ABCDE 2019

By: Admasu Asfaw Maruta

Dr. Clifford Afoakwah

Improving Children’s Health through Interventions:

A Quasi-Experiment of GAVI

GAVI, created in 2000, is an international organizationwhich aims bringing together public and private sectors tocreate equal access to new and underused vaccines forchildren living in the poorest countries.

It is funded by government donors, the Bill and MelindaGates Foundation, International Finance Facility forImmunization, GAVI Matching Fund, and PneumococcalAMC (Godal, 2000).

0.000

0.200

0.400

0.600

0.800

1.000

1.200

1.400

1.600

2007 2008 2009 2010 2011 2013

GA

VI (

$US

billi

on)

Year

Health-targeted aid through GAVI ($US billion)

Year Health-targeted aidthrough GAVI ($US billion)

2007 0.313

2008 0.445

2009 0.392

2010 0.533

2011 0.691

2013 1.362

Statistics of health outcomes pre and post GAVI

Pre-2000 Post-2000

Non-GAVI countries

GAVI countries

DiffNon-GAVI

countries

GAVI countri

esDiff

Neonatal mortality

19.99 39.65 -19.69 14.82 28.99 -14.18

Stunting 21.59 44.73 -23.138 18.47 33.51 -15.04

Infant mortality

56.72 67.35 -10.63 37.58 46.43 -8.85

Under-five mortality

48.16 127.59 -79.43 30.15 81.98 -51.83

1020

3040

50Ne

onat

al m

orta

lity ra

tes

19801981

19821983

19841985

19861987

19881989

19901991

19921993

19941995

19961997

19981999

20002001

20022003

20042005

20062007

20082009

20102011

20122013

2014

Year

GAVI countries non-GAVI countries

2040

6080

100

Infa

nt m

orta

lity ra

tes

19801981

19821983

19841985

19861987

19881989

19901991

19921993

19941995

19961997

19981999

20002001

20022003

20042005

20062007

20082009

20102011

20122013

2014

Year

GAVI countries non-GAVI countries

050

100

150

200

Child

mor

tality

rate

s

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

Year

GAVI countries non-GAVI countries

Pre-20000

24

6

2.8 3 3.2 3.4 3.6 3.8

GAVI countries Non-GAVI countries

Neonatal mortality

Post-2000

0.5

11.

52

2.5 3 3.5 4

GAVI countries Non-GAVI countries

Neonatal mortality

Pre-20000

12

34

3.8 4 4.2 4.4

GAVI countries Non-GAVI countries

Infant mortality

Post-2000

.2.4

.6.8

11.

23 3.5 4 4.5 5

GAVI countries Non-GAVI countries

Infant mortality

01

23

4

3.5 4 4.5 5

GAVI countries Non-GAVI countries

Under-five mortality

Pre-2000

.2.4

.6.8

11.

23 3.5 4 4.5 5 5.5

GAVI countries Non-GAVI countries

Under-five mortality

Post-2000

Empirical methodology

We used difference-in-difference estimation toanalyse both short and long-term impacts of theGAVI intervention on children’s health outcomes

Short-term health outcomes proxies (neonataland infant mortality rates and stunting)

Long-term health outcomes proxy (under-fivemortality rate)

𝑪𝑪𝑪𝑪𝒊𝒊𝒊𝒊 = 𝜷𝜷 𝑮𝑮𝑨𝑨𝑨𝑨𝑨𝑨𝒊𝒊 ∗ 𝑷𝑷𝑷𝑷𝑷𝑷𝒊𝒊_𝟏𝟏𝟏𝟏𝟏𝟏𝟏𝟏𝒊𝒊 + ∅𝑪𝑪𝑨𝑨𝒊𝒊𝒊𝒊−𝟏𝟏 + 𝝋𝝋𝑷𝑷𝑪𝑪𝑨𝑨𝒊𝒊𝒊𝒊 + 𝝉𝝉𝑿𝑿𝒊𝒊𝒊𝒊′ + 𝜶𝜶𝒊𝒊 + 𝜸𝜸𝒊𝒊 + 𝜺𝜺𝒊𝒊𝒊𝒊

𝑪𝑪𝑪𝑪𝒊𝒊𝒊𝒊 is child health outcome

𝑪𝑪𝑨𝑨𝒊𝒊𝒊𝒊−𝟏𝟏 is lagged log health aid

𝑷𝑷𝑪𝑪𝑨𝑨𝒊𝒊𝒊𝒊 is log per capita income

𝑿𝑿𝒊𝒊𝒊𝒊′ shows other control variablesincluding fertility rate, log ofpopulation, and CO2 emissions inmetric tons per capita

Main findings

Neonatal mortality Stunting

GAVI*Post_1999-0.113***

(0.012)-0.107***

(0.013)-0.076***

(0.014)0.027

(0.042)0.008

(0.047)-0.001(0.05)

𝑪𝑪𝑨𝑨𝒊𝒊𝒊𝒊−𝟏𝟏 No Yes Yes No Yes Yes

𝑷𝑷𝑪𝑪𝑨𝑨𝒊𝒊𝒊𝒊 No No Yes No No Yes

Country fixed effect Yes Yes Yes Yes Yes Yes

Years fixed effect Yes Yes Yes Yes Yes Yes

N 1656 1326 1281 309 264 257

R-Squared 0.945 0.951 0.953 0.916 0.922 0.920

Infant mortality Under-five mortality

GAVI*Post_1999-0.08***

(0.013)-0.09***

(0.013)-0.07***(0.012)

-0.16***(0.015)

-0.15***(0.016)

-0.11***(0.02)

𝑪𝑪𝑨𝑨𝒊𝒊𝒊𝒊−𝟏𝟏 No Yes Yes No Yes Yes

𝑷𝑷𝑪𝑪𝑨𝑨𝒊𝒊𝒊𝒊 No No Yes No No Yes

Country fixed effect

Yes Yes Yes Yes Yes Yes

Years fixed effect Yes Yes Yes Yes Yes Yes

N 1714 1368 1323 1714 1714 1323

R-Squared 0.953 0.959 0.965 0.957 0.963 0.967

Robustness checks

The effect of GAVI across income groups

The impact of GAVI:

Below 25th

percentile value of PCI

Between 25th and 75th percentile

value of PCI

Above 75th

percentile value of PCI

Neonatal mortality

-0.129***(0.033)

-0.103***(0.014)

-0.081***(-3.17)

Stunting-0.172***

(0.075)0.052

(0.061)0.207

(0.158)

GAVI*Post_1999-0.048(0.033)

-0.107***(0.017)

0.045(0.036)

Under-five mortality

-0.177***(0.037)

-0.145***(0.018)

-0.067**(0.033)

Falsification test

We redefined the pre-GAVI time period as 1960-1979 and post-GAVI time period as 1980-1999.

Variables Neonatal Stunting Infant mortality Under-five mortality

DID-0.056(0.036)

0.716(0.857)

-0.183(0.114)

-0.098(0.070)

Controls, country and year fixed

Yes Yes Yes Yes

N 614 120 614 614

R-Squared 0.982 0.96 0.984 0.980

Testing for omitted variable bias

We use the Oster (2017) to formally test for therelevance of unobservable variables such ashealth governance, institutional development,and health systems factors

We find that the effect of GAVI is notcompromised by omitted variable bias.

GAVI improves children’s health indeveloping countries through thefollowing potential direct andindirect channels.

Its direct investment in vaccines and immunisation in recipient countries

It indirectly creates a competition for the demand for vaccines which reduce vaccine prices.

It improves in global standards for safe injection

Contributions of the study

It supports the growing evidence that theeffectiveness of foreign aid needs to beexamined in a much-disaggregatedapproach.

Thanks