Economics 172 - University of California, Berkeleywebfac/emiguel/e172_s07/lecture6.pdf · Economics 172 Issues in African Economic Development Professor Ted Miguel Department of Economics
Post on 31-May-2020
2 Views
Preview:
Transcript
Econ
omic
s 17
2Is
sues
in A
frica
n E
cono
mic
Dev
elop
men
t
Pro
fess
or T
ed M
igue
lD
epar
tmen
t of E
cono
mic
sU
nive
rsity
of C
alifo
rnia
, Ber
kele
y
Eco
nom
ics
172:
Lec
ture
64
Hea
lth a
nd w
ealth
: cau
se o
r effe
ct?
�H
ow c
an w
e de
term
ine
whe
ther
poo
r hea
lth is
the
caus
e of
pov
erty
(as
Blo
om a
nd S
achs
ass
ert)
or v
ice
vers
a?�
This
is a
diff
icul
t pro
blem
�M
ore
gene
rally
how
to in
terp
ret C
orr(
A, B
) > 0
?1.
�A c
ause
s B
�: A
!B
2. �B
cau
ses
A�:
B !
A3.
A !
Ban
d B
!A
sim
ulta
neou
sly
4. S
ome
othe
r fac
tor C
cau
ses
both
: C !
Aan
d C
!B
5. T
he a
ssoc
iatio
n is
pur
ely
coin
cide
ntal
(but
regr
essi
on
conf
iden
ce in
terv
als
help
add
ress
this
)
Eco
nom
ics
172:
Lec
ture
65
Ano
ther
app
roac
h: a
naly
sis
with
�mic
ro-d
ata�
�B
oth
Blo
om a
nd S
achs
(199
8) a
nd A
JR (2
001)
focu
s on
br
oad
coun
try-le
vel h
isto
rical
tren
ds�
But
est
ablis
hing
cau
salit
y an
d th
eore
tical
cha
nnel
s is
ex
ceed
ingl
y di
fficu
lt in
that
set
ting
�A
noth
er a
ppro
ach
uses
dat
a at
the
leve
l of i
ndiv
idua
ls,
com
mun
ities
, or f
irms
to te
st th
eorie
s ab
out t
he li
nk
betw
een
heal
th a
nd w
ealth
�P
robl
em S
et #
1 fe
atur
es s
ome
anal
ysis
of t
his
kind
Eco
nom
ics
172:
Lec
ture
67
Mig
uel (
2005
) / M
igue
l and
Kre
mer
(200
4)
�E
duca
tion
is a
pos
sibl
e ch
anne
l lin
king
hea
lth, i
ncom
e
Poo
r hea
lthLo
wer
inco
me
Eco
nom
ics
172:
Lec
ture
68
Mig
uel (
2005
) / M
igue
l and
Kre
mer
(200
4)
�E
duca
tion
is a
pos
sibl
e ch
anne
l lin
king
hea
lth, i
ncom
e
Edu
catio
n
Poo
r hea
lthLo
wer
inco
me
Eco
nom
ics
172:
Lec
ture
69
Mig
uel (
2005
) / M
igue
l and
Kre
mer
(200
4)
�W
orm
s ar
e am
ong
the
wor
ld�s
mos
t pre
vale
nt d
isea
ses:
Par
asite
# in
fect
ions
glo
bally
Hoo
kwor
m1.
3 bi
llion
Rou
ndw
orm
1.3
billi
onW
hipw
orm
900
mill
ion
Sch
isto
som
iasi
s20
0 m
illio
n
Eco
nom
ics
172:
Lec
ture
611
Mig
uel (
2005
) / M
igue
l and
Kre
mer
(200
4)
�W
orm
s ar
e am
ong
the
wor
ld�s
mos
t pre
vale
nt d
isea
ses:
Par
asite
# in
fect
ions
glo
bally
Hoo
kwor
m1.
3 bi
llion
Rou
ndw
orm
1.3
billi
onW
hipw
orm
900
mill
ion
Sch
isto
som
iasi
s20
0 m
illio
n
�H
ealth
and
nut
ritio
nal c
onse
quen
ces
�an
emia
, w
eakn
ess,
list
less
ness
, stu
ntin
g, w
astin
g, s
tom
ach
pain
�
espe
cial
ly fo
r hea
vy in
fect
ions
Eco
nom
ics
172:
Lec
ture
612
Mig
uel (
2005
) / M
igue
l and
Kre
mer
(200
4)
�Tr
ansm
issi
on o
f wor
m in
fect
ions
thro
ugh
poor
hyg
iene
, sa
nita
tion.
Elim
inat
ing
one
pers
on�s
infe
ctio
n m
ay re
duce
tra
nsm
issi
on to
oth
ers:
�ext
erna
litie
s� /
spill
over
s
Eco
nom
ics
172:
Lec
ture
613
Mig
uel (
2005
) / M
igue
l and
Kre
mer
(200
4)
�Tr
ansm
issi
on o
f wor
m in
fect
ions
thro
ugh
poor
hyg
iene
, sa
nita
tion.
Elim
inat
ing
one
pers
on�s
infe
ctio
n m
ay re
duce
tra
nsm
issi
on to
oth
ers:
�ext
erna
litie
s� /
spill
over
s
�Tr
eatm
ent i
s ch
eap
(<U
S$1
per
yea
r)�
Dru
gs: a
lben
dazo
le, p
razi
quan
tel
Eco
nom
ics
172:
Lec
ture
614
Hea
lth a
nd e
duca
tion:
cau
se o
r effe
ct?
�W
hat i
s th
e im
pact
of w
orm
infe
ctio
ns o
n ed
ucat
ion?
Eco
nom
ics
172:
Lec
ture
615
Hea
lth a
nd e
duca
tion:
cau
se o
r effe
ct?
�W
hat i
s th
e im
pact
of w
orm
infe
ctio
ns o
n ed
ucat
ion?
�In
dat
a, c
hild
ren
with
wor
se w
orm
infe
ctio
ns te
nd to
hav
e w
orse
edu
catio
nal p
erfo
rman
ce th
an o
ther
chi
ldre
n
�H
ow to
inte
rpre
t Cor
r(H
ealth
, Edu
catio
n) >
0 ?
1. �A
cau
ses
B�:
Hea
lth!
Edu
catio
n2.
�B c
ause
s A
�: E
duca
tion !
Hea
lth3.
A !
Ban
d B
!A
sim
ulta
neou
sly
4. S
ome
othe
r fac
tor C
cau
ses
both
: C !
Aan
d C
!B
5. T
he a
ssoc
iatio
n is
pur
ely
coin
cide
ntal
(but
regr
essi
on
conf
iden
ce in
terv
als
help
add
ress
this
)
Eco
nom
ics
172:
Lec
ture
616
Hea
lth a
nd e
duca
tion:
cau
se o
r effe
ct?
�W
hat i
s th
e im
pact
of w
orm
infe
ctio
ns o
n ed
ucat
ion?
�In
dat
a, c
hild
ren
with
wor
se w
orm
infe
ctio
ns te
nd to
hav
e w
orse
edu
catio
nal p
erfo
rman
ce th
an o
ther
chi
ldre
n
�H
ow to
inte
rpre
t Cor
r(H
ealth
, Edu
catio
n) >
0 ?
1. �A
cau
ses
B�:
Hea
lth!
Edu
catio
n2.
�B c
ause
s A
�: E
duca
tion !
Hea
lth3.
A !
Ban
d B
!A
sim
ulta
neou
sly
4. S
ome
othe
r fac
tor C
cau
ses
both
: C !
Aan
d C
!B
5. T
he a
ssoc
iatio
n is
pur
ely
coin
cide
ntal
(but
regr
essi
on
conf
iden
ce in
terv
als
help
add
ress
this
)
Eco
nom
ics
172:
Lec
ture
619
Om
itted
var
iabl
e bi
as in
OLS
(1)
Y
i= a
+ b
Wi+
cX
i+
e i
Y: e
duca
tiona
l out
com
e (e
.g.,
scho
ol a
ttend
ance
)W
: ind
icat
or v
aria
ble
(=0
or 1
) for
hav
ing
wor
ms
X: c
hild
cha
ract
eris
tic (e
.g.,
hom
e so
cioe
cono
mic
sta
tus)
Eco
nom
ics
172:
Lec
ture
620
Om
itted
var
iabl
e bi
as in
OLS
(1)
Y
i= a
+ b
Wi+
cX
i+
e i
Y: e
duca
tiona
l out
com
e (e
.g.,
scho
ol a
ttend
ance
)W
: ind
icat
or v
aria
ble
(=0
or 1
) for
hav
ing
wor
ms
X: c
hild
cha
ract
eris
tic (e
.g.,
hom
e so
cioe
cono
mic
sta
tus)
a, b
, c: p
aram
eter
s / �
coef
ficie
nts�
to b
e es
timat
ed
Eco
nom
ics
172:
Lec
ture
621
Om
itted
var
iabl
e bi
as in
OLS
(1)
Y
i= a
+ b
Wi+
cX
i+
e i
Y: e
duca
tiona
l out
com
e (e
.g.,
scho
ol a
ttend
ance
)W
: ind
icat
or v
aria
ble
(=0
or 1
) for
hav
ing
wor
ms
X: c
hild
cha
ract
eris
tic (e
.g.,
hom
e so
cioe
cono
mic
sta
tus)
a, b
, c: p
aram
eter
s / �
coef
ficie
nts�
to b
e es
timat
ed
e: w
hite
noi
se �e
rror
/dis
turb
ance
� ter
m, E
(e) =
0i:
deno
tes
pers
on �i
� in
the
popu
latio
n, i
from
1, �
, N
Eco
nom
ics
172:
Lec
ture
622
Om
itted
var
iabl
e bi
as in
OLS
(1)
Y
i= a
+ b
Wi+
cX
i+
e i
Exp
ecta
tions
(�av
erag
e�) o
f the
out
com
e fo
r diff
eren
t cas
es:
1.E
(Yi|
Wi=
1, X
i=1)
= ?
Eco
nom
ics
172:
Lec
ture
623
Om
itted
var
iabl
e bi
as in
OLS
(1)
Y
i= a
+ b
Wi+
cX
i+
e i
Exp
ecta
tions
(�av
erag
e�) o
f the
out
com
e fo
r diff
eren
t cas
es:
1. E
(Yi|
Wi=
1, X
i=1)
= a
+ b
E(W
i| W
i=1,
Xi=
1)+
cE(X
i| W
i=1,
Xi=
1) +
E(e
i| W
i=1,
Xi=
1)
Eco
nom
ics
172:
Lec
ture
624
Om
itted
var
iabl
e bi
as in
OLS
(1)
Y
i= a
+ b
Wi+
cX
i+
e i
Exp
ecta
tions
(�av
erag
e�) o
f the
out
com
e fo
r diff
eren
t cas
es:
1.E
(Yi|
Wi=
1, X
i=1)
= a
+ b
+ c
+ E
(ei|
Wi=
1, X
i=1)
Eco
nom
ics
172:
Lec
ture
625
Om
itted
var
iabl
e bi
as in
OLS
(1)
Y
i= a
+ b
Wi+
cX
i+
e i
Exp
ecta
tions
(�av
erag
e�) o
f the
out
com
e fo
r diff
eren
t cas
es:
1.E
(Yi|
Wi=
1, X
i=1)
= a
+ b
+ c
+ E
(ei|
Wi=
1, X
i=1)
= a+
b+c
Eco
nom
ics
172:
Lec
ture
626
Om
itted
var
iabl
e bi
as in
OLS
(1)
Y
i= a
+ b
Wi+
cX
i+
e i
Exp
ecta
tions
(�av
erag
e�) o
f the
out
com
e fo
r diff
eren
t cas
es:
1.E
(Yi|
Wi=
1, X
i=1)
= a
+ b
+ c
+ E
(ei|
Wi=
1, X
i=1)
= a+
b+c
2.E
(Yi|
Wi=
1, X
i=0)
= a
+ b
+ E
(ei|
Wi=
1, X
i=1)
= a
+b
Eco
nom
ics
172:
Lec
ture
627
Om
itted
var
iabl
e bi
as in
OLS
(1)
Y
i= a
+ b
Wi+
cX
i+
e i
Exp
ecta
tions
(�av
erag
e�) o
f the
out
com
e fo
r diff
eren
t cas
es:
1.E
(Yi|
Wi=
1, X
i=1)
= a
+ b
+ c
+ E
(ei|
Wi=
1, X
i=1)
= a+
b+c
2.E
(Yi|
Wi=
1, X
i=0)
= a
+ b
+ E
(ei|
Wi=
1, X
i=1)
= a
+b3.
E(Y
i| W
i=0,
Xi=
1) =
a +
c +
E(e
i| W
i=1,
Xi=
1) =
a+c
Eco
nom
ics
172:
Lec
ture
628
Om
itted
var
iabl
e bi
as in
OLS
(1)
Y
i= a
+ b
Wi+
cX
i+
e i
Exp
ecta
tions
(�av
erag
e�) o
f the
out
com
e fo
r diff
eren
t cas
es:
1.E
(Yi|
Wi=
1, X
i=1)
= a
+ b
+ c
+ E
(ei|
Wi=
1, X
i=1)
= a+
b+c
2.E
(Yi|
Wi=
1, X
i=0)
= a
+ b
+ E
(ei|
Wi=
1, X
i=1)
= a
+b3.
E(Y
i| W
i=0,
Xi=
1) =
a +
c +
E(e
i| W
i=1,
Xi=
1) =
a+c
4.E
(Yi|
Wi=
0, X
i=0)
= a
+ E
(ei|
Wi=
1, X
i=1)
= c
Eco
nom
ics
172:
Lec
ture
630
Om
itted
var
iabl
e bi
as in
OLS
(1)
Yi=
a +
bW
i+ c
Xi
+ e i
(2)
E(Y
i| W
i=1)
�E
(Yi|
Wi=
0)
= [a
+ b
+ c
E(X
i| W
i=1)
+ E
(ei|
Wi=
1)]
�[a
+ 0
+ c
E(X
i| W
i=0)
+ E
(ei|
Wi=
0)]
Eco
nom
ics
172:
Lec
ture
631
Om
itted
var
iabl
e bi
as in
OLS
(1)
Yi=
a +
bW
i+ c
Xi
+ e i
(2)
E(Y
i| W
i=1)
�E
(Yi|
Wi=
0)
= [a
+ b
+ c
E(X
i| W
i=1)
+ E
(ei|
Wi=
1)]
�[a
+ 0
+ c
E(X
i| W
i=0)
+ E
(ei|
Wi=
0)]
= b
+ c
[E(X
i| W
i=1)
�E
(Xi|
Wi=
0)]
Eco
nom
ics
172:
Lec
ture
632
Om
itted
var
iabl
e bi
as in
OLS
(1)
Yi=
a +
bW
i+ c
Xi
+ e i
(2)
E(Y
i| W
i=1)
�E
(Yi|
Wi=
0)
= [a
+ b
+ c
E(X
i| W
i=1)
+ E
(ei|
Wi=
1)]
�[a
+ 0
+ c
E(X
i| W
i=0)
+ E
(ei|
Wi=
0)]
= b
+ c
[E(X
i| W
i=1)
�E
(Xi|
Wi=
0)]
True
effe
ct�O
mitt
ed v
aria
ble/
sele
ctio
n bi
as� t
erm
Eco
nom
ics
172:
Lec
ture
634
Dea
ling
with
om
itted
var
iabl
e bi
as�
Whe
n is
om
itted
var
iabl
e bi
as n
ot a
pro
blem
?1)
Col
lect
info
rmat
ion
on X
Eco
nom
ics
172:
Lec
ture
635
Dea
ling
with
om
itted
var
iabl
e bi
as�
Whe
n is
om
itted
var
iabl
e bi
as n
ot a
pro
blem
?1)
Col
lect
info
rmat
ion
on X
Bet
ter d
ata
colle
ctio
n re
duce
s O
VB
, few
er X
�s
Eco
nom
ics
172:
Lec
ture
636
Dea
ling
with
om
itted
var
iabl
e bi
as�
Whe
n is
om
itted
var
iabl
e bi
as n
ot a
pro
blem
?1)
Col
lect
info
rmat
ion
on X
Bet
ter d
ata
colle
ctio
n re
duce
s O
VB
, few
er X
�s2)
The
om
itted
var
iabl
e do
es n
ot a
ffect
the
outc
ome
Eco
nom
ics
172:
Lec
ture
637
Dea
ling
with
om
itted
var
iabl
e bi
as�
Whe
n is
om
itted
var
iabl
e bi
as n
ot a
pro
blem
?1)
Col
lect
info
rmat
ion
on X
Bet
ter d
ata
colle
ctio
n re
duce
s O
VB
, few
er X
�s2)
The
om
itted
var
iabl
e do
es n
ot a
ffect
the
outc
ome
c =
0
Eco
nom
ics
172:
Lec
ture
638
Om
itted
var
iabl
e bi
as in
OLS
(1)
Yi=
a +
bW
i+ c
Xi
+ e i
(2)
E(Y
i| W
i=1)
�E
(Yi|
Wi=
0)
= [a
+ b
+ c
E(X
i| W
i=1)
+ E
(ei|
Wi=
1)]
�[a
+ 0
+ c
E(X
i| W
i=0)
+ E
(ei|
Wi=
0)]
= b
+ c
[E(X
i| W
i=1)
�E
(Xi|
Wi=
0)]
True
effe
ct�O
mitt
ed v
aria
ble/
sele
ctio
n bi
as� t
erm
Eco
nom
ics
172:
Lec
ture
639
Dea
ling
with
om
itted
var
iabl
e bi
as�
Whe
n is
om
itted
var
iabl
e bi
as n
ot a
pro
blem
?1)
Col
lect
info
rmat
ion
on X
Bet
ter d
ata
colle
ctio
n re
duce
s O
VB
, few
er X
�s2)
The
om
itted
var
iabl
e do
es n
ot a
ffect
the
outc
ome
c =
03)
The
om
itted
var
iabl
es a
re n
ot c
orre
late
d w
ith th
e ex
plan
ator
y va
riabl
e of
inte
rest
(her
e, W
)
Eco
nom
ics
172:
Lec
ture
640
Dea
ling
with
om
itted
var
iabl
e bi
as�
Whe
n is
om
itted
var
iabl
e bi
as n
ot a
pro
blem
?1)
Col
lect
info
rmat
ion
on X
Bet
ter d
ata
colle
ctio
n re
duce
s O
VB
, few
er X
�s2)
The
om
itted
var
iabl
e do
es n
ot a
ffect
the
outc
ome
c =
03)
The
om
itted
var
iabl
es a
re n
ot c
orre
late
d w
ith th
e ex
plan
ator
y va
riabl
e of
inte
rest
(her
e, W
)�E
xper
imen
tal�
varia
tion
in v
aria
bles
of i
nter
est
Eco
nom
ics
172:
Lec
ture
641
Dea
ling
with
om
itted
var
iabl
e bi
as�
Whe
n is
om
itted
var
iabl
e bi
as n
ot a
pro
blem
?1)
Col
lect
info
rmat
ion
on X
Bet
ter d
ata
colle
ctio
n re
duce
s O
VB
, few
er X
�s2)
The
om
itted
var
iabl
e do
es n
ot a
ffect
the
outc
ome
c =
03)
The
om
itted
var
iabl
es a
re n
ot c
orre
late
d w
ith th
e ex
plan
ator
y va
riabl
e of
inte
rest
(her
e, W
)�E
xper
imen
tal�
varia
tion
in v
aria
bles
of i
nter
est
�If
treat
men
t with
dew
orm
ing
drug
s is
ass
igne
d �r
ando
mly
�, an
d re
duce
s w
orm
infe
ctio
n, th
en th
e re
duct
ion
in w
orm
s sh
ould
be
unco
rrel
ated
with
X
Eco
nom
ics
172:
Lec
ture
642
Mig
uel (
2005
) / M
igue
l and
Kre
mer
(200
4)
�P
rimar
y S
choo
l Dew
orm
ing
Pro
ject
(PS
DP
) in
Ken
ya
Eco
nom
ics
172:
Lec
ture
643
Mig
uel (
2005
) / M
igue
l and
Kre
mer
(200
4)
�P
rimar
y S
choo
l Dew
orm
ing
Pro
ject
(PS
DP
) in
Ken
ya
�R
ando
miz
ed e
valu
atio
n de
sign
: Thr
ee s
teps
(i) S
choo
ls d
ivid
ed b
y ge
ogra
phic
zon
e, (i
i) al
phab
etiz
ed,
(iii)
divi
ded
into
thre
e gr
oups
(1-2
-3, 1
-2-3
, etc
.)
Eco
nom
ics
172:
Lec
ture
644
Mig
uel (
2005
) / M
igue
l and
Kre
mer
(200
4)
�P
rimar
y S
choo
l Dew
orm
ing
Pro
ject
(PS
DP
) in
Ken
ya
�R
ando
miz
ed e
valu
atio
n de
sign
: Thr
ee s
teps
(i) S
choo
ls d
ivid
ed b
y ge
ogra
phic
zon
e, (i
i) al
phab
etiz
ed,
(iii)
divi
ded
into
thre
e gr
oups
(1-2
-3, 1
-2-3
, etc
.)
1998
1999
Gro
up 1
(25
scho
ols)
Trea
tmen
tTr
eatm
ent
Gro
up 2
(25
scho
ols)
Com
paris
onTr
eatm
ent
Gro
up 3
(25
scho
ols)
Com
paris
on C
ompa
rison
Eco
nom
ics
172:
Lec
ture
647
PS
DP
Bas
elin
e di
ffere
nces
(199
8)A
vera
geA
vera
geA
vera
geG
roup
1G
roup
2G
roup
3(N
=25)
(N=2
5)(N
=25)
Eco
nom
ics
172:
Lec
ture
648
PS
DP
Bas
elin
e di
ffere
nces
(199
8)A
vera
geA
vera
geA
vera
geG
roup
1G
roup
2G
roup
3(N
=25)
(N=2
5)(N
=25)
Latri
ne a
t hom
e0.
820.
810.
82Li
vest
ock
at h
ome
0.66
0.67
0.66
Chi
ld s
ick
ofte
n0.
100.
100.
08(s
elf-r
epor
ted)
1996
exa
m s
core
-0.1
00.
090.
01(n
orm
aliz
ed m
ean
0, s
.d. 1
)
Eco
nom
ics
172:
Lec
ture
649
Pro
gram
impa
cts
�Th
ree
type
s of
ana
lysi
s:(1
) Dire
ct tr
eatm
ent e
ffect
s: s
impl
e di
ffere
nce
betw
een
treat
men
t and
com
paris
on s
choo
ls(2
) With
in-s
choo
l ext
erna
lity
impa
cts
(3) C
ross
-sch
ool e
xter
nalit
y im
pact
s
Eco
nom
ics
172:
Lec
ture
650
Hea
lth, n
utrit
ion
impa
cts
(199
9)
Gro
up 1
Gro
up 2
G1�
G2
Rat
e of
mod
erat
e-he
avy
infe
ctio
n0.
270.
52-0
.25*
Sic
k in
pas
t wee
k 0.
410.
45-0
.04*
Hei
ght f
or a
ge-1
.13
-1.2
20.
09*
(Z-s
core
)
Hem
oglo
bin
(Hb)
124.
812
3.2
1.6
Eco
nom
ics
172:
Lec
ture
651
With
in-s
choo
l inf
ectio
n ex
tern
aliti
es (1
999)
Gro
up 1
Gro
up 1
Gro
up 2
(Tre
ated
)(U
ntre
ated
)
Rat
e of
mod
erat
e-he
avy
Infe
ctio
n, 1
999
0.24
0.34
0.52
Eco
nom
ics
172:
Lec
ture
652
With
in-s
choo
l inf
ectio
n ex
tern
aliti
es (1
999)
Gro
up 1
Gro
up 1
Gro
up 2
(Tre
ated
)(U
ntre
ated
)
Rat
e of
mod
erat
e-he
avy
Infe
ctio
n, 1
999
0.24
0.34
0.52
Rat
e of
mod
erat
e-he
avy
Infe
ctio
n, 1
998
0.39
0.41
-
Eco
nom
ics
172:
Lec
ture
653
Cro
ss-s
choo
l inf
ectio
n ex
tern
aliti
es (1
999)
�La
rge
redu
ctio
ns in
mod
erat
e-he
avy
infe
ctio
n le
vels
w
ithin
6 k
m o
f tre
atm
ent s
choo
ls in
199
9
Eco
nom
ics
172:
Lec
ture
656
Cro
ss-s
choo
l inf
ectio
n ex
tern
aliti
es (1
999)
�La
rge
redu
ctio
ns in
mod
erat
e-he
avy
infe
ctio
n le
vels
w
ithin
6 k
m o
f tre
atm
ent s
choo
ls in
199
9
�A
n av
erag
e re
duct
ion
in m
oder
ate-
heav
y in
fect
ions
of 2
3 pe
rcen
tage
poi
nts
in th
e st
udy
area
can
be
attri
bute
d to
cr
oss-
scho
ol e
xter
nalit
ies
Eco
nom
ics
172:
Lec
ture
657
Edu
catio
nal i
mpa
cts
�sc
hool
par
ticip
atio
n
��S
choo
l par
ticip
atio
n� d
ata
colle
cted
by
enum
erat
ors
durin
g un
anno
unce
d pr
imar
y sc
hool
vis
its
Eco
nom
ics
172:
Lec
ture
658
Edu
catio
nal i
mpa
cts
�sc
hool
par
ticip
atio
n
��S
choo
l par
ticip
atio
n� d
ata
colle
cted
by
enum
erat
ors
durin
g un
anno
unce
d pr
imar
y sc
hool
vis
itsG
roup
1 (T
)G
roup
2 (C
)G
roup
3 (C
)Y
oung
er g
irls,
and
all
boys
0.84
0.73
0.77
Old
er g
irls
(≥13
yea
rs)
0.86
0.80
0.81
Pre
-sch
ool,
Gra
des
1-2
0.80
0.69
0.70
Gra
des
6-8
0.93
0.86
0.89
Eco
nom
ics
172:
Lec
ture
659
Edu
catio
nal i
mpa
cts
�ac
adem
ic te
sts
�S
tand
ardi
zed
acad
emic
exa
ms
adm
inis
tere
d in
gra
des
3-8
in 1
998
and
1999
�W
hy m
ight
dew
orm
ing
affe
ct te
st s
core
s?
Eco
nom
ics
172:
Lec
ture
660
Edu
catio
nal i
mpa
cts
�ac
adem
ic te
sts
�S
tand
ardi
zed
acad
emic
exa
ms
adm
inis
tere
d in
gra
des
3-8
in 1
998
and
1999
�W
hy m
ight
dew
orm
ing
affe
ct te
st s
core
s?1)
Incr
ease
d tim
e in
sch
ool (
+)2)
Gre
ater
effi
cien
cy o
f lea
rnin
g (+
)
Eco
nom
ics
172:
Lec
ture
661
Edu
catio
nal i
mpa
cts
�ac
adem
ic te
sts
�S
tand
ardi
zed
acad
emic
exa
ms
adm
inis
tere
d in
gra
des
3-8
in 1
998
and
1999
�W
hy m
ight
dew
orm
ing
affe
ct te
st s
core
s?1)
Incr
ease
d tim
e in
sch
ool (
+)2)
Gre
ater
effi
cien
cy o
f lea
rnin
g (+
)
�B
ut th
e av
erag
e te
st g
ain
from
dew
orm
ing
is z
ero.
Why
?
Eco
nom
ics
172:
Lec
ture
662
Edu
catio
nal i
mpa
cts
�ac
adem
ic te
sts
�S
tand
ardi
zed
acad
emic
exa
ms
adm
inis
tere
d in
gra
des
3-8
in 1
998
and
1999
�W
hy m
ight
dew
orm
ing
affe
ct te
st s
core
s?1)
Incr
ease
d tim
e in
sch
ool (
+)2)
Gre
ater
effi
cien
cy o
f lea
rnin
g (+
)
�B
ut th
e av
erag
e te
st g
ain
from
dew
orm
ing
is z
ero.
Why
?�
Con
gest
ion
effe
cts
in th
e cl
assr
oom
�Ti
me
lags
Eco
nom
ics
172:
Lec
ture
664
Cos
t-ben
efit
calc
ulat
ions
�C
ost o
f thi
s pr
ogra
m: U
S$1
.46
per p
upil
per y
ear
�C
ost o
f a la
rger
-sca
le p
rogr
am in
nei
ghbo
ring
Tanz
ania
: on
ly U
S$0
.49
per p
upil
per y
ear
Eco
nom
ics
172:
Lec
ture
665
Cos
t-ben
efit
calc
ulat
ions
�C
ost o
f thi
s pr
ogra
m: U
S$1
.46
per p
upil
per y
ear
�C
ost o
f a la
rger
-sca
le p
rogr
am in
nei
ghbo
ring
Tanz
ania
: on
ly U
S$0
.49
per p
upil
per y
ear
�D
ewor
min
gas
a h
uman
cap
ital i
nves
tmen
t:H
ealth
gai
ns !
Mor
e sc
hool
ing !
Hig
her a
dult
wag
es
Eco
nom
ics
172:
Lec
ture
666
Cos
t-ben
efit
calc
ulat
ions
�D
ewor
min
gas
a h
uman
cap
ital i
nves
tmen
t:H
ealth
gai
ns !
Mor
e sc
hool
ing !
Hig
her a
dult
wag
es
�D
ewor
min
gle
d to
7%
gai
n in
sch
ool p
artic
ipat
ion
�P
revi
ous
stud
y: e
ach
year
of s
choo
l !7%
hig
her w
ages
�Ta
ke th
ese
gain
s in
wag
es (7
% x
7%
) ove
r 40
year
s in
th
e w
orkf
orce
, dis
coun
ted
5% p
er y
ear
Eco
nom
ics
172:
Lec
ture
667
Cos
t-ben
efit
calc
ulat
ions
�D
ewor
min
gas
a h
uman
cap
ital i
nves
tmen
t:H
ealth
gai
ns !
Mor
e sc
hool
ing !
Hig
her a
dult
wag
es
�D
ewor
min
gle
d to
7%
gai
n in
sch
ool p
artic
ipat
ion
�P
revi
ous
stud
y: e
ach
year
of s
choo
l !7%
hig
her w
ages
�Ta
ke th
ese
gain
s in
wag
es (7
% x
7%
) ove
r 40
year
s in
th
e w
orkf
orce
, dis
coun
ted
5% p
er y
ear
!D
ewor
min
gbe
nefit
s ar
e at
leas
t thr
ee ti
mes
(3x)
as
larg
e as
trea
tmen
t cos
ts (u
sing
the
Tanz
ania
cos
ts)
Eco
nom
ics
172:
Lec
ture
669
Giv
en th
e re
turn
s, w
hy is
take
-up
not 1
00%
?
�P
ossi
ble
expl
anat
ions
:(1
) Fre
e-rid
ing
/ ext
erna
litie
s--
Stro
ng e
vide
nce
peop
le le
arne
d th
roug
h th
eir s
ocia
l ne
twor
k th
at th
e dr
ugs
wer
e �n
ot e
ffect
ive�
Eco
nom
ics
172:
Lec
ture
670
Giv
en th
e re
turn
s, w
hy is
take
-up
not 1
00%
?
�P
ossi
ble
expl
anat
ions
:(1
) Fre
e-rid
ing
/ ext
erna
litie
s--
Stro
ng e
vide
nce
peop
le le
arne
d th
roug
h th
eir s
ocia
l ne
twor
k th
at th
e dr
ugs
wer
e �n
ot e
ffect
ive�
(2) S
ocio
-cul
tura
l exp
lana
tions
/ re
sist
ance
to n
ew
tech
nolo
gies
(evi
denc
e fro
m a
nthr
opol
ogy)
Eco
nom
ics
172:
Lec
ture
671
The
Impa
ct o
f Hig
her D
rug
Cos
ts
�In
199
8, 1
999,
200
0 de
wor
min
gw
as g
iven
for f
ree
�In
200
1, p
aren
ts in
25
rand
omly
cho
sen
Gro
up 1
and
G
roup
2 s
choo
ls p
aid
US
$0.1
0-0.
30 p
er c
hild
Eco
nom
ics
172:
Lec
ture
672
The
Impa
ct o
f Hig
her D
rug
Cos
ts
�In
199
8, 1
999,
200
0 de
wor
min
gw
as g
iven
for f
ree
�In
200
1, p
aren
ts in
25
rand
omly
cho
sen
Gro
up 1
and
G
roup
2 s
choo
ls p
aid
US
$0.1
0-0.
30 p
er c
hild
�20
01 d
ewor
min
gta
ke-u
p:Fr
ee-tr
eatm
ent s
choo
ls:7
5%C
ost-s
harin
g sc
hool
s:18
%
top related