3,908
Number of households interviewed 4,245 3,890
Household response rate (%)a 91.5 99.5
Women of reproductive age (15-49 years) 
Number of eligible women 4,172 4,011
Number of eligible women interviewed 4,172 4,011
Eligible women response rate (%)b 100 100
Primary adult female decision-makers (18+ years) 
Number of eligible women 3,466 3,252
Number of eligible women interviewed 3,466 3,252
Eligible women response rate (%)b 100 100
Primary adult male decision-makers (18+ years) 
Number of eligible men 2,998 2,527
Number of eligible men interviewed 2,998 2,527
Eligible men response rate (%)b 100 100
Children under 5 years of age 
Number of eligible children 2,571 2,209
Number of caregivers of eligible children interviewed 1,956 1,673
Eligible children response rate (%)b 100 100
Children under 2 years of age 
Number of eligible children 869 759
Number of caregivers of eligible children interviewed 847 743
Eligible children response rate (%)b 100 100 a Household response rates are calculated based on the result codes of Module 1, the household roster, and are defined as the number of 
households interviewed divided by the number of households occupied. Households that were found to be vacant, not a dwelling unit, or 
destroyed were considered unoccupied and thus excluded from the response rates.
b Individual response rates are calculated based on the result codes in the relevant individual modules (Modules 4, 5, and 6). These rates are 
defined as the number of eligible individuals interviewed divided by the number of eligible individuals. Eligibility determination for Modules 4, 5, 
and 6 is initiated in the household roster and confirmed in the respective module. (Note that for children under 5 years of age [Module 5], the 
primary caregivers of the children served as the respondents, not the children directly.)
Sources: Feed the Future Ethiopia ZOI Survey 2013; Feed the Future Ethiopia ZOI Survey 2018
2.1.2 Comparability of Data Sources Used for the Feed the Future 
Phase One Baseline and End-line ZOI Survey Assessment
This section discusses the comparability across data sources for the Feed the Future PI-ZOI 
indicators, and comparability of data sources used for the baseline and end-line surveys. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 28
Seasonality 
The Feed the Future Ethiopia ZOI end-line survey was conducted during Sept. 7-Oct. 4, 2018, 
while the baseline survey was conducted during June 14-July 12, 2013. To ascertain the potential 
impact of seasonality, it is essential to compare access to food during June/July with September 
(during which the bulk of the end-line survey took place). The three months fall under the “hunger 
season,” particularly in crop-producing highlands.5 Also important is the fact that most of the ZOI
as well as the sample are located in these highland areas. Outcome variables across the two 
surveys are thus reasonably comparable. 
Table 2.3: Season in which data collection took place for Feed the Future phase one 
baseline and end-line, by indicator
Indicator
Season of data collection
Baseline 
(2013)
End-line
(2018)
Daily per capita expenditures (2010 USD) June – July Sept. – Oct.
Prevalence of poverty: Percent of people living on less than $1.25/day (2005 PPP) June – July Sept. – Oct.
Depth of poverty: Mean percentage shortfall of the poor relative to the $1.25/day 
(2005 PPP) poverty line June – July Sept. – Oct.
Abbreviated Women’s Empowerment in Agriculture Index June – July Sept. – Oct.
Prevalence of moderate and severe hunger June – July Sept. – Oct.
Prevalence of exclusive breastfeeding of children under 6 months of age June – July Sept. – Oct.
Percent of children 6-23 months of age receiving a minimum acceptable diet June – July Sept. – Oct.
Prevalence of women of reproductive age consuming a diet of minimum diversity June – July Sept. – Oct.
Prevalence of stunted children under 5 years of age June – July Sept. – Oct.
Prevalence of wasted children under 5 years of age June – July Sept. – Oct.
Prevalence of underweight children under 5 years of age June – July Sept. – Oct.
Prevalence of underweight women of reproductive age June – July Sept. – Oct.
PPP=purchasing power parity
Sources: Feed the Future Ethiopia ZOI Baseline Survey 2013; Feed the Future Ethiopia ZOI End-line Survey 2018
Other Issues Regarding Comparability
As indicated above, IFPRI prepared to conduct the Feed the Future end-line survey in the 56 
woredas in the Feed the Future ZOI. However, due to considerable insecurity in the Somali 
region during the period leading up to the survey, IFPRI, with the advice of the CSA, decided to 
exclude the five woredas in that region from the end-line survey. Consequently, the Feed the 
Future end-line survey covered 51 of the 56 Feed the Future ZOI woredas and therefore the 
Feed the Future end-line survey dataset excludes the Somali region. Consequently, all outcome 
variables provided in Sections 3 through 7 exclude the Somali region for both baseline (2013) and 
end-line (2018) periods. 
5 See for instance: LIU (2010). An Atlas of Ethiopian Livelihoods. Livelihoods Integration Unit (LIU), Disaster Risk 
Management and Food Security Sector, Ministry of Agriculture, Addis Ababa, Ethiopia. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 29
2.2 Measures and Reporting Conventions Used 
Throughout This Report
2.2.1 Standard Disaggregates
A standard set of disaggregate variables is used in tables throughout this report. This section lists 
each of the standard disaggregate variables and defines how each variable is calculated.
Age in Years
Data on respondent’s age in years is collected in the household roster. For women ages 15-49 
and children under 6 years, more detailed age data are collected in subsequent questionnaire 
modules to confirm eligibility to respond to the module questions; these more detailed age data 
are used where available. Age is generally presented in the tables in 5- or 10-year age groups.
Age in Months
The age of children in months is collected in the child nutrition survey module, rather than in the 
household roster, so that the child’s parent or primary caregiver can be prompted to provide 
the most accurate age possible. Children’s age in months is presented by age categories for the 
children’s dietary intake and anthropometry tables. For example, for the minimum acceptable 
diet table (Table 5.7), presents children’s age disaggregated into 6-month age groups as follows: 
6-11 months, 12-17 months and 18-23 months. For the children’s anthropometry tables (Tables 
6.3, 6.4, and 6.5), which present the prevalence of stunting, wasting and underweight for all 
children under 5 years of age, children’s age is disaggregated into 12-month age groups as follows: 
0-11 months, 12-23 months, 24-35 months, 36-47 months, and 48-59 months.
Sex
Sex—male or female—is a standard disaggregate for the tables presenting children’s indicators, 
e.g., children’s anthropometry (Tables 6.3, 6.4, and 6.5). 
Educational Attainment (Household)
Household educational attainment reflects the highest level of education attained by any member 
of the household, as reported in the household roster. This variable is used in tables that present 
household-level data, and comprises four categories: no education (households with no member 
who has received formal education); less than primary (households with at least one member 
who has received formal schooling, but with no member who has completed primary or grade 
8); primary (households with at least one member who has completed primary, but with no 
member who has completed secondary); and secondary or more (households with at least one 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 30
member who has completed secondary education). Households are categorized in only one of 
the four categories.
Educational Attainment (Individual)
Educational attainment at the individual level reflects the highest level of education attained by 
individual household members, as reported in the household roster. This variable comprises four 
categories: no education (those who have not received any formal education), less than primary 
(those who have received formal education but who have not complete primary); primary (those 
who have completed primary but have not completed secondary); and secondary or more (those 
who have completed secondary education).
Gendered Household Type 
The USAID’s Feed the Future M&E Guidance Series Volume 6: Measuring the Gender Impact of 
Feed the Future notes that household-level indicators should be disaggregated by “gendered 
household types” – that is: 1) households in which members include both male and female adults
(adult is defined as age 18 or older); 2) households in which members include male adult(s), but
no female adults; 3) households in which members include female adult(s), but no male adults; 
and 4) households in which all members are under age 18 (children), i.e., households with children 
only and no adult members. This approach to conceptualizing household type is distinct from the 
standard “head of household” approach, which is embedded with presumptions about household 
gender dynamics and may perpetuate existing social inequalities and prioritization of household 
responsibilities that may be detrimental to women (USAID 2014:1).6 This variable is calculated 
using data on age and sex collected in the household roster of the survey questionnaire. 
Household Hunger
As described in greater detail in Section 5.1 of this report, the household hunger scale (HHS) 
characterizes households according to three categories of hunger severity: little to no household 
hunger, moderate household hunger, and severe household hunger. For the purposes of serving 
as a disaggregate in selected tables, the HHS is converted to a dichotomous measure reflecting
households that report little to no household hunger, and households that report moderate or 
severe household hunger. 
Household Size 
For the ZOI Survey, household size is defined as the total number of people (adult and children) 
that live together in the same dwelling unit. They can be related or unrelated, but they 1) 
6 United States Agency for International Development (USAID). (2014). Feed the Future M&E Guidance Series. 
Volume 6: Measuring the Gender Impact of FTF, March. Accessed 27 March 2015 at 
http://www.feedthefuture.gov/resource/volume-6-feed-future-measuring-gender-impact-guidance.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 31
acknowledge the same person or persons as lead decision-makers for the household; 2) share 
the same housekeeping and cooking arrangements; 3) share the same contiguous roof; and 4) 
who stayed in the household the night before the survey (de facto) as opposed to the usual 
household residents (de jure). This ordinal household size variable is recoded into a categorical 
variable as follows: small households (1-5 members), medium households (6-10 members), and 
large households (11 or more members). Note that the definition of household members used 
in the ZOI Survey may be slightly different from that used in other household surveys.
2.2.2 Reporting Conventions
The Feed the Future Ethiopia Zone of Influence Survey 2018-2019 End-line Assessment is 
primarily descriptive in nature. This section provides an overview of the conventions used in 
reporting these descriptive results.
• In the tables throughout this report, weighted point estimates and unweighted sample sizes 
(denoted by “n”) are presented. 
• Most estimates are shown to one decimal place, with the specific exceptions of per capita 
expenditures, women’s dietary diversity, WEAI, five domains of empowerment (5DE), and 
gender parity index (GPI) indicators, which are shown to two decimal places. Unweighted 
sample sizes in all tables and the population estimates in Tables 1.1 and 1.2 are shown as 
whole numbers. Values in the tables are suppressed when the unweighted sample size is 
insufficient to calculate a reliable point estimate (n<30); this is denoted by the use of the 
symbol “^” in the designated row and an explanatory footnote.
Bivariate relationships are described using cross tabulation, and the strength and direction of the 
relationships are assessed using statistical tests. For continuous variables, a t-test, or test of 
differences in mean, is used. For categorical variables, Pearson’s chi-squared test, or test of 
differences in proportions, is used. Explanatory table footnotes clarify the meaning of the 
significance test annotation, and statistically significant results are highlighted in the narrative 
throughout the report. Analyses are performed in Stata using “svy” commands to handle features 
of data collected through the use of complex survey designs, including sampling weights, cluster 
sampling and stratification. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 32
3. ZOI Survey 2018 P1-ZOI Population
This chapter describes the background characteristics of the P1-ZOI population using data from 
the Feed the Future Ethiopia ZOI Survey 2018 and documents any changes in demographic and 
household characteristics that occurred since the baseline survey in 2013. The demographic 
characteristics of the P1-ZOI population are expected to remain relatively similar over time. 
3.1 Demographics
Table 3.1a presents demographic characteristics of the households in the P1-ZOI for all 
households and by gendered household type. This table presents the average household size, as 
well as the average number of female adults and children within the household. Estimates are also 
presented by household education, defined as the highest level of education of any member of 
the household. An average household in the ZOI has five members. All household members of 
about 17 percent of the households had no education, while at least one member of 0.7 percent 
households has completed secondary school. A large majority of 56 percent of the households 
have at least one member with less than primary education while 26 percent have at least one 
member with primary education. 
Table 3.1a Household demographic characteristics in the P1-ZOI 
Characteristic
Total
(All 
households)
By gendered household type
Male and 
female 
adult
Female 
adult(s) 
only
Male 
adult(s) 
only
Child 
only
Mean household size 5.1 5.6 2.9 2.5 ^
Mean number of adult female household members1,2 1.2 1.2 1.2 n/a ^ 
Mean number of children (<2 years)1 0.3 0.4 0.1 0.1 ^ 
Mean number of children (0-4 years)1 0.6 0.7 0.3 0.2 ^ 
Mean number of children (5-17 years)1 2.1 2.3 1.4 1.0 ^ 
Mean percentage of adults who are female (%)1,2 55.5 48.3 100 n/a ^
Highest education level attained (%) 
No education 17.4 11.9 38.4 39.7 ^ 
Less than primary 56.0 57.1 49.9 44.0 ^ 
Primary 26.0 30.1 11.6 16.4 ^ 
Secondary or more 0.7 0.8 0.0 0.0 ^
n3 3,890 3,106 661 116 7 
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia.
^ Results not statistically reliable, n<30. 1
 The estimate is based on household members with known age. 2
 Feed the Future defines adult as an individual age 18 or older. Females age 15-17 are of reproductive age but are not considered adults by 
this definition. 3
 Sample n is the unweighted count of all households that responded to the survey. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 33
Table 3.1b compares the demographic characteristics of the households in the P1-ZOI in the 
phase one baseline and end-line ZOI Surveys. Estimates are shown for all households and include
the average household size, as well as the average number of female adults, male adults and 
children within the household. The percentage of households with a male primary adult decision￾maker and the percentage of households with a female primary adult decision-maker are also 
provided. Household education, defined as the highest level of education of any member of the 
household, is also presented in this table. The average household size is slightly higher in 2018 
relative to the average in the baseline (2013). Similarly, the average number of adult females and 
5-17 years old children is higher in the end-line relative to the respective averages in the baseline. 
Relative to the baseline, the end-line has a relatively lower proportion of households with a male 
primary adult decision-makers and a relatively higher proportion of households with a female 
primary adult decision-maker. However, these changes are not statistically significant. 
Table 3.1b Comparison of household demographic characteristics between the Feed 
the Future phase one baseline and end-line ZOI Surveys 
Baseline (2013) End-line (2018)
Diff.
p￾value1 Est. 95% CI Est. 95% CI Sig.
Mean household size 4.7 4.6-4.8 5.1 5.0- 5.2 0.0 0.000 ***
Mean number of adult male household 
members2,3 1.1 1.0-1.1 1.2 1.1-1.2 0.0 0.000 ***
Mean number of adult female household 
members2,3,4 1.1 1.1-1.1 1.2 1.2-1.2 0.0 0.000 ***
Mean number of children (<2 years)2 0.3 0.31-0.34 0.3 0.33-0.4 0.0 0.191 n/s
Mean number of children (0-4 years)2 0.6 0.6-0.7 0.6 0.6-0.7 0.0 0.888 n/s
Mean number of children (5-17 years)2 1.9 1.8-1.9 2.1 2.1-2.1 0.0 0.026 **
Mean percentage of adults who are male (%)2,3 44.3 43.5-45.1 44.5 43.6-45.1 0.2 0.622 n/s
Mean percentage of adults who are female (%)2,3,4 55.7 54.9-56.4 55.5 54.6-56.1 -0.2 0.622 n/s
Mean percentage of households with a male 
primary adult decision-maker (%)5 70.6 69.2-72.0 65.0 63.5-66.5 -5.6 0.000 ***
Mean percentage of households with a female 
primary adult decision-maker (%)5 81.6 80.8-83.2 83.6 81.9-84.3 2.0 0.387 n/s
Highest education level attained (%)
No education 31.1 29.7-32.4 17.4 16.0-18.4 -14.0 0.000 ***
Less than primary 52.3 50.8-53.7 56.0 53.9-57.1 4.0 0.000 ***
Primary 16.2 15.1-17.2 26.0 25.2-27.9 10.0 0.000 ***
Secondary or more 0.4 0.2-0.6 0.7 0.4-0.9 0.0 0.138 n/s
n6 4,245 3,890 
Sources: Feed the Future Baseline Survey 2013, Ethiopia; Feed the Future End-line ZOI Survey 2018, Ethiopia.
^ Results not statistically reliable, n<30. 1 A t-test of statistical significance of differences between the baseline and end-line estimates of means for continuous variables, and a Pearson’s 
chi-squared test of statistical significance of differences between the baseline and end-line estimates of proportions for categorical variables
differences found to be statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001. 2 The estimate is based on household members with known age. 3 Feed the Future defines adult as an individual age 18 or older. 
4 Females age 15-17 are of reproductive age but are not considered adults by this definition. 5 Primary decision-maker is the adult in the household who makes the socioeconomic decisions in the household. There can be both a female 
and male primary decision-maker in the same household though there may only be one or the other – or neither if there are no adults in the 
household. 6 Sample n is the unweighted count of all households that responded to the survey.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 34
Overall, there is improvement in educational status in the end-line relative to the baseline. The 
proportion of households with all-illiterate members declined by 14 percentage points. Similarly, 
the proportion with a member that completed primary education increased by 10 percentage 
points. 
Tables 3.2a and 3.2b present characteristics of the female and male primary adult decision￾makers in the sampled households in the P1-ZOI and changes in their characteristics since the 
baseline survey. The female and male primary adult decision-makers are household members age 
18 or older who self-identify as the primary adult male or primary adult female responsible for 
both social and economic decision making within the household. When both exist within a single 
household, female and male primary adult decision-makers are typically, but not necessarily, 
husband and wife. Table 3.2a and table 3.2b shows the age group, marital status, literacy status, 
educational attainment and participation in economic activity for these household members. 
These characteristics are shown by gender of primary adult decision-makers. 
Out of all primary decision-makers in the PI-ZOI end-line survey, about 56.3 percent are female 
and 43.7 percent male (see Table 2.2). Alternatively, 83.6 percent of sample households have a 
primary adult female decision-maker, while only 65 percent have a primary male adult decision￾maker (see Table 3.1b). The age distributions of female and male decision-makers are almost 
identical with 68 percent both being under 50 years of age. The majority of primary decision￾makers are married, though more female ones are either widowed or separated (26 percent). 
Literacy rates among male decision-makers (37.9 percent) is more than twice as high as that 
among female decision-makers (15.1 percent). Comparable fractions of female (93.8 percent) and 
male (97.5 percent) decision-makers participate in some form of economic activity. The 
proportion of female primary decision-makers that participated in nonfarm and wage labor 
activities (17.2 percent) is considerably lower than that of males (25.9 percent). 
The age profile of primary decision-makers has shifted up essentially because the same households 
are surveyed both in the baseline and end-line. The proportion of primary decision-makers with 
no formal education declined in the end-line relative to the baseline, although the decline was 
smaller among female primary decision-makers. The likely explanations for this decline include 
expanding adult literacy programs. Marital status of primary decision-makers has a similar pattern 
in the baseline and end-line, with little changes across the period. Compared to the baseline, the 
proportion of both female and male primary decision-makers participating in some form of 
economic activity increased in the end-line, with the rise more than twice as high for the former 
(10.1 percentage points) relative to the later (4.7 percentage points). Moreover, the participation 
rate of male primary decision-makers in nonfarm/wage activities grew faster than that 
experienced by their female counterparts, while the opposite is true for farm activities.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 35
Table 3.2a.: Comparison of characteristics of primary adult female decision-makers in 
the PI-ZOI, Feed the Future phase one baseline and end-line ZOI Surveys
Characteristic
Baseline 
 (2013) 
End-line 
(2018) p￾valueb Sig.c % 95% CI na % 95% CI na Diff.
Age 
18-24 15.9 14.7-17.2 553 5.7 4.9-6.5 187 -10.2 0.000 ***
25-29 20.0 18.6-21.3 694 14.2 12.9-15.3 460 -5.9 0.000 ***
30-39 26.0 24.5-27.5 902 30.8 29.2-32.3 1,000 4.8 0.790 n/s
40-49 12.3 11.1-13.3 425 17.3 15.9-18.5 561 5.0 0.000 ***
50-59 15.3 14.1-16.5 530 17.9 16.6-19.2 584 2.6 0.016 * 
60+ 10.4 9.4-11.4 362 14.2 12.9-15.3 460 3.8 0.000 ***
Marital status 
Married 72.2 70.9-73.9 2,510 72.6 70.9-73.9 2,355 0.4 0.020 **
Living in a consensual union 0.7 0.5-1.0 28 0.2 0.03-0.3 6 -0.5 0.879 n/s
Widowed 17.7 16.2-18.7 606 19.1 17.6-20.3 617 1.4 0.654 n/s
Divorced or separated 8.3 7.2-9.1 283 6.9 6.1-7.8 226 -1.4 0.003 **
Never married or in a union 1.1 0.7-1.4 39 1.3 1.0-1.8 48 0.2 0.040 **
Education 
No education 86.4 84.7-87.1 2,967 83.9 81.9-84.5 2,701 -2.5 0.002 ***
Less than primary 11.2 10.6-12.7 403 12.8 12.1-14.8 432 1.6 0.040 * 
Completed primary 2.3 1.8-2.8 80 3.2 2.7-4.0 110 0.9 0.008 **
Completed secondary ^ ^ 3 ^ ^ 2 ^ ^ ^ 
Economic activityd 
Participates in some form of 
economic activity 83.7 82.4-84.8 2,900 93.8 92.6-94.3 3,039 10.1 0.000 ***
Participation in economic activity by typee
Farm 82.2 80.9-83.5 2,851 92.1 91.1-92.9 2,993 9.9 0.000 ***
Nonfarm 9.7 8.7-10.7 338 11.1 9.9-12.1 358 1.4 0.090 * 
Wage/salaried 4.6 3.9-5.3 160 6.1 5.2-6.8 197 1.5 0.000 ***
Number of primary adult 
female decision-makers 3,466 3252
^ Results not statistically reliable, n<30
a Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample size 
reflects this loss in observations; therefore, the sum of disaggregate sample sizes may not equal the overall sample size. b Significance tests were performed to determine whether a difference exists between the baseline and end-line estimates. c Differences found to be statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001.
d Both paid and unpaid types of economic activity are included. Domestic work, such as caring for children and the elderly or cooking and 
cleaning, is not included.
e Farm work includes food crop farming, cash crop farming, livestock raising or fishing/fishpond culture; nonfarm work includes running small 
businesses or self-employment; and wage/salaried employment includes both agriculture and non-agriculture-based work that is salaried. 
Percentages do not add up to 100 percent because individuals can engage in more than one type of economic activity.
Notes: Estimates are based on primary adult female decision-makers who are de jure household members. 
Estimates are sample-weighted; numbers of observations are unweighted. 
Sources: Feed the Future Ethiopia ZOI Survey 2013; Feed the Future Ethiopia ZOI Survey 2018
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 36
Table 3.2b. Comparison of characteristics of primary adult male decision-makers in the 
PI-ZOI, Feed the Future phase one baseline and end-line ZOI Surveys
Characteristic
Baseline 
 (2013) 
End-line 
(2018) p￾valueb Sig.c % 95% CI na % 95% CI na Diff.
Age 
18-24 7.1 6.1-7.9 211 4.0 3.2-4.7 102 -6.2 0.000 ***
25-29 15.7 14.4-16.9 470 6.7 5.7-7.6 169 -7.9 0.000 ***
30-39 35.8 34.0-37.5 1,073 34.9 33.1-36.7 881 3.2 0.016 * 
40-49 16.8 15.4-18.1 503 23.2 21.5-24.8 586 5.5 0.000 ***
50-59 10.9 9.7-12.0 327 13.2 11.9-14.5 334 2.5 0.001 ***
60+ 13.8 12.6-15.0 414 18.0 16.5-19.5 455 3.1 0.000 ***
Marital status 
Married 89.7 88.5-90.7 2,687 90.1 88.5-90 2,266 0.4 0.000 ***
Living in a consensual union ^ ^ 28 ^ ^ 7 ^ ^ ^ 
Widowed 1.9 1.4-2.3 57 1.8 1.4-2.4 49 -0.1 0.213 n/s
Divorced or separated 2.7 2.1-3.2 80 2.7 2.1-3.4 71 0.0 0.543 n/s
Never married or in a 
union 4.9 4.1-5.6 146 5.0 4.4-6.2 134 0.1 0.000 ***
Education 
No education 61.9 57.6-61 2037 58.8 57.5-60.1 1,503 -3.1 0.000 ***
Less than primary 32.8 32-33.9 1081 35.1 33.9-36.4 1,033 2.3 0.000 ***
Completed primary 5.1 5.0-6.4 169 5.8 5.2-6.4 185 0.7 0.007 ***
Completed secondary ^ ^ 3 ^ ^ 5 0.19 0.328 n/s
Economic activityd 
Participates in some form 
of economic activity 92.8 91.4-93.3 2,769 97.5 96.6-97.9 2,460 4.7 0.000 ***
Participation in economic activity by typee
Farm 91.3 90.2-92.3 2,737 96.2 95.5-96.9 2,433 4.9 0.000 ***
Nonfarm 11.2 10.0-12.3 335 13.9 12.5-15.3 352 2.7 0.022 **
Wage/salaried 9.7 8.6-10.7 290 12 10.7-13.3 305 2.3 0.004 ***
Number of primary adult 
male decision-makers 2,998 2,527
^ Results not statistically reliable, n<30
a Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample size 
reflects this loss in observations; therefore, the sum of disaggregate sample sizes may not equal the overall sample size. b Significance tests were performed to determine whether a difference exists between the baseline and end-line estimates. c Differences found to be statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001.
d Both paid and unpaid types of economic activity are included. Domestic work, such as caring for children and the elderly or cooking and 
cleaning, is not included. e Farm work includes food crop farming, cash crop farming, livestock raising or fishing/fishpond culture; nonfarm work includes running small 
businesses or self-employment; and wage/salaried employment includes both agriculture and non-agriculture-based work that is salaried. 
Percentages do not add up to 100 percent because individuals can engage in more than one type of economic activity.
Notes: Estimates are based on primary adult male decision-makers who are de jure household members. 
Estimates are sample-weighted; numbers of observations are unweighted.
Sources: Feed the Future Ethiopia ZOI Survey 2013; Feed the Future Ethiopia ZOI Survey 2018
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 37
3.2 Living Conditions
Table 3.3a shows dwelling characteristics of the households in the P1-ZOI for the Feed the 
Future phase one baseline and end-line surveys. Many of these measures align with the Sustainable 
Development Goals (SDGs) definitions (UNSTATS, n.d.). The table presents the percentage of 
households who have access to an improved water source, improved sanitation, electricity and 
solid cooking fuel. The percentages of households practicing a correct water treatment practice 
or technology and those practicing open defecation are also included. The average number of 
people per sleeping room, as well as roof, exterior wall and floor materials are also presented. 
Estimates are shown for all households. Because changes in these characteristics over time may 
also influence changes in the Feed the Future indicators, the results of the test of statistical 
differences between the baseline and end-line surveys are also provided.
Table 3.3a reveals that 66 percent of the households have access to improved water sources in 
the end-line and this was nearly 10 percentage points higher than the proportion in the baseline. 
Only 10 percent of the households use a correct water treatment practice in the end-line; this is 
a small percentage point change, but it is a statistically significant increase from baseline. Similarly, 
the proportion with access to improved sanitation that isn’t shared with others declined slightly 
(62 percent), while the proportion practicing open defecation (37 percent) increased slightly in 
the end-line; both percentage point changes are small but statistically significant. Virtually all 
households used solid fuel for cooking both in the baseline and end-line. The proportion with 
access to electricity in the end-line (22 percent) was more than 3 folds of the proportion in the 
baseline. There was a 10 percent increase in the average number of people per sleeping room in 
the end-line relative to the baseline. 
Residences of about one-thirds of the households have natural (mostly straw) roofs and the 
proportion is lower in the end-line relative to the baseline. About two-thirds of the houses have 
finished roofs and this proportion increased during the period. Over 95 percent of the 
households live in houses with natural walls and floors, and these proportions did not change 
across the two rounds.
7
 
7 The proportion of houses with exterior walls made of the third category of materials (Rudimentary) could not be 
computed due to differences in coding of building materials in the baseline (2013) and end-line (2018) surveys. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 38
Table 3.3a Comparison of household dwelling characteristics between the Feed the 
Future phase one baseline and end-line ZOI Surveys 
Characteristic Baseline (2013) End-line (2018) 
Diff.
p￾value1 Sig.1 Est. 95% CI Est. 95% CI
Percent with improved water source (%)2 57.1 55.4-58.4 66.2 65.2-68.2 9.1 0.000 ***
Percent using a correct water treatment 
practice or technology (%)3 8.1 7.5-9.2 10.2 9.1-10.9 2.1 0.009 **
Percent with improved sanitation, not 
shared (%)4 63.1 61.3-64.2 61.5 59.5-62.5 -1.6 0.097 * 
Percent with improved sanitation, shared 
(%)4 0.1 0.1-0.1 0.1 0.1-0.1 0.0 0.871 n/s
Percent with unimproved sanitation (%)5 1.0 0.5-1.1 0.9 0.5-1.1 -0.1 0.728 n/s
Percent practicing open defecation (%)6 35.6 34.5-37.4 37.4 36.4-39.5 1.8 0.054 * 
Percent using solid fuel for cooking (%)7 98.8 98.5-99.1 98.9 98.5-99.2 0.1 0.926 n/s
Percent with access to electricity (%) 6.3 5.9-7.4 21.8 20.0-22.6 15.5 0.000 ***
Mean persons per sleeping room8 4.7 4.6- 4.9 5.1 5.1-5.2 0.4 0.000 ***
Household roof materials (%)9
Natural 43.2 42.4-45.4 32.4 31.6-34.6 -10.8 0.000 ***
Rudimentary 1.2 0.9-1.6 0.3 0.1-0.5 -0.9 0.005 ***
Finished 55.6 53.3-56.3 67.2 64.9-67.9 11.6 0.000 ***
Household exterior wall materials (%)10 
Natural 96.2 93.8-95.2 96.7 94.8-96.1 0.5 0.501 n/s
Finished 3.7 4.8-6.1 3.3 3.8-5.2 -0.4 0.502 n/s
Household floor materials (%)11
Natural 95.7 94.7-95.9 95.3 94.6- 95.9 -0.4 0.765 n/s
Rudimentary 2.8 2.6-3.7 1.1 0.8-1.4 -1.7 0.001 ***
Finished 1.4 1.1- 1.7 3.6 3.1- 4.2 2.2 0.008 ***
Sources: Feed the Future Baseline Survey 2013, Ethiopia; Feed the Future End-line ZOI Survey 2018, Ethiopia 2019 1 A t-test of statistical significance of differences between the baseline and end-line estimates of means for continuous variables, and a Pearson’s 
chi-squared test of statistical significance of differences between the baseline and end-line estimates of proportions for categorical variables.
Differences found to be statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001.
2 Improved water sources include piped water into the dwelling, piped water into the yard, public tap/standpipe, tube well/borehole, protected dug well, 
protected spring, and rainwater (WHO and UNICEF 2006). The proportion of the population with sustainable access to an improved water 
source is the SDG indicator 6.1.1 (UN ECOSOC 2016). The measure includes regularity in access to the water source – namely, that (a) water 
is available from this source all year round and (b) water from this source was available every day in the 2 weeks preceding the survey. 3
 Correct water treatment practice or technology refers to methods that effectively kill or remove pathogens. This includes boiling the water, 
adding bleach or chlorine, using a water filter (ceramic, sand, composite) and solar disinfection (WHO and UNICEF 2006). Practices such as straining 
through a cloth and letting it stand and settle are not considered effective approaches of water treatment. 4
 Improved sanitation facilities are those that separate human excreta from human contact and include the categories flush to piped sewer system, 
flush to septic tank, flush/pour flush to pit, composting toilet, ventilated improved pit latrine and pit latrine with a slab. Because shared and public 
facilities are often less hygienic than private facilities, shared or public sanitation facilities are not counted as improved (WHO and UNICEF 
2006). The proportion of the population with access to improved sanitation is the SDG indicator 6.2.1 (UN ECOSOC 2016). 5
 A sanitation facility is considered unimproved if human excreta is not adequately separated from human contact. This includes the following: 
flush/pour flush elsewhere; pit latrine without a slab/open pit; bucket; and hanging toilet. Households that report having no sanitation facility, or using 
the bush or field are considered as using an unimproved sanitation facility (WHO and UNICEF 2006). 6
 Households that report having no sanitation facility, or using the bush or field are considered as practicing open defecation. 7 Solid fuel is defined as charcoal, wood, animal dung, straw/shrubs/grass and agriculture crop residue. The other and no food cooked in household
categories are removed from percentages. SDG indicator 7.1.2 focuses on the proportion of population with primary reliance on clean fuels 
and technology so the indicator is not directly comparable to the SDG indicator on type of fuel used by households (UN ECOSOC 2016). 8 The average number of persons per sleeping room is a common indicator of crowding (UNDP 2003). 9
 Natural roof includes no roof, thatch (palm leaf, straw, reed), and sod or bamboo. Rudimentary roof includes wood planks and cardboard. Finished
roofs include metal, wood, calamine/cement fiber, ceramic tiles, cement and roofing shingles. The other category is removed from percentages. 10 Natural wall includes no walls, cane/palm/tree trunks, dirt, bamboo with mud and stone with mud. Rudimentary walls include plywood, cardboard, 
reused wood and unbaked bricks. Finished walls include cement, stone with lime/cement, bricks, cement blocks, unbaked bricks covered with plaster 
and wood planks/shingles. The other category is removed from percentages. 11 Natural floors include earth/sand, dung and palm leaves. Rudimentary floors include wood planks and bamboo slats. Finished floors include 
parquet/polished wood, vinyl or asphalt strips, ceramic tiles, cement and wall-to-wall carpet. The other category is removed from percentages.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 39
Tables 3.3b compares the dwelling characteristics of households in the P1-ZOI by gendered 
household type between the phase one baseline and end-line ZOI Surveys. The results of the test 
of statistical differences between the baseline and end-line surveys are also provided.
Table 3.3b reveals that the proportion using an improved water source was about 10 percentage 
points higher in the end-line relative to the baseline in households with male and female adults as 
well as in households with only female adults. Similarly, access to electricity increased by over 12 
percentage points in the latter households. The proportion of households with only male adults 
that use solid fuel was slightly higher in the end-line. 
Table 3.3b. Comparison of household dwelling characteristics, by gendered household 
type, between the Feed the Future phase one baseline and end-line ZOI 
Surveys 
Household 
characteristic
Baseline (2013) End-line (2018)
Diff. p-value1 Est. 95% CI Est. 95% CI Sig.
Households using an improved water source2
Male and female adult(s) 57.2 55.3-58.7 66.8 65.8-69.1 9.6 0.000 ***
Female adult(s) only 55.4 51.2-58.2 63.6 60.1-67.4 8.2 0.010 **
Male adult(s) only 62.1 55.5-68.8 66.2 57.7-75.1 4.1 0.492 n/s
Child(ren) only ^ ^ ^ ^ - - -
Households using a correct water treatment practice or technology3
Male and female adult(s) 8.3 7.5-9.4 10.8 9.5-11.7 2.5 0.142 n/s
Female adult(s) only 7.5 6.0-9.8 8 5.8-9.9 0.5 0.808 n/s
Male adult(s) only 7.2 4.1-11.3 6.7 1.6-10.4 -0.5 0.888 n/s
Child(ren) only ^ ^ ^ ^ - - -
Households using an improved sanitation facility, non-shared4
Male and female adult(s) 65.6 63.4-66.7 64 61.6-65.1 -1.6 0.444 n/s
Female adult(s) only 54.7 51.8-58.8 51.2 47.6-55.2 -3.5 0.211 n/s
Male adult(s) only 54.9 47.7-61.3 52.6 42.5-60.9 -2.3 0.67 n/s
Child(ren) only ^ ^ ^ ^ - - -
Households using an improved sanitation facility, shared4
Male and female adult(s) 0.2 0.09-0.2 0.1 0.01-0.2 -0.1 0.301 n/s
Female adult(s) only 0.0 - 0.3 0.1-0.7 0.3 0.159 n/s
Male adult(s) only 0.8 0.3-2.2 1.2 0.6-4.1 0.4 0.713 n/s
Child(ren) only ^ ^ ^ ^ - - -
Households using an unimproved sanitation facility5
Male and female adult(s) 0.7 0.4-0.9 0.8 0.5-1.1 0.1 0.575 n/s
Female adult(s) only 0.7 0.7-1.2 0.9 0.3-1.8 0.2 0.672 n/s
Male adult(s) only 2.5 0.04-3.7 1.1 0.8-2.5 -1.4 0.395 n/s
Child(ren) only ^ ^ ^ ^ - - -
Households practicing open defecation6
Male and female adult(s) 33.2 32.1-35.3 35.1 34.1-37.5 1.9 0.361 n/s
Female adult(s) only 44.2 40.1-47.1 47.5 43.4-51.1 3.3 0.218 n/s
Male adult(s) only 41.4 35.3-48.8 45.1 36.5-54.9 3.7 0.502 n/s
Child(ren) only ^ ^ ^ ^ - - -
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 40
Households using solid cooking fuels7
Male and female adult(s) 99.2 98.7-99.4 98.9 98.4-99.2 -0.3 0.519 n/s
Female adult(s) only 98.7 97.9-99.5 99.1 98.1-99.7 0.4 0.493 n/s
Male adult(s) only 94.7 92.3-98.1 98.6 95.9-100 3.9 0.04 **
Child(ren) only ^ ^ ^ ^ - - -
Households with access to electricity
Male and female adult(s) 6.1 5.6-7.3 22.7 20.7-23.6 16.6 0.000 ***
Female adult(s) only 6.5 5.1-8.6 19.1 15.7-21.7 12.6 0.000 ***
Male adult(s) only 8.8 4.7-12.4 13.3 6.7-19.1 4.5 0.308 n/s
Child(ren) only ^ ^ ^ ^ - - -
Sources: Feed the Future Baseline Survey, Ethiopia 2013; Feed the Future End-line ZOI Survey 2018, Ethiopia. 1 A Pearson’s chi-squared test of statistical significance of differences between the baseline and end-line estimates of proportions. Differences 
found to be statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001.
2 Improved water sources include piped water into the dwelling, piped water into the yard, public tap/standpipe, tube well/borehole, protected dug well, 
protected spring and rainwater (WHO and UNICEF 2006). The proportion of the population with sustainable access to an improved water 
source is the SDG indicator 6.1.1 (UN ECOSOC 2016). The measure includes regularity in access to the water source – namely, that (a) water 
is available from this source all year round and (b) water from this source was available every day in the 2 weeks preceding the survey. 3
 Correct water treatment practice or technology refers to methods that effectively kill or remove pathogens. This includes boiling the water, 
adding bleach or chlorine, using a water filter (ceramic, sand, composite) and solar disinfection (WHO and UNICEF 2006). Practices such as straining 
through a cloth and letting it stand and settle are not considered effective approaches of water treatment. 4
 Improved sanitation facilities are those that separate human excreta from human contact and include the categories flush to piped sewer system, 
flush to septic tank, flush/pour flush to pit, composting toilet, ventilated improved pit latrine and pit latrine with a slab. Because shared and public 
facilities are often less hygienic than private facilities, shared or public sanitation facilities are not counted as improved (WHO and UNICEF 
2006). The proportion of the population with access to improved sanitation is the SDG indicator 6.2.1 (UN ECOSOC 2016). 5
 A sanitation facility is considered unimproved if human excreta is not adequately separated from human contact. This includes the following: 
flush/pour flush elsewhere; pit latrine without a slab/open pit; bucket; and hanging toilet. Households that report having no sanitation facility, or using 
the bush or field are considered as using an unimproved sanitation facility (WHO and UNICEF 2006). 6
 Households that report having no sanitation facility, or using the bush or field are considered as practicing open defecation. 7 Solid fuel is defined as charcoal, wood, animal dung, straw/shrubs/grass and agriculture crop residue. The other and no food cooked in household
categories are removed from percentages. SDG indicator 7.1.2 focuses on the proportion of population with primary reliance on clean fuels 
and technology so the indicator is not directly comparable to the SDG indicator on type of fuel used by households (UN ECOSOC 2016).
3.3 Education
Table 3.4a and Table 3.4b present school attendance, educational attainment and literacy in 
the P1-ZOI. Table 3.4a presents the percentage of all household members between ages 5 and 
24 who currently attend school, the percentage of household members over age 9 who have 
completed primary school, and the percentage of household members over age 4 who are 
literate. Table 3.4b presents the same percentages separately for females and males. Table 3.4a 
also includes sex ratios for school attendance, attainment of primary education and literacy. 
In Ethiopia, primary education is defined as 8 years of education (in grades 1 through 8). 
Tables 3.4a and 3.4b reveal that the proportion of children aged 5-9 attending school is low at 30 
percent. This may indicate that children start attending schools late, particularly when comparing 
this to 77 and 68 percent of children aged 10-14 and 15-19, respectively, that are attending 
schools. Thirty percent of children aged 15-19 completed primary school and 89 percent are 
literate while primary school completion was higher (47 percent) and literacy was slightly lower 
(78 percent) among 20–24-year-olds. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 41
The results also reveal that the proportion of females under 20 that are attending schools and 
are literate is about the same as the proportion of males while the proportion of females under 
20 that completed primary schools is considerably higher. However, the proportion of females 
aged 20 and older that completed primary school and that are literate are lower than the 
proportion of males, which may indicate that households are now enrolling a higher proportion 
of girls into schools or that females stop schooling as they get older or both.
Table 3.4a. School attendance, educational attainment and literacy in the P1-ZOI 
Age group 
(years)
Percent Female to male ratio
n 
Attending 
school1
Attained a 
primary 
level of 
education2 Literate3
Attendin
g school1
Attained a 
primary 
level of 
education2 Literate3
5-9 29.7 n/a 31.5 1.0 0.0 1.0 3,282
10-14 77.3 1.8 83.8 1.0 2.4 1.0 3,502
15-19 68.2 30.0 88.9 1.0 1.3 1.0 2,229
20-24 33.6 47.0 78.4 0.6 0.9 0.8 1,168
25-29 0.0 36.9 53.1 0.0 0.7 0.5 1,083
30-34 0.0 24.5 33.5 0.0 0.5 0.4 1,194
35-54 0.0 14.4 30.4 0.0 0.7 0.3 3,251
55+ 0.0 6.7 13.6 0.0 0.4 0.1 1,644
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 1 Estimates include only household members who are 5-24 years old. 2 Estimates include only household members who are 10 years old or older. 3 Estimates include only household members who are 5 years or older.
Table 3.4b. School attendance, educational attainment and literacy in the P1-ZOI, by 
age and sex 
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 
n/a = not applicable
1 Estimates include only household members who are 5-24 years old. 2 Estimates include only household members who are 10 years or older. 3 Estimates include only household members who are 5 years or older.
Age group 
(years)
Female
 
Male
Attending 
school1
Attained a 
primary level 
of education2
Literate3
n 
Attending 
school1
Attained a 
primary 
level of 
education2
Literate3
n 
(%) (%) (%) (%) (%) (%)
5-9 30.4 n/a 32.1 1673 29.6 n/a 31.7 1609
10-14 77.3 2.7 83.6 1768 75.7 1.1 83 1734
15-19 67.7 34.6 89.1 944 67.9 27.6 88.9 1285
20-24 24.4 45.5 71.1 454 38.2 48.8 84.2 714
25-29 0.0 30.3 41 669 0.0 42.9 74.9 414
30-34 0.0 15.8 20.7 706 0.0 28.7 54.9 488
35-54 0.0 11.6 14.1 1640 0.0 16.6 48.3 1611
55+ 0.0 3.2 3.5 885 0.0 7.5 26.4 759
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 42
Table 3.5 provides a comparison of a household member’s highest educational attainment 
between the phase one baseline and end-line ZOI Surveys for all individuals and by sex of 
individuals. Generally, the highest education achieved by any member of the household improved 
over the period, and the change is statistically significant. The percentage of household members 
with no education was 45 percent lower in the end-line relative to the baseline. Looking at the 
educational status by sex of the household members, relative to males a higher percentage of 
female members were illiterate during both baseline and end-line ZOI Surveys. Moreover, the 
decline in the proportion of illiterate household members was relatively higher among males than 
females. However, level of education has improved over the two period for both male and female 
members. 
Table 3.5. Comparison of household member educational attainment in the PI-ZOI, 
in total and by sex, Feed the Future phase one baseline and end-line ZOI 
Surveys 
Sex and highest 
educational 
attainmenta
Baseline (2013) End-line (2018)
Diff.
p￾valueb Sig.c % 95% CI n % 95% CI n 
Total 
No education 31.1 29.7-32.4 1440 17 16.0-18.4 670 -13.9 0.000 ***
Less than primary 52.3 50.8-53.7 2420 56 53.9-57.1 2159 3.7 0.003 ***
Completed primary 16.2 15.1-17.2 748 26 25.2-27.9 1034 9.8 0.000 ***
Completed secondary ^ ^ 20 ^ ^ 26 - - - 
Female 
No education 54 52.5-55.5 2438 39 37.1- 40.1 1469 -15.4 0.000 ***
Less than primary 38.7 37.3-40.1 1747 48 46.5-49.6 1830 9.4 0.000 ***
Completed primary 7.1 6.3- 7.8 320 13 12.1-14.2 501 6.1 0.000 ***
Completed secondary ^ ^ 8 ^ ^ 5 - - - 
Male 
No education 37.7 36.3-39.2 1601 23 21.1- 23.8 797 -15.2 0.000 ***
Less than primary 49.4 47.9-50.9 2093 57 55.1-58.3 2009 7.3 0.000 ***
Completed primary 12.4 11.4-13.4 529 20 18.8 -21.5 715 7.8 0.000 ***
Completed secondary ^ ^ 13 ^ ^ 21 - - - 
^ Results not statistically reliable, n<30
Sources: Feed the Future Baseline Survey 2013, Ethiopia; Feed the Future End-line ZOI Survey 2018, Ethiopia. a Estimates include only de facto household members who are 5 years of age or older. b Significance tests were performed to determine whether a difference exists between the baseline and end-line estimates. c Differences found to be statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001.
Note: Estimates are sample-weighted; numbers of observations are unweighted.
Feed the Future Ethiopia ZOI Survey 2013; Feed the Future Ethiopia ZOI Survey 2018
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 43
4. Household Economic Status 
This chapter includes a background discussion of monetary poverty in Ethiopia, including the logic 
of the Living Standard Measurement Survey (LSMS)8 and consumption expenditures methodology. 
Appendix A2.2 provides an overview of the methodology for calculating the Feed the Future PI￾ZOI poverty indicators. Additional details are provided in the Feed the Future Guide to Statistics. 
The Household Roster and Household Consumption Expenditure survey modules are used to 
calculate the per capita expenditures and prevalence of poverty indicators. The household 
consumption expenditure module is similar to the LSMS, in which households’ consumption of 
various food and non-food items is measured to infer household income and well-being. 
Individuals’ per capita expenditures are then derived by dividing total household expenditures by 
the number of household members. From these data, household expenditure totals are calculated 
and used as a proxy for household incomes, based on the assumption that a household’s 
consumption is closely related to its income. Household consumption and expenditures are often 
preferred to income when measuring poverty due to the difficulty in accurately measuring 
income. According to Deaton, expenditure data are less prone to error, easier to recall and more 
stable over time than income data.9
 
Ethiopia recorded a rapid economic growth over the last 15 years since 2004/05. The continued 
growth has contributed to the decline in poverty. Official statistics indicate that poverty declined 
nationwide from about 44 percent in 1996 to 23 percent in 2015. Consistent with these figures, 
we see an increase in per capita expenditure and a decline in poverty among households in the 
Feed the Future ZOI. 
4.1 Daily Per Capita Consumption Expenditures
Table 4.1a presents the Feed the Future ZOI indicator daily per capita consumption 
expenditures in constant 2010 USD. Daily per capita consumption expenditures serve as a proxy 
for income. 
Table 4.1a includes the overall mean daily per capita consumption expenditures and the mean 
daily per capita consumption expenditure for selected percentiles of households. The percentiles 
provide information on the distribution of consumption expenditures. As is typical of 
consumption expenditure and income data, these estimates are positively skewed, with the 
majority of the population consuming or spending very little, and a small portion of the population 
consuming much more (see Figure 4.1 also). The mean, therefore, is affected by these outliers
who consume much more. A summary statistic more robust to outliers is the median (or the 
8
 Grosh, Margaret and Paul Glewwe. 1995. “A Guide to Living Standards Measurement Study Surveys and Their Data Sets.” Living Standards 
Measurement Study Group. Working paper No. 120. The World Bank, Washington, DC. 9
 Deaton, A. 2008. The Analysis of Household Surveys: A microeconomic approach to development policy. Baltimore: The Johns Hopkins 
University Press.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 44
50th percentile). Estimates in Table 4.1a are shown for all households as well as disaggregated by 
household characteristics, including gendered household type, household size, and household 
educational attainment. 
Average daily per capita consumption expenditure was $2.19 in constant 2010 USD in the P1-
ZOI at the time of the 2018 Ethiopia ZOI Survey. Median (50th percentile) daily per capita 
consumption expenditure was equivalent to $1.60, indicating that 50 percent of all individuals in 
the P1-ZOI had daily per capita consumption expenditures less than $1.60 in 2005 PPP. 
Table 4.1a Daily per capita consumption expenditures by household characteristic (in 
constant 2010 USD1
) in the P1-ZOI
Estimate (weighted) 
(constant 2005 PPP)
Characteristic Mean
Percentile
n2 10th 25th 50th 75th 90th
All households 2.19 0.57 0.95 1.60 2.72 4.34 3,876
Gendered household type
Male and female adults 2.22 0.58 0.97 1.64 2.76 4.40 3,089
Female adult(s) only 1.95 0.47 0.79 1.31 2.35 4.05 665
Male adult(s) only 2.05 0.59 0.87 1.60 2.50 3.71 115
Children only ^ ^ ^ ^ ^ ^ 7 
Household size
Small (1-5 members) 2.37 0.61 1.01 1.68 2.93 4.66 2,257 
Medium (6-10 members) 2.08 0.58 0.91 1.55 2.57 4.18 1,574 
Large (11+ members) 1.51 0.14 0.24 1.07 2.46 3.00 45
Household educational attainment
No education 1.97 0.43 0.80 1.49 2.50 3.74 664
Less than primary 2.13 0.57 0.90 1.54 2.62 4.26 2,153 
Primary 2.36 0.61 1.05 1.78 2.92 4.72 1,032 
Secondary or more ^ ^ ^ ^ ^ ^ 26
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 1 Daily per capita consumption expenditures measured in Birr were converted to constant 2010 USD using the 2005 and 2010 Consumer Price
Index (CPI) and the 2005 Purchasing Power Parity (PPP) Index estimated by the World Bank. We used the formula: (Per capita expenditures 
LCU)/ (2005 PPP LCU* (2018 CPI LCU/2005 CPI LCU) * (2010 CPI USD/2005 CPI USD)). The 2005 PPP LCU for Ethiopia is 2.75, 2018 CPI 
LCU/2005 CPI LCU=5.68. The conversion factor is 17.4. 2 Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample size 
reflects this loss in observations; therefore, subsample sizes may not add-up to the total sample size. 
Estimates in Table 4.1a also indicate that the median (50th percentile) per capita consumption 
expenditure for households was 2.8 times higher than the average of that of households in the 
bottom 10 percentile by consumption expenditures. Similarly, mean per capita consumption 
expenditure of households in the top 90 percentile is 7.6 times higher relative to those in the 
lowest 10 percentile. Mean consumption expenditure of households with male and female adults 
was the highest and that of those with female adults only the lowest. Moreover, this expenditure 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 45
level appears to decline with increase in household size and increase with educational attainment 
of households.
Table 4.1b compares average daily per capita expenditures between the phase one baseline and 
end-line ZOI surveys. The average daily per capita consumption expenditure has increased by 
23.7 percent from $1.77 in the baseline to $2.19 in the end-line and this difference is highly 
statistically significant. The increment is in real terms, meaning the impact of inflation has been 
accounted for. 
Table 4.1b Comparison of daily per capita consumption expenditures in constant 2010 
USD at 2005 PPP between the Feed the Future phase one baseline and end￾line ZOI Surveys1
Indicator
Baseline 
(2013)
End-line 
(2018) p-value2
(sig)
Daily per capita consumption expenditures 1.77 2.19 0.0000 (***)
Source: Feed the Future Baseline Survey 2013, Ethiopia; Feed the Future End-line ZOI Survey 2018, Ethiopia. 1 Daily per capita consumption expenditures measured in Birr were converted to constant 2010 USD using the 2005 and 2010 Consumer Price 
Index (CPI) and the 2005 Purchasing Power Parity (PPP) Index estimated by the World Bank. We used the formula: (Per capita expenditures 
LCU)/ (2005 PPP LCU* (2018 CPI LCU/2005 CPI LCU) * (2010 CPI USD/2005 CPI USD)). The 2005 PPP LCU for Ethiopia is 2.75, 2018 CPI 
LCU/2005 CPI LCU=5.68. The conversion factor is 17.4 for the end-line survey. The 2013 level, computed similarly using corresponding 
parameters, is found to be 12.17. 2
 A t-test of statistical significance of differences between the baseline and end-line estimates of mean expenditure. Differences found to be 
statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001.
Figure 4.1 shows the share of total consumption per quintile in the ZOI. Inequality is high in the 
ZOI. The richest 20 percent of the population consume nearly half of total consumption in the 
ZOI. In contrast, the poorest 20 percent consume only about 5 percent of the total consumption.
Figure 4.1. Share of consumption per quintile: Feed the Future P1-ZOI 
1 The poorest quintile is determined as the bottom fifth of the population by consumption expenditure. The poorest quintile’s share of total 
consumption is calculated by dividing the consumption of the poorest quintile by total consumption within the P1-ZOI.
4.8
9.6
14.5
22.1
49.0
0
10
20
30
40
50
60
1st (Poorest) 2nd 3rd 4th 5th (Richest)
Share of consumption within the PI￾ZOI (percentage)
Quintile
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 46
Table 4.1c compares average daily per capita consumption expenditures in constant 2010 USD 
between the phase one baseline and end-line ZOI surveys for all households in the P1-ZOI, as 
well as disaggregated by gendered household type, household educational attainment, and 
household hunger scale. The table shows that average daily per capita consumption expenditure
in 2010 USD has statistically significantly increased for “Male and female adults” households, 
households with some education, and households which experienced “little to no hunger.” 
Table 4.1c Comparison of mean daily per capita consumption expenditures in 
Constant 2010 USD at 2005 PPP in the PI-ZOI, in total and by selected 
household characteristics, Feed the Future phase one baseline and end-line 
ZOI Surveys
 Baseline 
(2013) 
End-line 
(2018) 
Diff. p-valuea Sig. Characteristic Mean b 95% CI N Mean 95% CI n 
All households 1.77 1.70-1.84 4,618 2.19 2.10-2.27 3,876 0.42 0.000 ***
Gendered household type
Male and female adults 1.73 1.65-1.81 3,534 2.21 2.12-2.31 3,089 0.48 0.000 ***
Female adults only 1.87 1.74-1.99 855 1.95 1.79-2.11 665 0.08 0.383 n/s
Male adults only 2.35 2.06-2.63 224 2.05 1.53-2.56 115 -0.35 0.250 n/s
Children only ^ ^ 2 ^ ^ 7 ^ ^ ^ 
Household education
No education 1.82 1.73-1.92 1,434 2.96 1.81-2.12 664 0.14 0.171 n/s
Less than primary 1.74 1.63-1.85 2,412 2.13 2.01-2.26 2,153 0.39 0.000 ***
Completed primary 1.79 1.66-1.92 746 2.36 2.22-2.49 1,032 0.57 0.000 ***
Secondary or more ^ ^ 20 ^ ^ 26 ^ ^ ^ 
Household hunger 
Little to no hunger 1.78 1.71-1.86 4,378 2.21 2.12-2.29 3,747 0.43 0.000 ***
Moderate hunger 1.47 1.29-1.65 198 1.54 1.21-1.88 105 0.07 0.698 n/s
Severe hunger 1.16 0.77-1.55 34 ^ ^ 23 ^ ^ ^ 
^ Results not statistically reliable, n<30
a Significance tests were performed to determine whether a difference exists between the baseline and end-line estimates. b Differences found to be statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001.
Notes: Estimates are sample-weighted; numbers of observations are unweighted. Estimates are based on de jure household members.
Daily per capita consumption expenditures measured in Birr were converted to constant 2010 USD using the 2005 and 2010 Consumer Price 
Index (CPI) and the 2005 Purchasing Power Parity (PPP) Index estimated by the World Bank. We used the formula: (Per capita expenditures 
LCU)/ (2005 PPP LCU* (2018 CPI LCU/2005 CPI LCU) * (2010 CPI USD/2005 CPI USD)). The 2005 PPP LCU for Ethiopia is 2.75, 2018 CPI 
LCU/2005 CPI LCU=5.68. The conversion factor is 17.4 for the end-line survey. The 2013 level, computed similarly using corresponding 
parameters, is found to be 12.17. 
Sources: Feed the Future Ethiopia ZOI Survey 2013; Feed the Future Ethiopia ZOI Survey 2018.
4.2 Prevalence and Depth of Poverty in the P1-ZOI
The prevalence of poverty, sometimes called the poverty headcount ratio, is measured by the 
percentage of individuals living below a poverty threshold.10 Estimates of poverty prevalence are 
10 Note that consumption expenditure data are not collected at the individual level but rather at the level of the 
household; daily per capita consumption expenditures are then derived by dividing total household expenditures 
by the number of household members. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 47
sensitive to the poverty thresholds used to identify the poor. A standardized poverty threshold 
of $1.25 per person per day in adjusted11 2005 PPP USD is used to track global changes in poverty 
across countries. The $1.25 threshold is in effect the extreme poverty threshold and represents 
the poverty line typical of the world’s poorest countries.
12 Poverty estimates may also be 
presented for an individual country’s own poverty and extreme poverty thresholds. 
Although poverty prevalence indicates how many individuals are impacted by poverty, it does not 
speak to how much people are impacted by poverty. In other words, it implicitly assumes that all 
those below the poverty line are equally poor. The depth of poverty, often called the poverty 
gap, is a useful poverty estimate because it captures the intensity of poverty. This measure 
indicates the average gap between consumption expenditure levels of the poor and the poverty 
line, with the non-poor counted as having a gap of zero. The measure is expressed as a proportion 
of the poverty line. The depth of poverty or poverty gap represents the entire ZOI population. 
The average consumption expenditure shortfall of the poor, in contrast, is estimated for only 
those individuals living below the poverty line. 
4.2.1 The $1.25 Poverty Threshold 
Table 4.2a presents poverty estimates at the $1.25 per person per day (2005 PPP) threshold. 
The prevalence of poverty and depth of poverty at the $1.25 per person per day poverty line are 
Feed the Future PI-ZOI indicators. Similar to the daily per capita consumption expenditures table, 
this table presents poverty estimates for all households in the P1-ZOI, as well as disaggregated 
by gendered household type, household size and household educational attainment. 
Poverty Prevalence 
According to the estimates in Table 4.2, prevalence of poverty, which is measured by percent of 
people living below $1.25 per person per day in 2005 PPP, is 32.3 percent. The prevalence of 
poverty is lowest for male and female adult households and highest for female adult only 
households; it increases with household size and declines with educational attainment. 
Using this measure, poverty declined by over 7 percentage points from the baseline estimate of 
about 39.9 percent (Table 4.3a). This difference is statistically significant. When disaggregated by 
household characteristics, the prevalence of poverty has considerably and statistically significantly
11 Per capita expenditure measured in Birr (the local currency unit (LCU)) were converted to 2005 USD PPP using 
the PPP conversion factor estimated (and annually updated until 2011) by the World Bank and the Consumer 
Price Index (CPI) computed by the Central Statistical Agency (CSA). We used the formula (Per capita expenditures
in 2018 LCU)/ (2005 USD PPP* (2018 CPI LCU/2005 CPI LCU)). The 2005 USD PPP (in 2005 prices) for Ethiopia 
is 2.75, (2018 CPI LCU/2005 CPI LCU)=5.68 and the product of the two gives us the adjusted conversion factor 
as 15.62. The 2013 levels were computed similarly using corresponding parameters including the (2013 CPI 
LCU/2005 CPI LCU)=3.96. The 2005 international poverty line is $1.25 USD PPP, while that of 2011 is 1.9 USD 
PPP. 12 
United Nations Statistics Division. nd. Millennium Development Goals Indicators. 
https://millenniumindicators.un.org/unsd/mdg/Metadata.aspx?IndicatorId=0&SeriesId=580 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 48
declined for male and female adult households while there was a slight but statistically insignificant 
increase in the prevalence of poverty for female adult only and male adult only households. 
Table 4.2. Poverty in the ZOI at the international poverty line of $1.25 (2005 PPP) - 2018
 Prevalence of 
Poverty2,5
Depth of 
Poverty3,5
Average consumption shortfall 
of the poor4,5
Characteristic
Percent
n5
Percent 
of 
poverty 
lineb
n5 In USD 
2005 PPP n5
population
Total (All households) 32.3 3876 12.2 3876 0.5 1,230
Gendered household type 
Male and female adults 31.3 3,089 11.8 3,089 0.5 920
Female adult(s) only 40.9 665 15.8 665 0.5 265
Male adult(s) only 36.3 115 12.5 115 0.4 42
Child(ren) only (no adults) ^ 7 ^ 7 ^ 3 
Household size 
Small (1-5 members) 29.3 2,267 10.4 2,267 0.4 685
Medium (6-10 members) 33.6 1,578 12.5 1,578 0.5 525
Large (11+ members) 51.5 45 32 45 ^ 20
Household education 
No education 37.3 670 15.6 670 0.5 247
Less than primary 33.7 2,159 12.4 2,159 0.5 704
Primary 30.7 1,034 10.7 1,034 0.5 270
Secondary or more 27.2 26 10.5 26 ^ 9 
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 
^ Results not statistically reliable, n<30. 1 The poverty headcount ratio is computed as the percentage of the ZOI population living on less than $1.25 a day at 2005 PPP prices. 2 The prevalence of poverty is the percentage of individuals living below the national poverty line. Poverty prevalence is sometimes referred to 
as the poverty incidence or poverty headcount ratio. 3 The depth of poverty, or poverty gap, is the average consumption shortfall as a percentage of the poverty line. 4 The average consumption shortfall of the poor is the average amount below the poverty threshold of a person in poverty. This value is 
estimated only among individuals living in households that fall below the poverty threshold. 5 Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample size 
reflects this loss in observations; therefore, subsamples may not total to the aggregate sample size. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 49
Table 4.3a: Comparison of prevalence of poverty at the $1.25 (2005 PPP) per person per day 
threshold in the PI-ZOI, in total and by selected household characteristics, Feed the 
Future phase one baseline and end-line ZOI Surveys
 
Baseline 
(2013) 
End-line 
(2018)
Diff.
p￾valueb
 
Sig.c Characteristic % 95% CI na % 95% CI na
All households 39.9 38.5-41.4 4,618 32.3 30.8-33.8 3,876 -7.6 0.000 ***
Gendered household type 
Male and female adults 40.4 38.8-42.0 3,534 31.3 29.6-32.9 3,089 -9.1 0.000 ***
Female adults only 38.9 35.6-42.2 855 40.8 37.1-44.6 665 1.9 0.504 n/s
Male adults only 31.0 24.9-37.1 224 36.3 27.3-45.1 115 5.3 0.445 n/s
Children only ^ ^ 2 ^ ^ 7 ^ ^ ^ 
Household education 
No education 43.5 41.0-46.1 1,434 37.7 34.0-41.3 664 -5.8 0.026 * 
Less than primary 39.4 37.4-41.3 2,412 33.7 31.7-35.7 2,153 -5.7 0.000 ***
Completed primary 36.8 33.3-40.2 746 27.5 24.8-30.2 1,032 -9.3 0.000 ***
Secondary or more ^ ^ 20 ^ ^ 26 ^ ^ ^ 
Household hunger 
Little to no hunger 39.2 37.7-40.6 4,378 31.3 29.8-32.8 3,747 -7.9 0.000 ***
Moderate hunger 55.4 48.4-62.4 198 59.5 50.0-69.1 105 4.1 0.550 n/s
Severe hunger 58.9 41.5-76.3 34 ^ ^ 23 ^ ^ ^ 
^ Results not statistically reliable, n<30
a The “n” reflects the unweighted number of households—not the number of household members—even though the prevalence of poverty 
measures the percentage of individuals living below a poverty threshold. Records missing information for the disaggregate variables have been 
excluded from the disaggregated estimates. The unweighted sample size reflects this loss in observations; therefore, disaggregate sample sizes 
may not total to the aggregated sample size.
b Significance tests were performed to determine whether a difference exists between the baseline and end-line estimates. c Differences found to be statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001.
Notes: Estimates are sample-weighted; numbers of observations are unweighted. 
Estimates are based on de jure household members.
The prevalence of poverty, sometimes referred to as the poverty incidence or poverty headcount ratio, is the percentage of individuals living 
below the $1.25 (2005 PPP) per person per day threshold. 
Depth of Poverty 
Depth of poverty for the Feed the Future ZOI in the end-line data using $1.25 and 2005 PPP is 
12.2 percent (Table 4.3b). The decline is only a percentage point when compared with the 
baseline estimate. 
The depth of poverty provides an indication of the amount of resource transfers that, if perfectly
targeted to poor households, would be needed to bring everyone in the ZOI up to the poverty 
line. With a P1-ZOI population of 13.8 million (in adult equivalent units), a poverty threshold of 
$1.25 per person per day (2005 PPP), poverty head count of 32.3 and a poverty gap of 12.2 
percent, $2,113,207 (2005 PPP) per day would need to be transferred to the poor to bring their 
income or expenditures up to the poverty threshold13. 
13 Since the mean depth of poverty in the ZOI is 12.2 percent, the poverty threshold is $1.25/day, and the population 
is 13.8 million; the average daily cost of raising the income or consumption expenditures of the poor up to the 
poverty line is calculated as: [(12.2 ÷ 100) * $1.25/day] *13857097 = $2,113,207/day.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 50
Table 4.3b: Comparison of depth of poverty at the $1.25 (2005 PPP) per person per day threshold 
in the PI-ZOI, in total and by selected household characteristics, Feed the Future 
phase one baseline and end-line ZOI Surveys
 
Baseline 
(2013) 
End-line 
(2018) 
Diff.
p￾valueb Sig.c Characteristic
% of 
poverty 
line 95% CI na
% of 
poverty 
line 95% CI na
All households 13.2 12.6-13.8 4,618 12.2 11.4-12.8 3,876 -1.0 0.054 n/s
Gendered household type 
Male and female adults 13.3 12.6-14.0 3,534 11.7 10.9-12.5 3,089 -1.6 0.010 * 
Female adults only 13.2 11.7-14.6 855 15.8 13.9-17.6 665 2.6 0.058 n/s
Male adults only 9.7 7.3-12.1 224 12.5 8.5-16.4 115 2.8 0.279 n/s
Children only ^ ^ 2 ^ ^ 7 ^ ^ ^ 
Household education 
No education 14.6 13.4-15.7 1,434 15.7 13.8-17.7 664 1.1 0.381 n/s
Less than primary 13.2 12.4-14.1 2,412 12.5 11.5-13.4 2,153 -0.7 0.327 n/s
Completed primary 11.4 9.9-12.8 746 10.0 8.7-11.2 1,032 -1.4 0.212 n/s
Secondary or more ^ ^ 20 ^ ^ 26 ^ ^ ^ 
Household hunger
Little to no hunger 12.8 12.2-13.4 4,378 11.6 10.9-12.3 3,747 -1.2 0.039 * 
Moderate hunger 20.3 16.7-24.0 198 27.7 21.8-33.6 105 7.4 0.073 n/s
Severe hunger 31.7 21.0-42.4 34 ^ ^ 23 ^ ^ ^ 
^ Results not statistically reliable, n<30 a The “n” reflects the unweighted number of households—not the number of household members—even though the depth of poverty measures 
the average gap between the consumption level of individuals and the poverty line. Records missing information for the disaggregate variables 
have been excluded from the disaggregated estimates. The unweighted sample size reflects this loss in observations; therefore, disaggregate 
sample sizes may not total to the aggregated sample size.
b Significance tests were performed to determine whether a difference exists between the baseline and end-line estimates. c Differences found to be statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001.
Notes: Estimates are sample-weighted; numbers of observations are unweighted. 
Estimates are based on de jure household members.
The depth of poverty, or poverty gap, is the average consumption shortfall multiplied by the prevalence of poverty. The depth of poverty was 
calculated by first subtracting the per capita daily consumption expenditure in local currency units (LCU) for the year and month of the ZOI 
Survey of each household in the sample from the USD $1.25 per day poverty line in LCU for the year and month of the ZOI Survey, with the 
resulting difference set to zero for all non-poor households. The figure was then divided by the USD $1.25 per day poverty line in LCU for the 
year and month of the ZOI Survey. Finally, after applying the household member sampling weight, the value for each household was summed 
across all households and then divided by the sum of the number of all sampled households with consumption data. 
Sources: Feed the Future Ethiopia ZOI Survey 2013; Feed the Future Ethiopia ZOI Survey 2018
The decline in the incidence of poverty was not replicated in the depth of poverty. The fall 
recorded in the latter is not statistically significant. Consumption expenditure rose significantly
for households with male and female adults, with some formal education, and with little to no 
hunger experience. It did not for other household types. Correspondingly, these are also the 
household types which experienced reduction in poverty. In fact, depth of poverty also fell 
marginally for households with male and female adults and for those with little to no hunger 
experience. Thus, the pattern of poverty reduction largely left the depth of poverty unchanged. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 51
4.2.3 The National Poverty Threshold
Table 4.4 presents poverty estimates at the national poverty threshold for Ethiopia. Similar to 
prior poverty tables, this table presents poverty estimates for all households in the P1-ZOI
(excluding as before the Somali sub-sample), as well as disaggregated by characteristics, including 
gendered household type, household size and household educational attainment.
The source of the national poverty line is the 2016 national poverty interim report (NPC 
(September 2017)). The 2011 poverty line is 3781 Birr per adult equivalent per year 2011 prices. 
If we divide it by 365, we will get the daily national poverty line in 2011 prices to be 10.36 Birr. 
Then, the only adjustment that is needed to arrive at the 2013 and 2018 poverty lines is to allow 
for annual inflation. Using the November 2018 retail price report, the cumulative inflation from 
2011 to 2013 is 125.68 and from 2011 to 2018 is 180.1. With this adjustment we get the poverty 
line to be Birr13.02 and Birr18.66 per adult equivalent per day for 2013 and 2018, respectively. 
Table 4.4 Total (absolute) poverty line in Birr (average price)
Year 1995/96 2010/11 2015/16
Absolute poverty line (Birr) 1075 3781 7184
Source: Computed using HICE survey 1995/96, 2010/11 and 2015/16 and MoFEC (2015). 
Note: The absolute poverty lines for 2010/11 and 2015/16 are respectively in December 2005 and December 2011 prices (NPC (September 
2017). 
Poverty Prevalence 
Based on the national poverty line of 18.66 Birr per adult equivalent per day, the prevalence of 
poverty is 30.0 percent at the end-line. The rate is less than that under the international poverty 
line because the national poverty line is slightly lower when converted into 2005 PPP prices. The 
conversion involves dividing the national poverty line in Birr by the product of the 2005 USD PPP 
for Ethiopia (= 2.75) and the CPI ratio 2018 CPI LCU/2005 CPI LCU (=5.68) – or the adjusted 
conversion factor of 15.62. The result is $1.19. Second, this poverty line implies that poverty in 
the ZOI declined by 6.4 percentage points during Feed the Future phase one – a change very 
close to that recorded using the international poverty line. 
Table 4.5 compares the prevalence of poverty at the national threshold between the phase one 
baseline and end-line ZOI surveys. The table shows that poverty at the national threshold 
declined by over 6 percentage points, from the baseline estimate of 36.4 percent. This difference 
is statistically significant. Households with male and female adults experienced a statistically 
significant decline in the prevalence of poverty during the period. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 52
Depth of Poverty 
The overall depth of poverty is 11.3 percent (Table 4.5). The depth of poverty provides an 
indication of the amount of resource transfers that, if perfectly targeted to poor households, 
would be needed to bring everyone in the ZOI up to the poverty line. With a P1-ZOI population 
of 13.8 million (in adult equivalent units), a poverty threshold of Birr18.66 per adult equivalent 
per day, and a poverty gap of 11.3 percent, Birr 29,218,798 per day would need to be transferred 
to the poor to bring their income or expenditures up to the poverty threshold14. Given the 2005 
PPP conversion factor for 2005 (2.75) and the inflation factor of 5.68, this amounts to $1,870,602 
per day in 2005 PPP. The difference with the figure obtained above using the international poverty 
line of $1.25 is accounted for by the lower national poverty line and poverty gap, which led to 
the smaller average daily consumption shortfall of the poor which amount to $0.16 and $0.13
using the international and national poverty lines, respectively.
Table 4.5: Comparison of prevalence of poverty at the national threshold of 13.02 Birr (baseline) 
18.66 Birr (end-line) in the PI-ZOI, in total and by selected household characteristics, 
Feed the Future phase one baseline and end-line ZOI Surveys
 
Baseline 
(2013) 
End-line 
(2018) 
Diff.
p￾valueb Sig.c Characteristic % 95% CI na % 95% CI na
All households 36.4 35.0-37.8 4,618 30.0 28.5 -31.4 3,876 -6.4 0.00 ***
Gendered household type 
Male and female adults 36.8 35.2-38.4 3,534 29.0 27.4-30.6 3,089 -7.8 0.000 ***
Female adults only 35.6 32.4-38.9 855 38.7 35.0-42.5 665 3.1 0.261 n/s
Male adults only 27.2 21.3-33.1 224 32.2 23.6-40.9 115 5.0 0.432 n/s
Children only ^ ^ 2 ^ ^ 7 ^ ^ ^ 
Household education
No education 39.9 37.4-42.4 1,434 35.6 31.9-39.2 664 -4.3 0.098 n/s
Less than primary 36.3 34.4-38.2 2,412 31.3 29.3-33.2 2,153 -5.0 0.000 ***
Completed primary 32.1 28.7-35.4 746 25.4 22.7-28.1 1,032 -6.7 0.006 **
Secondary or more ^ ^ 20 ^ ^ 26 ^ ^ ^ 
Household hunger
Little to no hunger 35.6 34.2-37.0 4,378 29.1 27.6-30.5 3,747 -6.5 0.000 ***
Moderate hunger 50.9 43.8-57.9 198 56.7 47.1-66.4 105 5.8 0.394 n/s
Severe hunger 58.9 41.5-76.3 34 ^ ^ 23 ^ ^ ^ 
Sources: Feed the Future Ethiopia ZOI Survey 2013; Feed the Future Ethiopia ZOI Survey 2018; National Planning commission 
^ Results not statistically reliable, n<30
a The “n” reflects the unweighted number of households—not the number of household members—even though the prevalence of poverty 
measures the percentage of individuals living below a poverty threshold. Records missing information for the disaggregate variables have been 
excluded from the disaggregated estimates. The unweighted sample size reflects this loss in observations; therefore, disaggregate sample sizes 
may not total to the aggregated sample size.
b Significance tests were performed to determine whether a difference exists between the baseline and end-line estimates. c Differences found to be statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001.
Notes: Estimates are sample-weighted; numbers of observations are unweighted. Estimates are based on de jure household members.
14 Since the mean depth of poverty in the ZOI is 11.3 percent, the poverty threshold is Birr18.66 per adult equivalent 
per day, and the population is 13.8 million in adult equivalent units; the average daily cost of raising the consumption 
expenditures of the poor up to the poverty line is calculated as: [(11.3 ÷ 100) * Birr18.66/day] *13857097 = 
Birr29,218,798 /day.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 53
5. Hunger and Dietary Intake 
This chapter presents findings related to hunger and the dietary intake of women and young 
children in the P1-ZOI. 
5.1 Household Hunger
The household hunger scale (HHS) is used to calculate the prevalence of households in the 
Ethiopia P1-ZOI experiencing moderate or severe hunger. The HHS was developed by the 
USAID-funded Food and Nutrition Technical Assistance II Project (FANTA-2/FHI 360) in 
collaboration with the United Nations Food and Agriculture Organization. It has been cross￾culturally validated to allow comparison across different food-insecure contexts. The HHS is used 
to assess, geographically target, monitor and evaluate settings affected by substantial food 
insecurity. The HHS is used to estimate the percentage of households affected by severity: little 
to no household hunger (HHS score 0-1); moderate household hunger (HHS score 2-3); and 
severe household hunger (HHS score 4-6). The HHS should be measured at the same time each 
year, and ideally at the most vulnerable time of year (i.e., right before the harvest, during the dry 
season, etc.).15,16
The hungry season in Ethiopia varies across the regions but generally occurs between May and 
September. The end-line data for the HHS were collected in September and early October. 
Tables 5.1a and 5.1b present estimates of household hunger for all households in the P1-ZOI, 
as well as by gendered household type and household educational attainment.
The hunger scale suggests low prevalence of hunger in Ethiopia. Only 3.3 percent of the 
households reported moderate or severe hunger. The prevalence of moderate and severe hunger 
is the highest for female only households and lowest for male and female adults households while 
it is observed to decline with educational attainment of households whereby the proportion of 
households with primary education that reported moderate or severe hunger is about one￾quarter of the proportion of households with no education. The table also indicates that the 
prevalence of households that reported moderate or severe hunger declined by 1.6 percentage 
points in the end-line compared to the baseline estimate of about 5 percent or the prevalence of 
households that experienced moderate or severe hunger declined by 33 percent during the 
period. The proportion of male and female adults households that reported moderate or severe 
hunger declined relatively faster between the baseline and end-line. Furthermore, the decline in 
the prevalence of moderate or severe hunger between the baseline and end-line periods increases 
with educational attainment of households. The proportion that reported moderate or severe 
15 Deitschler, Ballard, Swindale & Coates (2011). 16 For further description of the household hunger indicator and its calculation, please refer to the 2016 edition of 
the Feed the Future Indicator Handbook, available at https://www.agrilinks.org/post/feed-future-indicator￾handbook.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 54
hunger in the end-line was less than half the proportion in the baseline for households with 
primary education, it declined by about one-quarter for households with some primary education,
while it changed little for households with no education.
It is clear that the extent of hunger implied by the HHS scores are rather low for rural Ethiopia. 
However, it is worth noting that the household hunger scale may not be an appropriate measure 
of food security in the Ethiopian context (see Maxwell, Coates and Vaitla 2013).
Table 5.1a. Comparison of household hunger in the PI-ZOI, by severity, in total and by 
selected household characteristics, Feed the Future phase one baseline and 
end-line ZOI Surveys
 Baseline (2013) End-line (2018)
Characteristic
Little to 
none
(%)
Moderate
(%)
Severe 
(%) na 
Little to 
none
(%)
Moderate
(%)
Severe 
(%) na
All households 95.1 4.1 0.8 4,224 96.7 2.7 0.6 3,882
Gendered household type 
Male and female adults 96.3 3.2 0.6 3,235 97.5 2.2 0.3 3,105
Female adults only 90.8 7.7 1.5 776 93.2 5.3 1.5 661
Male adults only 93.0 6.0 1.0 209 93.4 3.8 2.8 116
Children only ^ ^ ^ 2 ^ ^ ^ 7 
Household education 
No education 93.5 5.4 1.1 1,217 93.8 4.9 1.3 670
Less than primary 95.5 3.8 0.7 2,251 96.8 2.7 0.6 2,158
Primary 96.5 3.0 0.6 731 98.4 1.4 0.2 1,034
Secondary or more ^ ^ ^ 20 ^ ^ ^ 26
^ Results not statistically reliable, n<30
a Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample size 
reflects this loss in observations; therefore, disaggregate sample sizes may not total to the aggregated sample size. 
Notes: Estimates are sample-weighted; numbers of observations are unweighted.
Estimates are based on primary adult decision-makers who are de jure household members.
Sources: Feed the Future Ethiopia ZOI Survey 2013; Feed the Future Ethiopia ZOI Survey 2018
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 55
Table 5.1b. Comparison of household hunger between the Feed the Future phase one 
baseline and end-line ZOI Surveys
 Baseline (2013) End-line (2018)
Diff.
P￾valueb Sig.c Characteristic
Moderate 
or severe
(%) 95% CI na 
Moderate 
or severe
(%) 95% CI
na
All households 4.9 3.8-6.2 4,224 3.3 2.6-4.2 3,882 -1.6 0.019 **
Gendered household type 
Male and female adults 3.8 2.8-5.0 3,235 2.5 1.9-3.3 3,105 -1.3
Female adults only 7.9 7.0-12.1 776 6.8 5.0-9.3 661 -1.1
Male adults only 7.0 4.0-11.8 209 6.6 3.1-13.4 116 -0.4
Children only ^ ^ 2 ^ ^ 7 - 
Household education 
No education 6.5 4.5-9.3 1,217 5.2 4.2-9.1 670 -1.3
Less than primary 4.5 3.4-5.8 2,251 3.3 2.4-4.3 2,158 -1.2
Primary 3.6 2.3-5.4 731 1.6 1.0-2.6 1,034 -2.0
Secondary or more ^ ^ 20 ^ ^ 26 - 
^ Results not statistically reliable, n<30
a Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample size 
reflects this loss in observations; therefore, disaggregate sample sizes may not total to the aggregated sample size. 
Notes: Estimates are sample-weighted; numbers of observations are unweighted.
Estimates are based on primary adult decision-makers who are de jure household members. b Significance tests were performed to determine whether a difference exists between the baseline and end-line estimates. c Differences found to be statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001.
5.2 Dietary Intake
This section presents information on the dietary diversity of women of reproductive age and on 
infant and young child feeding in the P1-ZOI. 
5.2.1 Dietary Diversity among Women Age 15-49 Years
Women of reproductive age (15-49 years) are at risk of multiple micronutrient deficiencies, which 
can jeopardize their health and their ability to care for their children and participate in income￾generating activities (c.f. Darnton-Hill et al. 2005). The Feed the Future women’s dietary diversity 
indicator is a proxy for the micronutrient adequacy of women’s diets. The dietary diversity 
indicator reports the mean number of food groups consumed in the day preceding the survey by 
non-pregnant women of reproductive age. 
For the Feed the Future phase one end-line/baseline ZOI Survey analysis, two dietary diversity 
indicators for women are calculated: the Mean Women’s Dietary Diversity Score (WDDS) and 
Women’s Minimum Dietary Diversity (MDD-W).17
17 The Feed the Future indicators list does not include MDD-W. It is added here to exploit the potential benefit due 
to its greater disaggregation. See next footnote. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 56
Mean Women’s Dietary Diversity Score
The Feed the Future mean WDDS indicator, presented in Table 5.2, is based on nine food groups: 
(1) grains, roots and tubers; (2) legumes and nuts; (3) dairy products; (4) organ meat; (5) eggs; 
(6) flesh food and small animal protein; (7) vitamin A-rich dark green leafy vegetables; (8) other 
vitamin A-rich vegetables and fruits; and (9) other fruits and vegetables. The number of food 
groups that each woman consumed the day preceding the survey is averaged across all women 
of reproductive age in the sample for whom dietary diversity data were collected to produce the 
mean WDDS. 
Table 5.2 shows the mean and median WDDS for all women of reproductive age in the P1-ZOI, 
and by individual- and household-level characteristics. Individual-level characteristics include 
women’s age groups and educational attainment. Household-level characteristics include 
gendered household type, household size and household hunger.
Dietary diversity is extremely low in ZOI. An average woman of reproductive age consumed 
from 2.33 food groups in the previous day. Dietary diversity increases with education but only 
marginally. We see little difference in WDDS between different generations of women. The 
dietary diversity is slightly higher for medium sized households (6-10 members), for households 
reporting little or no hunger, and households composed of both female and male adults.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 57
Table 5.2. Women’s mean and median dietary diversity scores in the P1-ZOI
Characteristic Mean
Median n1 Est. 95% CI
All women of reproductive age 2.33 2.24-2.41 2.00 3,956
Age
15-19 2.41 2.30-2.53 2.00 919
20-24 2.28 2.14-2.42 2.00 445
25-29 2.27 2.14-2.40 2.00 665
30-34 2.36 2.25-2.46 2.00 702
35-39 2.24 2.10-2.38 2.00 559
40-44 2.36 2.23-2.49 2.00 423
45-49 2.33 2.15-2.52 2.00 243
Educational attainment
No education 2.22 2.13-2.31 2.00 2,257
Less than primary 2.44 2.32-2.56 2.00 1,139
Primary 2.59 2.41-2.77 3.00 337
Secondary or more 2.65 2.16-3.13 2.00 38
Gendered household type
Male and female adults 2.35 2.27-2.44 2.00 3,476
Female adult(s) only 2.13 2.97-2.29 2.00 465
Male adult(s) only ^ ^ ^ 10
Child(ren) only (no adults) ^ ^ ^ 5
Household size
Small (1-5 members) 2.22 2.11-2.32 2.00 1,687
Medium (6-10 members) 2.42 2.33-2.52 2.00 2,161
Large (11+ members) 2.18 1.61-2.74 2.00 108
Household hunger
Little to no hunger 2.34 2.26-2.43 2.00 3,842
Moderate or severe hunger 1.78 1.50-2.06 2.00 114
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 
^ Results not statistically reliable, n<30. 1 Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted 
sample size reflects this loss in observations; therefore, subsamples may not total to the aggregate sample size
Women’s Minimum Dietary Diversity
The Feed the Future MDD-W indicator uses the following 10 food groups: (1) grains, roots and 
tubers; (2) legumes and beans; (3) nuts and seeds; (4) dairy products; (5) eggs; (6) flesh foods, 
including organ meat and miscellaneous small animal protein; (7) vitamin A-rich dark green leafy 
vegetables; (8) other vitamin A-rich vegetables and fruits; (9) other fruits; and (10) other 
vegetables.18 Achievement of women’s minimum dietary diversity is defined as having consumed 
foods from 5 of the 10 food groups in the past 24 hours. Thus, this indicator is a dichotomous 
18 The differences between the nine food groups used for the WDDS (Table 5.2), the standard FTFMS indicator, and 
the 10 food groups used for the new MDD-W measure (Table 5.3) include: (1) legumes and beans are separated 
from nuts and seeds; (2) meat (flesh foods) and organ meat are combined into one group; and (3) other fruits and 
other vegetables are separated into two groups.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 58
variable, and the measure is reported as the percentage of women who achieve a minimum 
dietary diversity.
Table 5.3 shows the percentage of all women of reproductive age in the P1-ZOI who have 
achieved the MDD-W threshold by individual-level and household-level characteristics. Individual￾level characteristics include women’s age and educational attainment. Household-level 
characteristics include of gendered household type, household size and household hunger. Close 
to 5 percent of women consume from 5 or more food groups. The following differences in MDD￾W are observed. The proportion of households that attained the MDD-W increases with 
educational attainment of households. However, the proportion that attained dietary diversity 
remains extremely low even among the most educated women. The proportion of women in 
household that reported little or no hunger that achieved MDD-W is considerably higher relative 
to those in households with moderate or severe hunger. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 59
Table 5.3. Percentage of women of reproductive age achieving minimum dietary 
diversity in the P1-ZOI 
Characteristic Percent 
n1 Est. 95% CI
All women of reproductive age 4.7 3.7-5.9 3,956
Age
15-19 5.1 3.6-7.3 919
20-24 5.7 4.0-8.2 445
25-29 5.1 3.6-7.3 665
30-34 3.7 2.5-5.3 702
35-39 4.0 2.4-6.6 559
40-44 4.2 2.6-6.9 423
45-49 5.3 3.0-9.3 243
Educational attainment
No education 3.6 2.7-4.9 2,257
Less than primary 5.7 4.1-7.9 1,139
Primary 8.5 5.9-12.0 337
Secondary or more 10.3 4.0-14.1 38
Gendered household type
Male and female adults 4.7 3.7-6.0 3,476
Female adult(s) only 4.6 2.7-7.7 465
Male adult(s) only ^ ^ 10
Child(ren) only (no adults) ^ ^ 5
Household size
Small (1-5 members) 3.5 2.5-4.8 1,687
Medium (6-10 members) 5.5 4.3-7.0 2,161
Large (11+ members) 6.9 2.3-18.8 108
Household hunger
Little to no hunger 4.8 3.8-6.0 3,842
Moderate or severe hunger 0.0 0.0-0.0 114
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 
^ Results not statistically reliable, n<30. 1 Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample size 
reflects this loss in observations; therefore, subsamples may not total to the aggregated sample size. 
Table 5.4 shows the percentages of women age 15-49 years who consumed each of the 9 
WDDS food groups by dietary diversity achievement status. The percentage of all women who 
consumed each of the 9 food groups is shown (the “Overall” column), as well as the percentages 
among women who achieve a minimum dietary diversity and among women who do not achieve 
a minimum dietary diversity.
The consumption of animal source foods (dairy, organ meat and flesh foods) is much more 
prevalent among the women who achieved minimum dietary diversity. Similarly, women who 
achieved minimum dietary diversity are more likely to consume fruits and vegetables than those 
who did not achieve minimum dietary diversity. Grains, roots and tubers as well as legumes/nuts 
are prominent food categories among women.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 60
Table 5.4. Percentage of women who consumed foods in the P1-ZOI, by achievement 
of minimum dietary diversity status
Food group
Percent
Achieved 
minimum dietary 
diversity
Did not achieve 
minimum 
dietary diversity Overall
Est. 95% CI Est. 95% CI Est. 95% CI
Grains, roots and tubers 97.3 93.0-90.0 81.2 77.3-84.6 82.0 78.2-85.3
Legumes nuts 82.1 74.2-87.9 52.6 47.1-58.7 54.3 48.5-59.9
Dairy products 78.3 70.0-84.9 21.1 18.7-23.7 23.8 21.3-26.4
Organ meats 20.0 12.1-31.2 1.0 0.7-1.5 1.9 1.3-2.8
Eggs 53.3 44.1-62.2 4.6 3.7-5.7 6.9 5.7-8.2
Flesh foods and other misc. small animal protein 59.2 48.7-69.0 12.7 10.5-15.3 14.9 12.5-17.6
Vitamin A-rich dark green leafy vegetables 71.2 61.4-79.4 30.0 25.3-35.2 32.0 27.3-37.0
Other Vitamin A-rich vegetables and fruits 61.0 50.6-70.4 12.2 10.3-14.3 14.5 12.5-16.7
Other fruits and vegetables 27.0 19.5-36.0 2.2 1.6-3.1 3.4 2.5-4.5
n 188 3,768 3.956
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 
Table 5.5 compares the WDDS and MDD-W indicators between the Feed the Future phase 
one baseline and end-line ZOI Surveys. Women’s dietary diversity score was 1.59 at the baseline 
while it increased to 2.33 in the end-line. This is an increase in WDDS of 46.5 percent, although 
starting from a very low base. Similarly, the percentage of women achieving minimum dietary 
diversity increased from less than 2 percent in the baseline to 4.6 percent at the end-line. That 
is, the proportion achieving MDD-W more than doubled during the period. Both of these changes 
are statistically significant.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 61
Table 5.5. Comparison of women’s dietary diversity between the Feed the Future 
phase one baseline and end-line ZOI Surveys 
 Baseline (2013) End-line (2018) p￾valueb Sig. Characteristic Mean c 95% CI na Mean 95% CI na
Women’s Mean Dietary Diversity Score (Mean number of food groups consumed)
All women of reproductive age 1.59 1.46-1.72 3,978 2.33 2.24-2.41 3,956 0.000 ***
Woman’s age
15-19 1.73 1.54-1.91 798 2.41 2.30-2.53 919
20-24 1.54 1.40-1.68 706 2.28 2.14-2.42 445
25-29 1.58 1.43-1.72 900 2.27 2.14-2.40 665
30-34 1.56 1.40-1.73 579 2.36 2.25-2.46 702
35-39 1.53 1.38-1.68 525 2.24 2.10-2.38 559
40-44 1.38 1.38-1.77 256 2.36 2.23-2.49 423
45-49 1.55 1.35-1.75 223 2.33 2.15-2.52 243
Woman’s education
No education 1.47 1.34-1.60 2,712 2.22 2.13-2.31 2,257
Less than primary 1.79 1.64-1.93 933 2.44 2.32-2.56 1,139
Primary 2.00 1.73-2.26 238 2.59 2.41-2.77 337
Secondary or more ^ ^ 22 2.65 2.16-3.13 38
Gendered household type
Male and female adults 1.60 1.47-1.72 3,333 2.35 2.27-2.44 3,476
Female adults only 1.57 1.39-1.74 605 2.13 2.97-2.29 465
Male adults only 1.55 1.11-1.99 39 ^ ^ 10
Children only ^ ^ 1 ^ ^ 5 
Women’s Minimum Dietary Diversity (Percent of women achieving minimum dietary diversity)
All women of reproductive age 1.8 1.2-2.9 3,978 4.7 3.7- 5.9 3,956 0.000 ***
Woman’s age
15-19 3.1 1.6-5.9 789 5.1 3.6-7.3 919
20-24 1.8 1.1-3.0 706 5.7 4.0-8.2 445
25-29 1.8 1.0-3.2 900 5.1 3.6-7.3 665
30-34 1.6 0.6-3.8 579 3.7 2.5-5.3 702
35-39 1.5 0.7-3.2 525 4.0 2.4-6.6 559
40-44 0.9 0.3-2.8 256 4.2 2.6-6.9 423
45-49 0.4 0.1-2.7 223 5.3 3.0-9.3 243
Woman’s education
No education 1.3 0.8-2.1 2,712 3.6 2.7-4.9 2,257
Less than primary 2.4 1.3- 4.4 933 5.7 4.1-7.9 1,139
Primary 5.7 3.2- 9.9 238 8.5 5.9-12.0 337
Secondary or more ^ ^ 22 10.3 4.0-24.1 38
Gendered household type
Male and female adults 1.8 1.2-2.9 3,333 4.7 3.7-6.0 3,476
Female adults only 1.8 0.7-4.8 605 4.6 2.7-7.7 465
Male adults only 2.8 0.4-17.8 39 ^ ^ 10
Children only ^ ^ 1 ^ ^ 5 
Source: Feed the Future Baseline Survey 2013, Ethiopia; Feed the Future End-line ZOI Survey, Ethiopia 2018. a Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample size 
reflects this loss in observations; therefore, disaggregate sample sizes may not total to the aggregated sample size.
b Significance tests were performed to determine whether a difference exists between the baseline and end-line estimates at the overall 
indicator level. A test of statistical significance in differences between the baseline and end-line estimates was conducted. A t-test of 
differences in means was used for WDDS, and a Pearson’s chi-squared test of differences in proportions was used for MDD-W. c Differences found to be statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 62
5.2.2 Infant and Young Child Feeding
This section presents young children’s dietary intake measures, including the prevalence of 
exclusive breastfeeding among children 0-5 months and the prevalence of children 6-23 months
consuming a minimum acceptable diet. 
Exclusive Breastfeeding 
Exclusive breastfeeding provides children with significant health and nutrition benefits, including 
protection from gastrointestinal infections and reduced risk of mortality due to infectious disease. 
Exclusive breastfeeding means the infant receives breast milk (including expressed breast milk or 
breast milk from a wet nurse) and may receive oral rehydration salts, vitamins, minerals or 
medicines, but does not receive any other food or liquid. This indicator measures the percentage 
of children 0-5 months who were exclusively breastfed during the day preceding the survey.
Table 5.6 shows the prevalence of exclusive breastfeeding among children 0-5 months in the 
P1-ZOI. Estimates are shown for all children, as well as by children’s sex and by educational 
attainment of the child’s primary caregiver. Note that the data are collected for the self-identified
primary caregiver and not strictly for the biological mother, although it is often the same person. 
Nearly three-quarters of all children aged 0-5 months are exclusively breastfed. The percentage 
is marginally higher for female children. 
Table 5.6. Prevalence of exclusive breastfeeding among children 0-5 months in the 
P1-ZOI 
Characteristic
Percent
n1 Est. 95% CI
All children 0-5 months 74.5 63.4-83.1 111
Child sex
Male 73.9 58.9-84.8 57
Female 75.1 62.0-84.8 54
Caregiver educational attainment2
No education 74.2 63.0-82.9 110
Less than primary ^ ^ 0
Primary ^ ^ 0
Secondary or more ^ ^ 0
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 1 Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample 
size reflects this loss in observations; therefore, subsamples may not total to the aggregate sample size. 2 The ZOI Survey identifies the primary caregiver of each age-eligible child. This person is likely, but not necessarily, the child’s biological 
mother. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 63
Minimum Acceptable Diet (MAD)
The prevalence of children 6-23 months consuming a minimum acceptable diet (MAD) measures 
the proportion of young children who receive a MAD, taking their breastfeeding status into 
consideration. This composite indicator measures both the minimum feeding frequency and 
minimum dietary diversity based on caregiver reports of the frequency with which the child was 
fed and the foods that the child consumed during the 24 hours preceding the survey. Calculation 
of the indicator requires data on children’s age in months, breastfeeding status, dietary diversity, 
number of semi-solid or solid feeds, and number of milk feeds.
Table 5.7 presents the Feed the Future MAD indicator for children 6-23 months in the P1-ZOI. 
Estimates are shown for all children, as well as by characteristics of the child, caregiver and 
household. Child characteristics include sex and age. Caregiver characteristics includes 
educational attainment. Household characteristics include gendered household type, household 
size and household hunger. 
Less than 10 percent of the children aged 6-23 months received a MAD. A marginally higher 
percentage of female children received a MAD. A higher proportion of older children aged 18-
23 months received a MAD, with the proportion of older children receiving a MAD about 50 
percent and three times higher relative to children aged 12-17 and 6-11 months, respectively. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 64
Table 5.7. Prevalence of children 6-23 months who receive a minimum acceptable 
diet in the P1-ZOI 
Characteristic
Percent
n1
Est. 95% CI
All children 6-23 months 8.4 6.4-11.0 639
Child sex
Male 7.5 5.0-11.1 324
Female 9.3 6.4-13.1 315
Child age
6-11 months 3.9 1.7-8.3 184
12-17 months 8.7 5.8-12.8 273
18-23 months 12.4 8.3-18.1 182
Caregiver educational attainment2
No education 8.4 6.4-11.0 637
Less than primary ^ ^ 2
Primary ^ ^ 0 
Secondary or more ^ ^ 0 
Gendered household type
Male and female adults 8.4 6.3-11.1 591
Female adult(s) only 8.3 3.1-20.7 46
Male adult(s) only ^ ^ 1 
Child(ren) only (no adults) ^ ^ 1
Household size
Small (1-5 members) 7.6 4.7-12.0 225
Medium (6-10 members) 9.0 6.2-12.9 395
Large (11+ members) ^ ^ 19
Household hunger
Little to no hunger 8.5 6.4-11.2 616
Moderate or severe hunger ^ ^ 23
Source: Feed the Future End-line ZOI Survey, Ethiopia 2018. 
^ Results not statistically reliable, n<30. 1 Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample size 
reflects this loss in observations; therefore, subsamples may not total to the aggregate sample size. 2 The ZOI Survey identifies the primary caregiver of each age-eligible child. This person is likely, but not necessarily, the child’s biological mother.
Table 5.8 presents the percentages of children achieving the minimum meal frequency and 
minimum dietary diversity, as well as the percentages of children consuming each of the food 
groups included in the MAD indicator. Estimates are shown for all children, as well as by child’s 
age, and are presented separately for breastfed children and non-breastfed children.
Table 5.8 shows that a large proportion (over 80 percent) of breastfed children are achieving the 
minimum meal frequency requirement but dietary diversity is extremely low among all children 
in the ZOI. The percentage of children who are not breastfed that achieved minimum meal 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 65
frequency requirement of MAD is less than 20 percent. Most of these children (both breastfed 
and non-breastfed) consume staples (grains, roots and tubers), but fruits and vegetables and 
animal source food consumption is low, especially meat and eggs.
Table 5.8.: Comparison of percent of children 6-23 months of age in the PI-ZOI 
achieving minimum feeding frequency, dietary diversity, and consuming 
specified foods, in total and by breastfeeding status and age, Feed the 
Future phase one baseline and end-line ZOI Surveys
Breastfeeding status and 
food group consumed
Baseline (2013) End-line (2018) 
All Child age (months) All Child age (months)
children 6-11 12-17 18-23 children 6-11 12-17 18-23
All children 6-23 months of age 
Achieving minimum meal 
frequency 41.2 28.7 45.1 48.1 75.1 66.2 81.8 73.7
Achieving minimum dietary 
diversity 3.5 3.2 2.6 4.8 11.1 7.3 10.0 16.6
Consuming: 
Grains, roots and tubers 40.2 28.9 41.6 49.1 61.4 49.4 62.8 71.4
Legumes and nuts 26.7 16.1 28.1 34.7 37.4 22.7 39.4 48.9
Dairy products 22.1 20.1 24.1 21.6 29.4 26.0 31.8 29.2
Flesh foods 1.9 2.4 1.0 2.5 13.6 8.2 13.4 19.3
Eggs 7.1 6.0 5.3 10.4 10.5 9.8 9.4 13.1
Vitamin A-rich fruits and 
vegetables 13.8 8.6 15.8 16.0 25.1 17.4 26.3 31.2
Other fruits and vegetables 3.2 1.9 4.6 2.8 5.8 6.6 5.1 6.0
Number of children 658 193 262 203 648 188 277 183
Breastfed children 
Achieving minimum meal 
frequency 44.0 29.4 47.1 55.6 81.2 68.4 86.8 86.7
Achieving minimum dietary 
diversity 3.4 3.3 2.8 4.4 11.1 7.5 9.2 18.1
Consuming: 
Grains, roots and tubers 38.8 28.2 40.5 47.8 60.2 49.3 61.8 70.1
Legumes and nuts 27.1 15.2 29.3 36.9 37.7 22.9 39.8 51.1
Dairy products 21.7 20.3 23.7 20.5 28.7 26.6 29.4 30.1
Flesh foods 1.6 2.2 1.0 1.8 14.1 8.5 13.4 21.7
Eggs 6.9 6.1 5.6 9.8 10.4 10.1 9.4 12.5
Vitamin A-rich fruits and 
vegetables 13.5 8.9 15.4 15.8 24.8 17.3 26.1 31.3
Other fruits and vegetables 3.4 2.0 4.8 3.0 6.0 6.8 5.1 6.4
Number of breastfed children 588 181 241 166 587 182 254 151
Non-breastfed children 
Achieving minimum meal 
frequency 7.5 ^ ^ 8.6 13.3 ^ ^ 6.9
Achieving minimum milk 
feeding frequency 7.5 ^ ^ 8.6 15.0 ^ ^ 6.9
Achieving minimum dietary 
diversity 4.7 ^ ^ 7.2 11.8 ^ ^ 8.9
Consuming: 
Grains, roots and tubers 57.8 ^ ^ 55.7 74.2 ^ ^ 78.0
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 66
Breastfeeding status and 
food group consumed
Baseline (2013) End-line (2018) 
All Child age (months) All Child age (months)
children 6-11 12-17 18-23 children 6-11 12-17 18-23
Legumes and nuts 21.3 ^ ^ 23.4 34.4 ^ ^ 37.9
Dairy products 27.1 ^ ^ 27.2 36.5 ^ ^ 24.2
Flesh foods 5.4 ^ ^ 6.3 8.7 ^ ^ 7.1
Eggs 8.8 ^ ^ 13.6 11.9 ^ ^ 16.5
Vitamin A-rich fruits and 
vegetables 17.1 ^ ^ 17.3 28.5 ^ ^ 30.6
Other fruits and vegetables 1.2 ^ ^ 1.9 3.8 ^ ^ 3.6
Number of non-breastfed 
children 49 5 12 32 61 6 23 32
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 
^ Results not statistically reliable, n<30.
Table 5.9 compares the prevalence of exclusive breastfeeding among children 0-5 months and 
the prevalence of children 6-23 months receiving a MAD between the phase one baseline and 
end-line ZOI Surveys. 
Although there is a marginal increase in the proportion of exclusively breastfed children (0-5 
months) in the end-line relative to the baseline, the change is not statistically significant in both 
female and male children. There is an increase in the end-line of the proportion of children 
achieving a MAD from the baseline (from about 3 percent to 7 percent for male children and to 
9 percent for female children). That is, the proportion that received a MAD more than doubled 
in both female and male children albeit from very low initial levels. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 67
Table 5.9. Comparison of children’s dietary intake between the phase one baseline 
and end-line ZOI Surveys
 Baseline (2013) End-line (2018) p￾valueb Sig.c Characteristic % 95% CI na % 95% CI na
Exclusive breastfeeding children 0-5 months
All children 0-5 months of age 72.2 65.3-78.3 204 74.5 63.5-83.0 111 0.709 n/s
Child’s sex 
Male 71.5 61.5-79.7 96 73.9 58.9-84.8 57
Female 72.9 62.2-81.4 108 75.1 62.0- 84.8 54
Caregiver’s educationd 
No education 72.2 65.3-78.3 204 74.2 63.0-82.9 110
Less than primary ^ ^ 0 ^ ^ 0 
Primary ^ ^ 0 ^ ^ 0 
Secondary or more ^ ^ 0 ^ ^ 0 
Gendered household type
Male and female adults 73.8 66.5-79.9 177 75.3 63.9-84.0 107
Female adults only ^ ^ 20 ^ ^ 3 
Male adults only ^ ^ 7 ^ ^ 0 
Children only ^ ^ 0 ^ ^ 0 
Minimum acceptable diet children 6-23 months
All children 0-5 months of age 3.4 2.0-5.7 633 8.4 6.4-11.0 639 0.003 **
Child’s sex 
Male 3.5 1.6-7.6 328 7.5 5.0-11.1 324
Female 3.3 1.7-6.4 305 9.3 6.4-13.1 315
Caregiver’s educationd 
No education 3.4 2.0-5.8 633 8.4 6.4-11.0 637
Less than primary ^ ^ 0 ^ ^ 2 
Primary ^ ^ 0 ^ ^ 0 
Secondary or more ^ ^ 0 ^ ^ 0 
Gendered household type
Male and female adults 3.1 1.8-5.4 571 8.4 6.3-11.1 591
Female adults only 2.1 0.3-13.9 52 8.3 3.1-20.7 46
Male adults only ^ ^ 10 ^ ^ 1 
Children only ^ ^ 0 ^ ^ 1 
Source: Feed the Future Baseline Survey 2013, Ethiopia; Feed the Future End-line ZOI Survey, Ethiopia 2018. a Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample size 
reflects this loss in observations; therefore, disaggregate sample sizes may not total to the aggregated sample size. b Significance tests were performed to determine whether a difference exists between the baseline and end-line estimates at the overall 
indicator level. A Pearson’s chi-squared test of statistical significance in differences in proportions between the baseline and 2018-2019
estimates was conducted. c Differences found to be statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 68
6. Nutritional Status of Women and Children
This chapter presents findings on the nutritional status of women and children, including the Feed 
the Future phase one anthropometry indicators. 
6.1 Body Mass Index of Women Age 15-49 Years
Body mass index (BMI) is a calculation used to understand nutritional status, particularly of adults. 
BMI is the weight of the individual in kilograms divided by their height in meters squared 
(weight[kg] / height[m]2
). BMI is an inexpensive and easy-to-perform method of screening for 
weight category: underweight, normal or healthy weight, overweight, and obese. BMI is
interpreted directly using categories with specific cut-off points, which is useful when assessing 
the nutritional status of adults. A high BMI can be an indicator of high body fatness, but BMI is 
not diagnostic of the body fatness or health of an individual. To determine if a high BMI is a health 
risk, a healthcare provider would need to perform further assessments.
Table 6.1 presents women’s mean BMI and the percentage of women by BMI category: 
underweight (BMI < 18.5), normal weight (18.5 ≤ BMI < 25.0), overweight (25.0 ≤ BMI < 30.0), 
and obese (BMI ≥ 30.0). Estimates are shown for all non-pregnant women of reproductive age, 
as well as disaggregated by individual- and household-level characteristics. Individual 
characteristics include age and educational attainment. Household characteristics include 
gendered household type, household size and household hunger.
BMI of women of reproductive age averaged 20.3. The underweight prevalence is 22 percent 
while about 4 percent of the women were over-weight or obese. There exist little differences in 
average BMI and proportion of women in different BMI categories across age, educational 
achievement and other categories of women. For instance, although mean BMI appears to 
increase with age the difference between the highest and lowest mean BMIs is less than 0.7. 
Similarly, relative to women in households with moderate or severe hunger, the proportion 
underweight is higher and the proportion with normal weight is lower for those in households 
with no hunger. 
Table 6.2 compares women’s anthropometry results between the phase one baseline and end￾line ZOI Surveys. The table indicate that there was a slight increase (0.23) in mean BMI of women 
in the end-line relative to the baseline. The prevalence of severely underweight women in the 
end-line is 1.5 percentage points lower than the baseline while the prevalence of underweight 
women in the end-line is 4 percentage points lower. The prevalence of over-weight and obese 
women in the end-line was no different from the baseline.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 69
Table 6.1. Mean BMI and prevalence of underweight, normal weight, overweight, and obese women in the P1-ZOI 
Characteristic
Mean BMI BMI category (percent)
n1
Under-weight Normal weight Over-weight Obese
Est. 95% CI Est. 95% CI Est. 95% CI Est. 95% CI Est. 95% CI
All non-pregnant women of 
reproductive age 20.3 20.2-20.5 21.9 19.9-24.0 74.3 72.2-76.3 3.4 2.8-4.2 0.4 0.2-0.7 3,604
Age
15-19 20.0 19.7-20.2 27.3 23.4-31.5 69.6 65.5-73.5 2.9 1.9-4.3 0.2 0.1-0.9 903
20-24 20.5 20.2-20.8 18.0 13.9-23.0 78.0 72.9-82.5 4.0 2.5-6.3 0.0 0.0-0.0 407
25-29 20.4 20.2-20.6 21.5 18.0-25.5 75.0 70.8-78.8 3.2 2.0-5.1 0.3 0.1-1.3 562
30-34 20.4 20.2-20.7 18.5 15.3-22.3 77.5 73.6-80.9 3.3 2.1-5.1 0.7 0.3-1.9 597
35-39 20.6 20.4-20.9 19.0 15.5-23.1 76.4 72.0-80.2 4.0 2.5-6.3 0.6 0.2-1.9 504
40-44 20.4 20.1-20.7 21.8 17.0-27.5 74.6 68.5-79.8 2.9 1.6-5.4 0.7 0.2-2.3 393
45-49 20.3 20.0-20.7 23.9 18.2-30.8 70.9 64.0-77.0 4.6 2.4-8.5 0.6 0.1-3.8 238
Educational attainment
No education 20.4 20.2-20.6 20.8 18.5-23.3 75.3 72.7-77.7 3.5 2.7-4.4 0.5 0.2-1.0 2,005
Less than primary 20.2 20.0-20.4 24.4 21.1-28.1 72.2 68.6-75.5 3.0 2.1-4.3 0.4 0.1-1.0 1,062
Primary 20.4 20.1-20.8 21.1 16.8-26.2 74.6 69.5-79.1 3.9 2.1-6.9 0.4 0.0-2.5 319
Secondary or more 21.4 20.6-22.2 18.2 8.7-34.2 78.7 62.9-89.0 3.1 0.4-18.5 0.0 0.0-0.0 37
Gendered household type
Male and female adults 20.4 20.2-20.5 21.8 19.7-23.9 74.4 72.3-76.4 3.4 2.7-4.1 0.5 0.3-0.8 3,143
Female adult(s) only 20.2 20.0-20.5 22.2 18.3-26.6 74.0 69.5-78.0 3.8 2.3-6.3 0.0 0.0-0.0 446
Male adult(s) only ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ 10
Child(ren) only (no adults) ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ 5
Household size
Small (1-5 members) 20.2 20.1-20.4 23.5 21.0-26.1 73.1 70.4-75.8 3.1 2.3-4.2 0.3 0.1-0.7 1,526
Medium (6-10 members) 20.4 20.3-20.6 20.4 18.1-22.9 75.5 73.0-77.9 3.6 2.8-4.6 0.4 0.2-0.8 1,981
Large (11+ members) 20.3 19.3-21.2 26.4 17.0-38.5 67.2 57.2-75.9 3.9 1.1-13.3 2.5 0.6-10.3 97
Household hunger
Little to no hunger 20.3 20.2-20.5 22.1 20.1-24.3 74.0 71.9-76.0 3.4 2.8-4.2 0.4 0.2-0.7 3,495
Moderate or severe hunger 20.4 20.0-20.9 14.1 8.2-23.0 84.0 73.4-90.9 2.0 0.5-7.3 0.0 0.0-0.0 109
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 
^ Results not statistically reliable, n<30. 1 Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample size reflects this loss in observations; therefore, 
subsamples may not total to the aggregate sample size
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 70
Table 6.2 Comparison of the nutritional status of women between the Feed the 
Future phase one baseline and end-line ZOI Surveys 
Characteristic
Baseline (2013) End-line (2018)
p-valueb Sig.c % 95% CI na % 95% CI na
Average BMI 
All non-pregnant women of 
reproductive age 20.1 20.0-20.3 3,517 20.3 20.2-20.5 3,604 0.007 **
Woman’s age 
15-19 19.9 19.7-20.2 742 20.0 19.7-20.2 903
20-24 20.1 19.9-20.4 597 20.5 20.2-20.8 407
25-29 20.3 20.1-20.5 747 20.4 20.2-20.6 562
30-34 20.2 19.9-20.4 510 20.4 20.2-20.7 597
35-39 20.2 20.0-20.5 470 20.6 20.4-20.9 504
40-44 19.9 19.4-20.3 238 20.4 20.1-20.7 393
45-49 20.0 19.6-20.4 213 20.3 20.0-20.7 238
Woman’s education 
No education 20.1 20.0-20.3 2,380 20.4 20.2-20.6 2,005
Less than primary 20.0 19.8-20.2 837 20.2 20.0-20.4 1.062
Primary 20.2 19.8-20.6 217 20.4 20.1-20.8 319
Secondary or more ^ ^ 22 21.4 20.6-22.2 37
Prevalence of severely underweight women (%)
All non-pregnant women of 
reproductive age 3.8 3.2-4.7 3,517 2.3 1.8-3.0 3,604 0.001 **
Woman’s age 
15-19 6.2 4.4-8.5 742 4.8 3.2-7.2 903
20-24 3.0 1.8-5.0 597 1.6 0.8-3.3 407
25-29 1.1 0.5-2.2 747 1.2 0.5-2.5 562
30-34 3.7 2.3-5.9 510 2.0 1.1-3.5 597
35-39 3.7 2.1-6.3 470 1.2 0.5-2.6 504
40-44 5.6 3.3-9.1 238 1.0 0.4-2.4 393
45-49 6.4 3.9-10.2 213 2.1 0.9-4.7 238
Woman’s education 
No education 3.7 2.9-4.7 2,380 1.7 1.2-2.4 2,005
Less than primary 4.5 3.3-6.3 837 3.7 2.4-5.7 1,062
Primary 3.5 1.7-6.9 217 2,4 1.2-4.8 319
Secondary or more ^ ^ 22 0.0 0.0-0.0 37
Prevalence of underweight women (%)1
All non-pregnant women of 
reproductive age 26.0 23.8-28.3 3,517 21.9 19.9-24.0 3,604 0.002 **
Woman’s age 
15-19 31.2 27.4-35.3 742 27.3 23.4-31.5 903
20-24 22.7 19.2-26.6 597 18.0 13.9-23.0 407
25-29 21.3 18.0-24.9 747 21.5 18.0-25.5 562
30-34 24.0 20.1-28.3 510 18.5 15.3-22.3 597
35-39 25.3 21.0-30.1 470 19.0 15.5-23.1 504
40-44 34.0 27.7-40.9 238 21.8 17.0-27.5 393
45-49 30.3 24.1-37.2 213 23.9 18.2-30.8 238
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 71
Characteristic
Baseline (2013) End-line (2018)
p-valueb Sig.c % 95% CI na % 95% CI na
Woman’s education 
No education 25.7 23.3-28.3 2,380 20.8 18.5-23.3 2,005
Less than primary 27.4 24.1-31.0 837 24.4 21.1-28.1 1,062
Primary 24.7 19.1-31.4 217 21.1 16.8-26.2 319
Secondary or more ^ ^ 22 18.2 8.7-34.2 37
Prevalence of overweight women (%)
All non-pregnant women of 
reproductive age 3.0 2.4-3.8 3,517 3.4 2.8-4.2 3,604 0.409 n/s
Woman’s age 
15-19 2.9 1.8-4.5 742 2.9 1.9-4.3 903
20-24 2.5 1.5-4.2 597 4.0 2.5-6.3 407
25-29 2.4 1.5-3.9 747 3.2 2.0-5.1 562
30-34 2.4 1.4-4.0 510 3.3 2.1-5.1 597
35-39 3.8 2.4-5.8 470 4.0 2.5-6.3 504
40-44 5.6 2.6-11.6 238 2.9 1.6-5.4 393
45-49 4.2 2.2-7.9 213 4.6 2.4-8.5 238
Woman’s education 
No education 2.8 2.1-3.8 2,380 3.4 2.7-4.4 2,005
Less than primary 3.6 2.5-5.2 837 3.0 2.1-4.3 1,062
Primary 2.9 1.2-6.5 217 3.9 2.2-6.9 319
Secondary or more ^ ^ 22 3.1 0.4-18.5 37
Prevalence of obese women (%)
All non-pregnant women of 
reproductive age 0.4 0.2-0.8 3,517 0.4 0.2-0.7 3,604 0.786 n/s
Woman’s age 
15-19 1.2 0.6-2.4 742 0.2 0.1-0.9 903
20-24 0.2 0.0-1.3 597 0.0 0.0-0.0 407
25-29 0.1 0.0-0.5 747 0.3 0.1-1.2 562
30-34 0.0 0.0-0.0 510 0.7 0.3-1.9 597
35-39 0.0 0.0-0.0 470 0.6 0.2-1.9 504
40-44 0.5 0.1-3.3 238 0.7 0.2-2.3 393
45-49 0.7 0.2-2.9 213 0.6 0.1-3.8 238
Woman’s education 
No education 0.3 0.1-0.7 2,380 0.5 0.2-1.0 2,005
Less than primary 0.4 0.1-1.1 837 0.4 0.1-1.0 1,062
Primary 1.6 0.5-5.5 217 0.4 0.0-2.5 319
Secondary or more ^ ^ 22 0.0 0.0- 0.0 37
Source: Feed the Future Baseline Survey 2013, Ethiopia; Feed the Future End-line ZOI Survey, Ethiopia 2018. 1
 Prevalence of underweight women is a Feed the Future ZOI indicator. a Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample size 
reflects this loss in observations; therefore, disaggregate sample sizes may not total to the aggregated sample size. b Significance tests were performed to determine whether a difference exists between the baseline and end-line estimates at the overall 
indicator level. A Pearson’s chi-squared test of differences was used for proportions while a t-test of differences in means was used for 
continuous indicators. c Differences found to be statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 72
6.2 Stunting, Wasting, Underweight among Children 
Under 5 Years
This section presents three indicators that rely on anthropometric measurements of children 
under five years in the P1-ZOI: stunting (low height-for-age), wasting (low weight-for-height) and 
underweight (low weight-for-age). 
6.2.1 Stunting (Low Height-for-Age)
Stunting is an indicator of linear growth retardation, most often due to a prolonged inadequate 
diet and poor health. Reducing the prevalence of stunting among children, particularly age 0-23 
months, is important because linear growth deficits accrued early in life are associated with 
cognitive impairments, poor educational performance and decreased work productivity as adults
(c.f. Black et al. 2008, Victora et al. 2008). Stunting is a height-for-age measurement that reflects 
chronic undernutrition. This indicator measures the percentage of children 0-59 months who are 
stunted, as defined by a height-for-age z-score more than two standard deviations (SDs) below 
the median of the 2006 WHO Child Growth Standard.19 The stunting measures presented below 
include the Feed the Future ZOI Survey indicator of stunting—moderate or severe (<-2 SD)—
and the indicator for severe stunting (<-3 SD). Mean z-scores are also presented. 
Table 6.3 shows the prevalence of stunting and severe stunting and mean height-for-weight z￾scores for children under 5 years in the P1-ZOI. Estimates are presented for all children and by 
child, caregiver and household characteristics. Child characteristics include sex and age. Caregiver
characteristics include educational attainment. Household characteristics include gendered 
household type, household size and household hunger.
Out of children of under 5 years of age nearly 39 percent are stunted, and 17.6 percent are 
severely stunted. Growth faltering occurs during the second 12 months of children whereby the 
proportion of children aged 12-23 months stunted and severely stunted is double the proportion 
aged 0-11 months. The HAZ score stabilizes after the second years of life. This is a typical pattern 
observed in many low-income countries. 
The proportion of male and female children stunted is similar. The proportion of stunted children 
in households with moderate or severe hunger is 10 percentage points lower than those in 
households with little to no hunger, which appears to be counter intuitive. Similarly, the 
proportion stunted and severely stunted in households with larger family size is over 10 
percentage points lower than those in medium and small family size. 
19 WHO. (2006).
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 73
Table 6.3. Prevalence of stunting and mean height-for-age z-scores among children 
under 5 years in the P1-ZOI 
Characteristic
Percent
Mean
z-score
n1
Stunted
(<-2 SD)
Severely stunted
(<-3 SD)
Est. 95% CI Est. 95% CI Est. 95% CI
All children under 5 years 38.9 36.1-41.8 17.6 15.6-19.8 -1.4 -1.5- -1.2 1,963
Child sex
Male 38.8 34.9-42.8 17.2 14.6-20.1 -1.4 -1.5- -1.2 976
Female 39.0 35.8-42.3 18.0 15.8-20.5 -1.3 -1.5- -1.2 987
Child age
0-11 months 19.6 14.5-25.9 8.3 5.2-13.0 -0.3 -0.5-0.0 234
12-23 months 41.0 35.8-46.4 17.8 14.0-22.3 -1.4 -1.6- -1.2 398
24-35 months 43.7 38.1-49.5 20.7 16.7-25.5 -1.5 -1.8- -1.3 406
36-47 months 42.1 37.0-47.5 18.7 15.3-22.5 -1.5 -1.7- -1.3 483
48-59 months 39.1 34.0-44.4 18.3 14.6-22.7 -1.5 -1.7- -1.4 442
Caregiver educational attainment2
No education 38.8 36.0-41.7 17.5 15.5-19.7 -1.4 -1.5- -1.2 1,958
Less than primary ^ ^ ^ ^ ^ ^ 4
Primary ^ ^ ^ ^ ^ ^ 0
Secondary or more ^ ^ ^ ^ ^ ^ 0
Gendered household type
Male and female adults 39.1 36.2-42.0 17.6 15.6-19.9 -1.4 -1.5- -1.2 1,808
Female adult(s) only 33.6 25.7-42.5 14.9 9.4-22.8 -1.1 -1.4- -0.8 137
Male adult(s) only ^ ^ ^ ^ ^ ^ 15
Child(ren) only (no adults) ^ ^ ^ ^ ^ ^ 2
Household size
Small (1-5 members) 39.9 35.9-44.1 16.1 13.3-19.3 -1.5 -1.6- -1.3 676
Medium (6-10 members) 38.9 35.4-42.5 19.1 16.7-21.8 -1.4 -1.5- -1.2 1,229
Large (11+ members) 26.5 15.2-42.0 5.8 2.1-15.0 -0.3 -1.0-0.4 57
Household hunger
Little to no hunger 39.2 36.3-42.1 17.6 15.6-19.9 -1.4 -1.5- -1.2 1,901
Moderate or severe hunger 28.4 18.7-40.7 17.2 8.9-30.5 -1.1 -1.6- -0.7 61
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia.
^ Results not statistically reliable, n<30. 1 Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample size 
reflects this loss in observations; therefore, subsamples may not total to the aggregate sample size. 2 The ZOI Survey identifies the primary caregiver of each age-eligible child. This person is likely, but not necessarily, the child’s biological mother. 
6.2.2 Wasting (Low Weight-for-Height)
Wasting is an indicator of acute malnutrition. Children who are wasted are too thin for their 
height and have a much greater risk of dying than children who are not wasted. The wasting ZOI 
indicator measures the percentage of children 0-59 months who are acutely malnourished, as 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 74
defined by a weight-for-height z-score20 more than 2 SDs below the median of the 2006 WHO 
Child Growth Standard, in the P1-ZOI. The wasting measures presented below include the Feed 
the Future ZOI Survey indicator of wasting—moderate or severe (<-2SD), the indicator for 
severe wasting (<-3SD), and the percentage of children who are overweight (>-2SD) and obese 
(>-3SD). Mean z-scores are also presented.
Table 6.4 shows the prevalence of wasting, severe wasting, and mean weight-for-height z-scores 
for children under 5 years in the P1-ZOI. The table also shows the prevalence of overweight and 
obese children under 5 years, as calculated using their weight-for-height z-scores, in the P1-ZOI. 
Estimates are presented for all children and by child, caregiver and household characteristics. 
Child characteristics include sex and age. Caregiver characteristics include educational 
attainment. Household characteristics include gendered household type, household size and 
household hunger.
Out of all children under 5 years of age, 9.7 percent are wasted, 3.7 percent are severely wasted, 
2.7 percent are overweight, and 1 percent obese. The prevalence of wasting is generally higher 
among children under 2 years of age relative to those 2 years or older. Moreover, the proportion 
of female children wasted is slightly (2.5 percentage points) lower than males while the proportion 
in households with only female adults is 3.5 percentage points higher than those in households 
with male and female adults. The proportion of children in households with moderate and severe 
hunger that are wasted is about twice the proportion in households with little to no hunger, 
respectively. 
20 A weight-for-length z-score is calculated for children 0-23 months of age and any other children who are measured 
lying down. A weight-for-height z-score is calculated for children 24-59 months of age who are measured standing 
up. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 75
Table 6.4. Prevalence of wasting and mean weight-for-height z-scores among children under 5 years in the P1-ZOI 
Characteristic
Percent 
Mean 
z-score
n1
Wasted
Severely wasted 
(<-3 SD)
Overweight 
(> +2SD)
Obese 
(> +3SD)
Est. 95% CI Est. 95% CI Est. 95% CI Est. 95% CI Est. 95% CI
All children under 5 years 9.7 7.8-12.0 3.7 2.7-5.0 2.7 2.0-3.7 1.0 0.7-1.7 -0.3 -0.4- -0.2 1,916
Child sex
Male 11.0 8.7-13.9 4.9 3.6-6.8 2.5 1.7-3.6 0.9 0.4-1.7 -0.3 -0.4- -0.2 950
Female 8.4 6.3-11.2 2.4 1.5-3.9 3.0 1.9-4.5 1.2 0.7-2.3 -0.2 -0.4- -0.1 966
Child age
0-11 months 13.3 9.0-19.2 6.5 3.5-11.9 7.8 5.2-11.5 3.5 1.9-6.4 -0.2 -0.5-0.1 221
12-23 months 12.2 9.4-15.6 3.2 1.9-5.4 2.3 1.0-5.1 0.9 0.3-2.4 -0.4 -0.5- -0.2 391
24-35 months 7.5 4.7-11.8 3.1 1.5-6.2 2.7 1.4-5.0 0.5 0.1-2.0 -0.1 -0.3-0.0 396
36-47 months 7.7 5.1-11.4 4.2 2.6-6.7 2.8 1.6-5.0 1.1 0.5-2.7 -0.2 -0.4- -0.1 479
48-59 months 10.0 7.1-13.7 2.7 1.5-4.9 0.2 0.0-1.6 0.2 0.0-1.6 -0.4 -0.5- -0.2 429
Caregiver’s educational attainment2
No education 9.7 7.8-12.0 3.7 2.7-5.0 2.7 2.0-3.7 1.0 0.6-1.6 -0.3 -0.4- -0.2 1,913
Less than primary ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ 2
Primary ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ 0
Secondary or more ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ 0
Gendered household type
Male and female adults 9.5 7.6-11.8 3.7 2.7-5.0 2.9 2.1-4.0 1.1 0.7-1.8 -0.2 -0.3- -0.1 1,763
Female adult(s) only 12.9 7.7-20.8 3.7 1.6-8.5 0.7 0.0-4.9 0.0 0.0-0.0 -0.5 -0.8- -0.3 136
Male adult(s) only ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ 15
Child(ren) only (no adults) ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ 1
Household size
Small (1-5 members) 8.0 6.0-10.6 2.6 1.6-4.1 2.4 1.4-4.1 1.3 0.6-2.4 -0.2 -0.4- -0.1 658
Medium (6-10 members) 10.4 7.9-13.5 4.1 2.8-6.0 2.8 2.0-3.9 0.8 0.4-1.5 -0.3 -0.4- -0.1 1,205
Large (11+ members) 16.3 7.0-33.6 8.3 3.5-18.5 5.1 1.2-19.0 3.7 0.9-14.5 -0.5 -1.0-0.1 52
Household hunger
Little to no hunger 9.5 7.6-11.7 3.6 2.7-5.0 2.7 2.0-3.7 1.1 0.7-1.7 -0.3 -0.4- -0.2 1,858
Moderate or severe hunger 17.8 10.0-29.7 5.1 1.8-13.7 3.2 0.8-11.4 0.0 0.0-0.0 -0.1 -0.5- -0.4 57
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 
^ Results not statistically reliable, n<30. 1 Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample size reflects this loss in observations; therefore, 
subsamples may not total to the aggregate sample size. 2 The ZOI Survey identifies the primary caregiver of each age-eligible child. This person is likely, but not necessarily, the child’s biological mother. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 76
6.2.3 Underweight (Low Weight-for-Age)
Underweight is a weight-for-age measurement that reflects acute undernutrition, chronic 
undernutrition, or both. This indicator measures the percentage of children 0-59 months who 
are underweight, as defined by a weight-for-age z-score of more than two SDs below the median 
of the 2006 WHO Child Growth Standard. The underweight measures presented below include 
the Feed the Future phase one ZOI Survey indicator of underweight—moderate or severe (<-2 
SD), the indicator for severe underweight (<-3 SD), and mean weight-for-age z-scores. 
Table 6.5 shows the prevalence of underweight, severe underweight and mean weight-for-age 
z-scores for children under 5 years in the P1-ZOI. Estimates are presented for all children and 
by child, caregiver and household characteristics. Child characteristics include sex and age. 
Caregiver characteristics include educational attainment. Household characteristics include 
gendered household type, household size and household hunger.
Out of children under 5 years of age, 22.5 percent are underweight and 7 percent severely 
underweight. The proportion underweight is the lowest (15 percent) among children aged 0-11 
months, the highest (25 percent) among 12-23-month-olds and stabilizes after 2 years of age. 
Households with large family size have the highest proportion of underweight (27.5 percent) and 
severely underweight (21 percent) children and these proportions are over 50 percent and four 
times higher than the respective proportions for households with small family size. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 77
Table 6.5. Prevalence of underweight and mean weight-for-age z-scores among 
children under 5 years in the P1-ZOI
Characteristic
Percent 
Mean z-score
n1
Underweight
(<-2 SD)
Severely 
underweight
(<-3 SD)
Est. 95% CI Est. 95% CI Est. 95% CI
All children under 5 
years
22.5 20.5-24.6 7.1 5.8-8.6 -1.0 -1.1- -0.9 2,153
Child sex
Male 23.9 21.1-26.7 8.0 6.3-10.1 -1.0 -1.1- -0.9 1,071
Female 21.1 18.7-23.7 6.2 4.7-8.1 -1.0 -1.1- -0.9 1,082
Child age
0-11 months 14.8 10.8-20.0 5.9 3.4-9.8 -0.3 -0.5-0.0 273
12-23 months 25.3 21.5-29.5 7.8 5.8-10.5 -1.0 -1.2- -0.9 450
24-35 months 22.1 18.2-26.5 6.9 4.6-10.2 -1.0 -1.1- -0.8 442
36-47 months 22.3 18.9-26.0 7.7 5.7-10.3 -1.1 -1.2- -1.0 516
48-59 months 24.8 21.2-28.9 6.6 4.6-9.4 -1.3 -1.4- -1.2 472
Caregiver’s educational attainment2
No education 22.5 20.5-24.6 7.1 5.8-8.7 -1.0 -1.1- -0.9 2,148
Less than primary ^ ^ ^ ^ ^ ^ 4 
Primary ^ ^ ^ ^ ^ ^ 0
Secondary or more ^ ^ ^ ^ ^ ^ 0 
Gendered household type
Male and female adults 22.1 20.2-24.2 6.9 5.6-8.5 -1.0 -1.1- -0.9 1,976
Female adult(s) only 24.5 17.9-32.5 7.9 4.5-13.4 -1.2 -1.4- -1.0 157
Male adult(s) only ^ ^ ^ ^ ^ ^ 17
Child(ren) only (no 
adults)
^ ^ ^ ^ ^ ^ 2 
Household size
Small (1-5 members) 18.0 15.2-21.3 4.8 3.4-6.7 -1.0 -1.1- -0.8 732
Medium (6-10 members) 24.6 22.0-27.4 7.6 6.1-9.4 -1.0 -1.1- -0.9 1,346
Large (11+ members) 27.5 16.4-42.4 21.0 11.5-35.1 -0.9 -1.6- -0.3 74
Household hunger
Little to no hunger 22.5 10.5-24.6 7.2 5.9-8.8 -1.0 -1.1- -0.9 2,084
Moderate or severe 
hunger 20.7 13.5-30.5 4.1 1.4-11.1 -0.8 -1.1- -0.4 68
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 
^ Results not statistically reliable, n<30. 1 Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample size 
reflects this loss in observations; therefore, subsamples may not total to the aggregate sample size. 2 The ZOI Survey identifies the primary caregiver of each age-eligible child. This person is likely, but not necessarily, the child’s biological mother. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 78
Table 6.6 compares children’s anthropometry results between the phase one baseline and end￾line ZOI Surveys. 
Table 6.6 indicates that all children’s anthropometry measures showed considerable 
improvement in the end-line relative to the baseline. The prevalence of stunted children under 5 
years of age is 11.5 percentage points lower in the end-line than in the baseline or the prevalence 
of stunting declined by 23 percent during the period. Similarly, the proportion of wasted children 
in the end-line was 4.4 percentage points less than the proportion in the baseline, and this was a 
decline of 31 percent in the prevalence of wasting during the period. Moreover, the proportion 
of underweight children in the end-line was 9 percentage points less than that in the baseline or 
the prevalence of underweight children declined by 28.6 percent during the period. 
Table 6.6 Comparison of the nutritional status of children between the Feed the 
Future phase one baseline and end-line ZOI Surveys 
 Baseline (2013) End-line (2018) p￾valueb Sig.c Characteristic % 95% CI na % 95% CI na
Prevalence of stunted children under 5 years
All children under 5 
years of age 50.3 47.1-53.5 1,553 38.9 36.1-41.8 1,962 0.000 ***
Child’s sex
Male 52.9 48.7-57.1 778 38.8 34.9-42.8 976
Female 47.7 43.6-51.8 775 39.0 35.8-42.3 987
Child’s age 
0-11 months 39.8 29.6-50.9 106 19.6 14.5-25.9 234
12-23 months 52.0 46.0-57.9 274 41.0 35.8-46.4 398
24-35 months 53.3 47.1-59.4 331 43.7 38.1-49.5 406
36-47 months 52.9 48.0-57.7 406 42.1 37.0-47.5 483
48-59 months 47.0 41.2-53.0 436 39.1 34.0-44.4 442
Woman’s education 
No education 50.4 47.1-53.6 1,551 38.8 36.0-41.7 1,958
Less than primary ^ ^ 1 ^ ^ 4 
Primary ^ ^ 0 ^ ^ 0 
Secondary or more ^ ^ 0 ^ ^ 0 
Gendered household 
type
Male and female adults 50.2 46.7-53.6 1,360 39.1 36.2-42.0 1,808
Female adults only 50.8 42.3-59.3 162 33.6 25.7-42.5 137
Male adults only 54.0 36.4-70.6 31 ^ ^ 15
Children only ^ ^ 0 ^ ^ 2 
Prevalence of wasted children under 5 years
All children under 5 
years of age 14.1 11.9-16.6 1,471 9.7 7.8-12.0 1,916 0.010 * 
Child’s sex
Male 14.3 11.8-17.3 733 11.0 8.7-13.9 950
Female 13.9 11.1-17.2 738 8.4 6.3-11.2 966
Child’s age 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 79
 Baseline (2013) End-line (2018) p￾valueb Sig.c Characteristic % 95% CI na % 95% CI na
0-11 months 14.9 9.0-23.5 97 13.3 9.0-19.2 221
12-23 months 13.8 9.4-19.8 257 12.1 9.4-15.6 391
24-35 months 13.5 9.5-18.9 318 7.5 4.7-11.8 396
36-47 months 14.9 11.7-18.9 386 7.7 5.1-11.4 479
48-59 months 13.7 9.9-18.6 413 10.0 7.1-13.7 429
Woman’s education 
No education 14.1 11.9-16.6 1,469 9.7 7.8-12.0 1,913
Less than primary ^ ^ 1 ^ ^ 2 
Primary ^ ^ 0 ^ ^ 0 
Secondary or more ^ ^ 0 ^ ^ 0 
Gendered household 
type
Male and female adults 14.1 11.8-16.9 1,292 9.5 7.6-11.8 1,763
Female adults only 12.5 8.1-18.9 151 12.9 7.8-20.7 136
Male adults only ^ ^ 28 ^ ^ 15
Children only ^ ^ 0 ^ ^ 1 
Prevalence of underweight children under 5 years
All children under 5 
years of age 31.5 28.6-34.6 2,102 22.5 20.5-24.6 2,153 0.000 ***
Child’s sex
Male 32.2 29.1-35.3 1,050 23.9 21.2-26.7 1,071
Female 30.9 26.9-35.3 1,052 21.1 18.8-23.7 1,082
Child’s age 
0-11 months 28.7 21.4-37.2 163 14.8 10.8-20.0 273
12-23 months 33.2 28.6-38.3 397 25.3 21.5-29.5 450
24-35 months 33.0 28.4-37.9 478 22.1 18.2-26.5 442
36-47 months 31.0 26.9-35.5 534 22.3 18.9-26.0 516
48-59 months 30.4 26.1-35.0 530 24.8 21.2-28.9 472
Woman’s education 
No education 31.6 28.7-34.7 2,099 22.5 20.5-24.6 2,148
Less than primary ^ ^ 1 ^ ^ 4 
Primary ^ ^ 0 ^ ^ 0 
Secondary or more ^ ^ 0 ^ ^ 0 
Gendered household 
type
Male and female adults 31.8 28.8-35.0 1,866 22.1 20.2-24.2 1,976
Female adults only 29.2 22.9-36.3 199 24.5 17.9-32.5 157
Male adults only 29.5 16.3-47.4 37 ^ ^ 17
Children only ^ ^ 0 ^ ^ 2 
Source: Feed the Future Baseline Survey 2013, Ethiopia; Feed the Future End-line ZOI Survey, Ethiopia 2018. a Records missing information for the disaggregate variables have been excluded from the disaggregated estimates. The unweighted sample size 
reflects this loss in observations; therefore, subsamples may not total to the aggregate sample size. b
 A Pearson’s chi-squared test of statistical significance in differences in proportions between the baseline and 2018-2019 estimates was 
conducted. c
 Differences found to be statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 80
7. Women’s Empowerment in Agriculture
This chapter presents findings related to the WEAI. Although women play a prominent role in 
agriculture, they face persistent economic and social constraints. Because of this, women’s 
empowerment is a main focus of Feed the Future. Empowering women is particularly important 
to achieving the Feed the Future objectives of inclusive agriculture sector growth and improved 
nutritional status. The WEAI was originally developed to track the change in women’s 
empowerment that occurs as a direct or indirect result of interventions under Feed the Future
and as a programming tool to identify and address the constraints that limit women’s full 
engagement in the agriculture sector.21 Data to calculate the WEAI was collected in both the 
Feed the Future Ethiopia phase one baseline and end-line ZOI Surveys. All the results related to 
the WEAI are presented here. 
7.1 Overview
The WEAI measures empowerment across five domains, which are presented in Table 7.1, 
together with their definitions, corresponding empowerment indicators and what it means to 
achieve adequacy in each empowerment indicator. The Production domain reflects individuals’ 
ability to provide input and autonomously make decisions about agricultural production. The 
Resources domain reflects individuals’ ownership of productive resources and input into decisions 
related to credit. The Income domain reflects individuals’ ability to direct the financial resources 
derived from productive economic activities and durable goods. The Leadership domain reflects 
individuals’ social capital and comfort speaking in public within their community. The Time domain 
reflects individuals’ workload and satisfaction with leisure time. An individual is identified as 
“empowered” by WEAI if he or she achieves adequacy in at least 80 percent of the weighted 
indicators (equivalent to four out of five domains).
21 Alkire, S. Malapit, H., et al. (2013).
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 81
Table 7.1 WEAI domains, indicators and definitions of adequacy
Domain Definition of domain Indicators Percent with adequate achievement
Production
Sole or joint decision-making
over food and cash crop 
farming, livestock, and fisheries, 
and autonomy in agricultural 
production
Input in productive 
decisions
Respondents have at least some input into 
or can make own personal decisions in at 
least two decision-making areas
Autonomy in 
production
Respondent has an RAI (Relative Autonomy 
Indicator)22 greater than 1 in at least one of 
the four areas of decision-making.
Resources
Ownership, access to and 
decision-making power over 
productive resources such as 
land, livestock, agricultural 
equipment, consumer durables 
and credit
Ownership of 
assets
Respondents own alone or jointly at least 
two small assets or one large asset.
Purchase, sale or 
transfer of assets
Respondents decide to sell, give away or 
rent at least one asset, but not only chickens 
and non-mechanized farming equipment
Access to and 
decisions on credit
Respondents decide alone or jointly whether 
to borrow or what to do with the money or 
item borrowed.1
Income Sole or joint control over 
income and expenditures
Control over use 
of income
Respondents have some input into decisions 
on income, but not only minor household 
expenditures
Leadership
Membership in economic or 
social groups and comfort in 
speaking in public
Group member Respondents are active members of at least 
one group in their community.
2
Speaking in public Respondents feel comfortable speaking in at 
least one public setting
Time
Allocation of time to productive 
and domestic tasks and 
satisfaction with the available 
time for leisure activities
Workload
Respondents spend less than 10.5 hours 
performing work activities in a 24-hour 
period.3
Leisure Respondents rate satisfaction level as at least 
five out of 10 1 Respondents who live in households that did not access credit are considered inadequate on access to credit and decisions on credit. 2 Respondents who report that no groups exist in their communities or who are not aware of any groups in their community are 
considered inadequate on group membership. 3 Respondents who reported the 24 hours preceding the survey to being an atypical workday are excluded. 
The WEAI aggregates information collected for each domain into a single empowerment index 
indicator. The index comprises two sub-indices: the Five Domains of Empowerment Index (5DE), 
which measures the empowerment of women across the five WEAI domains, and the Gender 
Parity Index (GPI), which measures the relative empowerment of men and women within the 
household. The WEAI is computed as a weighted sum of the 5DE and GPI.
The WEAI survey module is administered to the male primary adult decision-maker and the 
female primary adult decision-maker in each household so that the 5DE profiles of women and 
men in the same household can be compared. The primary adult decision-makers are individuals 
22 The Relative Autonomy Indicator (RAI) measures the ability of a person to act on what he or she values, to have 
his or her own intrinsic motivations prevail over motivations to please others or avoid punishment, for example. 
This indicator probes the person’s own understanding of the situation and enables the respondent to easily explain 
the different motivations that influence activities (Alkire 2007).
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 82
age 18 or older who self-identify as the primary decision-makers during the collection of the 
household roster information.23 There can only be one female and one male primary adult 
decision-maker in each household; however, a household can have only a female primary adult 
decision-maker or only a male primary adult decision-maker. If there are no adults 18 or older 
in the household, the household will not have any primary adult decision-makers. The 
questionnaire for the WEAI section is administered separately to the primary and secondary 
respondent, where the primary respondent in this case refers to the primary female decision￾maker age 18 or older. We have a total of 3,252 female adult decision-makers and 2,527 male 
adult decision-makers (secondary respondent).
This chapter presents findings on the WEAI, the 10 WEAI empowerment indicators, and the two 
WEAI sub-indices. Appendix 2.3 presents more information on the WEAI, including the survey 
questions that are used to elicit the data required to establish adequacy for each WEAI 
empowerment indicator and the adequacy criteria definitions for each WEAI empowerment 
indicator. For additional details refer to IFPRI (2013).
24
Section 7.1 includes an overview of the empowerment results: 1) achievement of adequacy for 
each WEAI empowerment indicator; 2) 5DE, GPI and WEAI scores; and 3) contribution of each 
WEAI empowerment indicator to the disempowerment of women and men. Subsequent sections 
in this chapter present more detailed results by empowerment domain.
Table 7.2a and Table 7.2b present the average percentage of women and men who are 
disempowered and achieving adequacy in the 10 WEAI indicators at both baseline and at end-line
using censored headcounts. The purpose of reporting on the average percentage of women who 
are disempowered achieving adequacy across the 10 WEAI indicators overall, which is a Feed the 
Future context indicator, and for the individual indicators is two-fold: to bring greater attention 
to the composition of empowerment and disempowerment, and to identify the individual 
indicators that present the greatest constraints to empowerment for women and men. 
On average, 39 percent of female primary adult decision-makers are disempowered and achieved 
adequacy in the 10 WEAI indicators during the end-line. The corresponding figure (24 percent) 
is lower for male primary adult decision-makers. Relative to non-youths, a higher proportion of 
both female and male youth (18-29-year-olds) are disempowered and achieved adequacy in the 
10 WEAI indicators. 
Table 7.2a: Comparison of average percent of primary adult decision-makers in the PI￾ZOI achieving adequacy in the 10 WEAI indicators, using censored 
23 The respondents to the WEAI survey module are only the primary adult decision-makers in the household and, 
therefore, may not be representative of the entire female and male populations in the surveyed area. 24 IFPRI (2013). https://www.feedthefuture.gov/the-womens-empowerment-in-agriculture-index/.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 83
headcounts, by sex and age, Feed the Future phase one baseline and end￾line ZOI Surveys
Notes: Estimates are based on primary adult decision-makers who are de jure household members.
The Feed the Future ZOI context indicator, “Average percent of women achieving adequacy across the 10 indicators of the WEAI,” is 
calculated as the sum of the censored headcount ratios for primary adult female decision-makers for each of the 10 WEAI indicators, divided by 
10 (the number of indicators.) It shows the average across the 10 indicators of proportion of primary adult female decision-makers in the ZOI 
population who are disempowered but still achieved adequacy in an individual WEAI indicator.
Table 7.2b presents the censored headcount ratios of individuals achieving adequacy in the 10 
WEAI indicators, disaggregated by sex and age. 
The proportion of women who are not yet empowered but have adequate achievement is higher 
relative to men during the end-line for all WEAI indicators except speaking in public and for the 
most part, this pattern holds for both age groups as well as during the baseline. The results 
indicate that primary women decision-makers have adequacy levels of higher than 50 percent 
during the end-line in control over use of income, input in productive decisions, ownership of 
assets, purchase and sale or transfer of assets, and autonomy in production. Primary men adult 
decision-makers have adequacy levels of less than 50 percent in indicators. Both primary men and 
women adult decision-makers had the lowest achievement in access to and input into decisions 
on credit. 
WEAI indicator
Baseline (2013) End-line(2018)
% 95% CI n % 95% CI n 
Womenc 
Averaged
 40 38-42 3,466 39 39-39 3,251
Age category 
18-29 39 37-42 1,247 40 40-40 647
30+ 40 38-41 2,219 38 38 -38 2,604
Mene 
Averaged 32 29-34 2,998 24 24-24 2,542
Age category 
18-29 35 32-37 682 32 32-32 270
30+ 31 28-33 2,316 23 23-23 2,272
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 84
Table 7.2b: Comparison of adequate achievement in each WEAI indicator in PI-ZOI 
using censored headcount ratios, by sex and age, Feed the Future phase 
one baseline and end-line ZOI Surveys
WEAI indicator and characteristic Baseline (2013) End-line (2018) 
% 95% CI n % 95% CI n Diff.
Women 
Input in productive decisions 67.1 65.1-68.3 2,178 63.4 61.7-65.2 1,943 -3.7
Age category
18-29 66.7 63.8-69.2 794 64.8 61.1-68.7 397 -1.9
30+ 67.3 64.8-68.8 1,384 63.1 61.6-65.4 1,546 -4.2
Ownership of assets 55.7 53.6-57.0 1,806 62.6 60.5-63.9 1,906 6.9
Age category
18-29 55.5 52.2-57.8 657 67.1 62.7-70.2 407 11.6
30+ 55.8 53.3-57.6 1,149 61.5 59.6-63.5 1,499 5.7
Access to and decisions on credit 5.9 5.1-6.7 193 11.1 9.4-11.6 322 5.2
Age category
18-29 6.1 4.6-7.2 71 14.2 10.4-15.7 80 8.1
30+ 5.8 4.9-6.9 122 10.3 8.7-11.1 242 4.5
Control over income 67.3 65.4-68.6 2,187 63.9 62.3-65.7 1,959 -3.4
Age category
18-29 68.7 65.6-70.9 815 67.2 63.6-71.0 412 -1.5
30+ 66.6 64.2-68.3 1,372 63.2 61.6-65.5 1,547 -3.4
Group membership 28.2 27.0-30.1 933 22.3 21.0-23.9 689 -5.9
Age category
18-29 27.7 26.1-31.2 342 22.2 19.2-25.8 138 -5.5
30+ 28.4 26.6-30.5 591 22.3 20.9-24.3 551 -6.1
Workload 35.2 33.5-36.7 1,148 42.3 40.9-44.5 1,309 7.1
Age category
18-29 31.7 29.3-34.6 382 37.7 34.6-42.4 236 6 
30+ 37.3 34.9-39.1 766 43.5 42.1-46.1 1,073 6.2
Autonomy in production 54 52.6-56.0 1,773 52.7 50.5-54.1 1,602 -1.3
Age category
18-29 53.5 51.5-57.2 649 56.9 52.9-60.8 348 3.4
30+ 54.3 51.5-57.2 1,124 51.7 49.5-53.5 1,254 -2.6
Purchase, sale or transfer of assets 52.7 50.3-53.7 1,699 61.3 59.0-62.5 1,860 8.6
Age category
18-29 50.6 46.9-52.6 594 64.4 59.7-67.4 389 13.8
30+ 53.9 51.2-55.5 1,105 60.5 58.5-62.4 1,471 6.6
Speaking in public 31.2 30.2-33.4 1,039 27.2 26.4-29.6 857 -4 
Age category
18-29 29.8 28.3-33.5 369 30.3 27.5-34.9 191 0.5
30+ 32.0 30.3-34.4 670 26.4 25.6-29.1 666 -5.6
Leisure 32.7 30.5-33.8 1,051 20.6 19.3-22.2 635 -12.1
Age category
18-29 36.5 32.9-38.4 426 23.6 21.4-28.3 152 -12.9
30+ 30.1 28.2-32.2 625 19.9 18.3-21.4 483 -10.2
Men
Input in productive decisions 57.9 55.4- 58.9 1,656 41.9 40.6-44.5 1,054 -16
Age category
18-29 62.4 58.6-66.1 408 52 47.3-59.5 139 -10.4
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 85
30+ 56.6 53.6-57.7 1,248 40.7 39.2-43.3 915 -15.9
Ownership of assets 40.5 38.4-42.0 1,165 36.9 35.7-39.5 931 -3,6
Age category
18-29 45.7 42.9-50.5 408 40.8 35.9-47.9 109 -4.9
30+ 38.9 36.3-40.3 1,248 36.4 35.1-39.1 822 -2.5
Access to and decisions on credit 3.5 2.6-3.9 93 5.5 4.5-6.2 133 2 
Age category
18-29 3.5 2.1-4.9 23 7.8 4.4-10.9 20 4.3
30+ 3.5 2.4 -3.8 70 5.2 4.2-6.0 113 1.7
Control over income 56.9 54.3-57.9 1,625 41.6 40.3-44.2 1,046 -15.3
Age category
18-29 61.9 58.4-65.9 407 51.8 46.9-59.2 138 -10.1
30+ 55.4 52.3-56.4 1,218 40.5 38.9-43.0 908 -14.9
Group membership 22.6 20.7-23.7 643 12.5 11.2-13.8 310 -10.1
Age category
18-29 25.5 22.7-29.5 171 16.7 11.6-20.6 42 -8.8
30+ 21.7 19.4-22.7 472 11.9 10.7-13.4 268 -9.8
Workload 35.4 33.1-36.6 1,010 33.5 32.4-36.2 850 -1.9
Age category
18-29 38.0 34.1-41.5 248 48.6 44.6-56.8 132 10.6
30+ 34.7 32.0-35.9 762 31.7 30.4-34.3 718 -3 
Autonomy in production 39.7 37.9-41.5 1,151 29.8 28.4-32.1 749
Age category
18-29 44.5 41.8-49.5 299 42.1 37.0-49.1 112 -2.4
30+ 38.2 36.0-40.0 852 28.3 26.8-30.6 637 -9.9
Purchase, sale or transfer of assets 35.6 33.3-36.8 1,015 35.2 34.0-37.8 889 -0.4
Age category
18-29 37.2 33.4-40.8 243 33.1 28.4-40.0 89 -4.1
30+ 35.2 32.5-36.4 772 35.5 34.1-38.1 800 0.3
Speaking in public 40.4 38.6-42.2 1,169 27.9 27.0-30.6 714 -12.5
Age category
18-29 42.1 38.8-46.4 279 36.2 31.4-43.2 97 -5.9
30+ 39.9 37.7-41.7 890 26.9 25.9-29.7 617 -13
Leisure 25.2 22.6-25.7 701 11.9 10.7-13.3 298 -13.3
Age category
18-29 30.3 26.3-33.2 195 21.3 17.2-27.4 58 -9 
30+ 23.6 20.8-24.3 506 10.8 9.5-12.1 240 -12.8
Note: Estimates are based on primary adult decision-makers who are de jure household members.
Source: Feed the Future Ethiopia ZOI Survey 2013 and 2018
Table 7.2c compares the 5DE scores for female and male primary adult decision-makers 
between the phase one baseline and end-line ZOI Surveys. The 5DE shows that 29 percent of 
women are empowered in the end-line. The 71 percent that are not yet empowered have 
acquired adequacy in 63 percent of the indicators. Thus, the women’s empowerment index (5DE) 
is 0.738. The 5DE for men indicates that 61 percent of men are empowered and the adequacy 
score for the men that are not empowered is 67 percent. The 5DE scores in the end-line are 
higher than those in the baseline for both men and the women. Based on the five domains of 
empowerment, the proportion achieving empowerment is lower and the level of 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 86
disempowerment is higher among women compared to men at both the baseline and end-line. 
Furthermore, the decrease in the disempowerment index during the baseline-end-line period is 
more than twice higher for men than for women while growth in proportion of men achieving 
empowerment was almost 50 percent higher than women. 
Table 7.2c. Comparison of WEAI, 5DE and GPI scores, and average empowerment gap 
in the PI-ZOI, Feed the Future phase one baseline and end-line ZOI 
Surveys 
Indicator Baseline (2013) End-line (2018)
Diff.
p￾valuea Sig.b Est. 95% CI n Est. 95% CI n 
Womenc 
WEAI score 0.70 0.69-0.72 3265 0.75 0.75-0.76 3,047 0.05 0.000 ***
Age category
18-29 0.71 0.68-0.71 1194 0.72 0.72-0.73 612 0.02 0.000 ***
30+ 0.69 0.69-0.72 2071 0.76 0.76-0.77 2,435 0.07 0.000 ***
5DE score 0.68 0.74 0.06 
Women achieving 
empowerment (%) 22.3 718 29.2 884 6.9 0.000 ***
Weighted indicators in which 
disempowered individuals have 
adequate achievements (%)d
59.0 3,466 63.0 3,047 -1.0 0.000 ***
Number of women 3466 3252
GPI score 0.87 4436 0.90 3,726 0.03
Women achieving gender parity 
(%) 44.8 1437 48.7 1,463 3.9 0.099 * 
Age category
18-29 43.4 513 42.3 257 -1.1 0.77 n/s
30+ 45.9 932 50.8 1,218 4.9 0.066 * 
Average empowerment gap 23.5 19.3 0 ***
Number of dual-adult 
householdse 2,218 1,863
Menf
5DE score 0.80 0.87 0.07
Men achieving empowerment (%) 42.2 912 61.0 1,116 18.8
Weighted indicators in which 
disempowered individuals have 
adequate achievements (%)d
65.0 2,998 67.0 2,542 -8.0
Number of men 2,218 1,863
Source: Feed the Future Feedback ZOI Survey, Ethiopia 2013; Feed the Future End-line ZOI Survey 2018, Ethiopia. a Significance tests were performed to determine whether a difference exists between the baseline and end-line estimates. An F-test of 
statistical significance in differences in proportions between the baseline and 2018 estimates was conducted. b Differences found to be statistically significant are indicated by level: * p<0.05, ** p<0.01, *** p<0.001. c Estimates for women exclude households that do not have a primary adult female decision-maker or that have missing or incomplete indicator 
data.
d Also referred to as the average adequacy score. e Number of households with both a male primary adult decision-maker and a female primary adult decision-maker. f Estimates for men exclude households that do not have a primary adult male decision-maker or that have missing or incomplete indicator data.
Notes: 
Estimates are sample-weighted; numbers of observations are unweighted.
Estimates are based on primary adult decision-makers who are de jure household members.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 87
The GPI shows that 49 percent of the women have gender parity with the primary males in their 
households at the end-line. This has shown a 4-percentage points improvement from the baseline 
value. The 51 percent women that did not achieve gender parity have an empowerment gap of 
19.3 percent with the males in their households. This gap is 4.2 percentage points lower in the 
end-line relative to the 23.5 percent at the baseline. Thus, the overall GPI is 0.901 in the end-line, 
with a 0.031 percentage points (3.6 percent) improvement over the baseline. The WEAI, which 
is a weighted average of the 5DE sub-index value of 0.738 and the GPI sub-index value of 0.901, 
is 0.754, which is an improvement of 0.052 (7 percent) in the end-line relative to its baseline value 
of 0.702.
Figure 7.1 shows the contribution of each empowerment indicator to the disempowerment of 
women and men. The indicators that contribute most to women’s disempowerment are lack of 
leisure, discomfort with speaking in public, lack of participation in groups, and lack of access to 
and decisions on credit. The indicators that contribute to men’s disempowerment are roughly 
similar with those that contribute to women’s disempowerment. Lack of leisure, lack of 
participation in groups, and lack of access to and decisions on credit are the three most important 
contributors to men’s disempowerment. Heavy work time requirements and discomfort speaking 
in public also contribute to disempowerment, but less so for men than for women.
Figure 7.1: Contribution of each indicator to the disempowerment of women and men in 
the P1-ZOI end-line Survey 
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 
Figure 7.2a shows the contribution of each of the five domains to the disempowerment of 
women. Figure 7.2b shows the same for men. According to Figure 7.2a, the domains that 
contribute most to women’s disempowerment are time, leadership and resources, accounting 
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0.20
0.22
0.24
0.26
0.28
Women Men
Disempowerment index (1 
– 5DE)
Leisure
Workload
Speaking in public
Group membership
Control over use of income
Access to and decisions on credt
Purchase, sale or transfer of assets
Ownership of assets
Autonomy in production
Input in productive decisions
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 88
for 35, 30, and 20 percent, respectively, of disempowerment of women. The composition of men 
disempowerment is similar to that of the women with time, leadership and resource domains 
accounting for roughly similar percentages to women’s disempowerment. 
Figure 7.2a: Contribution of each of the five domains to the disempowerment of women 
in the P1-ZOI end-line Survey 
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 
Figure 7.2b: Contribution of each of the five domains to the disempowerment of men in 
the P1-ZOI end-line Survey 
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 
7.2 Production
Adequacy in Production is measured by input into decisions about agricultural activities in which 
an individual participates and autonomy in agricultural production. Respondents are considered 
adequate in production if they make decisions alone, have input into most or all decisions, or feel 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 89
that they could make decisions if they wanted to for at least two agricultural activities and 
respondent has a relative autonomy index (RAI) greater than 1 in at least one of the four areas 
of agricultural decision-making.
Table 7.3 presents the percentages of female and male primary adult decision-makers who are 
involved in agricultural activities (i.e., food crop farming, cash crop farming, livestock raising or 
fishing), nonfarm economic activities, and wage or salaried employment. This table also presents 
the percentages of female and male primary adult decision-makers who have input into the 
decisions made about specific activities. 
During the end-line, around 84 percent of women participate in food crop farming, 73 percent in 
livestock raising and 59 percent in cash crop farming. The proportion of men that participated in 
the respective activities is relatively higher at around 94 percent, 78 percent and 67 percent. 
Although participation in nonfarm economic activities and wage employment is generally low, the 
percentage of men participating in nonfarm economic activities is much higher (almost double) 
compared to the women. Fishing is a very rare activity for both men and female decision-makers. 
The percentage of women making decisions in each activity is generally high (88 percent or higher
for all activities except fishing and nonfarm economic activities) but lower than the percentage of 
men that have input into decision-making. The gap between the percentage of men and women 
that participate in making decisions is the highest for nonfarm economic activities.
Table 7.3. Participation in economic activities and input into activity decision-making 
among female and male primary adult decision-makers in the P1-ZOI end￾line Survey 
Activity
Participates in activity Has input1 into decisions about 
activity
Female Male Female Male
Percent n2 Percent n3 Percent n4 Percent n4
All primary adult 
decision-makers 93.9 3,252 97.6 2,527 93.6 3,039 97.6 2,460
Type of agricultural activity
Food crop farming 83.9 3,252 94.4 2,527 88.9 2,691 96.3 2,367
Cash crop farming 58.9 3,252 67.3 2,527 89.7 1,863 96.3 1,659
Livestock raising 73.0 3,252 77.6 2,527 89.9 2,341 96.0 1,942
Fishing or fishpond culture 0.8 3,252 1.2 2,527 64.4 21 76.9 29
Nonfarm economic activities 11.1 3,252 13.9 2,527 87.9 351 92.1 350
Wage or salaried employment 5.8 3,252 12.1 2,527 79.2 185 95 300
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 1 Having input means that the individual reported having input into most or all decisions regarding the activity. 2 Estimates exclude households that do not have a female primary adult decision-maker or that have missing or incomplete indicator data. 3 Estimates exclude households that do not have a male primary adult decision-maker or that have missing or incomplete indicator data. 4 Estimates exclude individuals who do not participate in an activity or report that no decision was made.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 90
7.3 Resources
Adequacy in Resources is measured by three indicators: (1) ownership of assets, (2) Purchase, sale 
or transfer of assets, and (3) access to and decisions related to credit. Respondents are 
considered adequate in asset ownership if they own alone or jointly at least two small assets or 
one large asset. Respondents are considered adequate in the purchase, sale or transfer of assets 
if they decide to sell, give away or rent at least one asset, but not only chickens and non￾mechanized farming equipment. Respondents are considered adequate to access to credit if they 
decide alone or jointly whether to borrow cash or in-kind or what to do with the money or item 
borrowed. Table 7.4a presents the findings for ownership of productive resources. 
Agricultural land, large livestock and non-mechanized farm equipment are among the most 
commonly owned productive assets in the end-line with 91.5 percent, 77.4 percent and 71.5 
percent of the households owning the assets, respectively. A relatively lower proportion of the 
households own means of transportation (1.4 percent), mechanized farm equipment (0.8 
percent), and fishpond or fishing equipment (0.3 percent). A higher percentage of men own most 
assets compared to women. This gap is especially very large for cell phones. Around 39 percent 
of the men own cellphones while only 17.7 percent of women own cellphones. On the other 
hand, a higher percentage of women own poultry (chickens, ducks, turkeys and pigeons) and small 
consumer durables. Ownership of land is critical for livelihood in these rural communities. 
Surprisingly, the proportion of men that own agricultural land is only slightly (4.2 percentage 
points) higher than women while the proportion of women that own non-agricultural land is 
higher by about the same proportion.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 91
Table 7.4a. Household and female and male primary adult decision-maker ownership 
of productive resources in the P1-ZOI end-line Survey 
Type of resource
Someone in the 
household owns 
item
 Female owns 
solely or jointly Male owns 
solely or jointly
Percent n1 Percent n2 Percent n3
Agricultural land 91.5 5,779 79.1 3,252 83.3 2,527
Large livestock 77.4 5,779 69.5 3,252 76.1 2,527
Small livestock 48.0 5,779 42.5 3,252 46.6 2,527
Chickens, ducks, turkeys and pigeons 50.4 5,779 41.3 3,252 35.3 2,527
Fish pond or fishing equipment 0.3 5,779 0.2 3,252 0.3 2,527
Non-mechanized farm equipment 71.5 5,779 45.9 3,252 71.1 2,527
Mechanized farm equipment 0.8 5,779 0.5 3,252 1.1 2,527
Nonfarm business equipment 3.1 5,779 2.3 3,252 2.7 2,527
House or other structures 36.8 5,779 33.9 3,252 33.7 2,527
Large consumer durables 3.6 5,779 3.3 3,252 3.9 2,527
Small consumer durables 62.1 5,779 54.4 3,252 44.4 2,527
Cell phone 40.4 5,779 17.7 3,252 38.5 2,527
Non-agricultural land 2.9 5,779 2.6 3,252 3.0 2,527
Means of transportation 1.4 5,779 0.9 3,252 1.5 2,527
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 1 Estimates exclude households that have no primary adult decision-maker and that have missing or incomplete indicator data. Respondents 
who indicated “not applicable” are also excluded. 2 Estimates exclude households that do not have a female primary adult decision-maker or that have missing or incomplete indicator data. 3 Estimates exclude households that do not have a male primary adult decision-maker or that have missing or incomplete indicator data. 
Table 7.4b presents the percentage of women who can make a decision to purchase or to sell, 
give away, or rent owned items. Women are considered as having the ability to make a decision 
if they can solely make a decision or if they can make these decisions with others with any degree 
of input. The percentage of women and men who have input in decisions to purchase, sell, give 
or rent items is comparable for most items during the end-line. A relatively higher percentage of 
women have input in decisions related to poultry, nonfarm business equipment and small 
consumer durables. On the other hand, more men make decisions on farm equipment, cellphones 
and means of transportation compared to women. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 92
Table 7.4b. Household and female and male primary adult decision-maker control 
over productive resources in the P1-ZOI end-line Survey 
Type of resource
Someone in the 
household owns 
item
Female can decide to 
purchase/sell/give/rent 
items
Male can decide to 
purchase/sell/give/rent 
items
Percent n1 Percent n1 Percent n1
Agricultural land 91.5 5,779 90.9 2,954 92.7 2,319
Large livestock 77.4 5,779 92.5 2,488 94.1 2,057
Small livestock 48.0 5,779 90.7 1,558 92.3 1,312
Chickens, ducks, turkeys and pigeons 50.4 5,779 84.7 1,622 73.8 1,263
Fish pond or fishing equipment 0.3 5,779 80.5 9 90.6 9 
Non-mechanized farm equipment 71.5 5,779 69.9 2,173 94.3 1,900
Mechanized farm equipment 0.8 5,779 71.8 24 89.3 34
Nonfarm business equipment 3.1 5,779 77.7 102 67.3 99
House or other structures 36.8 5,779 94.5 1,152 94.3 881
Large consumer durables 3.6 5,779 93.2 132 94.2 122
Small consumer durables 62.1 5,779 90.1 1,980 71.5 1,602
Cell phone 40.4 5,779 48.9 1,334 81.3 1,231
Non-agricultural land 2.9 5,779 90.5 89 87.1 88
Means of transportation 1.4 5,779 63.7 47 83.1 48
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 1 Estimates exclude households that have no primary adult decision-maker and that have missing or incomplete indicator data. Respondents 
who indicated “not applicable” are also excluded.
Tables 7.5a and 7.5b show the third indicator of the Resources domain—access to and decision￾making on credit. Table 7.5a presents the percentage of female primary adult decision-makers 
who report that a member of her household received any loan in the past 12 months—overall 
and disaggregated by source. The percentages of households that received an in-kind loan (e.g., 
food items or raw materials) or a cash loan are also presented. The in-kind and cash loan 
categories are not mutually exclusive; a household could have received both types of loans. For 
female primary adult decision-makers living in households that received a loan, the table also 
presents the percentages who report having contributed to the decision to take the loan and the 
subsequent decisions on how to use the loan. Table 7.5b presents the same information for male 
primary adult decision-makers.
About 25 percent of women reported that their household received a loan from non￾governmental organization, informal or formal lenders, friends or relatives, or group-based 
microfinances during the end-line. The main sources of credit are friends or relatives, and group￾based microfinance. The majority (92 percent) of women contribute to credit decisions. 
Particularly, a slightly higher proportion contributed on the decision how to use the loan. Table 
7.5b indicates that about 29 percent of men reported that their households have taken credit. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 93
For both women and men, the major sources of credit are friends or relatives, and group-based 
microfinance. About the same proportion of men participated in the decision regarding loans as 
that of women. 
Table 7.5a. Access to credit among female primary adult decision-makers in the P-ZOI 
end-line Survey 
 
Credit source (percent)1
Any 
source 
(percent)
Non￾governmental 
organization
Informal 
lender
Formal 
lender
Friends or 
relatives
Group￾based 
microfinan
ce
Household received a loan 
Any loan 25.4 1.9 2.5 1.3 8.4 12.8
In-kind loan 2.1 1.3 0.4 0.2 0.4 0.5
Cash loan 23.7 0.7 2.3 1.2 8.1 12.7
n2 3,239 3,239 3,239 3,239 3,239 3,239
Primary adult decision-maker contributed to credit decision 
Any decision 92.0 73.9 88.1 94.8 88.4 95.1
On whether to borrow 85.9 59.9 84.6 92.6 82.1 90.3
On how to use loan 91.5 73.9 88.1 94.8 88.4 95.1
n3 755 64 82 41 276 380
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 1
 Percentages sum to more than 100 because loans may have been received from more than one source. 2 Estimates exclude households that do not have a female primary adult decision-maker or that have missing or incomplete indicator data. 3 Estimates exclude households that do not have a female primary adult decision-maker, that did not receive a loan, or that have missing or 
incomplete indicator data. 
Table 7.5b. Access to credit among male primary adult decision-makers in the P1-ZOI 
end-line Survey 
 
 Credit source (percent)1
Any source 
(percent)
Non￾governmental 
organization
Informa
l lender
Formal 
lender
Friends or 
relatives
Group￾based 
microfina
nce
Household received a loan 
Any loan 29.2 2.8 2.9 1.4 9.2 14.3
In-kind loan 2.4 1.3 2.8 1.3 8.9 14.1
Cash loan 27.1 1.6 0.2 0.1 0.4 0.4
n2 2,523 2,523 2,523 2,523 2,523 2,523
Primary adult decision-maker contributed to credit decision 
Any decision 91.7 74.3 96.8 84.6 93.1 93.9
On whether to borrow 87.1 63.2 95.3 84.6 91.3 87.6
On how to use loan 91.2 74.3 95.8 84.6 92.6 93.2
n3 674 70 73 31 228 336
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 1
 Percentages sum to more than 100 because loans may have been received from more than one source. 2 Estimates exclude households that do not have a male primary adult decision-maker or that have missing or incomplete indicator data. 3 Estimates exclude households that do not have a male primary adult decision-maker, that did not receive a loan, or that have missing or 
incomplete indicator data.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 94
7.4 Income
Adequacy in Income is measured by input into decisions related to income. Respondents are 
considered adequate if they have substantial input into most or all decisions or feel they can make 
a decision for at least one economic activity. Table 7.6 shows the percentages of female and 
male primary adult decision-makers who have input into the decisions made regarding the use of 
income derived from an economic activity. Ninety-three percent of the women who participated 
in one or more of the economic activities had input into decision-making on use of income from 
the activity. While the percentage of women making income decisions is always lower relative to 
the percentage of men, it is very close in most activities except for fishing and nonfarm economic 
activities. 
Table 7.6. Input into decision-making on use of income among male and female 
primary adult decision-makers in the P1-ZOI end-line Survey 
Activity
Has input1 into income from activity
Female Male
Percent n2 Percent n3
Total 93.2 3,201 97.5 2,515
Food crop farming 89.3 2,687 95.8 2,366
Cash crop farming 90.1 1,866 96.3 1,657
Livestock raising 90.3 2,330 96.6 1,932
Fishing or fishpond culture 73.9 20 83.0 28
Nonfarm economic activities 80.0 349 93.8 350
Wage or salaried employment 88.1 184 91.9 297
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 1 Having input means that the individual reported having input into most or all decisions regarding the use of income generated from the 
activity. 2 Estimates exclude households that do not have a female primary adult decision-maker or that have missing or incomplete data. Estimates 
also exclude respondents who do not participate in the activity or who report that no decision was made regarding the activity. 3 Estimates exclude households that do not have a male primary adult decision-maker or that have missing or incomplete data. Estimates also 
exclude respondents who do not participate in the activity or who report that no decision was made regarding the activity. 
7.5 Leadership 
Adequacy in Leadership is measured through an individual’s active involvement with community 
organizations and speaking in public. The first indicator of the Leadership domain is an individual’s 
participation in a community organization. Respondents are considered adequate in this indicator 
if they are active members of at least one community organization. The second indicator is an 
individual’s ease speaking in public and the respondent is considered adequate if she/he feels 
comfortable speaking in at least one public setting. 
Table 7.7a shows the percentages of female and male primary adult decision-makers who are 
active members of organizations in their community. About 44 percent of women and 55 percent 
of men are involved in some type of community group. The highest percentage of women and 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 95
men participate in religious groups at 33.5 and 40 percent, respectively. The percentage of men 
participating in all groups is higher than women; particularly, the difference is highest in 
membership of local government. 
Table 7.7a. Group membership among female and male primary adult decision-makers 
in the P1-ZOI end-line Survey 
Type of group
Is an active group member
Female Male
Percent1 n2 Percent1 n3
All primary adult decision-makers 43.9 3,252 55.2 2,527
Agricultural producers’ group 0.7 3,252 2.3 2,527
Water users’ group 9.6 3,252 13.7 2,527
Forest users’ group 0.9 3,252 3.3 2,527
Credit or microfinance group 3.7 3,252 5.3 2,527
Mutual help or insurance group 11.0 3,252 16.3 2,527
Trade and business association 0.2 3,252 0.5 2,527
Civic or charitable group 0.4 3,252 1.4 2,527
Local government 6.1 3,252 14.7 2,527
Religious group 33.5 3,252 40.0 2,527
Other women’s group 7.6 3,252 6.0 2,527
Other 0.0 3,252 0.2 2,527
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia 2019 1 The denominator for this percentage includes all interviewed individuals, even those who reported that no group exists or that they are 
unaware of the existence of a group in their community. Individuals who report that none of the groups exist or who are unaware of any 
groups are counted as having inadequate achievement of this empowerment indicator. 2 Estimates exclude households that do not have a female primary adult decision-maker or that have missing or incomplete data. 3 Estimates exclude households that do not have a male primary adult decision-maker or that have missing or incomplete data.
The second indicator of the domain is an individual’s ease speaking in public, which is measured 
by three questions related to the level of difficulty an individual faces when voicing her/his opinion 
regarding community decisions. Table 7.7b shows that the proportion of males who feel 
comfortable speaking in public about community related issues (79 percent) is 32 percentage 
points higher than females (47 percent). The difference between the percentage of females and 
males with ease of public speaking in any one of the three specific issues is about the same as the 
difference above. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 96
Table 7.7b. Comfort with speaking in public among female and male primary adult 
decision-makers in the P1-ZOI end-line Survey 
Topics for public discussion
Comfortable speaking in public about 
selected topics
Female Male
percent n1 percent n2
All primary adult decision-makers 47.0 3,252 78.6 2,527
Topics 
To help decide on infrastructure to be built in the community 43.4 3,252 75.7 2,527
To ensure proper payment of wages for public works or 
other similar programs 41.8 3,252 71.9 2,527
To protest the misbehavior of authorities or elected officials 42.0 3,252 71.8 2,527
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia 2019 1 Estimates exclude households that do not have a female primary adult decision-maker or that have missing or incomplete data. 2 Estimates exclude households that do not have a male primary adult decision-maker or that have missing or incomplete data.
7.6 Time
Adequacy in the last domain, Time, assesses the workloads of female and male primary adult 
decision-makers, as directly measured using a time allocation log. Respondents are considered 
adequate if they spend 10.5 hours or less performing work activities from 5:00 a.m. in the morning 
to 9:00 p.m. in the evening. Table 7.8a shows the percentages of female primary adult decision￾makers that performed the listed activities the day prior to the survey and the average number 
of hours that they spent performing each activity. Table 7.8b shows the same for male primary 
adult decision-makers. The percentages indicate those individuals who reported performing the
activity, irrespective of the length of time that they spent performing the activity. The average 
hours spent performing an activity is the average across all individuals, assigning zero hours to 
individuals who did not perform an activity. Both primary and secondary activities are presented. 
Individuals were allowed to report up to two activities per time use increment (30 minutes). If 
the individual reported performing two activities, one was designated as the primary activity, and 
the other as a secondary activity. Some individuals may not have reported a secondary activity 
for each 30-minute period.
Cooking and domestic cleaning duties take a lion share of women's time on average. Women 
also contribute a considerable share of their time to farm and livestock-related activities. In 
contrast, the majority of men’s time is spent on farming and livestock-related activities followed 
by time with family.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 97
Table 7.8a. Time allocation among female primary adult decision-makers in the P1-
ZOI end-line Survey 
Activity
Primary activity Secondary activity1
Percent of 
women
Mean hours 
devoted
Percent of 
women
Mean hours 
devoted
Sleeping and resting 99.9 3.66 12.1 2.17
Eating and drinking 98.6 2.33 19.7 1.31
Personal care 59.2 0.58 8.0 0.35
School and homework 1.3 0.02 0.4 0.01
Work as employed 1.2 0.06 0.3 0.06
Own business work 2.3 0.11 0.5 0.08
Farming/livestock/fishing 42.8 1.79 5.5 1.05
Shopping/getting services 6.7 0.23 2.0 0.13
Weaving, sewing, textile care 0.5 0.01 1.2 0.01
Cooking 77.7 2.05 16.1 1.08
Domestic work (cleaning) 75.9 3.20 19.5 1.68
Gathering fuelwood/fetching water 24.7 0.34 3.8 0.19
Care for children/adults/elderly 15.3 0.39 10.2 0.39
Travel and commuting 8.7 0.18 1.6 0.12
Watching TV/listening to radio/reading 1.2 0.02 0.6 0.01
Sitting with family 42.1 1.00 11.4 0.59
Hobbies/sports/exercising/reading 3.0 0.06 0.6 0.04
Religious activities 18.0 0.43 3.1 0.26
Social/community activities 16.3 0.44 2.4 0.25
Selling 2.7 0.08 0.4 0.00
Other 0.4 0.01 0.0 0.05
n 3,252 3,252 3,252 3,252
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 1 Respondents were allowed to report up to two activities per time use increment (30 minutes) from 5:00 a.m. to 9:00 p.m. prior to the survey. 
If two activities were reported, one was designated as the primary activity, and the other as a secondary activity. Some women may not have 
reported a secondary activity for each 30-minute period. 2 Estimates exclude households that do not have a female primary adult decision-maker or that have missing or incomplete data.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 98
Table 7.8b. Time allocation among male primary adult decision-makers in the P1-ZOI 
end-line Survey 
Activity
Primary activity Secondary activity1
Percent of 
men
Mean hours 
devoted
Percent of 
men
Mean hours 
devoted
Sleeping and resting 99.5 3.74 13.8 3.14
Eating and drinking 98.5 2.49 19.2 2.15
Personal care 62.3 0.73 8.4 0.62
School and homework 0.9 0.02 0.5 0.02
Work as employed 3.6 0.23 0.8 0.19
Own business work 3.2 0.16 0.4 0.13
Farming/livestock/fishing 81.3 5.38 9.2 4.64
Shopping/getting services 5.0 0.22 2.1 0.18
Weaving, sewing, textile care 0.7 0.02 1.5 0.01
Cooking 5.0 0.10 1.1 0.07
Domestic work (cleaning) 8.3 0.22 1.7 0.17
Gathering fuelwood/fetching water 5.8 0.09 0.8 0.07
Care for children/adults/elderly 3.8 0.06 1.5 0.05
Travel and commuting 13.3 0.33 2.0 0.27
Watching TV/listening to radio/reading 3.3 0.06 2.1 0.06
Sitting with family 53.7 1.42 12.8 1.26
Hobbies/sports/exercising/reading 5.2 0.10 1.0 0.08
Religious activities 22.5 0.54 4.7 0.46
Social/community activities 26.7 0.90 3.7 0.79
Selling 0.2 0.15 0.1 0.13
Other 0.7 0.01 0.0 0.01
n 2,527 2,527 2,527 2,527
Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 1 Respondents were allowed to report up to two activities per time use increment (30 minutes) from 5:00 a.m. to 9:00 p.m. prior to the 
survey. If two activities were reported, one was designated as the primary activity, and the other as a secondary activity. Some women may 
not have reported a secondary activity for each 30-minute period.
Estimates exclude households that do not have a male primary adult decision-maker or that have missing or incomplete data.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 99
8. Summary and Conclusions 
This report provides information about the ZOI indicators during the 2013-2018 (baseline-end￾line) period. The assessment is designed for use as a monitoring tool, and as such provides point 
estimates of the indicators with an acceptable level of statistical precision. However, Feed the 
Future ZOI sample calculations are not designed to support conclusions of causality or program 
attribution, nor is the end-line assessment designed to measure change from the baseline. 
The Ethiopia ZOI end-line survey was conducted by International Food Policy Research Institute
(IFPRI) in conjunction with its data collection partner, BST Consulting PLC. Fieldwork for the 
ZOI Survey took place between September and October 2018.
Twelve Feed the Future indicators were included in this assessment: (1) Daily per capita 
expenditures (as a proxy for income) in USG-assisted areas; (2) Prevalence of poverty; (3) Depth 
of poverty; (4) Prevalence of households with moderate or severe hunger; (5) Women’s dietary 
diversity; (6) Prevalence of children 6-23 months receiving a minimum acceptable diet (MAD); (7) 
Prevalence of exclusive breastfeeding among children under 6 months of age; (8) Prevalence of 
underweight women; (9) Prevalence of stunted children under 5 years of age; (10) Prevalence of 
wasted children under 5 years of age; (11) Prevalence of underweight children under 5 years of 
age; and (12) Women's Empowerment in Agricultural Index (WEAI).
Household Economic Status
The daily per capita expenditures in the Feed the Future ZOI was $2.19 (in 2010 USD) in 2018 
and 32.3 percent of the population fell below the $1.25/day (2005 PPP) poverty line. The depth 
of poverty is estimated at 12.2 percent. Relative to the baseline, daily per capita expenditure is 
higher, and the prevalence and depth of poverty is lower in the end-line. 
Hunger and Dietary Intake
Prevalence of households with moderate or severe hunger is 3.3 percent and this is 30 percent 
lower than the proportion in the baseline (4.9 percent), though this rather very low rate may be 
due to the non-validity of the measure in Ethiopia. Dietary diversity is extremely low. Women of 
reproductive age consume from an average 2.3 food groups in a day, and only 4.6 percent of 
women achieve the minimum dietary diversity or consume from at least 5 food groups. Only 8.4 
percent of children 6-23 months receive a MAD. Prevalence of exclusive breastfeeding among 
children under 6 months of age is higher at about 74 percent. The proportion of children 6-23 
months that receive a MAD, and the mean and minimum dietary diversity of women of 
reproductive age in the end-line were statistically significantly higher relative to the baseline. 
Nutritional Status of Women and Children
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 100
Nearly 22 percent of women of reproductive age are underweight. Prevalence of stunted children 
under 5 years of age is about 39 percent. The prevalence of wasting is about 10 percent while 
22.5 percent of children under 5 years of age are underweight. The proportion of underweight 
women of reproductive age, and the proportion of underweight, stunted, and wasted children 
under 5 years of age statistically significantly declined during the baseline-end-line period. 
Women’s Empowerment in Agriculture 
Women’s empowerment is measured through the full Women Empowerment in Agricultural 
Index (WEAI). In 2018, the WEAI for the ZOI in Ethiopia is estimated as 0.75. Small 
improvements across all domains are recorded. Access to and input into decisions on credit is a 
major constraint to women's empowerment. Lack of leisure time, group membership and 
speaking in public are also among important contributors to women’s disempowerment. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 101
References
Alkire, S., Malapit, H., Meinzen-Dick, R., Peterman, A., Quisumbing, A., Seymour, G., and Vaz, A. 
(2013). Instructional Guide on the Women’s Empowerment in Agriculture Index. International 
Food Policy Research Institute (IFPRI). (2013). Retrieved from 
http://www.ifpri.org/publication/womens-empowerment-agriculture-index. 
Alkire, S., Meinzen-Dick, R., Peterman, A., Quisumbing, A., Seymour, G., and Vaz, A. (2013). The 
Women’s Empowerment in Agriculture Index. World Development, 52(C), 71-91. 
Ballard, T.; Coates, J.; Swindale, A.; and Deitchler, M. (2011). Household Hunger Scale: Indicator 
Definition and Measurement Guide. Washington, DC: Food and Nutrition Technical 
Assistance II Project, FHI 360.
Black, R.E., et al. (2008) Maternal and Child Undernutrition: Global and Regional Exposures and 
Health Consequences. The Lancet. 371(9608):243-260. 
Darnton-Hill, I., et al. (2005) Micronutrient deficiencies and gender: social and economic costs. 
American Journal of Clinical Nutrition, May 2005, 81(Supplement): 1198S-1205S. 
Deaton, A. (2008). The Analysis of Household Surveys: A microeconomic approach to development 
policy. Baltimore: The Johns Hopkins University Press.
Deaton, A. and Zaidi, S. (2002). “Guidelines for constructing consumption aggregates for welfare 
analysis, Working Paper No. 135. Washington, DC: The World Bank.
Deitchler, M., Ballard, T., Swindale, A., and Coates, J. (2011). FANTA Technical Note No. 12: 
Introducing a Simple Measure of Household Hunger for Cross-Cultural Use. Washington, DC: 
USAID. 
Foster, J., Suman S., Lokshin, M. and Sajaia, Z. (2013). A Unified Approach to Measuring Poverty and 
Inequality: Theory and Practice. Washington, DC: The World Bank.115-118. 
Grosh, M.E. and Munoz, J. (1996). A manual for planning and implementing the living standards 
measurement study survey. Living Standards Measurement Study Group Working Paper No. 
126. Washington, DC: The World Bank.
Grosh, M. and Glewwe, P. (1995). A Guide to Living Standards Measurement Study Surveys and 
Their Data Sets. Living Standards Measurement Study Group Working Paper No. 120. 
Washington, DC: The World Bank.
Haughton, J. and Khandker, S. (2009). Handbook on poverty and inequality. Washington, DC: The 
World Bank.
Hoenig, John M., and Heisey, Dennis M. (2001). “The Abuse of Power: The Pervasive Fallacy of 
Power Calculations for Data Analysis,” The American Statistician, 55(1); 19-24. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 102
Kaplinsky, R. and Morris, M. A Handbook for Value Chain Analysis. Ottawa, Canada: International 
Development Research Center.
Maxwell, D., Coates, J., and Vaitla, B. (2013). How Do Different Indicators of Household Food Security 
Compare? Empirical Evidence from Tigray. Feinstein International Center, Tufts University: 
Medford, USA.
Mckenzie, David, and Ozier, O. (2019) “Why ex-post power using estimated effect sizes is bad, 
but an ex-post MDE is not,” at https://blogs.worldbank.org/impactevaluations/why-ex-post￾power-using-estimated-effect-sizes-bad-ex-post-mde-not (May 16, 2019).
NPC (September 2017). “Ethiopia’s Progress Towards Eradicating Poverty - An Interim Report
on 2015/16 Poverty Analysis Study,” National Planning Commission (NPC), Federal 
Democratic Republic of Ethiopia, Addis Ababa, Ethiopia. 
United Nations Development Group (UNDP). (2003). Indicators for monitoring the Millennium 
Development Goals: definitions, rationale, concepts and sources. New York: United Nations.
University of Oxford. (2013). Alkire Foster Method: OPHI’s method for multidimensional measurement. 
Oxford Poverty & Human Development Initiative (OPHI). Retrieved from 
http://www.ophi.org.uk/research/multidimensional-poverty/alkire-foster-method. 
UNSTATS. N.d. SDG Indicators – Metadata repository. Available at: 
https://unstats.un.org/sdgs/metadata/. 
USAID. (2013). Feed the Future Indicator Handbook: Definition Sheets (updated 
October 18, 2014). 
Victora, C.G., et al. (2008). Maternal and Child Undernutrition: Consequences for Adult Health 
and Human Capital. The Lancet. 371(9608):340-357. 
Webber, C.M. and Labaste, P. (2010). Building Competitiveness in Africa's Agriculture: A Guide to Value 
Chain Concepts and Applications. Washington, DC: The World Bank. 
https://openknowledge.worldbank.org/handle/10986/2401 
WHO and UNICEF. (2006). WHO Child Growth Standards and the Identification of Severe Acute 
Malnutrition in Infants and Children. World Health Organization and United Nations 
Children’s Fund.
WHO/UNICEF/USAID/AED/FANTA 2/UC DAVIS/IFPRI/UNICEF. (2010). Indicators for Assessing 
Infant and Young Child Practices (Part 2 Measurements).
Zhang, L.C. (1999). A note on post-stratification when analyzing binary survey data subject to 
nonresponse. Journal of Official Statistics, 15(2): 329-334. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 103
Appendix 1. Supplementary Data and Figures
Table A1.1. ZOI Survey 2018 Feed the Future indicator estimates 
Unweighted sample sizes, point estimates, standard deviations, confidence intervals, design effects and 
nonresponse rates for the Feed the Future indicators for the ZOI. 
Indicator Estimate SE
95% CI DEFF Nonresp
onse 
rate (%)1
Unweight
ed 
number
Weighted 
number Lower Upper
Daily per capita consumption expenditures 2010 USD 
All households 2.19 0.0422 2.10 2.27 2.5 3.0 3,876 3,876
Male and female 2.21 0.048 2.12 2.31 2.3 3.6 3,089 3,091
Female adult(s) only 1.95 0.081 1.79 2.11 1.3 0.0 665 668
Male adult(s) only 2.05 0.259 1.53 2.56 1.0 0.9 115 110
Prevalence of Poverty: Percent of people living on less than $1.25/day (2005) 
All households 32.3 0.003 30.8 33.8 3.8 3.0 3,876 3,876
Male and female 31.3 0.008 29.6 32.9 3.6 3.6 3,089 3,091
Female adult(s) only 40.8 0.020 37.1 44.6 1.5 0.0 665 668
Male adult(s) only 36.3 0.045 27.3 45.1 1.0 0.9 115 110
Depth of Poverty: Mean percent shortfall relative to the $1.25/day poverty 
 
 
All households 12.2 0.003 11.4 12.8 1.5 3.0 3,876 3,876
Male and female 11.7 0.003 10.9 12.5 1.7 3.6 3,089 3,091
Female adult(s) only 15.8 0.009 13.9 17.6 1.0 0.0 665 668
Male adult(s) only 12.5 0.020 8.5 16.4 1.0 0.9 115 110
Prevalence of moderate and severe hunger 
All households 3.3 0.003 2.6 4.2 1.9 0.2 3,882 3,882
Male and female 2.5 0.003 1.9 3.3 1.7 0.3 3,098 3,100
Female adult(s) only 6.8 0.009 5.0 9.3 1.2 0.0 661 665
Male adult(s) only 6.6 0.022 3.1 13.4 1.1 0.0 116 110
Women dietary diversity: Mean number of food groups consumed by women 
 All women ages 15-49 2.3 0.018 2.2 2.4 5.4 1.4 3,956 3,971
Prevalence of underweight women of reproductive age 
All non-pregnant women 
 
21.9 0.007 19.9 24.0 2.3 0.0 3,604 3,616
Prevalence of exclusive breastfeeding among children under 6 months of age 
All children 74.5 0.042 63.4 83.1 1.4 0.0 111 111
Male children 73.9 0.719 58.9 84.8 1.2 0.0 57 56
Female children 75.1 0.060 62.0 84.8 1.0 0.0 54 55
Prevalence of children 6-23 months receiving a minimum acceptable diet (%) 
All children 8.4 0.011 6.4 11.0 1.1 1.4 639 633
Male children 7.5 0.015 5.0 11.1 1.1 1.2 324 323
Female children 9.3 0.016 6.4 13.1 1.0 1.6 315 310
Prevalence of stunted children under 5 
 
 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 104
All children 38.9 0.011 36.1 41.8 1.7 11.1 1,963 1,954
Male children 38.8 0.015 34.9 42.8 1.7 11.8 976 972
Female children 39.0 0.015 35.8 42.3 1.1 10.5 987 981
Prevalence of wasted children under 5 
 
 
All children 9.7 0.006 7.8 12.0 2.4 11.7 1,916 1,910
Male children 11.0 0.010 8.7 13.9 1.7 12.5 950 951
Female children 8.4 0.008 6.3 11.2 1.9 11.0 966 959
Prevalence of underweight children under 5 years of age 
All children 22.5 0.009 20.5 24.6 1.3 2.5 2,153 2,138
Male children 23.9 0.013 21.2 26.7 1.1 3.2 1,071 1,066
Female children 21.1 0.012 18.7 23.7 1.1 1.9 1,082 1,073
Women’s Empowerment in Agriculture Index (score) 
All women 0.75 0.0007 0.75 0.76 1.9 0.0 3,252 3252
Women ages 18-29 0.72 0.003 0.72 0.73 1.3 0.0 647 643
Women ages 30 and 0.76 0.002 0.76 0.77 1.7 0.0 2,605 2608
Source: Source: Feed the Future End-line ZOI Survey 2018, Ethiopia. 
Note. The number of child-headed households in the sample is too small to obtain any valid estimate. SE = standard error; CI = confidence interval; DEFF = 
design effect. 1 The nonresponse rates are calculated as the fraction of the expected number of respondents for each indicator and respondent type.
Table A1.2. Women’s Empowerment in Agriculture Index results for indicators that 
compose the 5 domains of empowerment, using uncensored headcount ratios 
Domain Indicator
Baseline End-line 
p-value2 Percent n1 Percent n1
Production
Input in productive 
decisions 88.3 3,401 91.7 3,167 0.001
Autonomy in production 74.7 3,416 80.6 3,209 0.012
Resources
Ownership of assets 80.0 3,464 90.8 3,252 0.000
Purchase, sale or transfer of 
assets
74.5 3,464 88.7 3,252 0.000
Access to and input into 
decisions on credit 15.0 3,431 23.4 3,239 0.000
Income Control over use of income 89.0 3,391 92.1 3,201 0.001
Leadership Group membership 44.8 3,463 44.7 3,191 0.985
Speaking in public 48.0 3,463 47.0 3,252 0.771
Time Workload 52.3 3,445 67.0 3,252 0.000
Leisure 49.7 3,461 40.7 3,252 0.003
Source: Feed the Future Feedback ZOI Survey, Ethiopia 2013; Feed the Future End-line ZOI Survey 2018, Ethiopia. 1 Estimates exclude households that do not have a female primary adult decision-maker or that have missing or incomplete indicator data. ‘ 
2 An F-test of statistical significance in differences in proportions between the baseline and 2018 estimates was conducted.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 105
Appendix 2. Methodology
A2.1 Sampling and Weighting25
The Feed the Future Ethiopia End-line ZOI Survey 2018 is the third round of the Feed the Future
surveys designed and implemented starting at the baseline in 2013. It is a panel survey covering 
the same households first surveyed at baseline (2013). 
Recall that the original objective was to conduct an impact evaluation of Feed the Future
investments. It was agreed at the time that, given the design of the program and limits to the 
budget for evaluation studies, the most feasible evaluation strategy is matching based on samples 
of households within and outside of the Feed the Future ZOI. Sample size with sub-samples were 
determined accordingly using the relevant power calculations using incidence of poverty (poverty 
headcount) as the primary indicator (see “Sampling at baseline” below).
At end-line, this objective was modified to tracking the difference in the levels of high-level 
program indicators estimated at baseline and end-line for households in the Feed the Future ZOI 
only. Households outside of the ZOI were dropped from the sample as a consequence. 
Accordingly, the level of the indicators (all proportions other than two – daily per capita 
consumption expenditure and WEAI) were estimated for the baseline and the end-line. 
Subsequently, the two are compared for each indicator to ascertain whether there is a statistically 
significant difference between them. 
Nevertheless, the change in objective of comparison and sample at end-line implies that the ex￾ante power calculations prior to the baseline is no longer applicable. 
What are the options? 
The most obvious choice is to redo the power calculations ex-post – or compute post-hoc 
power. However, there appears to be a consensus that this is neither meaningful nor helpful (see 
Mckenzie and Ozier (2019), Hoenig and Heisey (2001), and references therein. In short, post￾hoc power has a one-to-one inverse relationship with corresponding p-values and thus cannot 
provide any additional information. It may even be misleading. 
Another option suggested by Mckenzie and Ozier (2019) is compute the minimum detectable 
size effect (MDE) using the following formula MDE= (zα/2+ z1-β)*SE, where SE is the standard 
error of the outcome. For the poverty headcount this comes to 4.5 percentage points. This 
magnitude is smaller than the envisaged reduction for this indicator at the beginning of the study
as well as what is estimated at end-line. 
25 This section is taken from the baseline report except with appropriate updates and changes.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 106
Sampling at baseline (Given that the end-line was not implemented as an impact 
evaluation, this information is provided for background only.)
The baseline survey was conducted in woredas in USAID/Ethiopia’s ZOI and also in woredas not 
within the ZOI. By interviewing households at baseline and at end-line both inside and outside of 
the ZOI and by using a non-experimental impact estimator such as matching, it will be possible 
to undertake an impact evaluation that determines whether improvements in Feed the Future
performance indicators in the ZOI can be attributed to the totality of Feed the Future activities. 
If these surveys collect information on who participates in the various Feed the Future
interventions, it will also be possible to assess both the direct and spillover impacts of Feed the 
Future. The direct effects are estimated by comparing changes in households that take up Feed 
the Future interventions with matched households outside the ZOI who, given their 
characteristics, would have taken up the intervention had it been available. The spillover effects 
are estimated by comparing changes in households in the ZOI that did not take up Feed the 
Future interventions with matched households outside the ZOI who, given their characteristics, 
would not have taken up the intervention even if it had been available. We discuss below whether 
our sample design can detect other impacts.
Determining sample size
The size of the sample depends on a number of considerations. First, is the purpose of the survey 
to monitor Feed the Future performance indicators or is it to both monitor Feed the Future
performance indicators and provide baseline information for impact evaluation? The survey is 
designed and conducted to achieve the latter. A note of clarification on indicator tracking is 
appropriate at this point. The RFA requests disaggregating indicators by household gender and 
age composition. Specifically, the incidence of poverty, pattern of per capita expenditure and 
prevalence of hunger (i.e., Indicators 1, 2, and 8) are to be disaggregated in this way. Similarly, 
Indicator 11 will be disaggregated by gender of children while indicator 9 will be disaggregated by 
gender of children as well as wealth quintiles of households. The sample size required to track 
indicators and detect impact at these levels of disaggregation would be rather large. Instead, it 
was agreed that the survey should be designed in such a way that key indicators are, as 
appropriate, disaggregated by household demographic and wealth characteristics and tracked, 
though without necessarily aspiring to causal impact evaluation. Obviously, impact at the 
household level will be assessed in the manner described in the evaluation design section above. 
The survey was designed and implemented accordingly. 
Second, sample size is affected by the desired level of statistical significance (the sample has to be 
sufficiently large to minimize the chance of detecting an effect that does not exist) and desired 
statistical power (the sample has to be sufficiently large to minimize the chance of not detecting 
an effect that does exist). Following standard practice, these were set at a target level of 
significance of 5 percent (two-tailed) and statistical power of 80 percent. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 107
Sample size also depends on the minimum level of impact the survey is desired to detect in the 
relevant indicator. For example, should the sample size be large enough to detect that the 
intervention has reduced poverty by 5 percentage points, by 10 percentage points or by 20 
percentage points? These levels of impact, known as minimum detectable effect sizes, are 
inversely related to sample size. Smaller effect sizes require larger samples; conversely, larger 
effect sizes require smaller samples. The size of the sample also depends heavily on which Feed 
the Future indicator is being considered. This is important because required sample sizes are 
affected by the variability of the indicator. Where the indicator(s) is (are) characterized by high 
levels of variability, larger sample sizes are needed. It is also affected by what is called the design 
effect, loosely defined as the extent to which the indicator is correlated across households or 
individuals within a geographic locality. Higher correlations mean that larger samples are needed.
In addition to all these considerations, the size of the sample depends on precisely what is meant 
by “Feed the Future impact.” Is Feed the Future impact defined in terms of a particular 
intervention or is it defined in terms of whether the totality of Feed the Future activities in the 
ZOI leads to changes in performance indicators that can be attributed to Feed the Future? 
Finally, we need to take into account the fact that over time some households will move, all 
members will disperse to other households or the household will choose not to continue to be 
interviewed. Based on our experiences with other longitudinal household surveys in rural 
Ethiopia, we assume that 10 percent of the sample will attrite between baseline and end-line. 
A central high-level objective of the Feed the Future initiative is poverty reduction. In light the 
broad outline of Feed the Future targets in Ethiopia, it is reasonable to opt for a sample size large 
enough to detect a 10-percentage point reduction in the incidence of poverty linked to Feed the 
Future.26 The sample size was thus chosen to be large enough to detect this level of impact. This 
minimum detectable size effect is equivalent to a 22 percent (0.22) standard deviation of poverty 
reduction in Feed the Future areas over and above that achieved in comparable but non-Feed the 
Future areas.27 The sample is divided into two-third treated (Feed the Future ZOI) and one-third 
control (non-Feed the Future ZOI) woredas. The sample is clustered at the woreda level. The 
aim is ensuring that by the end-line, there are on average 75 households interviewed per woreda, 
with these allocated across three Enumeration Area (EAs) each containing 25 households. We 
calculated the design effect as equaling 8.4. 
In summary:
26 The incidence of poverty measured by the head count ratio calculated using the PPP poverty line of $1.25 USD
per day. Note also that in the Feed the Future’s guidance notes on target setting, it is stated that in Ethiopia, Feed 
the Future should reduce the prevalence of poverty from 39.0 to 27.3 percent over a 5-year period. 27 Calculations using Ethiopian Household Income and Expenditure surveys show that the standard deviation of 
poverty incidence is around 0.45 and so a 22 percent reduction in this is equivalent to a 10-percentage point 
reduction in poverty.
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 108
 Minimum detectable effect size – 10 percentage point reduction in the incidence of 
poverty linked to Feed the Future
 Statistical significance – 5 percent
 Statistical power – 80 percent 
 Design effect – 8.4 
 Enumeration Areas (EA) – 3 per woreda
 Attrition – 10 percent with 75 households per woreda in the end-line survey 
Given these features and assumptions, 56 woredas in the Feed the Future ZOI and 28 woredas
outside the ZOI are required. In the baseline, 84 households were selected for interview per 
woreda or 28 households per EA; given an assumed rate of attrition of 10 percent, this will mean 
that on average, at end-line, there will be 75 households interviewed in each woreda. Therefore, 
the baseline survey was planned to collect information from 4,704 households in the ZOI (56 
woredas x 3 EAs per woreda x 28 households per EA) and 2,352 households outside the ZOI 
(28 woredas x 3 EAs per woreda x 28 households per EA) giving a total baseline sample of 7,056 
households residing in 252 EAs (with each EA located in a corresponding kebele).
Heterogeneity 
A number of sources of heterogeneity were considered in the design of the sample. First, the 
Feed the Future’s ZOI spans woredas with diverse agro-ecological potential. In line with this 
heterogeneity, the non-ZOI woredas were drawn from a set purposively defined to have 
characteristics similar to woredas found in the ZOI. Moreover, the woreda composition of the 
sample was made to reflect the distribution of Feed the Future-supported projects by using the 
percentage of Feed the Future’s ZOI (or woredas) each major Feed the Future project covers as 
a basis of its share in the sample (see Table A2.2 below). 
Second, households within a locality differ in terms of dimensions that may be relevant to the 
performance of Feed the Future-related investments. Two such dimensions can be specifically 
important – the gender and age of household heads. Accordingly, the household composition of 
the sample in each EA will be determined by the distribution of household types in the community 
as defined by the gender and age of household heads (see the discussion in the next paragraph 
for further detail). This is particularly relevant for the Women Empowerment in Agriculture Index 
(WEAI).
The EA-level sample is divided into female- and male-headed households, each group further 
divided into youth-headed and mature-headed households. Thus, the EA sample is divided into a 
total of 4 age-gender groups. The share of each in the sample is determined by the corresponding 
shares reported by CSA’s Population Census of 2007. Census 2007 data show the distribution of 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 109
household heads by age and gender reported in columns 2-3 of Table 1.2. Columns 4-5 of the 
same table report the composition of the sample households.
Table A2.1: Household composition of the EA sample
 
Share in the population of rural 
household heads – Census 2007 (%)
Male Female
Young (15-34 years of age) 29.6 (8) 5.4 (2)
Mature (35 years of age or older) 48.9 (13) 16.1 (5)
Source: Authors’ calculation using CSA data. 
Note: The numbers in brackets are implied (columns 2-3) number of sample households in an EA (with the 
total being a predetermined 28). 
Table A2.2: Feed the Future sample woredas, grouped by Feed the Future program 
Program
Number 
of 
Sample 
woredas
Total 
number of 
woredas in 
the 
program
Share in 
the ZOI 
(%)*
Share 
in the 
Feed 
the 
Future
Sample 
(%)*
Number of 
sample 
woredas in a 
program as a 
fraction of 
total in the 
program (%)
AGP 48 111 74.5 85.7 43.2
PRIME 7 30 20.1 12.5 23.3
GRAD 6 16 10.4 11.1 37.5
AGP+GRAD 5
AGP+ENGINE 12
Total Number of FTF Sample 
Woredas 56
Total Number of Control 
Sample Woredas 28
Total Number of Sample 
Woredas 84
Both treatment and control woredas were randomly selected using proportions derived from 
population size and project coverage. The regional distribution of the planned sample is 
summarized in the next table. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 110
Table A2.3: Feed the Future sample households, planned by region 
Number of sample households Amhara Oromia SNNP Somali Tigray Total
Initially planned to be interviewed 1848 2436 1680 420 672 7056
Actually interviewed 1848 2414 1677 400 672 7011
Notes: * Shares sum over 100% for two reasons. First, 8 woredas (or about 5%) are covered by more than one Feed the Future-supported 
program – five of these are in the sample. There are 149 woredas in the Feed the Future’s ZOI. Second, AGP and ENGINE woredas overlap, 
though all AGP woredas were covered by ENGINE at the time of the baseline – there are 12 AGP-ENGINE woredas in the sample. Note also 
that 11 AGP-LMD woredas are included under the AGP heading in this table. 
Selection of the sample
Sample selection included three steps. The first step involved the systematic random selection of 
woredas from the Feed the Future project categories identified above with probability of 
selection proportional to size (i.e., combining project coverage and population size). Three EAs 
were subsequently selected from each woreda randomly with selection probability proportionate 
to size. In the third step, sample households were chosen from a freshly compiled complete listing 
of households in each selected using systematic random sampling. The selection probabilities 
combined the ratio of 28 households to the total number of households in the EA and the 
proportions of the four household types formed by the age and gender of the household head. 
The product of the inverse of the selection probabilities at the three stages form the sample 
weight of each household for the purpose of estimating the relevant indicators. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 111
A2.2 Poverty Prevalence and Consumption Expenditure 
Methods
Data Source 
The Feed the Future baseline and end-line surveys conducted in 2013 and 2018 are the sources 
of data for the estimation of consumption expenditures and poverty levels. Data on household 
food and non-food expenditures as well as local prices were collected by these surveys. 
Measuring poverty and comparing poverty across groups 
Poverty Indices
Since the work of Sen (1976) on the axiomatic approach to measurement of poverty, several 
indices of poverty have been developed. 
The most widely used poverty indices are the percentage of the poor (headcount index), the 
aggregate poverty gap (poverty gap index), and the distribution of income among the poor 
(poverty severity index). The poverty measure itself is a statistical function that translates the 
comparison of the indicator of household well-being and the chosen poverty line into one 
aggregate number for the population as a whole or a population subgroup. Many alternative 
measures exist, but the three measures described below are the ones most commonly used.
Incidence of poverty (headcount index). This is the share of the population whose income or 
consumption is below the poverty line, that is, the share of the population that cannot afford to 
buy a basic basket of goods. 
Depth of poverty (poverty gap). This provides information regarding how far households are 
from the poverty line. This measure captures the mean aggregate income or consumption 
shortfall relative to the poverty line across the whole population. It is obtained by adding up all 
the shortfalls of the poor (assuming that the non-poor have a shortfall of zero) and dividing the 
total by the population. In other words, it estimates the total resources needed to bring all the 
poor to the level of the poverty line (divided by the number of individuals in the population). 
Poverty severity (squared poverty gap). This takes into account not only the distance 
separating the poor from the poverty line (the poverty gap), but also the inequality among the 
poor, that is, a higher weight is placed on those households further away from the poverty line. 
Method of aggregating poverty and hypothesis testing 
We used Foster, Greer, and Thorbecke (1984) Pα class of poverty measures. Defining the per￾adult (per capita) consumption expenditure of household i by Yi, and ranking Yi, as: 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 112
12 1 ......... ........ , qq N YY Y Z Y + ≤ ≤ ≤< ≤
where Z is poverty line, N is the total population, and q is the number of poor. Consequently, 
we can classify an individual poor if the real per adult consumption expenditure is less than to Z 
(Y<Z) and non-poor if the real per adult consumption expenditure is greater than or equal to Z 
(Y>=Z). Once we classify an individual into poor and non-poor, we can aggregate poverty at 
national and sub-national levels using the Pα class of poverty measures given by
1
1 ) ; 0, for .
q
i
i
ZY P YZ
NZ
α
α α =
  − =   ≥<   ∑
When α=0, the corresponding poverty index is called the headcount index (P0). Hence P0 
corresponds to the fraction of individuals falling below the poverty line. If α=1, the poverty index 
is called the poverty gap index (P1) and it measures the aggregate poverty deficit of the poor 
relative to the poverty line, where α=2 it measures the squared proportional shortfalls from the 
poverty line and is commonly known as an index of the severity of poverty. 
Comparing Poverty between Groups 
There are two ways of comparing poverty indices across groups or over time. The first way to 
compare poverty indices between, say, two groups (group 1 and group 2) is to conduct a 
statistical test or means separation test. If the poverty measures are estimated from unit record 
data (i.e., on the basis of sample observations), it is possible to test whether the observed 
differences in their values are statistically significant. The hypothesis test developed by Kakwani 
(1993) can be used to test whether poverty indices (Pα) differ significantly between groups and 
over time. The standard error of Pα is calculated using the following formula (Ravallion 1992):
2 () () , PP SE P
n
αα
α
− =
where SE (.) is the standard error. Consequently, the standard error (SE) of the difference in 
poverty index between group 1 and group 2 (SE (Pα1 - Pα2), having a random sample n1 and n2, 
respectively, is given by 
22
12
12
12
() , ss SE P P
nn α α −=+
where s1 and s2 are the sample estimator of the variance of the asymptotic distribution of 
1 1 22 α and α nPnP , 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 113
such that: 
22
12 1 2 SE P P SE P SE P ( ) ( ( )) ( ( )) . α α −= + α α
The test statistic (t) is given by: 
12
12
() . ()
PP
t
SE P P
α α
α α
− = −
This is asymptotically normally distributed with zero mean and unit variance. In a large sample, if 
the calculated value of t (the test statistics) has an absolute value less than 1.96 (2.58), then the 
difference in the poverty indices between two groups or dates is not statistically significant at the 
5 percent (1 percent) level, using a two-tail test. 
Data Preparation 
Currency Conversions using CPI and PPP 
The following steps were involved in converting the 2005 PPP international poverty line of $1.25
to one in Ethiopian Birr at 2013 and 2018 prices. Purchasing power parity conversion factors
(PPP) and consumer price indices (CPI) are used for the purpose. 
Conversion: To make it comparable with the expenditure data collected in 2013 and 2018 in 
Ethiopia, the poverty line has to be converted into Ethiopian Birr in 2013 and 2018 prices. These 
required a number of inputs/steps.
i. PPP Conversion factor – Initial: Converts PPP dollars into Ethiopian Birr in 2005 prices
using the PPP conversion Factor for Ethiopia of 2.75; 
ii. PPP Conversion factor – Updates (2005-2011 and 2005-2018): The conversion factor is 
thus extrapolated by adjusting it for local inflation using CSA’s Consumer Price Index with 
2005 as the base year. The following ratios are used:
o 2013 CPI/2005 CPI = 3.96;
o 2018 CPI/2005 CPI = 5.68;
Poverty Thresholds 
Using these elements, the new poverty line for the Feed the Future baseline became Birr13.61 
per adult equivalent per day in 2013 prices. Similarly, the end-line poverty line for the Feed the 
Future was Birr19.53 per adult equivalent per day in 2018 prices. 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 114
A2.3 Criteria for Achieving Adequacy for Women’s 
Empowerment in Agriculture Indicators
The table below presents the five Women’s Empowerment in Agriculture (WEAI) domains of 
empowerment, their corresponding empowerment indicators, the survey questions that are used to elicit 
the data required to establish adequacy for each empowerment indicator, and how adequacy criteria 
are defined for each empowerment indicator. For additional details refer to the Feed the Future Guide 
to Statistics.
Dimension Indicator 
name Survey questions Aggregation of adequacy criteria Inadequacy 
criteria
Production Input in 
productive 
decisions
G2.02 A-C, F How much 
input did you have in making 
decisions about: food crop 
farming, cash crop farming, 
livestock raising, fish culture; 
G5.02 A-D To what extent 
do you feel you can make 
your own personal decisions 
regarding these aspects of 
household life if you want(ed) 
to: agriculture production, 
what inputs to buy, what 
types of crops to grow for 
agricultural production, when 
or who would take crops to 
market, livestock raising
Must have at least some input into or 
can make own personal decisions in 
at least two decision-making areas 
Inadequate if 
individual 
participates 
BUT does 
not have at 
least some 
input in 
decisions; or 
she does 
not make 
the 
decisions 
nor feels she 
could.
Resources Ownership 
of assets
G3.02 A-N Who would you 
say owns most of the [ITEM]? 
Agricultural land, Large 
livestock, Small livestock, 
chicks etc.; Fish pond/equip; 
Farm equipment (non￾mechanized); Farm equip 
(mechanized); Nonfarm 
business equipment; House; 
Large durables; Small 
durables; Cell phone; Non￾agricultural land (any); 
Transport
Must own at least one asset, but not 
only one small asset (chickens, non￾mechanized equipment, or small 
consumer durables)
Inadequate if 
household 
does not 
own any 
asset or 
only owns 
one small 
asset, or if 
household 
owns the 
type of asset 
BUT she 
does not 
own most of 
it alone
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 115
Dimension Indicator 
name Survey questions Aggregation of adequacy criteria Inadequacy 
criteria
 Purchase, 
sale or 
transfer of 
assets
G3.03-G3.05 A-G Who would 
you say can decide whether 
to sell, give away, 
rent/mortgage [ITEM] most of 
the time? G3.06 A-G Who 
contributes most to decisions 
regarding a new purchase of 
[ITEM]? Ag land; Large
livestock, Small livestock; 
Chickens etc.; Fish pond; 
Farm equipment (non￾mechanized); Farm equipment
(mechanized) 
Must be able to decide to sell, give 
away, or rent at least one asset, but 
not only chickens and non￾mechanized farming equipment 
Inadequate if 
household 
does not 
own any 
asset or 
only owns 
one small 
asset, or 
household 
owns the 
type of asset 
BUT she 
does not 
participate 
in the 
decisions 
(exchange 
or buy) 
about it
 Access to 
and 
decisions 
on credit
G3.08-G3.09 A-E Who made 
the decision to borrow/what 
to do with money/item 
borrowed from [SOURCE]?
Non-governmental 
organization (NGO); Informal 
lender; Formal lender (bank); 
Friends or relatives; ROSCA 
(savings/credit group)
Must have made the decision to 
borrow or what to do with credit 
from at least one source 
Inadequate if 
household 
has no 
credit OR 
used a 
source of 
credit BUT 
she did not 
participate 
in ANY 
decisions 
about it
Income Control 
over use 
of income
G2.03 A-F How much input 
did you have in decisions on 
the use of income generated 
from: Food crop, Cash crop, 
Livestock, Nonfarm activities, 
Wage & salary, Fish culture; 
G5.02 E-G To what extent do 
you feel you can make your 
own personal decisions 
regarding these aspects of 
household life if you want(ed) 
to: Your own wage or salary 
employment? Minor 
household expenditures?
Must have some input into decisions 
on income, but not only minor 
household expenditures
Inadequate if 
participates 
in activity 
BUT she has 
no input or 
little input 
on decisions 
about 
income 
generated 
Feed the Future Ethiopia Zone of Influence Phase One End-line Assessment 116
Dimension Indicator 
name Survey questions Aggregation of adequacy criteria Inadequacy 
criteria
Leadership Group 
member 
G4.05 A-K Are you a member 
of any: Agricultural / livestock/ 
fisheries producer/ market 
group; Water, forest users’,
credit or microfinance group; 
Mutual help or insurance 
group (including burial 
societies); Trade and business 
association; Civic/charitable 
group; Local government; 
Religious group; Other 
women’s group; Other group. 
Must be an active member of at least 
one group 
Inadequate if 
not an active 
member of a 
group or if 
unaware of 
any group in 
the 
community 
or if no 
group in 
community
 Speaking in 
public
G4.01 – G4.03 Do you feel 
comfortable speaking up in 
public: To help decide on 
infrastructure (like small wells, 
roads) to be built? To ensure 
proper payment of wages for 
public work or other similar 
programs? To protest the 
misbehavior of authorities or 
elected officials? 
Must feel comfortable speaking in at 
least one public setting 
Inadequate if 
not at all 
comfortable 
speaking in 
public
Time Workload G6 Worked more than 10.5 
hours in previous 24 hours. 
Total summed hours spent toward 
labor must be less than 10.5
Inadequate if 
works more 
than 10.5 
hours a day
Leisure G6.02 How would you rate 
your satisfaction with your 
available time for leisure 
activities like visiting 
neighbors, watching TV, 
listening to radio, seeing 
movies or doing sports?
Must rate satisfaction level as at least 
five out of 10
Inadequate if 
not satisfied 
(<5)
1 Time poverty is defined as performing productive or domestic work for more than 10.5 hours in a 24-hour period