Transcript The long-run economic costs of AIDS: With an application to South

The Economic Impact of
HIV/AIDS
Shanta Devarajan
How HIV/AIDS affects the
economy

Labor supply (e.g., South African
labor force expected to decline by
12.8 percent by 2010)
• But: a 13 percent decline in labor
supply reduces GDP by only 8 percent
or so, implying that GDP per capita rises
How HIV/AIDS affects the
economy (cont’d)

Productivity losses (absenteeism,
retraining workers, death benefits)
• Estimated to add upto 15 percent to
companies’ wage bill (South Africa, Cote
d’Ivoire)
• But: Large companies especially are
able to adapt, reducing productivity
losses
How HIV/AIDS affects the
economy (cont’d)

Public finances
• Increased health expenditures, reducing
public investment
• Reduces GDP growth
• But: Assumes that public investment
would have been productive
Previous estimates of the
impact of AIDS on GDP growth (%)
Arndt and
Lewis (2000)
South Africa
-0.8 to –1.0
Bonnel (2000)
47 countries
-0.7
Kambou,
Devarajan and
Over (1992)
Over (1992)
Cameroon
-0.5 to –1.2
30 countries
-0.3 to –0.6
Sackey and
Swaziland,
Raparla (2000) Lesotho,
Namibia
-0.8 to –1.5
How HIV/AIDS affects the
economy (cont’d)

Human capital
• AIDS kills young adults
• Reduces incentive and means to invest
in children’s education
• Reduces parents’ transmission of
knowledge to their children
Implications



Children’s ability to invest in their
children’s education is lower, and so
on…
Vicious cycle
Previous estimates of impact of AIDS
may seriously underestimate the
long-run impact
Overlapping-generations model
Λt = family’s human capital at t
s=state of family (father dies, mother
dies, etc.)
z(s)Λt(s) = transmission of knowledge
from parents to children
e=amount of education
Then: Λt+1 = z(s)Λtf(e) + 1
Overlapping generations model
(cont’d)
Note that e is the result of the family
solving :
Max EU (ct, Λt+1) s.t.
y=α[Λt + 2(1-e)]
Increase in probability of premature
adult mortality lowers both Λt and
Λt+1
Λt+1
1
45
Λ*
Λt
South Africa: Probabilities
1990
(no
AIDS)
Both
alive
Father
dead
Mother
dead
Both
dead
0.855
0.101
0.039
0.005
South Africa: Probabilities
Both
alive
Father
dead
Mother
dead
Both
dead
1990
(no
AIDS)
0.855
0.101
0.039
0.005
2010
(with
AIDS)
0.294
0.168
0.347
0.194
Effect of AIDS (with pooling)
No AIDS
year
Human
capital
Education
Household
income
1960
2.62
0.50
19.5
1990
3.14
0.64
22.3
2020
4.32
0.97
29.6
2050
7.86
1.00
53.7
2080
13.85
1.00
94.7
Effect of AIDS
No AIDS
AIDS
Household
income
Human
Education
Household
income
0.50
19.5
2.62
0.5
19.5
3.14
0.64
22.3
3.14
0.2
26.4
2020
4.32
0.97
29.6
2.01
0
17.8
2050
7.90
1.00
53.7
1.00
0
12.9
2080
13.85
1.00
94.7
1.00
0
12.9
year
Human
capital
Education
1960
2.62
1990
capital
Policy responses

Spend on public goods to reduce
premature mortality
• Need additional spending of 3-4% of
GDP per year to restore growth

Lump-sum subsidies to families
• Modest growth restored

School-attendance subsidies
• Rapid growth restored
Kenya: Probability of premature
adult mortality
1990
2010
2030
2040
No
AIDS
0.148
0.080
0.0025
0.022
AIDS
0.353
0.359
0.154
0.111
Kenya
No AIDS
AIDS
year
Human
capital
Education
Household
income
Human
capital
Education
Household
income
1990
3.7
0.57
2715
3.7
0.44
2725
2000
3.9
0.66
2317
3.5
0.50
2210
2020
6.0
1.00
2921
4.5
0.78
2323
2030
8.1
1.00
3438
5.8
0.95
2776
2040
10.2
1.00
4168
7.4
1.00
3294
India



Overall prevalence rate <1%
BUT: Six states have generalized
epidemics (>1% prevalence reported in
prenatal clinics)
In one of these states (Tamil Nadu),
percentage of truckers paying for sex rose
in 2002 (to 21%)
• Only 37% of them used condoms

In Orissa, 61% of women have never
heard of AIDS
Conclusion




AIDS is different from other diseases since
it affects young adults
Economic impact of AIDS is not just on
this generation, but the next, whose
education will suffer
Economic costs can be huge, and felt
many years from now
Early and strong action can reduce these
costs