IMPROVING BASIC SERVICES FOR THE BOTTOM FORTY PERCENT LESSONS FROM - - PowerPoint PPT Presentation

IMPROVING BASIC SERVICES FOR THE BOTTOM FORTY PERCENT

LESSONS FROM ETHIOPIA

by by Qaiser ser Khan, Jean-Paul aul Faguet, t, Christ stopher

• pher Gaukl

kler, and Wendms dmsya yamrr mrregne egne Mekash asha Pre resent ented ed at Lo London n School

• ol of Econo

nomi mics cs and Politi tical cal Science nce Wednesda esday y Octob

• ber

er 8, 2014

OBJECTIVES OF THE STUDY

Esta tablis blish whether ether, , and to what t ext xtent nt, , there re exist ist dist stribu ributi tional

• nal pover

erty ty and social al outcomes comes in access ss to basic ic decentralized ntralized service ice deliv iver ery y in Eth thiopi iopia. a. Assess ess if wored reda level l spend nding ing helps ps in reachin hing g sector

• r

targe rgets ts, , parti ticula cularl rly for the poores est. t. Assess ess distr tribu ibuti tiona

• nal

l impa mpacts cts of woreda eda level l spendin nding g on: wealth lth spatial tial equa uali lity ty gender er hist storic

• rically

ally underser served ed groups. ups.

ETHIOPIA’S REGIONS

ETHIOPIA IS A BIG, DIVERSE, IMPORTANT COUNTRY

Popul ulat ation ion ~92 million ion 98 et ethni hnic c groups; s; 93 langua uage ges Sus ustai ained ned growt wth h rate >10% 0% in recent cent years ars

• Fastest growing country in Africa, and 3rd fastest in the world

One e of fa fastest est pover erty ty reduc ducer ers s in the e worl rld Rapid d strides ides towards ds MDGs

• Child mortality fell 123  88/thousand
• Primary NER rose 68%  82% in 2005-2010

 Intere eresting sting empiric irical l cont ntext xt to study udy dece centra ntraliz lized d service ice provis ision ion

ETHIOPIA’S DECENTRALIZATION FRAMEWORK

1995 constitution titution decentralized ntralized to 9 r regional ional stat ates es and 2 c cities. ies.  Full l administ inistrat ativ ive e decentr traliz alization ation with th ext xten ensiv sive intergo ergovern ernmental mental fiscal al transf ansfer ers. Regio gions ns receiv eive e bl block k grants nts via formula mula ba based ed on ne n needs ds – including uding addit itional ional fundin ing g for histor

• ric

ically ally less s develop loped ed areas. as. In 2002, decentralization ntralization exten ende ded d to ~850 ~850 woredas edas  expect ected ed to take e on t n the bu bulk k of service ice deliv iver ery y responsib ponsibilit ilities es. Woredas das receiv ive block k grants ants from m regio gional nal govern ernments, ments, also so govern erned d by formulas ulas (set t by regional ional govern ernments). ments).

INTRODUCTION

ETHIOPIA’S PROMOTION OF BASIC SERVICES (PBS) PROGRAM

Governmen ernment t & developmen

• pment

t partne ner cooperation peration to speed ed Ethiopia

• pia to MDG targets.

ts. Largest st donor suppor ported d program m in the world. Impl plement emented ed via local al governmen ernments. ts. Current nt (3 (3rd

rd)

) phase se of PBS runs from 2013 13-2018, 8, esti tima mated ed cost st US $6.3 billion

• n split

it 50:50 bet etween een GOE and developmen

• pment

t partne ners. s. Uses es intergo governmen ernmenta tal fiscal cal transf sfer er mech chan anism ism to finan ance ce basic sic services ces in educat cation,

• n,

health, th, agricu cultu ture, e, water suppl ply and rural roads s – constitu stituti tiona nally y decen entraliz tralized ed functi ctions

• ns imp

mpleme ement nted ed at the local level

• el. These

se transf sfer ers s finance nce recurrent nt costs ts (most stly y salari ries es of teacher hers, s, health th extensi nsion

• n worker

ers, s, & agri ricul cultur ture e extensi nsion

• n worker

ers). s). Primary respon

• nsib

sibilit ity for these se basic sic services ices lies s with h woreda da governmen ernments. ts. Ethiopia

• pia is makin

ing g record-brea eaki king ng progress ess to reach h the MDG targets. ts.

PBS APPROACH RELIES ON WDR 2004 MODEL OF SERVICE DELIVERY INTRODUCTION

Stronger Compact More Choice & Participation Stronger Voice

CITIZEN VOICE: FINANCIAL TRANSPARENCY & ACCOUNTABILITY

• Et

Ethiop iopia ia‐sp speci ecific ic FTA A tools ls developed eloped

• Explain budget process to citizens
• Elicit citizen feedback on service delivery
• Translated into local languages
• Shared as example of Africa‐wide budget transparency initiative
• Stron
• ng

g governm ernmen ent t buy uy‐in in of FTA A approa

• ache

ches s is now w institution titutionaliz lized d in the e PFM system em

• More

e than n 90% % of local l governm ernmen ents ts posti ting ng bud udge gets ts

• 50% of servi

vice ce deliv iver ery y un units s disclos closed d their eir plan, , targe gets s and achie ievem emen ents ts to the e pub ublic

• More

e than n 3000 00 local l governm ernment nt offici icials als traine ned d in FTA A tool

• ls

s (entire ntire coun untr try). y). More e than n 270,0 0,000 00 citizens zens traine ned d in budge udget t and bud udge get t process cess so fa far.

9

CITIZEN VOICE: PBS SOCIAL ACCOUNTABILITY COMPONENT (ESAP2) IMPLEMENTATION PROGRESS

From

• m a pilot
• t of abou
• ut

t 80 wored redas through

• ugh capacit

ity y bui uilding ding and awarene reness ss activi viti ties es un under der ES ESAP AP1 1 as of 2010

• SA

A is now w imp mpleme lement nted ed in 224 4 wore redas das throu

• ugh

gh 49 CSOs s us using g over er 5 SA A tool

• ls

s in five e sector

• rs

s [tools

• ls:

: CRCs, CSCs, Participatory Budgeting and Planning, Gender Based Budgeting and participatory PETS].

• A t

A tot

• tal

al of 15 wore reda das have comple pleted d applicati tion

• n of the

e tools ls and moved ed to the e level el of inter erface ce meeti tings ngs - a forum for service providers [373], citizens representatives [3.447] to discuss joint serviced improvement plans.

DATA

Fiscal cal & C Census sus Stat atistical istical analysis ysis using ng cross ss-time time pooled ed data a for all woredas edas.

• FY 2008-2011 so far. Will expand to 1992-2013.

Database abase consis sists ts of 727 woredas edas, , 85 zones es and 10 regions ions

• Demographic data from 2007 census  727 woredas
• Database covers the whole country – all people, all space

DH DHS Analysis ysis of DHS S data a from m 2006 2006-2011 allo lows ws a natural ural experiment periment on the impact act of health lth ext xtensi nsion n worker ers. s.

METHODOLOGY

METHODOLOGY

Three ree key y stages.

• es. Stage

age 1 examine mines the e effects ects of wored reda-le level el spendin nding g on results ts in Ed Educ ucat ation, ion, Health th and Agricu cult lture. ure. lnOmt

mt =

=  + + ln lnEmt

mt +

+ Rm + + Cm+ηlnKmt

mt +τt +

+ mt

mt

(1) (1) lnO = = different rent out utcome come variables ables: : net et enroll

• llment

ment rate, e, antena enata tal care us usage, e, capturin uring key y out utcom comes es in each h sect ctor

• r.

. E = E = yearly arly expend enditu ture re per capita ta in the e relevant ant sector;

• r;

K = capital l expen endit diture ure per capita, a, R = percen entage tage of rur ural popul ulati tion

• n in each

ch wored reda; ; C = vector

• r of demo

mogr grap aphi hic c contr ntrols

• ls (rur

ural/u al/urb rban, n, et ethnici hnicity ty, region ion). ). R and C are from

• m census

nsus and so time-invarian riant. t. Stages ges 2 & 3 3 analyze yze expen endit ditures ures and results ts by wealth alth qui uintil iles. es.

DISTRIBUTION OF WOREDA EXPENDITURES BY SECTOR

Effec ect t of Log of Per Capita ita Educatio ation Expen endi ditu ture e on Log of Educatio ion Outcomes mes Indep ep. . Var ariab iable le Depen enden dent Var ariab iable le/ / Indicat icator

• r

Coef efficien icient/ (SE) Signif ific ic ance Log Expend. Log of Net Enrollment Rate 0.2705 (.0281) *** Log

• f

Pupil- Teacher Ratio

• 0.2242

(.0203) ***

Notes es: Based ed on

• n Cr

Cross-ti time pooled

• oled da

data taset et fr from

• m

2008-2011. St Standa andard er error

• rs giv

given en in in parenthe arenthesis. Signif Signific icanc ance is is def define ined as as: *** *** at at 1% level el. Num umber of

• f Ob

Obser ervat ations ions: 2583 for

• r NER

ER and and 2695 for PTR.

Per capita ta expend penditure ture is strong

• ngly

ly associ

• ciat

ated ed with h rising ng net et enrolme

• lment

nt rate, , and also so fal falling ing pup upil teach cher er ratios.

• s.

One e dollar lar increa rease se in per capita a educ ucat ation ion spendin ending g at wore reda da level el increa rease ses net et enrolme

• lment

nt rate by 3.6%. 6%.

EDUCATION

Effect ect of Log of Pe Per Capit ita Healt lth h Expenditur nditure e on Health th Outcom

• mes

es Indic icat ator

• r

Coeffic icient/ ient/( SE) Signif nific ican an ce ce Log Log Expend nd. Log of Penta 3 vaccinations

• 0.0611

(0.0271) ** Log of Penta 3 vaccinations on expenditure lagged

• ne year

0.1050 (.0370) *** Log of Antenatal Care 0.0784 (0.0341) ** Log of Contraceptive Acceptance Rate

• 0.0250

(0.0404) NS Log of Contraceptive Acceptance on expenditure lagged

• ne year

0.1373 (.0487) *** Log of Deliveries by Skilled Birth Attendants 0.2438 (0.0732) ***

Notes

• tes: Based on
• n cross

cross-time me poole pooled dataset et from

• m 200

2008- 2011

• 2011. Stand

ndard er errors

• rs giv

given en in in parent parenthes esis

• is. Signi

ignific icance nce is is define ined as as: *** *** at at 1% lev level el and nd ** ** at at 5% lev level

• el. Number

umber of

• f

Ob Obser ervations ions: 1,664 664 for

• r Penta

enta 3, 2,277 277 for

• r ANC,

C, 2,243 243 for

• r

Cont Contracepti ceptive acce ccept ptanc nce rate and nd 2,154 154 for

• r Deliv

liver eries ies by by skill illed ed birth attend endants nts.

Matern ernal al mortal alit ity and child mortal ality ty imp mprove with h increasin asing g per capita ta spend ndin ing. g. One dollar r incre rease se in per r capit ital spend ndin ing g imp mproves es Penta ta-3 3 vaccin ccinat ation ion by 4.9%, %, Ac Access cess to ante-nat atal al care by 3.6%, Contracepti traceptive e accept eptan ance ce rates s by 6.4%, %, and Acces ess s to skilled ed birth th attenda ndant nts s by 11. 1.3%. %. (First st two imp mprove e child d mortality ality, latter er two imp mprove e matern ernal al mortal ality ty.)

HEALTH

Association iation of Log of Pe Per Capita ta spending ding on Agric ricult ulture ure Extension ion worker ers with Yield ld

Indep. . Var.

Depen endent dent Variab iable/ le/Indic Indicat ator

• r (Yield

eld in Quinta tales les/Hect ectare) are) Coeffic icie ie nt/(SE) E) Signif nific ic ance Log Linear ear Regres essio ion n Estimat ates es

Log

• g

Expend.

Log of Cereal yield 0.151 (0.0257) *** Log of Pulses Yield 0.016 (0.0455) NS Log of Root Crops yield 0.134 (0.1080) NS Log of Vegetables yield 0.177 (0.0674) *** Log of Oilseeds yield

• 0.075

(0.0851) NS Log of Enset Yield

• 0.604

(0.2315) *** Log of Fruits Yield 0.0233 (0.1360) * Log of Coffee Yield

• 0.057

(0.1302) NS

Per capita ta expend penditure ture is associ

• ciat

ated ed with h increasing reasing yiel elds ds for cereals reals, , vegeta tables bles and d ense set (1% % confide idenc nce inter erval). al). Impa mpact ct on root t crops ps, , pul ulses, es, and d oil-see seeds ds not signif ificant. cant. AGRICULTURE

Bot

• ttom
• m

40% Shar are Top

• p 20%

Shar are Multiple iple by whic ich h Bot

• ttom
• m qu

quintile ile excee eed d top qu quintile ile Educatio ion 56% 13% 2.7 Health alth 63% 10% 3.4 Combin ined ed Educatio ion and d Health alth 58% 12% 2.9

Bott ttom

• m 40% of
• f

populat ulation n receiv ives es onl nly 33% of tota tal l public ic educat ation ion spendin nding g . . 34 34% for health. lth. Woreda da level l spendi nding ng is pro-poor

• or,

, even while le

• verall

rall sector

• r spend

nding ing is not. INCIDENCE OF WOREDA HEALTH & EDUCATION SPENDING ON BOTTOM 40%

INCIDENCE OF HEALTH IMPROVEMENTS (DHS, 2006 – 2011)

Notes: Child and Under five mortality changes are multiplied by negative one to obtain overall improvement.

Effect ect of One Birr per Capit ita a Woreda da spending ding on Agric ricultu ulture re Extensio ion n worker ers on Improved ed Farming ing Techniq hnique ues by Plot

• t Size

Quintile, tile, 2011 Plot

• t size quintile

ntile Any improved technique Smalle llest and Po Poore rest 0.000322*** Second

• nd Smalles

llest 0.000344*** Middle dle 0.000524*** Fourth th 0.000667*** Larges gest and Riche hest 0.000818*** Based on Probit models. Standard errors given in parenthesis. Cross section data with 303,242

• bservations.

Significance is defined as: *** at 1% level, ** at 5% level and * at 10% level, and NS is not significant.

Agric icult ulture ure spending ding improves es techn hniq ique ues s in all landholding holding quintiles. ntiles. Impact act is more re pronounce nounced d for the e bett etter er off. INCIDENCE OF WOREDA AGRICULTURE SPENDING ON IMPROVEMENTS

Effect ects of Educati tion

• n Expendit

nditure ure on NER and NIR by Gender der

• Indep. Var.

Dependent Variable/Indi cator Coefficient/( SE) Significa nce Log Linear ear Regres essio ion n Estimat ates es

Log Log Expen end.

NER 1-8 Male 0.115 (0.0117) *** NER 1-8 Female 0.158 (0.0137) *** NIR 1-8 Male 0.340 (0.0240) *** NIR 1-8 Female 0.431 (0.0269) *** Notes: Significance is defined as: *** at 1% level, ** at 5% level and * at 10% level, and NS is not

• significant. N = 2,583 for NER; N= 2,464 for NIR.

Impac act of extension ion spending ding on Predic dicted ed Probab abilit ilities ies a Field ld using Improved ed Farming ing Techniq nique ues by Gender der, , 2011 Gender der Probability of field using improved technique Male headed ed 0.000027*** Female ale headed ed

• 0.000023***

Notes: Based on Probit models. Standard errors given in parenthesis. Significance is defined as: *** at 1% level. N = 303,242.

GENDER EQUITY COMPARED: EDUCATION AND AGRICULTURE

CONCLUSIONS

Stron

• ng

g associ sociat ation ion of woreda

• reda expenditure

penditures with h imp mproveme ements nts in educ ucati tion,

• n,

healt lth h and agricultu ture.

• re. Less

ess clear ar for agricu cultu ture. re. Wored reda spendin ending g appears s to be pro-poo poor and benef enefit its s the e bott

• ttom
• m 40%

% --

• - muc

uch h more re than n region ional l or federal deral spendin ending. g. Eq Equi uity effect ect of wored reda spendin ending g is very y strong

• ng for educat

ation ion and health.

• alth. Less

Less strong

• ng for agricultu

ture. re. Gende nder equi uity y appears s good

• d for healt

lth h and educat ation, ion, not

• t agricu

cultu ture. re. Targeti ting ng of lagging ng regions ions works ks excep cept t for Somali.

• mali. That

hat appears s driven en by fa fact t that t Soma

• mali

i region ion transf nsferred erred only ly 49% of federal deral grants nts to wore redas das,

• vs. 73% for all regions

gions (excluding cluding Addis is Ababa ba) ) dur uring ng 4 years s of this s study udy. [ Perhaps Somali region is paying teachers, HEWs and Das directly due to the lack of capacity or security in some woredas?]

THANK YOU

CITIZEN VOICE: DECENTRALIZATION Big ig D Dis iscre reti tion

• n: Expe

pendi diture ure and de d decis isio ion-maki making g authorit

• rity

y transfer erre red d from

• m upp

pper er to lower er tie iers of go gover ernme ment. t. Example amples: : How should uld res esour urce ces s be di e divid ided ed betw etwee een n in inves estme ment t and e d expe pendi diture re? ? How much ch should uld be e spe pent on sc schools? hools? Hospi pita tals? ls? Roads ds?  Sorts ts of que uesti tions

• ns the

e lit iter erat ature ure focuses

• cuses on

 Unf Unfortuna tunately ely, lack k of pr pre-de dece centra trali liza zati tion

• n da

data mea eans we c e can't t say anythin ing g about ut this is.

CITIZEN VOICE: DECENTRALIZATION

Small ll Discre reti tion

• n:

: Conc ncerns erns the e effec ects ts of woreda reda-le level el discre creti tion

• n and decision

cision- makin king on servi vice ce qua uality y and d appropriat priatene eness ss to local al condi nditi tions

• ns.

. Ex Examples les: On which ch side e of a road d or stream eam shou

• uld

ld a new w schoo

• ol be bui

uilt? t? Ho How w shou

• uld

ld a vaccina inati tion

• n campa

mpaign ign be targe geted? ed? When hen and how w exact ctly ly should ld a road d be mainta ntained ined?  Suc uch h decisions cisions can increas rease e the e efficien ciency cy of pub ublic c services vices by tailor

• ring

ing them em to highl hly y speci cific ic local l conditions nditions and needs eeds.  Squeez ueeze e more “bang” out of each public “buck”.  Ou Our det etailed ailed micro

• data

a allows ws us us to examine mine this.  Results ts sug ugges gest t local small l discre creti tion

• n is imp

mproving ving the e fit and approp

• pri

riat atene eness ss of servi vice ces s so as to make e them em more re effectiv ective, e, pro-poor poor, and gender nder-neu eutral.

• tral. Thi

his s may be helping ping speed ed Et Ethiopi

• pia

a towards ds its MDGs.

WDR TRIANGLE OF ACCOUNTABILITY AND PBS

Decen centr traliz lizat ation ion brings ngs service vice provide iders s and staff f un under er local al governme ernment nt contr ntrol,

• l, where

here people eople have e more re access cess to leader ders s . Cent nter er decides cides staffing ing level els s & allocat ations, ions, but ut staff f are locally ally hired ed and managed aged  appears s to reduce ce absent enteeis eism m grea eatly tly. Financial ancial transpare parency ncy and d accoun

• unta

tability bility makes es information

• rmation about
• ut local

al spend nding ing available ble to peop

• ple.
• le. Citiz

tizens ns receiv eive e trainin ning g on how w to provide ide feed edback back on bud udge gets ts. Grievance ance Redre ress ss Mecha hanism nism allows ws feedback eedback to the e Et Ethiopian iopian Ombuds budsman, man, including ding regional ional Ombuds budsmen en – independ dependen ent t of governm ernmen ent t and repor

• rts

ts to parliamen ament. t. Struc uctu tured red Social ial Accounta

• untabi

bility lity designe igned d to fa facilit itate us user r feed edba back ck is being ing rolled led out ut to over er 340 wore redas das covering ering 4.5 million ion service vice us users. s.

Outcomes Teachers & Health/ Agriculture Extension workers hired Spending for salaries of Teachers & Health/ Agriculture Extension workers Direct impact:

• Net enrollment ratio
• Pupil-teacher ratio
• Gains in literacy
• Increased life

expectancy

• Lower infant mortality
• Increased agricultural

income Outputs Delivery Inputs Catalytic impact:

• Number of children

vaccinated

• Women receiving ANC
• Number of people using

contraception

• Yield from agricultural

crops

METHODOLOGICAL APPROACH: STAGE II

Incid cidenc ence ana nalys ysis is of woreda da-le level el spending ing by quint intil iles es, using ng the wealth lth quint intile le breakd akdown wn from the DHS survey. We alloca

• cate

e imp mprovements ements in health th and d educati cation

• n outc

tcomes mes to each h quint intil ile e by pro-rat rating ing the expend enditu iture re increas creases es to each h quint intil ile e by the imp mprovem emen ent t in

• utcomes

comes achie ieved ed for that at quintil intile e cont ntrolling lling for the average rage imp mprovement ement for all groups

• ups. That

at is, ∆Ei = (∆Ii/∆I)*∆E (2) ∆Ei is the per capita increase in expenditure for quintile “i”, ∆E is the overall all increas rease e in expenditu nditure re per capita ita, ∆Ii is the change in outcome change for quintile “i”, and ∆I the outcom come chan ange e for all quint ntil iles.

• s. The results

lts indicat cator

• r for educati

cation

• n

was the net et enrol

• llment

lment rate which is direct ctly ly affected ed by woreda da recruit ruited ed teacher chers where e for health th the results lts indica cator

• r is averag

rage e of three e indica icator

• rs which are directly

tly affected ed by locally ally recruit ruited d health lth extens ension

• n work

rkers.

• s. These are:

e: increas rease e use of cont ntrac racep epti tion,

• n,

incre rease ased d rates of immun uniza izati tion

• n,

, incre reas ased ed use of pre-natal natal care re and d increased reased use skilled d birth h atten endants dants.

METHODOLOGICAL APPROACH: STAGE III

We exami mine e the link betw etween een woreda expendi nditu ture e and sect ctor

• ral

al results ts for differe erent nt wealth th qui quinti tiles, es, using ng DHS household ehold survey y data a and limited ed dependen ndent t variab able estim imat atio ions ns as follows ws: Pi,j = f(ϴ,Constan Constant) t) (3) (3) Pi,j = probab abilit ity of improvement for household “j” in the quintile “i” α = wheth ether er the househol ehold has had conta tact ct or not

• t with a health

th exten ensio sion n worker

• er. ϴ is “yes” or “no” variable, with “yes” being coded 1 and “no”

coded ed 0.

CITIZEN VOICE: IMPACT OF THE FINANCIAL TRANSPARENCY AND ACCOUNTABILITY ACTIVITY  A A 2013 h househ sehold

• ld survey

y shows ws citizens izens who know about ut public ic budget t incre reased ased from m 9% in 2009 to 28% in 2013  42.5% of respon pondents dents have seen n the FTA templat plates es posted ed  87% of resp spon

• nde

dent nts s conf nfirmed irmed that t Bud udge get t Literacy eracy Trainin aining g (BLT) T) encouraged raged them m to become me active e parti ticipa cipant nts s in the budget t process ess  37% of those

• se who had ever seen the budget

t inform rmat ation ion post sted ed in their ir area a had some me discussion ussion with h their ir respec pectiv tive e Woreda da/Ci /City ty officials ials; ; 26% with h other er citizens izens  Citi tizen ens s are asking ing for more re engagemen agement t in the budget t process ess

CITIZEN VOICE: GRIEVANCE REDRESS MECHANISM (GRM)

 Capacit ity y bui uilding ding of the e Et Ethiopian iopian Inst stitu tute e of Ombuds budsman man (EI EIO), O), its regional ional branc nches es as well ll as GRM officer icers

 Capacity building training for staff at federal and regional levels  Designing procedures, guidelines and manuals to have similar standards in the country

 Conducti nducting ng stud udy y on exist sting ing GRM system m at fede dera ral l and regional ional levels els and developi eloping ng standa ndardized dized GRM procedur cedural al manual uals s  Providing iding trainin ning g and awarene reness ss creati tion

• n on GRM for key

y stakeh eholde

• lders

s and for the e pub ublic  Sup uppor

• rt

t for the e esta stabl blishme ment nt of addi diti tion

• nal GRM offices

ces in regions ions and trainin ning g for staffs fs in governm ernment nt struc ucture ture on n how w to hear r and redres dress s grievance ances

CITIZEN VOICE: SOCIAL ACCOUNTABILITY IMPLEMENTATION PROGRESS

 Some me main n challenges lenges -  Low capacity among CSO and service providers  Need to continue current investments to reach a critical mass of beneficiaries to cause significant impact and generate representative evidence on sustainability.  Options to selectively seek options for linking SA outputs with public policy

SOURCES OF DATA

Budget t and expenditu nditure re data from m MOFED: D: 2008-200 0011 Poverty y rate data: 2011 1 HIES based PMT applied ed to census sus to develop p poverty ty map. 10% sampl mple e from

• m the last

t Ethi hiopi pian an Census sus: 2007 DHS HS Surveys ys from 2006-20 2011 Agricultur culture and Rural Samp mple Surveys ys for 2007-20 2011 1 (each year differe erent nt zones s are surveyed) d) and over four r years s all zones are covered) ed) Ad Admini nist strativ rative data in educati tion

• n and health:

th:

• Although there is a discrepancy between survey data and HMIS/EMIS data, survey

data is only available at an aggregated level. Over time administrative data is improving and results are

Resid ence Quinti tile le Contra ra cepti tive e Use

Measles es vaccina cinati

• n
• n

ANC with doctor ctor, , nurse e midwif ife e

• r HEW

Deliv liver ery y by doctor ctor, , nurse e midwif ife e or HEW

Urban Total 0.071* **

0.176** * 0.124** * 0.110***

Rural Total 0.094 ***

0.144** * 0.107** * 0.017**

Poorest 0.043 **

0.115** * 0.148** * NS

Second 0.078* **

0.106** 0.073** * NS

Middle 0.069 ***

0.123** * 0.097** * NS

Fourth 0.108* **

0.181** * 0.096** * 0.022*

Richest 0.127* *

0.189* 0.130** NS

Notes: Based on Probit models. Significance is defined as: *** at 1% level, ** at 5% level and * at 10% level, and NS is not significant. Number of observations varies by quintile, location and outcome.

Health th Ex Extension ension Worker ers s seem em to be effecti ective e in helping ping the e poor r access ss basi sic c health alth se servic vices. Ef Effect ect is more re pronou nounc nced ed

• n the

e up upper qui uintiles tiles. PREDICTED PROBABILITIES FOR “SUCCESSFUL” HEALTH OU OUTCOM COMES ES BY PLACE CE OF OF RESIDENC SIDENCE AND WEAL ALTH TH QU QUIN INTILE ILE IF HH WAS AS VISITED ITED BY AN HEW W IN THE PAST ST YEAR AR

SPATIAL VARIATION OF WOREDA EXPENDITURES PER CAPITA

ADDITIONAL QUESTIONS THAT NEED TO BE ANSWERED

With h exce cepti ption

• n of some

e service ice use data a for health th from the DHS there e is not

• t much

ch direct t infor

• rma

mati tion n on fa facil ility ty use Not

• t much

h eviden ence ce on qualit ity of services vices There e is not

• t much

h eviden ence ce on demand nd side e issue ues s which h also play a critica ical role and needs ds to be asses sessed sed The eviden ence ce on the imp mpact act of soci cial al accoun

• unta

tabilit ity on services ices needs s to be e estab ablished. shed.