University of Southern California Dissertations and Theses

Political determinants and economic effects of HIV/AIDS: a push for the multisectoral approach

(USC Thesis Other)

Political determinants and economic effects of HIV/AIDS: a push for the multisectoral approach

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content POLITICAL DETERMINANTS AND ECONOMIC EFFECTS OF HIV/AIDS:
A PUSH FOR THE MULTISECTORAL APPROACH

by

Dollie Davis

__________________________________________________________________

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(POLITICAL ECONOMY AND PUBLIC POLICY)

December 2008

Copyright 2008 Dollie Davis

ii
Dedication

To Erik, Mom, and Dad.

iii
Acknowledgements

I am grateful to Professors Jeffrey Nugent and Carol Wise for invaluable
guidance, insights, patience, critiques, opportunities, and encouragement. Professors
Iris Chi and Taha Al Sabea have provided me with much appreciated time and
valuable input. Many professors and staff at USC, in both the Economics and
International Relations Departments, have been extremely helpful with this process.
Along with Professor Nugent, I want to thank all of those who participated in his
weekly Development Group. Your time, comments, and encouragement have been
vital to the completion of my work. Finally, I would like to thank all the
interviewees, researchers, authors, and experts who have provided me with essential
information to complete this dissertation.
I would have never been able to begin or finish this work without the
astounding encouragement, patience, and back rubs from my husband, Erik Yosten.
My father, Charles Davis, provided perpetual sage advice, invaluable assistance, and
undying support. My Mother, Sheila Melen, was there for every need and phone
call, and somehow not a one of them ever stopped believing in me. To my family
and friends: thank you all for the endless patience, encouragement, inspiration, and
love. I truly cannot thank you all enough and cannot express how blessed I feel to
have you in my life.

iv
Table of Contents

Dedication ii

Acknowledgements iii

List of Tables v

List of Figures vii

Abstract viii

Chapter 1: Introduction 1
Conclusion 9

Chapter 2: Literature Review 11
Economic Effects of HIV/AIDS 12
Societal Determinants of HIV/AIDS 15
Political Determinants of HIV/AIDS 25
Multisectoral HIV/AIDS Programs 37
Conclusion 46

Chapter 3: Multisectoral HIV/AIDS Programs in Practice:
MAP Review, Interviews, Case Studies 48
MAP Review 48
Interviews 56
Case Studies 60
Conclusion 71

Chapter 4: Methodology, Models, Results, Difference-in-Difference
Approach 73
Methodology 73
Models and Results 83
Difference-in-Difference Approach 130
Discussion of Results 136
Conclusion 138

Chapter 5: Concluding Remarks 139
Reflection on the Findings 139
Future Research 143

Bibliography 145

v
List of Tables

Table 1: List of Models on Political and Societal determinants of
HIV/AIDS 74

Table 2: List of Models on economic effects of HIV/AIDS 75

Table 3: List of Models on determinants of a HIV/AIDS
multisectoral program 75

Table 4a: Explanation of Variables 76

Table 4b: Descriptive Statistics 78

Table 5: Results Model 1-Dependent Variable: 1997 HIV/AIDS Rate 84

Table 6: Results Model 2a-Dependent Variable: 2005 HIV/AIDS Rate 89

Table 7: Results Model 2b-Dependent Variable: 2005 HIV/AIDS Rate 93

Table 8: Wald Test – Model 2b 94

Table 9: Exogeneity Test – Model 2b 94

Table 10: Results Model 2c-Dependent Variable: 1995 HIV/AIDS Rate 98

Table 11: Results Model 3a-Dependent Variable: ΔHIV/AIDS Rate
(1995-2005) 103

Table 12: Results Model 3b-Dependent Variable: ΔHIV/AIDS Rate
(1995-2005) 107

Table 13: Wald Test – Model 3b 108

Table 14: Exogeneity Test – Model 3b 109

Table 15: Results Model 4-Dependent Variable: 2005 Expenditures
on HIV/AIDS 114

Table 16: Results Model 5 Dependent Variable: 2005 Women’s
Condom Knowledge 118

vi
Table 17: Results Model 6-Dependent Variable 1990-2005
GDP Growth Rate 122

Table 18: Results Model 7-Dependent Variable Has a Multisectoral
Program 126

Table 19: Difference in the Differences – I 131

Table 20: Difference in the Differences – II 132

Table 21: Differences in Differences – III 133

Table 22: Difference in the Differences – IV 134

Table 23: Difference in the Differences – V 135

Table 24: Difference in Difference Results 137

Table 25: t-Test: Two-Sample Assuming Equal Variances 138

vii

List of Figures

Figure 1: Motivation for Research on HIV/AIDS Source: Gauri
And Lieberman (2004) 6

Figure 2: Probabilities and effects from Risky HIV/AIDS
Behavior with Multisectoral Start Dates Source: Davis (2008) 22

Figure 3: National HIV/AIDS Council (NAC) Source: Ayvalikli,
Brown, and Mohammad (2004) 41

Figure 4: MAP Input-Level and Output-Level Results: Source:
Gorgens-Albino, Mohammad, Blankhart, and Odutolu (2007) 50

Figure 5: HIV/AIDS Rates in Uganda 1987-2002 Source:
Avert Charity 63

viii
Abstract

The proposed dissertation offers an explanation for the large differences in
HIV/AIDS rates among 89 low and middle-income countries throughout the Sub
Saharan African, Asian, and Latin American regions over a ten-year period (1995-
2005). The HIV/AIDS rates in these countries vary widely and seemingly
independently of economic wealth. One possible determinant of these differences is
the presence and degree of development of strong multisectoral programs aimed at
both prevention and cure of HIV/AIDS. The main hypothesis for this dissertation is:
“A country’s success in combating HIV/AIDS lies in the government’s ability to
implement an effective multisectoral program.” This hypothesis is explored through
quantitative models using data from the ten-year period (1995-2005). Results show
that the presence of a multisectoral program over the ten-year period is associated
with a significantly lower HIV/AIDS incidence rate by 2005. This effect is produced
by controlling for various political, economic, societal, and institutional factors.
Although there is some anecdotal evidence which suggests that multisectoral
programs help to improve the HIV/AIDS problem in developing countries, there has
been little if any empirical work done on this subject to date.

1
Chapter 1: Introduction

This dissertation seeks to explain why some low-income countries like
Zimbabwe and Cambodia continue to struggle with high HIV/AIDS rates while
many others like Senegal and the Philippines have rates that are both low and
declining. For example, Senegal and Zimbabwe both face widespread poverty and
illiteracy but the HIV/AIDS rate among Senegalese adults at the end of 2005 was
only 0.8% while 20.1% of adults were infected in Zimbabwe.
1
With rates like those
of Zimbabwe, and with such high poverty, it is clear that this epidemic poses a direct
threat to these economically fragile countries.
Recent projections convincingly show that the economies of those countries
increasingly affected by HIV will decline, as this disease hampers key inputs to
economic growth such as labor productivity, social cohesion, domestic savings and
human capital. Similarly, by preventing a youthful workforce from completing the
necessary education to work and save, sustainable growth in these high-HIV/AIDS
developing countries will be undermined. According to Bodiang (2001:3), “If the
epidemic continues at its present rate, the hardest-hit nations stand to lose up to 25
percent of their projected economic growth over the next 20 years.”

1
Joint United Nations Programme on HIV/AIDS, UNAIDS, “Uniting the World Against AIDS.”
(2007). Current HIV/AIDS prevalence rates from HIV INSITE.

2
Some developing countries have successfully organized to combat this
epidemic while others remain quite vulnerable to further spread of the disease.
2
A
popular explanation for this latter scenario, albeit a somewhat tautological one, has
to do with a lack of material resources and education. However, my argument is that
the nature of bureaucratic institutions within each country is a more important factor.
The empirical results presented in this dissertation suggest that a developing country
government that has adopted a particular policy known as a multisectoral approach
to HIV/AIDS has a much higher probability of successfully fighting this disease,
irrespective of its poverty rate and educational level.
For the purpose of this dissertation, the definition of a multisectoral program
derives from The Commonwealth Secretariat (2003:2):
A multisectoral response means involving all sectors of society—
governments (beyond just the health ministry: agriculture, finance,
education, tourism, etc.), business, civil society organizations,
communities and people living with HIV/AIDS in addressing the causes
and impact of the HIV/AIDS epidemic.

In other words, by incorporating all of the various sectors and representatives of civil
society into the fight against HIV/AIDS, a collective voice can better articulate the
kinds of reforms that are needed. This approach contrasts with that of relegating the
HIV/AIDS problem solely to the health ministry, a practice that is common among
many low-income countries. Although multisectoral HIV/AIDS programs are on the
rise, there remains a lack of data and information which has rendered program and
policy analysis especially challenging.

2
South Africa’s vulnerability is discussed in Poku et al. (2004).
3
Clearly, HIV/AIDS affects all facets of government and society and hence
the need for a multisectoral approach. As one analyst puts it: “There is a direct
relationship between HIV infection and poverty, inequality, the status of women in a
society, social disruption, illiteracy, human rights violations and all the other factors,
which define the context for development work” (Bodiang 2001:4). A multisectoral
program can work because
It creates a mechanism for coordination of information along with an
inclusion of all in society, regardless of the sector, to speak together as a
larger, influential voice. This is what is needed to convince decision
makers in national governments that something needs to be done (The
Commonwealth Secretariat 2003:2).

Furthermore, there is evidence that multisectoral programs managed by actors that
coordinate their goals are more successful than those that work in isolation.
3

A multisectoral approach is also important because it allows those vulnerable
and ignored groups in society (many of which have high AIDS incidence rates) to be
included in a national collective approach with one common goal. Through the work
of local non-governmental organizations (NGOs) and their established networks,
there is an inclusion of these fragile and otherwise inactive groups.
A typical multisectoral approach might involve the education ministry setting
up HIV/AIDS information programs for students and teachers. Also, families that
depend on agriculture for subsistence can be made aware of these programs and a
variety of government units such as the military, labor and other departments can be
called upon to foster comprehensive complementary HIV/AIDS prevention

3
UNAIDS (1999).
4
programs. NGOs can then work closely with the government to identify and educate
the vulnerable groups in society.
Thailand is an example of a low-income country with a successful
multisectoral HIV/AIDS program. The Multisectoral National AIDS Prevention and
Control Committee, established in 1991 by the Thai Prime Minister, united several
governmental sectors that worked with NGOs, businesses, and the media to set up an
extensive anti-AIDS/pro-condom campaign. The program's results were impressive:
Thailand’s rate of new HIV infections reportedly fell from 140,000 in 1992 to 30,000
in 2001. By contrast, South Africa’s government did not aggressively implement a
multisectoral program until 2000 while the HIV/AIDS prevalence rate rose from
3.2% around 1995 to 19.94% around 1999. Since the inception of the multisectoral
program, however, this trend has slowly been reversed and South Africa’s AIDS rate
had fallen to 18.8% in 2005.
4

There are other political determinants (besides a multisectoral program), as
well as societal factors, and these potentially influence HIV/AIDS rates in
developing countries. These variables will be thoroughly described in Chapter 4.
One possible political determinant of HIV/AIDS would be: a country’s involvement
in civil war, while examples of societal determinants would include the legalization
of prostitution, the percentage of the population that is Muslim and, the percentage of
the population that are males aged 15-30. In particular, the outbreak of civil wars
can exacerbate the spread of HIV/AIDS, as forced sex and sex with irregular partners

4
Current HIV/AIDS prevalence rates from HIV INSITE.
5
is more likely to occur under these circumstances. As for the other variables
mentioned, the legalization of prostitution, which normally requires the monitoring
of sexual behavior and social programs for sex workers, is expected to decrease the
incidence of AIDS. Similarly, for Muslim populations, sex outside of a marriage is
forbidden and is met with severe punishment thereby decreasing the likelihood of
exposure to multiple partners. Males of productive ages 15-30 are more likely to
engage in sexual behavior and therefore may be at higher risk of contracting the
disease. Additional political, societal, and economic variables such as a country’s
per capita GDP and level of income inequality will be introduced in upcoming
chapters.
The main hypothesis for this dissertation is that over the past ten years (1995-
2005) a country’s ability to combat HIV/AIDS lies in the extent to which its
government has implemented a multisectoral HIV/AIDS program. But several other
relationships will also be considered: the impetus for establishing a multisectoral
program, the determinants of condom awareness and those forces that prompt a
government to spend more on HIV/AIDS. Finally, the direct relationship between
HIV/AIDS and economic growth will be explored to see if the AIDS epidemic in
fact hinders economic growth. These relationships will be examined by using
multivariate regression analysis.
As seen in Figure 1 below, the highest HIV/AIDS incidence rates are
disproportionately grouped within the lowest income countries. Indeed, there is no
group that has rates as high as those of Sub Saharan Africa.
6

Figure 1: Motivation for Research on HIV/AIDS Source: Gauri and Lieberman
(2004)

There are currently 24.5 million adults and children living with HIV and 12
million orphans (caused by AIDS-related deaths) in Sub Saharan Africa.
5
These
incidence rates differ between countries, as Figure 1 shows, with rates running at
over 20% of the population in many countries in the southern part of the globe and
less than 1% in some of the western countries. For Sub Sahara Africa, prevalence
rates peaked in the late 1990’s before leveling off, but they did so at currently high
rates. With severe poverty, infrastructural shortages, and political instability in many
of these countries, even the most basic actions to help curb the epidemic are rarely
taken.

5
Avert Charity, “Sub Saharan Africa HIV & AIDS statistics.”

Motivation for Research on HIV/AIDS Motivation for Research on HIV/AIDS
Total: 34 – 46 million
>90 percent in developing countries
2.5-3.5 million died in 2003 alone
Western Europe
520 000 – 680 000
North Africa & Middle
East
470 000 – 730 000
Sub-Saharan Africa
25.0 – 28.2 million
Eastern Europe
& Central Asia
1.2 – 1.8 million
South
& South-East Asia
4.6 – 8.2 million
Australia
& New Zealand
12 000 – 18 000
North America
790 000 – 1.2 million
Caribbean
350 000 – 590 000
Latin America
1.3 – 1.9 million
East Asia & Pacific
700 000 – 1.3 million
Adults and children estimated to be living with HIV/AIDS as of e Adults and children estimated to be living with HIV/AIDS as of end 2003 nd 2003
Source: UNAIDS (2003)
7
Antiretroviral drugs (ARVs) can slow HIV infection rates and even prolong
the onset of AIDS for up to twenty years. This extra time is vital as it could be used
for further education and increased participation in the workforce. Even though
these drugs are cheaper in the African countries, there is an unmet need for much
higher access to ARVS throughout this region.
At least 85% (almost 900,000) of South Africans who needed
antiretroviral drugs were not yet receiving them by mid-2005; the same
applied to 90% or more of those in need in countries such as Ethiopia,
Ghana, Lesotho, Mozambique, Nigeria, the United Republic of Tanzania
and Zimbabwe (UNAIDS 2005:2).

Ideally, multisectoral HIV/AIDS programs can help in identifying and distributing
ARVs to those groups who are most in need of treatment.
A main focal point of this dissertation is the World Bank’s Multi-Country
HIV/AIDS Program for Africa (MAP). Launched in 2000, this program is slated for
12-15 years and aims to support the mobilization of the Sub Saharan African
countries against the HIV/AIDS epidemic.
6
Initially, the World Bank committed
US$500 million both in 2001 and 2002 to fund HIV/AIDS programs. Nearly every
country in Africa is able to access these funds as long as they provide proof that a
sound national anti-AIDS strategy is in place and endorsed by the government. The
prime prerequisite is that each country is required to establish a multisectoral
coordinating body. The World Bank directly funds various ministries and other
government agencies and it oversees the results using previously agreed on targets

6
Found in the Generic Operations Manual for MAP (2004).

8
and timetables. Today, 29 African countries are participating in the MAP and in
each of these countries a high-level body has been created to oversee the national
multisectoral HIV/AIDS program. As a result HIV/AIDS rates have started to fall in
many of these countries.
In Asia, the disease also reached high levels in the late 1990’s but, unlike Sub
Saharan Africa, has failed to level off. In 2003 alone, there were over one million
newly infected Asians. AIDS is now growing more rapidly in Asia, with 7.4 million
people living with HIV/AIDS.
7
For example, Indonesia, one of the most populous
countries in the world, has faced increasing infection rates primarily among
commercial sex workers (CSWs) and injection drug users (IDUs). An increase in the
HIV/AIDS rate is of particular concern in other highly populated countries,
specifically China and India. Here “the epidemic…may be masked by the large
population leading to low reported prevalence, which has resulted in considerable
complications among government and developmental stakeholders” (Hossain
2007:1). A potential reason that the World Bank has not introduced multisectoral
programs in the Asian region is that when taken at face value, the HIV/AIDS rates
don’t yet appear to be very high. Yet as the region’s HIV rates grow apace, action is
urgently needed to prevent the high economic and social costs of the epidemic.
Like Asia, Latin America has a relatively small but nevertheless rather
alarming incidence of HIV/AIDS throughout the region. Although it has the smallest
number of people estimated to be living with HIV/AIDS, namely 1.6 million people,

7
UNAIDS (2003).
9
prevalence rates in this region have generally not decreased over time. This
indicates that effective anti-AIDS programs have not been established in many of
these countries. The majority of those infected in Latin America are men who
engage in homosexual activities. This is a particularly vulnerable group because of
the homophobia and stigmatization intrinsic to this region. The good news is that
Latin American governments have effectively distributed ARVs across the region.
At the end of 2006, almost 76% of those in need had received these drugs.
Unfortunately, the majority of ARVs are dispensed in cities and not in the rural areas
where many people are also in need. Several Latin American countries have NGOs
with an explicit mandate to combat HIV/AIDS operating on the ground; but without
stronger political leadership in this realm, both the health and economic growth of
these countries are likely to fall victim to this epidemic.

Conclusion
The purpose of this dissertation is to examine the relationship between the
successful adoption of a multisectoral HIV/AIDS program and the degree to which
countries have halted the spread of this disease. The importance of this research lies
in the urgent need to combat AIDS as a goal in itself, but also because actively
addressing the AIDS problem is key to sustaining economic growth in those
countries plagued by this epidemic. A first step in my research is to identify those
variables that directly affect the HIV/AIDS rate in the AIDS-infected countries.
Again, my core argument is that over the past ten years (1995-2005) a country’s
10
success in effectively combating HIV/AIDS lies in the government’s ability to put in
place an effective multisectoral program.
Several studies have sought to explain the crisis of HIV/AIDS in developing
countries, particularly Sub-Saharan Africa, but from a different angle. In Chapter 2,
those studies most relevant to this dissertation are reviewed including: research on
the economic effects of HIV/AIDS, the societal determinants of HIV/AIDS, the
possible political determinants of HIV/AIDS, and literature on the multisectoral
HIV/AIDS approach. Chapter 3 includes a review of the current results from the
World Bank’s Multi-Country HIV/AIDS Program for Africa (MAP) as well as
information gleaned from interviews with experts directly involved with the design
and implementation of multisectoral HIV/AIDS programs; three case studies
regarding the success of multisectoral HIV/AIDS programs in various countries will
then be presented. Chapter 4 lays out the methodology, models, and results of my
quantitative analyses while Chapter 5 provides concluding remarks.

11
Chapter 2: Literature Review

Literature on the HIV/AIDS epidemic is broad and varies in scope from
simple biological facts to intricate personalized stories. For this literature review,
focus is placed on the economic effects as well as the social and political
determinants of the disease. There have been a number of previous studies regarding
the negative economic effects related to high HIV/AIDS rates. Some of these studies
concentrate on the economic effects of HIV/AIDS in one country while others
explore entire regions. Other studies look at societal determinants (such as education
and gender inequality) and how these may influence the HIV/AIDS rate. Few
studies, however, have been published on the political determinants of HIV/AIDS.
It is of particular interest to this dissertation to identify those underlying
political determinants that may account for the broad differences in the success or
failure of lowering HIV/AIDS rates across the developing countries. After
reviewing literature on the economic effects of HIV/AIDS, an examination of
literature on the societal determinants as well as political determinants of HIV/AIDS
will be undertaken. Finally, I will review the available literature on general aspects
of those multisectoral HIV/AIDS programs now in place in some countries. The
literature review will be divided into four parts:
1) Economic Effects of HIV/AIDS - the possible economic implications of
rising HIV/AIDS rates in developing countries.
12
2) Societal Determinants of HIV/AIDS - those societal factors that can affect
HIV/AIDS rates in developing countries.
3) Political Determinants of HIV/AIDS - those political variables that can
affect HIV/AIDS rates in developing countries.
4) Multisectoral HIV/AIDS Programs - those general variables that explain a
government’s adoption of a multisectoral HIV/AIDS program.

Economic Effects of HIV/AIDS
A number of studies have estimated the relationship between HIV/AIDS and
economic growth overtime to show that the disease does, in fact, hurt growth. Other
studies have employed simulation models to gauge what could happen to economic
growth in a country should the HIV/AIDS epidemic go unmanaged. Examples of
both types of studies are reviewed below.
In 1992, Mankiw et al. extended the standard Solow model of economic
growth to include human capital (skilled labor) H as well as the original inputs to
growth: capital K, labor L (and later an exogenous measure for technological
progress A). In doing so, they observed that the level of skilled labor differs among
rich and poor economies and this has important implications for economic growth.
In answering the question as to why some countries are richer than others Mankiw et
al. (1992) summarize a model that has been interpreted to mean: “Countries are rich
because they have high investment rates in physical capital [K}, spend a large
fraction of time accumulating skills [H], have low population growth rates, and have
high levels of technology [A]” (Jones 2002:57). While this growth model has proven
13
to have high predictive power, it does not include health as an input for economic
growth. My point: poor health as a factor could likely worsen growth through its
adverse impact on productivity, among other things.
Ten years after Mankiw et al. (1992) introduced human capital into the
standard economic growth equation, McDonald and Roberts (2002) incorporated a
variable for health and found it had a positive and significant impact on economic
growth. It seems that the next step in the economic growth literature will be to
include the actual incidence of an epidemic like HIV/AIDS into the growth model.
McDonald and Roberts (2006) introduce the problem of HIV/AIDS into their growth
model by measuring the disease’s impact on health as a key component of human
capital and population growth. Health capital for this study consists of several
determinants, including past per capita income, nutritional status, and the actual
prevalence of HIV/AIDS.
Quantitative models in McDonald and Roberts (2006) are estimated by 5 year
intervals over the years 1960 to 1998. One hundred and twelve countries from both
the developed and developing world are separated into six samples: Latin America
and the Caribbean, Asia, Africa, the OECD bloc, the developing world, and the
entire world. The dependent variable is the level of per capita income at the end of
each 5-year period. After executing a growth model, the authors find that
. . . the HIV impacts are negative, with Africa recording an average
reduction of 0.59 percent in income per capita for one percent increase in
HIV prevalence, while the World and Developing World sample indicate
negative impacts of 0.05 and 0.08 percent (McDonald and Roberts 2006:
14).

14
A similar study by Ukpolo (2004) looks at the effects of HIV/AIDS on
growth in several countries over time. Ukpolo (2004) uses an ordinary least squares
(OLS) regression (similar to the one used in this dissertation) to examine the
relationship between economic growth and HIV/AIDS in 17 African countries
during the period from 1990 to1996. The author finds that HIV/AIDS had a negative
impact on economic growth over the time period. Other results from Ukpolo (2004)
indicate that openness of the economy as well as an increase in the labor force has a
positive impact on economic growth.
Some of the many reasons that HIV/AIDS hurts economic growth are offered
below in an excerpt from a report by Malawi’s National AIDS Commission (2003):
Illness prevents the primary breadwinner from working, increases the
amount of money the household spends on health care, and requires
other household members to miss school or work in order to care for the
patient. Death of the patient results in a permanent loss of income,
either through lost wages and remittances, or through a decrease in
agricultural labour supply. Households must also bear the costs of
funerals and mourning, which in some settings are substantial. When
children are withdrawn from school in order to save on educational
expenses and increase the labour supply, the household suffers a severe
loss of future earning potential (National AIDS Commission 2003:35).

The above literature suggests that HIV/AIDS has a negative impact on
economic growth over a certain time period. A study by Lewis and Arndt (2000)
looks at this negative growth impact over a future time period. Their results are
based on a simulation model of the South African Economy that uses available data
such as labor supply and HIV/AIDS prevalence rates. They compare the GDP
effects of HIV/AIDS under two possible scenarios, one with the current HIV/AIDS
15
rate (which was 21.5% at the time of their study), and the other a zero rate of
HIV/AIDS. Running their model from 2000-2010, the authors find that the economic
toll from HIV/AIDS in South Africa could translate to a 17% drop in GDP; their
other predictions included a fall in population growth, labor productivity, total factor
productivity, and a spending shift toward health care by consumers and government.
To date, the HIV/AIDS rate remains high in South Africa, but (according to available
data) it did drop from its peak level 21.5% in 2002 to 18.8% in 2005. Economic
growth is currently running at around 4.2% in South Africa, although growth did
take a dip (2.74%); coincidentally around the time when the HIV/AIDS rate was at
its highest.
Other economic growth studies and simulation models (similar to those
reviewed in this section) have consistently found that HIV/AIDS has a negative
relationship with economic growth. While it is clear that the economic effects from
the disease are severe, it is also important to examine the determinants of HIV/AIDS.

Societal Determinants of HIV/AIDS
A number of studies seek to identify those key determinants that influence
HIV/AIDS rates in developing countries. The majority of these studies concentrate
primarily on societal variables, such as literacy rates and religion in explaining the
spread of the disease. Below I review those works that rely on societal variables as
determinants of the disease.
16
Mead Over (1998) offers a general look at the determinants of HIV/AIDS in
developing countries in a paper entitled “The Effects of Societal Variables on Urban
Rates of HIV Infection in Developing Countries.” Specifically, the author is
concerned with the high HIV/AIDS rates in the urban areas of some developing
countries as opposed to very low rates in others.
8
Available data from 1997 are used
to explain differences in HIV/AIDS prevalence rates throughout 72 countries in Sub
Saharan Africa, Latin America & the Caribbean, Asia, and the Middle East. The
HIV/AIDS rate is the dependent variable for this cross-sectional study and the
explanatory variables include: per capita GNP, percent of the population that is
foreign-born, the Muslim percentage of the population, the Gini index for income
inequality, the male to female literacy gap, urban male to female equality ratios, and
the military’s representation as a percentage of the urban population.
Over (1998) finds that all variables significantly affect the HIV/AIDS rate;
the strongest, negative determinants of HIV/AIDS in a country are the degree of
income inequality, poverty levels, and a large military presence in urban areas.
These results seem to indicate that HIV/AIDS policies which promote a more equal
distribution of income, a boost in employment, and a push for HIV/AIDS awareness
in the military are of utmost importance. These same variables are included in the
estimation results in Chapter 4 in a re-estimation of Over (1998) with the inclusion
of dummy variables for the countries which had multisectoral HIV/AIDS programs
by 1997.

8
HIV/AIDS rates are often higher in urban areas than in rural areas of developing countries.
17
Another study that looks at societal determinants of HIV/AIDS is
“Inequalities in Knowledge of HIV/AIDS Prevention: An Overview of Socio-
Economic and Gender Differentials in Developing Countries” by Gwatkin and
Deveshwar-Bahl (2001). These authors attempt to estimate the success of
HIV/AIDS education in reaching poorer segments of the population. For twenty-
three developing countries, they examine the inequalities of knowledge regarding
HIV/AIDS. The authors first compare the relationship between wealth differences
and knowledge about HIV/AIDS prevention and then they compare gender
differences and knowledge regarding HIV/AIDS prevention. After collecting and
analyzing World Bank data, the authors find that the richest strata of the population
across all countries has 22% higher knowledge regarding HIV/AIDS prevention than
the poorest segment of the population. Furthermore, men have a 20% higher rate of
knowledge about HIV/AIDS than do women. It is clear that, for this study, the
societal determinants of HIV/AIDS are poverty, lack of education, and gender
inequalities. This study reinforces the proposition that sectors and agencies beyond
the health ministry need to be included in programs to combat HIV/AIDS. From this
study, it appears that health education and women’s rights are vital factors to include
when designing a successful multisectoral program.
There are other societal determinants of HIV/AIDS besides the standard
explanations just seen (such as education and poverty), for example, the variables of
behavior and knowledge (such as an understanding of condom use and one’s age at
first intercourse). In their article entitled “Are Africans Practicing Safer Sex?
18
Evidence from Demographic and Health Surveys for Eight Countries,” Glick and
Sahn (2008) scrutinize Demographic and Health Surveys (DHSs) conducted over the
last 2 decades in eight Sub Saharan African countries. The purpose of their study is
to define the components of a successful HIV/AIDS prevention policy for the
African countries. Their project examines risky sexual behavior among men and
women at two different points in time and analyzes which groups take on risky
behavior more often according to wealth and education. Intuitively, an increase in
wealth and education should lead to less risky sexual behavior. However, previous
evidence has shown that multiple sexual partners is a given regardless of these
variables in many African countries. In fact, there is evidence in this study that the
number of sexual partners will increase with wealth amongst the male population.
Similarly, when more condoms are used (a sign of wealth), many times the number
of partners increases as well. Taking into account this ambiguous relationship
between health, education, and risky sexual behavior in Africa, Glick and Sahn
(2008) strive to find an appropriate HIV/AIDS reduction policy.
Results from the study are somewhat mixed, although positive on balance:
“Among men but not women, one tendency we have observed is for education and to
a lesser extent, wealth, to lead to higher risk by increasing the demand for additional
sex partners” (Glick and Sahn 2008:435). A more benign result is that education is
associated with greater condom use by both men and women. Finally, women with
higher wealth and education levels are more likely to delay their first sexual activity.
In sum, this study indicates that higher wealth and education in Africa leads to a
19
higher number of sexual partners among men but also to higher condom use, making
the overall effects less risky. For women, greater wealth and education lead both to
more condom use and less probability of an early sexual debut, both activities
decreasing the risk of contracting HIV/AIDS. Again, it seems that education and
wealth, especially for women, are important determinants of HIV/AIDS. In the case
of this study, higher wealth and education of women leads to more knowledge about
the disease and therefore less risky sexual behavior.
As will be shown below, a closer look at these studies reveals that Zambia
and Uganda seem to have experienced the most favorable changes in fighting
HIV/AIDS. Coincidentally, these countries are two out of eight in this study that
have the longest running multisectoral HIV/AIDS programs; both of which were
implemented before 2000 (all countries in Glick and Sahn (2008) except for Uganda,
Zambia, and Burkina Faso implemented a multisectoral HIV/AIDS program after
2000 and data from the DHSs do not go past 2003). Furthermore, the multisectoral
programs in Zambia and Uganda have garnered substantial support from
international donors and NGOs. Zambia, for example, was the first country ever to
have a “Joint UN Team on AIDS” which has a primary focus on facilitating a
multisectoral response to the disease.
9

Some specific results from Glick and Sahn (2008) are as follows: the change
in sexual activity of 15-19 year olds between 1998 and 2003 reveal that the
probability of first intercourse for females declined in both urban and rural Zambia

9
From United Nations Development Programme (UNDP) Zambia page.
20
over this period. The probability of first intercourse for 15-19 year old males also
declined in urban and rural Zambia, but not as significantly as that of females. The
results are the same for Uganda. In Burkina Faso, another country in which a
multisectoral HIV/AIDS program was implemented early on, the probability of first
intercourse for males and females declined over time in rural areas but no significant
changes were identified in the urban areas.
The probability of using a condom at their most recent intercourse for
females increased significantly overtime in the urban areas of all countries except
Kenya and Nigeria. Condom use among females in rural areas also increased,
especially in Zambia and Uganda but actually decreased in Kenya and Nigeria
between 1998 and 2003. For males, condom use increased over time in all countries
except for Kenya. Both Kenya and Nigeria joined the World Bank MAP and
implemented a multisectoral program in 2001. Although thorough results of the
MAP by country are scarce, interviews that I conducted with World Bank experts
revealed that Kenya and Nigeria both experienced problems early on due to a strictly
committed health minister and a highly populated environment respectfully.
10

Finally, for females, the effects of education and wealth reveal that more
education decreases the likelihood of intercourse with multiple partners in
Mozambique, Nigeria, Uganda, and Zambia. However, for males, higher education
increases the likelihood of promiscuity in all countries except for Uganda, Zambia,
and Kenya. In the case of wealth; more assets decreases the likelihood of intercourse

10
Information from interviews is provided in Chapter 3.
21
with multiple partners for females in all countries except Nigeria, where there is no
visible change. For males, wealth increases the likelihood of intercourse with
multiple partners in all countries except Uganda and Zambia. Few changes were
observed in Ghana and Nigeria. This difference in the likelihood of partners with
respect to wealth between males and females is due to reasons stated above: for
males, having multiple sexual partners is common in many African countries and so
the number of partners typically increases with wealth.
Results from the sample of countries analyzed in Glick and Sahn (2008)
reveal that Zambia and Uganda were most successful at fighting HIV/AIDS between
the late 1990’s and early 2000’s. These happen to be two Sub Saharan African
countries where the government implemented multisectoral programs early on and
maintained strong relationships with NGOs and international organizations. This
combination of a grounded multisectoral program and close interaction with
international organizations appears to be an effective policy tool to combat
HIV/AIDS. Figure 2, below, further summarizes some of the results from Glick and
Sahn (2008), and in particular, the changes in risky sexual behavior of urban females
(ages 15-19) overtime. These changes are presented along with dates of
multisectoral program implementation to show possible corresponding effects from
the program.
To read the table from left to right: 1) Country, 2) Date of Multisectoral
Program Implementation, 3) Probability of 1
st
sexual intercourse before age 15, 4)
Probability of using a condom at last intercourse, 5) the effect of education on having
22
more that one sexual partner, and 6) the effect of wealth on having more that one
partner. As mentioned before, two different points in time are examined for each
country to see if behaviors have changed overtime. These points in time differ by
country; in Figure 2 they are written inside the boxes: Benin-1996 and 2001, Burkina
Faso-1999 and 2003, Ghana-1998 and 2003, Kenya-1998 and 2003,
Mozambique-1997 and 2003, Nigeria-1999 and 2003, Uganda-1995 and 2001,
Zambia-1996 and 2001.

FEMALES
(Urban)
Country
M-S
Start
Probability of
1st
intercourse
before age 15
Probability
of Condom
use at last
intercourse
Education on
more than 1
partner
Wealth on
more than 1
partner
Benin
2002
up b/t 1996
and 2001
Up b/t 1996
and 2001 Up
down but not
significant
Burkina Faso 1990
down b/t 1999
and 2003
Up b/t 1999
and 2003 no change Down
Ghana 2000
down b/t 1998
and 2003
Up b/t 1998
and 2003 no change Down
Kenya 2001
down but not
significant
up but not
significant no change Down
Mozambique 2003
up but not
significant
Up b/t 1997
and 2003 Up Down
Nigeria 2001
up but not
significant
up but not
significant no change no change
Uganda 1992
down b/t 1995
and 2001
Up b/t 1995
and 2000 no change Down
Zambia 1999
down b/t 1996
and 2001
up b/t 1996
and 2000 Down Down

Figure 2: Probabilities and effects from Risky HIV/AIDS Behavior with
Multisectoral Start Dates Source: Davis (2008)

23
The results in green reflect improvements in risky behavior related to
contracting HIV/AIDS, those in red show a worsening of such behavior, and those in
black reveal insignificant or no changes. It is clear from the table that condom use is
improving among urban females throughout Africa. Similarly, higher wealth had led
to a decrease in the number of sexual partners for urban females in the majority of
the countries. It is important to note that this table represents just one example of
several comparisons made throughout the article. It is clear from this table as well as
the other results reported by Glick and Sahn (2008) that Zambia, with a different data
period between 1996 and 2001, has seen the largest improvements and these seem to
coincide with the implementation of their multisectoral HIV/AIDS program in 1999.
Other countries that experienced improvements in the table are that of Burkina Faso
(where a multisectoral program was implemented in 1990) and Ghana, where a
multisectoral program was launched in 2000. Although seemingly anecdotal, these
findings do indicate a connection between a reduction of risky HIV/AIDS behavior
and successful implementation of multisectoral HIV/AIDS programs. To test
whether multisectoral HIV/AIDS programs are useful in the promotion of condom
awareness (recall that a principal finding in Glick and Sahn (2008) is that condom
use has improved) I present a quantitative model that further probes the question in
Chapter 4.
It is these types of determinants from Glick and Sahn (2008), such as the
probability of having more than one sexual partner, which adds to the literature on
the societal determinants of HIV/AIDS. From the studies reviewed thus far it has
24
been seen that the main societal determinants of HIV/AIDS include poverty, lack of
education, gender inequalities, as well as certain risky sexual behavior including
multiple sexual partners, and minimal condom use.
Societal determinants of HIV/AIDS are further examined by Jeffrey Sachs
(2005) who asks why Africa, in particular, has been so vulnerable to disease and
slow in terms of economic growth. The author mentions that the region has been
problematic for centuries:
I began to suspect that the omnipresence of disease and death had
played a deep role in Africa’s prolonged inability to develop
economically. . . Even before the Industrial Revolution, Africa had the
lowest urbanization rate of any part of the world, and apparently the
world’s lowest living standards on the eve of the era of modern
economic growth (Sachs 2005:194).

Sachs (2005) also mentions that good governance can facilitate economic growth but
does not necessarily dictate such growth. As an example, he points to Senegal,
Uganda, Ethiopia, and Ghana (to name a few) as African countries that maintain
stable governance yet continue to face extreme poverty, debt, and illiteracy.
In exploring for other potential reasons for the extended economic troubles in
the African region Sachs (2005) suggests that “. . .it would be necessary to unravel
the interconnections between extreme poverty, rampant disease, unstable and harsh
climate conditions, high transport costs, chronic hunger, and inadequate food
production” (Sachs 2005:196). In his work on economic growth, Sachs (2005)
acknowledges that some of these problems could be improved via large investments
in infrastructure, medications, and agriculture. Unfortunately, no one societal or
25
political determinant for HIV/AIDS is offered by Sachs (2005). However, I would
argue that addressing these factors that hurt economic growth would be a good start
in fighting the disease.
From this literature review on the societal determinants of HIV/AIDS, it
becomes clear that such variables as poverty and a lack of education contribute to the
prevalence of the disease. In the following section I review those studies that point
to certain political determinants of HIV/AIDS, keeping in mind that societal and
political factors are far from mutually exclusive.

Political Determinants of HIV/AIDS
The identification of those forces that are driving the vast disparity in
HIV/AIDS rates across the developing world seems to require a further explanation
than those offered in the current literature. My intention is to approach this problem
by examining the possible political pitfalls (e.g., political instability such as that
currently witnessed in Zimbabwe, etc.) that may determine high HIV/AIDS
prevalence rates. Because little work has been published on the political
determinants of HIV/AIDS, just a handful of political variables are examined in this
dissertation. Political determinants of HIV/AIDS in this study consist of such
variables as: political strife (such as the outbreak of civil war), anti-AIDS policies
(such as a multisectoral program) and, other institutional factors which are likely to
foster the implementation of effective HIV/AIDS policies. Political determinants of
HIV/AIDS are of particular importance to this dissertation because identifying
26
political failures (such as not implementing a multisectoral HIV/AIDS program) in a
given country’s efforts to combat HIV/AIDS may clear the way for more effective
policies regarding the disease.
One of the few studies to examine the political determinants of the
HIV/AIDS rate is entitled “AIDS and the State: The Politics of Government
Responses to the Epidemics in Brazil and South Africa” by Gauri and Lieberman
(2004). The authors explore why some countries have responded to the HIV/AIDS
epidemic more aggressively than others and they focus on those specific strategies
and policies that governments have undertaken in the fight against HIV/AIDS.
Specifically, Gauri and Lieberman (2004) compare the response to the
HIV/AIDS epidemic in two middle-income countries: Brazil and South Africa. In
both countries the disease started among gay men and commercial sex workers and
soon thereafter both were warned that the consequences of ignoring the disease
would be disastrous. Yet, Brazil has now managed to control the epidemic while the
disease remains at a dangerously high level in South Africa. Gauri and Lieberman
(2004) argue that the principal factor is the difference in political institutions in each
country. For example, Brazil has an open and decentralized government with power-
sharing widely dispersed, while South Africa’s political system is more autocratic
and decision-making is more narrowly confined.
They thus conclude that:
National government action requires convincing decision-makers
that something needs to be done, that action is likely to be effective,
27
and that inaction will be politically damaging. We argue that such
pressure is more likely forthcoming in a decentralized institutional
environment. . . (Gauri and Lieberman 2004:6).

Because a disease like HIV/AIDS has a slow onset and is highly stigmatized,
national policymakers operating within a more centralized government tend to
concentrate on more immediate, challenges related to economic development; this is
precisely the pattern that has occurred in South Africa. In a more decentralized
political system like Brazil’s, policymakers at the sub-national level can concentrate
on a range of other problems and bring them to the attention of national level
decision-makers.
In the Brazilian case, for example, the authors emphasize the importance of
such health sector policies as condom promotion and ARV treatment. Recognizing
that access to these kinds of treatments does not reach all social sectors, especially
those groups most vulnerable to the disease, Brazilian policymakers constructed an
“AIDS Bureaucracy” and then implemented a multisectoral HIV/AIDS program:
Brazil established an AIDS bureaucracy in 1985, when a National AIDS
Program (NAP) was created within the Ministry of Health. In 1988 the
inter-ministerial National Commission to Control AIDS was established.
The Commission reports to the Ministry of Health but included
representatives from Ministries of Education, Labor, and Justice, the
principal association of lawyers in Brazil, various universities, and four
NGOs (Gauri and Lieberman 2004:9).

It is through this multisectoral program that better access is provided for far more
groups in Brazilian society. Moreover, it is the decentralized political environment
of Brazil that allows the program to create specific policies on the government’s
behalf.
28
In contrast, what explains the lack of an early, independent AIDS
bureaucracy and a multisectoral program in South Africa? Though a multisectoral
program was finally implemented in 2000, South Africa lacked an early institutional
environment conducive to fighting HIV/AIDS. According to Gauri and Lieberman
(2004:10):
Throughout the history of the AIDS epidemic, there has not been a clear
center of power for the direction of AIDS policy, as the various AIDS
policy-making structures have been granted limited autonomy, and have
often been contradicted by various national executives.

In 2000, for example, South African President Thabo Mbeki avoided the subject of
HIV/AIDS in his speech at the International AIDS Conference and defended those
scientists who claim that HIV does not cause AIDS. More recently, Mbeki has
removed a deputy health minister who, unlike the health minister (who has
recommended that eating garlic can cure the disease), was knowledgeable and active
in promoting an anti-HIV/AIDS policy framework.
11
Clearly, the political
environment in South Africa was not conducive to the creation of an AIDS
bureaucracy when the epidemic was on the rise and this stunted the implementation
of a successful multisectoral program. The authors point to this as a likely factor for
the difference in HIV/AIDS rates between these two countries.
According to Gauri and Lieberman (2004), the political determinants that
seemed to help in the curbing of HIV/AIDS are a decentralized, open political
environment with a voice given to all members of society. It is the argument of this

11
“Sacking the wrong health minister,” The Economist 16 Aug. 2007.

29
dissertation that the multisectoral HIV/AIDS program itself is an effective policy
tool that adds to those political determinants of HIV/AIDS.
The above article by Gauri and Lieberman (2004) provides a strong
foundation for this dissertation. However, whereas these authors compare political
institutions in the two countries as the basis for explaining different HIV/AIDS rates,
my research differs in several ways. First, the epidemic is explored both qualitatively
and quantitatively with a key independent variable being the presence of a
multisectoral HIV/AIDS program in 89 low and middle-income countries over a ten-
year period. Similar political determinants of the disease are explored but societal
determinants are included as well. Economic effects from the disease are reviewed
and information from interviews and case studies is provided. It is the political
determinants (such as decentralization of political decision-making) formulated by
Gauri and Lieberman (2004) that most inform my argument.
In their comparison of Brazil and South Africa, Gauri and Lieberman (2006)
coin the term “boundary institutions” to further explain Brazil’s success in
combating HIV/AIDS. Boundary institutions are another political determinant of
HIV/AIDS and are defined as “the rules and procedures, especially those
implemented by the state, which involve monitoring or regulating citizens according
to particular group identities” (Gauri and Lieberman 2006:47). According to the
authors, these boundary institutions are useful when a government wants to regulate
diverse populations for a variety of reasons, such as denying privileges or setting
right past wrongs. Strong boundary institutions exist when group labels (black, gay,
30
etc.) are widely and consistently used throughout a civil society. In turn, weak
boundary institutions exist when group labels are flexible or vague and used
inconsistently.
If a government reports information in terms of different groups, citizens tend
to process this information from the standpoint of their own group. Those citizens
who do not belong to a “high HIV/AIDS prevalence group” for example, are likely
to believe that they are at less risk because their particular group does not mingle
with the “high prevalence group.” Where boundary institutions are strong, group
members desire to maintain a positive group status and therefore will deny
knowledge of possible risks within their own group to ensure distance from the
disease. Where boundary institutions are weaker, inter-group mingling and
subsequent risky behavior are more likely to occur. The key point the authors are
making is that strong boundary institutions may lead to less aggressive HIV/AIDS
policies.
The authors argue that “. . . strong boundary institutions impeded AIDS
policy formulation and implementation in South Africa when judged against Brazil’s
response” (Gauri and Lieberman 2006:60). Both, Brazil and South Africa have large
“mixed” populations with the “black” population being more susceptible to the
disease. Although apartheid ended in 1994, the South African government obviously
faced huge challenges in terms of race. A main pitfall was the government’s claim
that HIV/AIDS was a problem confined to blacks as a group, rather than alerting the
entire population to this risk. Brazil, in contrast, encouraged sexual relations and
31
partnerships between races in the late 1800s in an effort to “whiten” the population
thus causing fewer divisions between racial groups.
The prior work by Gauri and Lieberman (2004) emphasized the importance
of an open and decentralized political environment as a key political determinant for
Brazil’s success in fighting HIV/AIDS. This follow-up article incorporates a new
political determinant of HIV/AIDS, i.e. the concept of boundary institutions. The
authors reiterate that because South Africa has strong boundary institutions, policies
to curb HIV/AIDS have been slow to emerge. Brazil’s weak boundary institutions
(and hence the propensity for more groups in society to perceive themselves to be at
risk) have fostered the implementation of more cohesive HIV/AIDS policies. The
earlier implementation of these policies in Brazil has further helped to curtail the
HIV/AIDS epidemic.
In addition to the concept of particular groups created by boundary
institutions, a recent article from the Economist suggests that location had something
to do with Brazil’s rapid and efficient response to HIV/AIDS. In Brazil, the disease
began amongst white males in the prosperous south-east region. There, the
government has been efficient at promoting the use of condoms and providing free
treatment, and NGOs have been able to hold the government to their promises. The
problem now is that the disease has spread throughout the north-east region where
these mechanisms are not as strong. Because the health care system in this region is
more erratic, less people are being diagnosed and properly treated. These difficulties
with disease prevention in the north-east are clearly problematic for the Brazilian
32
government and immediate measures need be taken to preserve its role as a
champion in fighting HIV/AIDS.
12

To reiterate, in both works by Gauri and Lieberman (2004, 2006) political
determinants in specific countries were offered as potential reasons for the differing
HIV/AIDS rates. Specifically, they suggest that a decentralized, open political
environment and weak boundary institutions are effective for lowering HIV/AIDS
rates. My quantitative models presented in Chapter 4 will incorporate a variable for
decentralization, however, no existing variable could be found for “boundary
institutions.” While there is little work to date that examines the political
determinants of HIV/AIDS, considerable work has been published on the political
determinants of economic growth. The political determinants of growth have been
studied by Rodrik (2003, 2007), in particular, and appear to be similar to some of the
determinants mentioned in Gauri and Lieberman (2004). I offer a review of Rodrik’s
(2003, 2007) work on the subject below.
Though standard variables sufficed in explaining economic growth in the
section on the economic effects of HIV/AIDS, Rodrik (2003, 2007) suggests that
competent political institutions are also a main contributor to economic growth.
Rodrik (2003) proposes three determinants of growth (other than the standard
variables seen in the section on economic effects): geography, trade integration, and

12
“A Portrait in Red,” The Economist 13 Mar. 2008.

33
institutions. Geography is attractive as an explanation for growth because it
influences factors that are independent of wealth, such as natural resources and
climate. Integration is important when looking at economic growth because trade
allows poorer countries to strengthen their ties to the world economy and thereby
increase their level of economic competitiveness.
While geography and trade are valuable additions to the economic growth
model, the variable most important for this dissertation is that of political
determinants, or what the author calls political institutions.
Institutions have received increasing attention in the growth literature
as it has become clear that property rights, appropriate regulatory
structures, the quality and independence of the judiciary, and the
bureaucratic capacity cannot be taken for granted in many settings and
that they are of utmost importance to initiating and sustaining economic
growth (Rodrik 2003:8).

It seems evident that good institutions are a necessary condition for both launching
and sustaining economic growth.
It is important to note that while certain qualities of institutions may prove
necessary for achieving economic growth; this does not necessarily mean that they
will be sufficient for fighting an epidemic like HIV/AIDS. Take the case of
Botswana: economic growth has been substantial over time, capital investment has
been consistent, and strict budget laws have been enforced since independence in
1966. These strong institutional factors have enabled Botswana to become one of
the few middle income countries on the African continent. Yet, Botswana also has
the second largest HIV/AIDS prevalence rate in the world. The institutional
framework of the government appears stable and economic progress has not been a
34
problem. However, Botswana has high income inequality which translates into an
unemployment rate of 23.8% and poor health care outside urban areas (female life
expectancy is 49.58 years).
13
These statistics threaten the future of Botswana’s
economic progress.
Botswana finally adopted a multisectoral program in 2000 and by 2006 the
HIV/AIDS rate had dropped from 35.8% to 24.1%.
14
Other factors undoubtedly
added to this progress, but I would argue that implementing a HIV/AIDS policy
based on a multisectoral program in a country with strong political institutions likely
had a significant effect on lowering the HIV/AIDS rate. It seems that two important
political determinants have worked in Botswana’s favor: strong political institutions
that spurred Botswana’s economic growth and the subsequent implementation of a
multisectoral HIV/AIDS program.
Rodrik’s (2007) emphasis on institutions as key determinants of economic
growth is in line with the goal of this dissertation: to identify those determinants that
work best for fighting HIV/AIDS. Whereas my emphasis will be on the effect of a
multisectoral HIV/AIDS program, Rodrik (2007) casts a much wider net in exploring
those political institutions that work best: “The rule of law, a high-quality judiciary,
representative political institutions, free elections, independent trade unions, social
partnerships, institutionalized representation of minority groups, and social
insurance. . .” (Rodrik 2007:161). These political institutions (for which data are
available) will be examined in my quantitative analysis conducted in Chapter 4.

13
Statistics from the CIA World Factbook, (2007).

14
Current HIV/AIDS prevalence rates from HIV INSITE.
35
Indeed, there is considerable overlap between the political determinants of
economic growth and those described in the MAP manual (reviewed below) for the
successful implementation of multisectoral HIV/AIDS programs. Similar
determinants were also described by Gauri and Lieberman (2004) regarding the
particular political environment in Brazil which supported the implementation a
successful multisectoral program. These analyses suggest that the political
determinants of HIV/AIDS as well as the political determinants of economic growth
may be complementary: both represent similar variables for successful policy
implementation such as a voice for vulnerable groups, representation from several
sectors, and a decentralized environment.
Rodrik (2007) goes on to explain how good institutions materialize and the
particular environment that is most conducive to their adoption.
I emphasize the importance of “local knowledge” and argue that a
strategy of institution building must not overemphasize best-practice
“blueprints” at the expense of local experimentation. I make the case
that participatory and decentralized political systems are the most
effective ones we have for processing and aggregating local knowledge
(Rodrik 2007:155).

As mentioned before, this type of decentralized, political institution is analogous to
the environment of Brazil (circa 1988) described by Gauri and Lieberman (2004).
This environment allowed local groups to communicate with policymakers at a sub-
national level, thereby, making clear the needs of different groups in society. Having
an already decentralized and plural environment made the adoption and
implementation of the multisectoral program in Brazil all the more effective.
36
Rodrik (2007) also adds that “Even under the best possible circumstances, an
important blueprint requires domestic expertise for successful implementation.
Alternatively, when local conditions differ greatly, it would be unwise to deny the
possible relevance of institutional examples from elsewhere” (Rodrik 2007:164). It
seems that even a well drawn plan for a good institution, like the World Bank’s
manual on implementing multisectoral programs, does not necessarily mean that the
institutional setting will be compatible. Political conditions vary in every country
and some attributes may be more conducive to adopting such a blueprint than others.
This is not to say that a concise plan on implementing a multisectoral HIV/AIDS
program is unnecessary, on the contrary, a blueprint is useful but may be more so
within the proper institutional environment. To be clear, the political determinants
of economic growth (such as the rule of law and decentralization) described by
Rodrik (2003; 2007) may also be strong political determinants for effective AIDS-
policy making and hence, a lower HIV/AIDS rate. This is speculation however and
so a quantitative model involving these variables will be explored in Chapter 4.
From literature reviewed thus far, it has been seen that even though
HIV/AIDS rates differ across all countries, the disease poses an economic threat to
all of them. Economic effects from the disease include among other things: a fall in
growth, labor productivity, and savings. It is then important to ask, what are the
main determinants of HIV/AIDS? From reviewed literature it seems that there are
both societal and political determinants of the disease. The main societal
determinants include poverty and gender inequality, which worsens the incidence of
37
HIV/AIDS, while political determinants such as decentralization tend to help in the
fight against HIV/AIDS. Some political determinants for economic growth (such as
a representation of minority groups in society) are similar to the determinants for
effective HIV/AIDS policy. For this dissertation, it is hypothesized that the
particular anti-AIDS policy; a multisectoral program, is a strong political
determinant of the disease. The brief literature regarding multisectoral HIV/AIDS
programs is reviewed below.

Multisectoral HIV/AIDS Programs
Just a few articles assess the role that multisectoral HIV/AIDS programs have
played in halting this disease. The literature on multisectoral HIV/AIDS programs
primarily concentrates on their importance, implementation strategies, and in some
cases, their pitfalls. In terms of this dissertation, none of the studies on multisectoral
HIV/AIDS programs attempts to quantify the effects of these programs across
developing countries in any way comparable to the objectives of this dissertation.
This is the main contribution of my research.
In Bodiang (2001), “HIV/AIDS: The Multisectoral Approach,” there is a
focus on Sub Saharan Africa and an emphasis on the need for a broad response to the
HIV/AIDS epidemic that reaches outside the health sector alone.
15
The main reason
proposed in Bodiang (2001) for a multisectoral approach is to coordinate the tackling

15
Bodiang currently works as project manager and technical expert in public health at the Swiss
Centre for International Health and prior to this she served as a physician at several hospitals and
clinics throughout Switzerland and several African countries.
38
of the disease with other development issues, such as inequality, illiteracy, poverty,
and women’s rights.
Bodiang (2001:6) points to the need for a stable institutional environment for
the successful implementation of a multisectoral program:
Multisectoral involvement is not completely new. In most cases,
however, early actions were limited to low-profile scattered activities
that lacked a systematic approach and originated from a personal
commitment of concerned individuals rather than a policy backed
by the institution itself.

The author offers some examples of how different sectors can be effective in
combating HIV/AIDS. The transportation sector, for instance, is vulnerable to the
contraction of HIV/AIDS due to frequent sexual activity among truck drivers in Sub
Saharan Africa. The prolonged time on the road increases the probability of high-
risk sexual relationships with multiple partners along the route. With a specific
HIV/AIDS awareness campaign, truck drivers can be directly educated about the
disease. They can then disseminate this information along their routes, thereby
raising the awareness of others within this sector.
The education sector is one of the hardest-hit by the epidemic, but also has
the potential to make some of the greatest strides in promoting HIV/AIDS
awareness. According to Bodiang (2001:10), “Where AIDS is widespread;
education-an essential building block of development-is being impaired. The
epidemic is eroding the supply of teachers and diluting the quality of education.”
Not only are teachers and students disappearing because of the disease, those
remaining in the schools aren’t receiving enough preventative training about the
39
epidemic. With a multisectoral program, resources are disbursed directly to the
education sector, thereby providing teachers with the knowledge and skills to convey
the necessary information to their students who, ideally, will pass it along to their
peers. Bodiang (2001) concludes by emphasizing the urgency for policymakers to
create the proper environment for a multisectoral HIV/AIDS program.
Around the time that Bodiang (2001) published this paper; the World Bank’s
Multi-Country HIV/AIDS Program for Africa (MAP) was launched. While this
program was groundbreaking in that it required each country to establish a
multisectoral HIV/AIDS program, specific World Bank guidelines for implementing
such a program were not published until 2004. In 2003, a general set of guidelines
was published by The Commonwealth Secretariat (2003) entitled “Guidelines for
Implementing a Multi-Sectoral Approach to HIV/AIDS in Commonwealth
Countries.” These guidelines re-emphasized the need for national policymakers to
work with vulnerable groups in society. “. . . Coordinated economic, political and
social efforts at a national level are needed to reduce the vulnerability of particular
groups and sections of society and must complement programmes and interventions
operating at the level of the individual and the community” (The Commonwealth
Secretariat 2003:2).
The Commonwealth guidelines set forth a broad list of requirements,
including a flexible approach to each country’s individual situation and the
appointment of precise roles to national leaders within ministries, the business sector,
and communities throughout the country in order to best meet the needs for treatment
40
and assign tasks to the appropriate ministry, NGO, etc. Response assessments are
then required to measure the progress of HIV/AIDS reduction and to determine
further needs for program implementation. One year after the release of the
Commonwealth Secretariat (2003) guidelines the World Bank published its official
operations manual for launching Multi-Country HIV/AIDS Programs in Africa
(MAP), entitled “Turning Bureaucrats into Warriors: Preparing and Implementing
multisectoral HIV/AIDS Programs in Africa, A Generic Operations Manual.”
The manual explicitly defines those necessary background and institutional
arrangements for fighting HIV/AIDS such as the role of government, the private
sector, NGOs, and decentralized agencies. Chapters on monitoring, evaluation, and
financial disbursement are also included. The document pays particular attention to
the organization of a multisectoral body, which it calls the National HIV/AIDS
Council (NAC):

In each MAP country, a high-level body has been created to oversee
the national multisectoral HIV/AIDS program. The NAC includes
representation from all principal stakeholders concerned with the
epidemic—public sector organizations, private business, NGOs, and
headed by the head of state, prime minister or other senior public
official. . . (Ayvalikli, Brown, and Mohammad 2004:21).

The primary goals set for the NACs are to revise and approve HIV/AIDS strategies,
policies, and budgets, as depicted in Figure 3 below.
41

Figure 3: National HIV/AIDS Council (NAC) Source: Ayvalikli, Brown, and
Mohammad (2004)

Specifically, all officials within the NACs are responsible for several
important actions including: directing the national strategic HIV/AIDS plans that
each MAP country creates; promoting the mainstreaming of the epidemic at all
levels, public and private; encouraging the involvement of people living with HIV/in
the participation of policymaking; and, the development of strong and effective
monitoring and evaluation system. Because NACs do not meet on a day-to-day
basis, many of these duties are carried out by a secretariat known as the National
AIDS Secretariat (NAS). The NAS is composed of a staff of specialists and
coordinators who serve as the technical support of NAC activities and the main
facilitators of the multisectoral HIV/AIDS program. The NAS frequently provides
reports to the NAC regarding their work on the implementation of monitoring and
training of those in line ministries, the financing of projects, as well as the
coordination of NGOs and other donors.
42
Although the NAC and the NAS are important models for the successful
implementation of a multisectoral HIV/AIDS approach, problems have arisen. A
common concern is that every country has its own unique norms, values, and,
conditions, therefore, one similarly structured model will not work effectively in
every country. Recently it has been mentioned that the roles of specific jobs within
the NACs are not well defined, thus causing inefficiency in implementing many
HIV/AIDS policies. “In some sub-Saharan African countries, NACs have evolved
into complex bureaucracies with the same constraints of other public sector bodies:
poor lines of accountability, staff recruitment and retention problems, lack of
incentives to improve performance, political interference, corruption and so on.”
16

In recognition of these challenges, the World Bank does encourage a 'learning by
doing' approach and is consistently looking for ways to improve. Already, it has
been suggested in World Bank publications that NACs can improve if roles and
responsibilities are re-defined and mundane tasks (such as data collection) are
contracted out. In Kenya, for example the NAC has become more effective due to
the creation of sub-committees to handle specific responsibilities such as finance and
program management.
As mentioned above, all countries are unique and even if the blueprint for a
successful multisectoral program for MAP countries may be well laid-out and
thorough, the World Bank cannot fully control how the actual plan is implemented
by the various governments. It is quite likely that some of these developing

16
HLSP Institute Technical Brief (2006:4).
43
countries that have implemented the MAP lack the necessary political capacities to
successfully implement a multisectoral program. Because many of the countries
responsible in the MAP sample have had tenuous political histories, their institutions
are still weak and problematic. One has to hope for the correct combination of
political institutions in each country (such as a decentralized environment and a
strong connection between government and NGOs/Civil Society) that can ideally
foster the adoption of a multisectoral HIV/AIDS program. This is not to argue that a
multisectoral program will not work in a weaker political environment: it should
merely work better in a strong one.
The guidelines provided by Bodiang (2001), the Commonwealth Secretariat
(2003), and the World Bank MAP Generic Operations Manual (2004) specify the
tools for governments to use in creating successful multisectoral HIV/AIDS
programs. All three works point to the need to work broadly with NGOs and
vulnerable groups in society, to establish representation at sub-national and national
levels, and to practice ongoing monitoring of each country’s multisectoral
HIV/AIDS program.
An assessment of the effectiveness of the World Bank’s MAP to date and
examples of those developing countries with successful HIV/AIDS multisectoral
programs are described in Chapter 3. Specifically, the programs in Thailand,
Ethiopia, and Uganda are highlighted as they have experienced successful
multisectoral HIV/AIDS programs in different regions with variations on program
implementation.
44
There are just a handful of articles that evaluate the implementation of
multisectoral HIV/AIDS programs. Most assessments suggest further strategies for
improving the multisectoral approach (current statistics show that 22 out of 28
countries in this dissertation database have experienced falling HIV/AIDS rates since
the MAP implementation).
17
At the same time, Galaty, Gavian, and Kombe (2006:
222) argue that “the multisectoral approach . . . needs a more substantial conceptual
framework as well as mechanisms for tracking, evaluating, and prioritizing
multisectoral interventions.” In other words, experts seem to confer that the
multisectoral HIV/AIDS approach is the best option for developing countries, but
changes and improvements can always be made.
Harman (2007) discusses some failures of the World Bank MAP. The author
credits the MAP for emphasizing the need for HIV/AIDS prevention to be a top
political priority in sub-Saharan Africa, while also noting that the World Bank has
set a precedent in making the first large multilateral commitment to fighting
HIV/AIDS via a multisectoral approach. According to the author, however, MAP
strategies still lack sufficient interaction with civil society organizations.
“Multisectorality was narrowly conceived to mean community-based organizations
and a select number of national NGOS and faith-based organizations only. The
funding directed to such organizations was often too little, too late” (Harman
2007:4). This inefficiency led to funds being distributed late to NGOs and to the
abandonment of some projects.

17
HIV/AIDS prevalence statistics were obtained from several websites. Most recent rates come from
HIVINSITE.
45
Harman (2007) also expresses how the NAC’s role as coordinator of a
national response has too frequently been derailed by bureaucratic tasks surrounding
implementation funding deadlines and tracking objectives. The fact that NACs deal
so much with funding causes hostility between them and the health ministries, who
were accustomed to having financial control before program implementation. From
this article, it seems that more attention and interaction with civil society
organizations is needed to improve the MAP and tighter boundaries must be set on
the roles that NACs perform. As mentioned above, according to several World Bank
publications, NAC improvement appears to be a significant concern of the experts
working on the MAP.
A different assessment of the multisectoral approach is provided by Putzel
(2004) in “Governance and AIDS in Africa: Assessing the International
Community's ‘Multisectoral’ Approach.” Rather than suggesting ways to improve
programs already in place, the author evaluates the structural layout of these
programs and provides an explanation for the various problems that have occurred
since their inception. The author explains that while a multisectoral approach is
necessary it may reach too deeply into the work of the health sector: “In the process
of elaborating a multisectoral response, there has been a tendency to minimize and
secondarise the medical dimensions that remain centrally important in any fight
against disease” (Putzel 2004:4). The author provides a framework that
reemphasizes the importance of the medical aspects (of MAPs) which are often
dismissed in the current generation of multisectoral HIV/AIDS programs.
46
Putzel (2004) cites Uganda as an example of having a successful
multisectoral HIV/AIDS program but also argues that it works well because it is
located within the ministry of health. The program is an innovative one created by
the president in 1992, although no parallel program consisting of sectors and
ministries was established. This framework has not changed, which means that the
multisectoral approach was simply tacked on to the health sector program. Putzel
(2004) expresses concern for the future of multisectoral programs like MAP because
their framework lacks this interconnectivity to the health sector.
It is clear that there is still plenty of room for additional studies on
multisectoral HIV/AIDS programs. To date much of the work has addressed their
importance and in some early cases, their success, e.g. in Uganda, Brazil, and
Thailand. My intention in this dissertation is not only to add to the limited amount of
literature on the subject but also to provide further evidence for the importance of
such multisectoral programs.

Conclusion
The literature on the economic effects of HIV/AIDS, the societal and political
determinants of HIV/AIDS, and multisectoral HIV/AIDS programs has been
reviewed in this chapter. Through this review it is evident that there are negative
economic effects from the disease, such as a fall in labor productivity caused by
illness and then death. Several determinants of HIV/AIDS have been identified,
including a lack of education and a closed political environment, both of which can
47
contribute to the further spread of the disease due to insufficient knowledge and a
lack of policymaking capacities at multiple levels. Finally, the literature on
multisectoral HIV/AIDS programs has provided guidelines for their implementation
and in some cases also suggestions for their improvement. Quantitative data
representing economic effects, societal and political determinants, and multisectoral
programs will be explored in Chapter 4. It is now important to examine
multisectoral HIV/AIDS programs in practice and explore those variables that seem
to be most important for their success.
48
Chapter 3: Multisectoral HIV/AIDS Programs in Practice:
MAP Review, Interviews, Case Studies
To date, a multisectoral program exists in 76 out of the 89 developing
countries that are included in the database for this dissertation. While the majority of
the 76 countries showed declining HIV/AIDS rates after the implementation of a
multisectoral program, some also showed little change. Of those countries with little
change after program implementation began, HIV/AIDS rates have remained low.
Based on the literature reviewed in the previous chapter, the apparent lack of success
in fighting HIV/AIDS in some countries may be due to negative societal and political
determinants that hinder an effective battle against the disease. The main hypothesis
for this dissertation is that a multisectoral HIV/AIDS program is an effective policy
tool for achieving lower HIV/AIDS rates. Below, reviews of those multisectoral
HIV/AIDS programs where information is available are presented.

MAP Review
While the Multi-Country HIV/AIDS Program in Africa (MAP) is an on-
going project, only a handful of reviews have been published. Most recently, MAP
results were published in “The Africa Multi-Country AIDS Program 2000-2006” by
Gorgens-Albino, Mohammad, Blankhart, and Odutolu (2007). Their results,
although possibly biased, appear to be exceptional. The main attributes that account
49
for MAP success are: the creation of targeted HIV/AIDS awareness campaigns;
prevention measures in such areas as medical services, the monitoring and better
supervision of such medical services, routine surveying and testing; and, an explicit
government framework for program implementation communication, training, and
oversight. To avoid bias, an overview of the MAP results from the Center for Global
Development (CGD) will also be included here.
According to the World Bank document by Gorgens-Albino, Mohammad,
Blankhart, and Odutolu (2007), MAP funding has helped to strengthen existing
government and civil society institutions, as well as the ties between these entities. It
has also fostered the creation of new anti-HIV/AIDS institutions at national and sub-
national levels. Evaluation teams have been assembled in most countries with MAP
funding to ensure proper use of the funds. MAP countries have also reached out to
grass-roots and faith-based organizations in an attempt to help those most vulnerable
segments of the population. Funding has provided these organizations with home-
based care, counseling, HIV testing, and information about treatment and prevention.
A more in-depth look at the use of MAP funds reveals that they are allocated
into three different groups: civil society, public (health and other ministries) and the
National AIDS Commissions (NACs) which were described in the previous chapter.
Statistics listed in the document reveal that funds disbursed into the public sector
helped to train 74,793 non-health ministry employees. Funds disbursed to civil
society organizations trained 474,391 employees/volunteers and 10,938
decentralized government bodies have implemented HIV/AIDS work plans.
50
Furthermore, 6,999,528 Africans have been tested for HIV, 4,041,973 condoms have
been distributed, and 3,012 sites provide ARVs. NAC funds have been primarily
used for improving the capacity for monitoring and program evaluation, coordinating
responses among groups, training civil society organizations on how to properly
manage MAP funds, and keeping track of resources.
The MAP review document provides general information on inputs and
results from the countries pooled together. Some inputs and results from funds
disbursed for “system strengthening” are depicted below in a table from Gorgens-
Albino, Mohammad, Blankhart, and Odutolu (2007:37).

MAP Input-Level and Output-Level Results in Countries in Africa with MAPs
Systems Strengthening for HIV/AIDS (estimated US $319 million disbursed)
Input Result
Non–MOH and local government staff trained with MAP funds 74,793 (23 countries)
Health Ministry staff trained with MAP funds 13,181 (23 countries)
Civil society staff trained with MAP funds 474,391 (23 countries)
Total staff trained with MAP funds 562,366 (23 countries)
Percentage of all staff and volunteer training funded by MAP 56%
Decentralized governments that implemented HIV work plans 10,938 (25 countries)
Employees in workplace reached with HIV programs 2,258,844 (23 countries)
Number of organizations provided with technical support 41,107 (25 countries)

Figure 4: MAP Input-Level and Output-Level Results. Source: Gorgens-Albino,
Mohammad, Blankhart, and Odutolu (2007)

With funds disbursed towards the strengthening of systems to fight HIV/AIDS, the
majority of MAP participating countries have experienced improvements. As shown
in the examples from Figure 4 above, at least 23 out of the 28 MAP countries in this
dissertation have seen increases in training at all levels as well as the implementation
51
of anti-AIDS plans and programs that were not in place before the MAP. To examine
these kinds of improvements, participating countries have been evaluated by internal
data collectors and a list is provided below. Naturally, detailed country information
and statistics from the World Bank MAP would be desirable in each case. But since
they are not available,
18
the list represents available quotations from Gorgens-
Albino, Mohammad, Blankhart, and Odutolu (2007) regarding new anti-AIDS
policies and overall country performance since MAP inception:
“In Madagascar, local political leaders demonstrated their commitment to
HIV by publicly going for an HIV test to motivate the population to be tested
as well” (Gorgens-Albino et al.: 40).
“In Nigeria, the MAP directly funded the State Action Committees on AIDS
and has also built their capacity for project and financial management. In
return, the local government has made funding available for district voluntary
counseling and testing centers” (Gorgens-Albino et al.: 46).
“MAP funds have helped Malawi to (1) increase the number of people
accessing counseling and testing services; (2) hold a very successful National
HIV Testing Week during which about 100,000 people were tested; (3)
increase the proportion of youth ages 15-19 years abstaining from sex; and
(4) improve capacity of local authorities to coordinate the national response
through personnel, transport, equipment, and operational support” (Gorgens-
Albino et al.: 42).

18
Reasons for incomplete country data are due to legality issues that are explained in the CGD report
discussed on page 53.
52
“In Ghana, the MAP is reported to have improved partnerships between
Muslims and Christians who support vulnerable, infected, and affected
persons in the community” (Gorgens-Albino et al.: 44).
“The Rwanda national police force has created eight anti-AIDS clubs with 30
members each, integrated HIV activities into the police’s national strategic
plan, and provided access to treatment and care to HIV-positive police force
members and spouses” (Gorgens-Albino et al.: 45).
“In Tanzania, all eligible civil society organizations were trained in proposal
writing, project management, and reporting” (Gorgens-Albino et al.: 47).
“In Sierra Leone, the district councils were empowered and procedures were
laid down for training and engagement of the community-based
organizations” (Gorgens-Albino et al.: 47).
“In Angola, it was reported that ‘focal support teams have been created and
trained in seven priority Government Ministries, i.e. Education, Interior,
Youth, Family, Social Assistance, Labour and Health,’ and 250 NGOs were
trained in service provision” (Gorgens-Albino et al.: 47).
“One Zambian district commissioner remarked that ‘income-generating
activities have made a very big difference. Before we were funded by the
MAP, other funders just used to give us food to give to our clients. When the
food ran out we had nothing to give our clients. Now we have this hammer
mill from MAP as an income-generating activity, so we always have some
income and food for our clients’” (Gorgens-Albino et al.:58).
53
“In Congo (Brazzaville), the MAP facilitated the emergence of an NGO that
specialized in the care of children affected by AIDS. This led to an
improvement of services delivered to this population” (Gorgens-Albino et al.:
43).
“In Guinea-Bissau, approximately one and a half years into its
implementation, it is reported that the MAP is beginning to change health
services in the country, empowering and enhancing regional and national
health care facilities as well as personnel in five priority regions in the
country” (Gorgens-Albino et al.: 56).

It was also reported in this MAP review document that MAPs in Rwanda, Senegal,
Sierra Leone, Uganda, and Zambia will be ending even though the financing for
current HIV/AIDS programs has not been completely disbursed. A possible reason
for this is that ministries of finance in these countries prefer to have more flexibility
in allocating funds than is offered by the MAP guidelines.
From the country examples provided by Gorgens-Albino, Mohammad,
Blankhart, and Odutolu (2007), it seems that requiring a multisectoral approach for
all countries has contributed to noticeable improvements in the fight against
HIV/AIDS. Many new policies have been implemented and vulnerable groups have
been targeted. Furthermore, political determinants for fighting HIV/AIDS such as
strong institutions have improved: “. . . approximately 32 percent of MAP funding
was allocated to build institutions to contribute to a multisectoral response and to
54
develop capacity to manage HIV responses at the national and decentralized levels”
(Gorgens-Albino, Mohammad, Blankhart, and Odutolu 2007: 9). A main political
innovation brought about by the MAP was the creation of the NAC, an entity that
several of the countries did not possess prior to implementing the program.
There are few other publications reviewing the MAP to date and most are
provided by the World Bank. The Center for Global Development, however,
published “Following the Funding for HIV/AIDS” by Oomman, Bernstein, and
Rosenzweig (2007). This MAP review begins by addressing the lack of available
data:
As the World Bank does not require funding data to be disaggregated
by program activities, the exact amounts spent on each activity are
unknown. . .There are also no consistent data across countries about
the amounts of money being received by different types of organizations,
as this information is not required by the World Bank (Oomman,
Bernstein, and Rosenzweig 2007:44).

This lack of data makes concise analysis of multisectoral programs implemented by
the MAP difficult. The CGD also requested data from each country’s NAC but was
told that, because of legally-enforced disclosure agreements, the World Bank was
unable to release it. Data on the dollar amounts given to each country are available,
however, and according to Oomman, Bernstein, and Rosenzweig (2007), funding is
distributed slowly as the first year of the MAP is typically dedicated to system
establishment (such as the scaling-up of management in the ministry of finance).
New disbursements require country reports on how the previous funding was used
before new funding is allocated. Slow funding has been a problem of the MAP:

55
MAP recipients are expected to request funding for six months of
activities, and then account for a portion of their expenditures before
receiving subsequent tranches of money. . . Such small amount can
impede the smooth flow of funds down the funding chain to implementers. . .
(Oomman, Bernstein, and Rosenzweig 2007:45).

Problems with overburdened NACs (as mentioned in Chapter 2) can also slow down
the funding process, causing less anti-AIDS programs to be implemented.
Although data availability and slow funding are cited as main shortcomings
of the World Bank’s MAP, Oomman, Bernstein, and Rosenzweig (2007:48) do
recognize the importance of funding a multisectoral response to the disease: “. . .
MAP money supports a multisectoral response, with MAP funding supporting at
least ten different ministries, many of which are unlikely to get AIDS funding from
other sources.” The authors conclude that the World Bank’s unique approach to
funding recipients (such as ministries and civil society organizations) rather than
types of programs (such as sexual education) has been effective. Funding to actual
recipients allows each recipient (such as the transportation minister) to use their
allotted budget to target a specific group or area thereby reaching more groups in
society.
To elaborate further on the inner workings of the MAP to date, I conducted
some interviews with frontline policymakers at the World Bank. In addition, I
compiled brief sketches of individual country experiences in the use of multisectoral
programs to fight HIV/AIDS. Although the MAP only applies to Sub Saharan
Africa, materials concerning the impact of multisectoral programs from countries in
56
the Asian, and Latin American and Caribbean regions are reviewed as well. These
interviews and country sketches are summarized in the following two sections.
From qualitative information thus far, it is clear that HIV/AIDS hurts
economic growth and there are some specific societal and political determinants that
are most conducive to lowering HIV/AIDS rates. Specifically, political determinants
such as a decentralized political environment are conducive to establishing efficient
anti-AIDS policies. Again, it is the hypothesis of this dissertation that an effective
policy tool to curbing HIV/AIDS rates is the implementation of a multisectoral
program. I will confirm this hypothesis through a summary of the multisectoral
approach taken from interviews with experts in the field, country case studies in this
chapter, and by conducting a multivariate regression analysis and a quantitative
difference-in-difference approach in Chapter 4.

Interviews
In order to get a more in-depth grasp on the progress of these multisectoral
programs, I conducted phone interviews with experts in the field of HIV/AIDS
prevention in the developing countries. The main interviewees were: those
responsible for compiling the World Bank’s Generic Operations Manual on
implementing the multisectoral HIV/AIDS MAP program; a World Bank Senior
Development Economist; a staff member in the World Bank’s Global HIV/AIDS
Program; and a past consultant at Development Alternatives, Inc. (DAI) who had
experience in working with the HIV/AIDS multisectoral programs. The following
57
summary of these interviews focuses on information not found thus far in the
literature.
My first set of questions solicited a general overview of the multisectoral
aspect of MAP and the related opinions of the respondents. The respondents were
unanimous about the importance of a multisectoral approach as a policy tool for
HIV/AIDS prevention, although there were differences of emphasis. The education,
social welfare, and transportation sectors were most frequently referred to as crucial
for fighting the epidemic. There was also consensus on how key it is to identify the
population set most affected by the epidemic, as well as determining which groups
were primarily causing its spread. Truck drivers were most cited as a high risk group
that fall victim to HIV/AIDS and most rapidly spread the disease. Hence, it is
imperative that the transportation sector is aggressively targeted by the MAP.
Nigeria, Burkina Faso, and Benin were most mentioned in my interviews as
countries that have used all sectors efficiently in the campaign against HIV/AIDS.
To monitor MAP performance, a task manager along with a group of experts from
the World Bank conducts evaluations twice a year, paying special attention to each
country's national strategy. Numerous examples were given by the respondents, the
point being that various strategies are more successful in some countries than in
others. Senegal, for example, has done especially well at identifying the vulnerable
population, monitoring prostitution and including faith-based organizations in its
overall outreach. Nigeria was said to have a strong national strategy, but because
there are several sub-national states, monitoring the efficient implementation of a
58
multisectoral program has been difficult. Madagascar was mentioned as having a
sound national strategy, and Eritrea was identified as providing superior monitoring
of its multisectoral program. Burkina Faso got high marks for having a very involved
president, which shows the significance of leadership for a strong strategy, a factor
that has also worked in Uganda’s favor. Malawi pools money from all donors into
one pot so as to avoid donors’ biases in formatting its own multisectoral approach.
Finally, Rwanda was noted for the professional management of its funding through
donor coordination.
When asked about the most important aspects of the multisectoral approach,
interviewees identified the need for strong political institutions and their ability to
interface effectively with NGOs and grass roots organizations. Emphasis was also
placed on the need for organizations that can perform work that goes beyond the
capability of the health ministry, for example, going door to door in villages and
offering medicine and information to those without access to care. Moreover, health
ministries in some countries will not reach out to or treat vulnerable groups in
society such as commercial sex workers and drug users. Hence, these groups will
only be identified and treated through multisectoral programs. Finally, the policy of
learning by doing (highlighting those parts of the program that work well in each
country and paring down those that do not) was also emphasized for program
success.
Going beyond the specific features of multisectoral HIV/AIDS program
within the MAP, my next questions addressed the allocation of World Bank funds to
59
the participating MAP countries. To receive World Bank public funding, ministers
are asked to come forward with proposals but are not formally assisted by the Bank
in proposal preparation. The idea is to engender a more entrepreneurial spirit on the
part of MAP applicants, although one interviewee mentioned that the World Bank
should have been slightly more involved in this proposal writing process due to the
widely varying capacities across participating countries. Kenya was given as an
example where the health minister initially refused funding because of the desire for
complete financial autonomy.
Some countries, such as Tanzania, were mentioned as having a strong
program due to the appointment of a multisectoral coordinator and team in all
districts. Botswana, which is too wealthy to qualify for World Bank MAP funding,
implemented a multisectoral program of its own in 2000. Apparently, this was an
about-face, as officials in Botswana had previously been reluctant to acknowledge
the extent of the disease or speak about it openly at that time. Finally, it was
mentioned that there is still a strong belief that multisectoral HIV/AIDS programs
are an important part of the policymaking process; however, no plan is ideal. The
difficulties of implementing a successful HIV/AIDS multisectoral program were
reiterated, as each country has individual characteristics and needs that defy a one-
size-fits-all strategy.
The above information from MAP reviews and interviews may be biased, but
the overall results of multisectoral HIV/AIDS program implemented in MAP
countries are positive. The majority of countries have experienced improvements in
60
many important areas of fighting HIV/AIDS including, reaching vulnerable groups
through NGOs and faith-based organizations, establishing educational centers
through the education system, and building testing and treatment centers through the
health and labor sectors. The multisectoral approach required for the MAP countries
clearly assisted and helped to facilitate these improvements. A closer look at
countries with successful multisectoral HIV/AIDS programs reinforces this view.

Case Studies
As mentioned at the outset, just a handful of countries have implemented
long-running multisectoral HIV/AIDS programs. Below is a review of three
successful long-running programs in Uganda, Ethiopia, and Thailand. The programs
have important similarities such as the provision of a voice for vulnerable groups in
society. These case studies are presented to further strengthen the argument for the
importance of multisectoral HIV/AIDS programs and identify other political and
societal determinants of HIV/AIDS.

Uganda: A Success
In 1986 a new president, Yoweri Museveni, took power in Uganda and
immediately declared that fighting AIDS was a patriotic duty from village to state
level. He also called for the abstention from sex before marriage, being faithful, and
condom use (also called the ABC approach). Ignoring any stigmatization, the health
minister announced the existence of HIV in Uganda and focused on the need for
61
education and assurance of safe blood. In 1992 the government adopted a
multisectoral approach and the Uganda AIDS Commission (UAC) was established to
oversee the multisectoral program. President Museveni also ensured political
openness early on by touring the country, speaking openly about the disease, and
allowing for complete media freedom so that information and advertisements
regarding the disease could reach all segments of society. The HIV/AIDS
prevalence rate in Uganda went from 15% in the early 1990’s to about 6% in 2005.
Even before the creation of the UAC in 1992, the government of Uganda
was taking a pro-active role in the fight against HIV/AIDS. In 1987 a psychosocial
support organization was established for those infected with and affected by the
disease. In 1990 an AIDS information center was created to disseminate information
to testing centers. Participation by religious and civil-society organizations as well
as ministries was encouraged through Uganda’s multisectoral approach and the UAC
today continues to coordinate and facilitate such participation. An example of the
work of the UAC includes the linkage of partners from the public and private sectors
along with various ministers to collaborate in identifying HIV/AIDS priority areas.
Once the priority areas are identified, monitoring and future planning of policies are
performed. Policies such as the emancipation of women, universal education, and
poverty eradication are geared towards reaching and aiding the most vulnerable
groups in society.
In 1997, Uganda endorsed a policy of governmental decentralization with a
transfer of power to the district level. By 2001, over 700 different HIV/AIDS
62
agencies were established throughout all regions of Uganda. Anti-AIDS
organizations founded among local communities make up 21.9% of these many
agencies; 17.1% of the HIV/AIDS agencies are government-based, 17.2% are NGOs,
and 16.2% are faith based organizations.
19
The early decentralization of the political
environment in Uganda clearly assisted in the creation of so many diverse
HIV/AIDS combating agencies throughout the country.
The Ugandan government and its subsequent multisectoral program provides
one of the most successful HIV/AIDS elimination stories to date, and they built on
this progress by adopting the MAP. Uganda was granted $47.5 million from the
World Bank’s lending arm, the International Development Association (IDA), in
2001 and with it has managed to scale up technical support, materials, expanding
services and reducing stigma. “Of the total project financing . . . 38% was channeled
directly to support the local response: $8.5 million to 233 district-based departments,
NGOs, and community-based organizations, and $12.5 million to 3,629 community-
led HIV and AIDS initiatives (CHAIs)” (Gorgens-Albino 2007:76). These MAP
funds have financed a wide range of materials including mattresses, school supplies,
and support services to those in need. The CHAIs are said to be a novel approach as
they disburse funds directly to communities that implement initiatives directed
towards their specific groups. This allows precise needs of those in the community
to be directly met.

19
Inventory of Agencies with HIV/AIDS Activities and HIV/AIDS Interventions in Uganda (2001).

63
Figure 5 shows HIV/AIDS prevalence in Uganda over time, including a sharp
decline after 1992 when the multisectoral program was implemented.

Figure 5: HIV/AIDS Rates in Uganda 1987-2002. Source: Avert Charity

Those political determinants that seem to contribute to Uganda’s success in
fighting HIV/AIDS are: a strong and influential leader, a decentralized political
environment, a voice and opportunities for vulnerable groups, a close relationship
between the government and community, and the implementation of a multisectoral
HIV/AIDS program. Societal determinants that seem to contribute to Uganda’s
success in fighting HIV/AIDS are: an active and strong presence of NGOs and
community-based organizations, and effective anti-stigma and anti-discrimination
campaigns.
64
Ethiopia: A Success
Ethiopia is another example of a country where the government established a
multisectoral HIV/AIDS program jointly with the implementation of the World
Bank’s MAP. Ethiopia’s efforts to combat HIV/AIDS began early, as a national
taskforce was created in 1985 to address the disease. In 1987, the ministry of health
implemented a control program and in 1989 surveillance activities began. These
programs were established to control and monitor the main causes of HIV/AIDS,
such as sharing needles, blood transfusions, unprotected sex, etc. Reportedly, the
main causes of HIV/AIDS in Ethiopia are heterosexual intercourse and mother-to-
child transmission. HIV/AIDS prevention and control activities were decentralized
in 1993 enabling regional health bureaus to become further involved in the fight
against HIV/AIDS.
Ethiopia launched a national HIV/AIDS policy in 1998 and promoted a
multisectoral effort to provide care to vulnerable groups and scale up centers for
testing and screening. Policies that targeted vulnerable groups such as women and
People Living With HIV/AIDS (PLWHA) were implemented. This multisectoral
effort paid special attention to community-based organizations such as the Dawn of
Hope Ethiopia Association, which continues to make progress in anti-AIDS
endeavors relying on peer education and counseling, home-based care, and anti-
stigma campaigns. HIV/AIDS prevalence was 10.63% around 1995 and dropped to
about 4.4% in 2005.
65
The Ethiopia Multisectoral AIDS project was one of the first initiated under
MAP. “The IDA credit of $59.7 million to the Government of Ethiopia was
approved on September 12, 2000. . . Ethiopia used the project to develop a
participatory, decentralized, and community-driven response to HIV/AIDS”
(Gorgens-Albino et al. 2007:65). The country took a local level, decentralized
approach by placing offices at all sub-national and national levels. Nearly half of the
MAP funding was distributed to offices at the local levels. These, in turn, were able
to reach vulnerable groups in society. Local groups now have the means to get
tested for HIV, receive ARV treatments, and discuss the disease openly in treatment
groups, thereby decreasing the stigma associated with the disease.
Furthermore, MAP has funded several faith-based organizations, including
the Dawn of Hope Ethiopia Association which has now successfully branched out to
offices in 13 other Ethiopian communities. This particular organization has been
effective in distributing a set amount of funds to those living with HIV. The
majority of the funds are used to start up small businesses in such ventures as
furniture and garment making. The organization also sets aside money for the
educational expenses of AIDS orphans. Finally, volunteers from the faith-based
organization provide weekly home-based care to those living with AIDS. In the case
of Ethiopia, the government seemed to be leaning towards a strategic multisectoral
approach early on and results from the MAP appear to have bolstered positive
outcomes.
66
Those political determinants that seem to contribute to Ethiopia’s success in
fighting HIV/AIDS are: a decentralized political environment, a voice and
opportunities for vulnerable groups, a close relationship between the government and
community, and the implementation of a multisectoral HIV/AIDS program. Societal
determinants that seem to contribute to Ethiopia’s success in fighting HIV/AIDS are:
active faith-based organizations, and effective anti-stigma and anti-discrimination
campaigns.

Thailand: A Success
While Uganda and Ethiopia had successful government-sponsored
multisectoral HIV/AIDS programs, their progress was further increased by the
assistance from the World Bank MAP. In Contrast, Thailand has not been part of
any MAP, the World Bank has not implemented such a program in Asia as of yet.
Nonetheless, Thailand has managed to preserve a successful multisectoral program
over a significant amount of time. Similar to the case of Uganda, Thailand elected a
new Prime Minister in 1991, Anand Panyarachun, who removed the existing AIDS
program from the health ministry and established a multisectoral National AIDS
Prevention and Control Committee. Prime Minister Panyarachun directly headed
this committee, thereby granting it greater political influence and funding.
This strong political leadership, along with a multisectoral program that
brought together NGOs, the Economic Development Board, and several key
ministries, helped to increase social awareness about HIV/AIDS. Cabinet members
67
assisted in promoting public information about HIV/AIDS via television and radio,
as well as complying schools to provide HIV/AIDS education. Because of
Thailand’s large population of commercial sex workers, “The 100% Condom
Program” was established. This program has worked because it did not try to
suppress commercial sex which could easily have led to an underground epidemic.
Rather, it merely involved all necessary parties at all levels and encouraged regular
condom use. By 1996, this HIV/AIDS-focused government was providing more
than $80 million to the AIDS Control Program anually.
An example of Thailand’s aggressive response to the disease is that of the
Mae Chan community. In 1988, a small community hospital in Mae Chan took active
measures to combat the disease including the implementation of ARV treatment,
HIV testing, counseling services, and nutrition aid, as well as working with local
faith-based groups to secure additional support. Soon, the entire Mae Chan
community became involved in the fight against HIV/AIDS through efforts such as
establishing education programs for youth and broadcasting HIV/AIDS information
via local radio.
This response to addressing and fighting the HIV/AIDS epidemic has been
recognized by the United Nations Joint Programme on AIDS (UNAIDS) as a “good
practice” example. As such, UNAIDS organized a workshop in 2000 to educate
government officials, religious bodies, school teachers, health workers, and social
68
workers from select rural communities in China, Vietnam, Lao People’s Democratic
Republic, and Cambodia on what they might learn from Mae Chan’s example.
20

The Mae Chan response is an example of a community-based multisectoral
approach within a country that implemented a successful multisectoral program at
the national level. Certainly, active and consistent responses such as these have
added to Thailand’s success in fighting HIV/AIDS. Since 2000, Thailand has
remained active in its fight against the epidemic by increasing the amount of ARVs
distributed, establishing programs to increase treatment for all members of the
population living with HIV/AIDS, and targeting teenagers for education about
condom use. All of Thailand’s anti-AIDS measures seemed to have worked over
time: after a peak of about 140,000 HIV/AIDS cases in 1991, infections were
estimated to have fallen to about 21,000 in 2003.
Those political determinants that seem to contribute to Thailand’s success in
fighting HIV/AIDS are: a strong and influential leader, political rights and programs
for vulnerable groups, media openness regarding the disease, a close relationship
between the government and community, and the implementation of a multisectoral
HIV/AIDS program. Societal determinants that seem to contribute to Thailand’s
success in fighting HIV/AIDS are: the enforcement of HIV/AIDS education in
schools and the legalizationa and monitoring of prostitution. Unfortunately,
complete data on every political and societal determinant of HIV/AIDS identified in

20
From Popline Document: Mae Chan Workshop on Integrated Community Mobilization towards
Effective Multisectoral HIV / AIDS Prevention and Care.
69
the case studies on Uganda, Ethiopia, and Thailand are unavailable. What data is
available, however, is examined in the regression analysis in Chapter 4.

Failures?
Few countries in the present study have been identified as those whose
HIV/AIDS rates have increased after implementation of a multisectoral program. In
most of these cases, the HIV/AIDS rate had been quite low and remains so, though
increasing very slightly. Precise starting dates for multisectoral programs, as well as
consistent reporting of HIV/AIDS prevalence rates, are difficult to obtain. Those
cited here are as accurate as possible.
Cameroon’s HIV/AIDS prevalence went from approximately 4.7% in
1995 to approximately 5.4% in 2005. A multisectoral program was
established in 1998 and the MAP was incorporated after 2000.
Madagascar’s HIV/AIDS prevalence went from approximately 0.15%
in 1995 to approximately 0.5% in 2005. A multisectoral program was
established with the MAP in 2001.
Nepal’s HIV/AIDS prevalence went from approximately 0.2% in
1995 to approximately 0.5% in 2005. A multisectoral program was
established in 1992.
Vietnam’s HIV/AIDS prevalence went from approximately 0.1% in
1995 to approximately 0.2% in 2005. A multisectoral program was
established in 1997.
70
El Salvador’s HIV/AIDS prevalence went from approximately 0.6%
in 1995 to approximately 0.7% in 2005. A multisectoral program was
established in 1989.
Jamaica’s HIV/AIDS prevalence went from approximately 0.7% in
1995 to approximately 1.2% in 2005. A multisectoral program was
established in 1988.
Peru’s HIV/AIDS prevalence went from approximately 0.35% in
1995 to approximately 0.5% in 2005. A multisectoral program was
established in 2002.
Tunisia’s HIV/AIDS prevalence went from approximately .04% in
1995 to approximately 0.1% in 2005. A multisectoral program was
established in 1987.
Most of the prevalence rates in these 8 countries started out and stayed low
(increasing by less than 0.5%). Most rates remained low although at a stable level,
though Cameroon’s rate increased from 4.7% to 5.4%. In-depth data for these
countries are sparse. Cameroon’s government appears to be dedicated to alleviating
HIV/AIDS, but it is hampered by a very feeble health system infrastructure.
Furthermore, many organizations are involved in HIV/AIDS programs in Cameroon
but coordination among them is reported to be weak. These downfalls may be
negatively affecting Cameroon’s multisectoral HIV/AIDS program.
Jamaica was one of the first countries in this study to implement a
multisectoral HIV/AIDS program in 1988, and while the HIV/AIDS prevalence rate
71
has remained fairly low, it has increased over time. Jamaica’s multisectoral
program, the National AIDS Committee (NAC), was established by the Health
Minister but it was not a part of the Jamaican government. The NAC is a private
organization with several members and its main function is that of advising the
health minister. Unlike the successful programs in Thailand and Uganda, the
Jamaican multisectoral HIV/AIDS program was not established strong government
leadership. In 2002, however, The HIV/AIDS Prevention and Control Project for
Jamaica was funded by the World Bank as part of their Multi-Country HIV/AIDS
Lending program for the Caribbean Region. This funding was used to assist the
Jamaican Government to further strengthen the multisectoral approach as well as to
scale up treatment and prevention activities. Hopefully, the collaboration between
this multisectoral program and the government will be beneficial to Jamaica’s fight
against HIV/AIDS.

Conclusion
Examples of multisectoral HIV/AIDS programs in practice from the MAP
countries, interviews, and case studies reveal that the majority of developing
countries have experienced many improvements since program inception. MAP
funds have been used to establish a multisectoral approach in many countries that, in
the past, had only health sector-based programs. This multisectoral approach
allowed new funds to be allocated to the scaling up of education, health care, and
more with a specific focus on vulnerable groups in the population. The interviews
72
and case studies have revealed that each country is unique and no one blueprint will
work in all countries. For example, in Uganda, the multisectoral program was
implemented under an active leader in a decentralized environment which enabled all
parts of the country (urban and rural) to quickly learn about the threat of the disease.
Ethiopia’s multisectoral program, a variation on this theme, works well through the
government’s communication and interaction with faith-based organizations which
reach vulnerable groups in the country.
From the available research presented reviewed thus far in this dissertation, it
is clear that various societal and political determinants contribute to the level of the
HIV/AIDS rate in a country. Research has also revealed that multisectoral
HIV/AIDS programs seem to be an effective policy tool for combating HIV/AIDS,
especially if they can be uniquely implemented to meet the specific needs of each
country. It is now necessary to quantify the previously reviewed economic effects,
societal, and political determinants and to more rigorously examine their relationship
to HIV/AIDS rates.
73
Chapter 4: Methodology, Models, Results, Difference-in-
Difference Approach
Qualitative information from reviewed literature, results from the World
Bank MAP, and country case studies have shown that a multisectoral HIV/AIDS
program is often an effective tool for combating HIV/AIDS. In an attempt to further
verify these qualitative results, quantitative data will be explored in this chapter.
Specifically, the answer to the dissertation question concerning whether and how the
presence of a multisectoral HIV/AIDS program influences the HIV/AIDS rates in
low and middle income countries will be explored using a multivariate regression
analysis. To confirm that multisectoral HIV/AIDS programs are an influential
political determinant for HIV/AIDS rate reduction, a difference-in-difference
approach will also be executed. For the regression analysis: seven different models
will be presented and several societal, political, and economic variables will be
explored below. A thorough explanation of variables and their sources will be
presented in Table 4a. Explanation and execution of the difference-in-difference
approach will immediately follow results from the regression analysis.
Methodology
The multivariate regression analysis consists of a pool of 89 low and middle
income countries within Sub Saharan Africa, Asia, and Latin America over the ten
year period 1995-2005. Seven different models are explored and presented in 3
sections:
74

A.) Political and Societal determinants of HIV/AIDS
B.) Economic Effects of HIV/AIDS
C.) Determinants of a HIV/AIDS multisectoral program
In section A.), political and societal determinants of HIV/AIDS will be
estimated over six models. These models are described in Table 1 below.

A.) Political and Societal Determinants of HIV/AIDS
Table 1: List of Models on Political and Societal determinants of HIV/AIDS

Model
No.
Dependent Variable / Purpose
1 HIV/AIDS rate 1997 / To review the original estimation by Over
(1998) and a re-estimation of Over (1998) including a variable for a
multisectoral HIV/AIDS program
2a HIV/AIDS rate 2005/ To identify significant political and societal
determinants of HIV/AIDS including a variable for a multisectoral
HIV/AIDS program
2b HIV/AIDS rate 2005/ A two-stage model of 2a with instrumental
variables for the multisectoral HIV/AIDS program variable to ensure
there is no reverse causality
2c HIV/AIDS rate 1995/ To identify significant political and societal determinants of
past HIV/AIDS including a variable for a multisectoral HIV/AIDS program
3a Change in HIV/AIDS between 1995 and 2005 / To identify significant political and
societal determinants of the change in HIV/AIDS including a variable for a
multisectoral HIV/AIDS program
3b Change in HIV/AIDS between 1995 and 2005 / A two-stage model of 3a with
instrumental variables for the multisectoral HIV/AIDS program variable to ensure
there is no reverse causality
4 Expenditures on HIV/AIDS 2005 / To identify significant political and societal
determinants of spending on HIV/AIDS including a variable for a multisectoral
HIV/AIDS program
5 Condom Awareness 2005 / To identify significant political and societal
determinants of condom awareness including a variable for a multisectoral
HIV/AIDS program

75
In section B.), economic effects of HIV/AIDS will be estimated. This model
is described in Table 2 below.

B.) Economic Effects of HIV/AIDS
Table 2: List of Models on economic effects of HIV/AIDS
Model
No.
Dependent Variable / Purpose
6 GDP growth rate 1990 to 2005 / To identify significant economic
effects caused by HIV/AIDS

In section C.), determinants of a HIV/AIDS multisectoral program will be
estimated. This model is described in Table 3 below.
C.) Determinants of a HIV/AIDS multisectoral program
Table 3: List of Models on determinants of a HIV/AIDS multisectoral program
Model
No.
Dependent Variable / Purpose
7 Has a multisectoral HIV/AIDS program as of 2005 / To identify
significant determinants of a HIV/AIDS multisectoral program

The 3 previous tables provide a brief overview the 7 different models that
will be presented in this chapter. Table 4a below provides a complete list,
description, and source of all variables used in the models. Table 4b provides a list
of descriptive statistics for all variables presented in Table 4a. Descriptive statistics
are necessary because they provide a summary of the data and its features. The
independent variables for these econometric models consist of a wide array of
political, societal, economic, and regional factors.

76
Table 4a: Explanation of Variables
Variable Measurement / Source Used in Model
No:
HIV/AIDS rate
1995 and 1997
Percentage of adults (aged 15-49) living with
HIV/AIDS. / CIA, UNAIDS, Avert, WHO
1, 2c, 3a
HIV/AIDS rate
2005
Percentage of adults (aged 15-49) living with
HIV/AIDS / HIV insite
2a, 2b
Change in
HIV/AIDS rate
1995 to 2005
Change in percentage of adults (aged 15-49) living with
HIV/AIDS between 1995 and 2005 / CIA, UNAIDS,
Avert, WHO, HIV insite
3a, 3b
Expenditures on
HIV/AIDS 2005
Per Capita government expenditures on HIV/AIDS /
UNAIDS
4
Multisectoral
HIV/AIDS
program 2005,
1997, 1995
Percent of time a country has had a multisectoral
program over the years 1995 to 2005 (^also calculated
as =1 if a country has a program and =0 if a country has
no program as of 1995, 1997, 2005)/ Davis (*)
^1, 2c, 7

2a, 3a, 4, 5
Condom
Awareness 2005
Percentage of all sexually active women (aged 15-49)
who know about condoms as a contraceptive method /
DHS
5
GDP p/c growth
1990 to 2005
Percentage growth in GDP per capita from 1990 to
2005 / World Bank
6
Gini Coefficient
2005
A ratio with values between 0 and 1: a low Gini
indicates a more equal wealth distribution, while a high
Gini indicates more unequal distribution. / CIA, UNDP
1, 2a, 2b, 2c, 3a,
3b, 5, 6, 7
GDP per capita
1995, 1997,2005
Per Capita Gross Domestic Product per capita in US
Dollars / World Bank
1, 2a, 2b, 2c,
3a,3b, 5
Muslim percentage Percentage of the population that is Muslim by religion
/ CIA
1, 2a, 2b, 2c, 5
Gender Parity
Index (literacy)
Ratio of female to male literacy rates. GPI less than1 =
Males more literate than Females, GPI=1 = perfect
equality. / UNESCO
1
Gender Inequality Arithmetic mean of early marriage, polygamy,
inheritance, parental authority indicator (0=equality;
1=max inequality). / OECD Development Centre
1
Male population
(15-30) 1995, 2005
Percentage of the male population (aged 15-30) / World
Resources Institute
2a, 2b, 2c, 3a, 3b,
4, 7
Literacy rate 2005 Percentage of males and females (aged 15 +) that are
literate / World Bank
2a, 2b, 2c, 3a, 3b,
5
Telephones 1995,
2005
Number of telephone mainlines per 100 people in the
population / World Bank
2a, 2b, 2c, 3a, 3b,
5, 6
HIV/AIDS rate of
neighbor 1995,
2005
Highest HIV/AIDS rate of all neighboring countries /
CIA, UNAIDS, Avert, WHO, HIV insite
2a, 2b, 2c, 3a, 3b,
5, 6, 7

77
Table 4a: Explanation of Variables (Continued)
Civil War Calculated as = 1 if a country has had a civil war
anytime between 1995 and 2005 and = 0 if there has
been no civil war over this time / CIA, US Dep’t. of
State
2a, 2b, 2c, 3a, 3b,
4, 6
Legalized
Prostitution
Calculated as = 1 if a country has had legalized
prostitution anytime between 1995 and 2005 and = 0 if
prostitution has been illegal over this time. /
politicalbase.com
2a, 2b, 2c, 5
World Bank MAP Calculated as = 1 if a country was a part of the MAP by
2005 and as = 0 if they were not. ^ also interacted with
variable for multisectoral program/ World Bank
3a, 3b,
^4
Gross National
Income
1995, 2005
Income measured in constant LCU. / World Bank 4, 7
Health Minister is
a MD as of 2005
Calculated as = 1 if a country’s health minister is a
medical doctor and = 0 if he/she is not. / Davis (*)
4
Decentralization
2005
Number of jurisdictions within each government tier in
a country. ^ also interacted with variable for
multisectoral program/ World Bank
4, 6, 7
^2a, 2c
Rule of Law 2005 An index between 1 and 100 representing the extent to
which people have confidence in and abide by the rules
of their society. ^ also interacted with variable for
multisectoral program/ World Bank
7
^2a, 2c
Government
Effectiveness 2005
An index between 1 and 100 representing the quality of
public services, the quality of the civil service and the
degree of its independence from political pressures. ^
also interacted with variable for multisectoral program/
World Bank
6, 7
^2a, 2c

New Political
Regime
Calculated as = 1 if a country has had a new leader
anytime between 1995 and 2005 and = 0 if there has
been no new leader over this time / CIA, US Dep’t. of
State
1c, 2b, 7

78
Table 4a: Explanation of Variables (Continued)
Secondary School
Enrollment 1995
Percentage of the population (aged 25 +) enrolled in
secondary school. / Barro-Lee
4, 6
Investment
averaged over 1995
to 2005
The share of total GDP that is devoted to investment
in fixed assets measured as Gross Fixed Capital
Formation (averaged 1995-2005) / OECD
Development Centre
6
Males in the
Military 1997
Percentage of males in the population that are in the
military. / World Bank
1

Africa Calculated as = 1 if a country is on the continent of
Africa and = 0 if a country is not.
2c, 3b, 7
Asia Calculated as = 1 if a country is on the continent of
Asia and = 0 if a country is not.
7
Latin America and
Caribbean
Calculated as = 1 if a country is in Latin America and
Caribbean and = 0 if a country is not.
7
(*) data compiled through multiple sources including UNAIDS, USAID, WHO, Avert, CIA, World
Bank.

The following is a list of descriptive statistics for all variables:

Table 4b: Descriptive Statistics

AIDS 1995 AIDS 2005 Change AIDS AIDS
Expend.
MS Program MS 1995
Mean 0.051952 0.044209 -0.004587 1.837947 0.482213 0.152174
Median 0.023600 0.017500 0.000000 1.886955 0.454545 0.000000
Maximum 0.283000 0.213000 0.182600 6.101439 1.000000 1.000000
Minimum 0.000100 0.000700 -0.118000 -2.302585 0.000000 0.000000
Std. Dev. 0.068779 0.058242 0.039177 1.850035 0.298062 0.363158
Skewness 1.699539 1.615054 1.686344 0.001994 0.265844 1.936728
Kurtosis 5.369921 4.565284 13.95882 2.964780 2.304501 4.750916
Jarque-B. 32.90967 24.69377 251.9857 0.002408 1.468957 34.63296
Probability 0.000000 0.000004 0.000000 0.998797 0.479756 0.000000
Obs. 46 46 46 46 46 46
MS 1997 Cond. Ed. GDP Grth GINI GDP p/c GDP p/c95
Mean 0.173913 0.816043 0.014587 0.441696 2.782715 844.7609
Median 0.000000 0.900500 0.014000 0.430000 2.643398 374.5000
Maximum 1.000000 0.989000 0.059000 0.740000 3.597476 4250.000
Minimum 0.000000 0.339000 -0.021000 0.290000 2.149219 90.00000
Std. Dev. 0.383223 0.178600 0.017226 0.091866 0.412890 935.6463
Skewness 1.720618 -1.353543 0.312572 0.834582 0.458805 1.901636
Kurtosis 3.960526 3.723676 3.169854 3.855425 2.006031 6.387662
Jarque-B. 24.46571 15.04971 0.804337 6.742573 3.507466 49.72051
Probability 0.000005 0.000540 0.668868 0.034345 0.173126 0.000000
Obs. 46 46 46 46 46 46

79
Table 4b: Descriptive Statistics (Continued)
Percent
Muslim
Gender Lit.
Gap
Male Pop. Male Pop.
1995
Literacy Rate Phones per
100 people
Mean 0.284543 0.736522 0.132576 0.141087 0.758696 0.048413
Median 0.131000 0.770000 0.136000 0.141500 0.790000 0.014500
Maximum 1.000000 1.010000 0.158000 0.189000 1.000000 0.230000
Minimum 0.000000 0.000000 0.040000 0.090000 0.240000 0.001000
Std. Dev. 0.342989 0.228144 0.020847 0.017063 0.211099 0.061737
Skewness 0.986041 -0.898495 -2.343567 0.051694 -0.727822 1.438313
Kurtosis 2.459540 3.642862 9.979522 4.930817 2.482187 3.982892
Jarque-B. 8.013980 6.981358 135.4757 7.165926 4.575139 17.71202
Probability 0.018188 0.030480 0.000000 0.027793 0.101513 0.000143
Obs. 46 46 46 46 46 46
Phones 95 AIDS N.95 AIDS N. Civil War LegalProst. WB MAP
Mean 0.026024 0.058135 0.060652 0.304348 0.391304 0.434783
Median 0.008000 0.042500 0.040500 0.000000 0.000000 0.000000
Maximum 0.181000 0.283000 0.334000 1.000000 1.000000 1.000000
Minimum 0.000100 0.000000 0.000000 0.000000 0.000000 0.000000
Std. Dev. 0.038888 0.063154 0.075079 0.465215 0.493435 0.501206
Skewness 2.159116 1.526975 1.923276 0.850420 0.445435 0.263117
Kurtosis 7.572950 5.316282 6.283076 1.723214 1.198413 1.069231
Jarque-B. 75.82142 28.15923 49.01789 8.669158 7.742121 7.675853
Probability 0.000000 0.000001 0.000000 0.013107 0.020836 0.021538
Obs. 46 46 46 46 46 46
GNI 95 GNI MD Decent. MS*Dec. MS*GovEf
Mean 818.9130 1147.174 0.563801 4.754132 2.706594 0.308449
Median 390.0000 570.0000 0.572727 5.141510 2.337047 0.134430
Maximum 3790.000 5010.000 1.000000 12.37871 12.37871 6.354207
Minimum 140.0000 160.0000 0.000000 0.000000 0.000000 0.000000
Std. Dev. 952.2727 1198.914 0.385842 2.839952 2.743012 0.923055
Skewness 2.055225 1.698749 -0.180745 -0.257102 1.406204 6.302741
Kurtosis 6.525677 5.225283 1.586087 3.078455 5.110941 41.80817
Jarque-B. 56.20853 31.61519 4.082165 0.518576 23.70094 3191.196
Probability 0.000000 0.000000 0.129888 0.771601 0.000007 0.000000
Obs. 46 46 46 46 46 46
MS*RuLw Gov. Eff. New Reg. School En. Avg. Inv. % in Mil.
Mean 0.156836 0.648049 0.673913 0.182717 0.194348 0.012609
Median 0.112591 0.352818 1.000000 0.129000 0.195000 0.010000
Maximum 0.571000 13.97927 1.000000 0.860000 0.340000 0.100000
Minimum 0.000000 0.061909 0.000000 0.005000 0.070000 0.000000
Std. Dev. 0.146837 2.015479 0.473960 0.163667 0.054473 0.019141
Skewness 0.964603 6.495955 -0.741982 1.977484 0.336447 2.871120
Kurtosis 3.033982 43.48235 1.550538 7.938496 3.108178 12.21391
Jarque-B. 7.135726 3464.586 8.247592 76.72514 0.890271 225.9163
Probability 0.028216 0.000000 0.016183 0.000000 0.640737 0.000000
Obs. 46 46 46 46 46 46

80
Table 4b: Descriptive Statistics (Continued)
AFRICA ASIA LA
Mean 0.608696 0.195652 0.173913
Median 1.000000 0.000000 0.000000
Maximum 1.000000 1.000000 1.000000
Minimum 0.000000 0.000000 0.000000
Std. Dev. 0.493435 0.401085 0.383223
Skewness -0.445435 1.534391 1.720618
Kurtosis 1.198413 3.354354 3.960526
Jarque-B. 7.742121 18.29072 24.46571
Probability 0.020836 0.000107 0.000005
Obs. 46 46 46

The variables listed in table 4a above are all used in the multivariate
regression analysis of this dissertation. Those variables that represent political
determinants of HIV/AIDS include:
1. A variable for a multisectoral HIV/AIDS program, this is measured as the
average amount of time the country has had a program over the ten year
period. It is important to note that there is no data available listing which
countries have multisectoral programs, the amount of time they have been
present, sector participation, etc. Therefore each of the 89 countries in this
dissertation was carefully researched and all available information was
indexed and applied.
2. The rule of law.
3. Government effectiveness.
4. Level of decentralization.
81
5. Whether the country has experienced a civil war anytime in the years 1995-
2005.
6. Whether a country has experienced a new political regime anytime in the
years 1995-2005.
7. Whether the country is a member of the World Bank MAP.

Those variables that represent societal determinants of HIV/AIDS include:
1. Condom awareness.
2. The percentage of the population that is in the military.
3. Legalized prostitution.
4. Percentage of the population that is Muslim.
5. Male population ages 15-30.
6. Adult literacy rate.
7. Highest HIV/AIDS prevalence rate of the neighbor country.
8. Whether health minister has an MD.
9. Secondary school enrollment rates.
10. Gender parity index of adult literacy.
11. Gender inequality.

Economic variables include:
1. The GINI coefficient for income inequality.
2. Number of telephones per 100 people.
82
3. Per capita GDP.
4. Gross National Income.
5. 2005 expenditures on HIV/AIDS.
6. Average Investment.
7. GDP per capita growth from 1990-2005.

Those variables that represent regions are:
1. Africa
2. Asia
3. Latin America and the Caribbean

Models 1, 2a, 2c, 3a, 4, 5, and 6 are all executed using the Pooled Least
Squares estimation technique which, like ordinary least squares (OLS), minimizes
the sum of squared residuals to find a model that best fits the data. Models 2b and 3b
use Two Stage Least Squares estimation to ensure that the independent variable:
“multisectoral program” is not correlated to the dependent variable: “HIV/AIDS
rate.” This requires the selection of instrumental variables (IVs) for the independent
variable (multisectoral program). Valid IVs must have a causal influence on the
existence of a multisectoral program but have no direct causal effect on the
dependent variable: the HIV/AIDS rate. Instruments used in these two stage models
have a positive and significant influence in the existence of multisectoral HIV/AIDS
programs but have no significant relationship with the HIV/AIDS rate. In a two
83
stage least squares estimation, an OLS model using the IVs in place of the original
multisectoral program variable is executed. To test that the instruments are valid, a
Wald coefficient test for relevance and a Residual OLS test for exogeneity are
preformed. Finally, Model 7 uses a Probit estimation which is similar to Ordinary
Least Squares except that the dependent variable is a latent or dummy variable taking
on the value either 0 or 1. This quantitative portion will be executed to help discern
those variables that carry the heaviest explanatory weight for the models below.

Models and Results
Model 1: Re-estimation of Over (1998)
Model 1 is a re-estimation of the model presented in Over (1998). The
original data in Over’s (1998) study could not be obtained but similar data was used
when available. One variable from Over (1998) could not be obtained: the percent of
foreign-born people in the population. Furthermore, for the majority of countries,
data on gender inequality and gender parity index are published for only one year.
The first execution of Model 1 (in column I) represents the results from Over (1998)
using available data and the second execution of Model 1 (in column II) represents
the results including a variable for multisectoral programs. The purpose of Model 1
is to re-estimate an early study that looked at the determinants of HIV/AIDS and to
see if the inclusion of the multisectoral program variable in this prior study holds any
weight. Below is an algebraic form of Over’s (1998) model:
84
HIV/AIDS rate (1997) = α + β
1
Gini + β
2
Muslim percentage + β
3
GNP per
capita1997 + β
4
Men in the military1997 + β
5
Gender Parity Index + β
6
Gender
Inequality + ε
Table 5: Results Model 1-Dependent Variable: 1997 HIV/AIDS Rate

Model 1
I I I
Constant -0.024531 -0.004710
(0.042207) (0.01834)
M-S Program 97 -0.029732**
(0.012979)
Gini 0.133297* 0.100637
(0.070098) (0.069455)
Percent Muslim -0.069201*** -0.069615***
(0.020308) (0.019695)
GDP P/C 97 -1.31E-05 -1.21E-05
(8.27E-06) (8.04E-06)
Military 97 -0.314673 -0.324432
(0.334467) (0.324432)
Gender Parity Index 0.013146 0.013993
(0.035210) (0.034146)
Gender Inequality 0.087046*** 0.089892***
(0.027515) (0.026711)

R-Squared 0.355609 0.403071
Observations 74 74

*Significant at 10% level, **Significant at 5% level, ***Significant at 1%

Results
Model 1 is a re-estimation of Mead Over’s (1998) study reviewed in Chapter
2. Unfortunately, problems were encountered when trying to locate identical data so
85
as much data as possible was collected for the re-estimation and the same countries
from Over’s (1998) original model were used. Some (though not all) identical
results were found, most likely due to differences in data sources and the non-
inclusion of the “foreign-born percentage” variable. Nonetheless, the signs of the
estimated coefficients with one exception (military population) are consistent with
those expected and two out of the six variables are also statistically significant.
Furthermore, when the variable for having a multisectoral HIV/AIDS program as of
1997 was included, the influence on the 1997 HIV/AIDS rate was found to be
negative and significant. Specifically, the results for all variables are as follows:
• The Gini coefficient in estimation I has a significant and positive influence
on the 1997 HIV/AIDS rate indicating that income inequality worsens
HIV/AIDS as a disproportionate amount of assistance and information for the
rich is likely to leave the poor more vulnerable to the disease (Gini is not
significant in estimation II).
• The Muslim percentage of the population in estimations I and II has a
significant and negative influence on the 1997 HIV/AIDS rate indicating that
the rules (no sex outside the marriage) and morals instilled when practicing
the religion help in curbing HIV/AIDS.
• GDP per capita 1997 in estimations I and II has an insignificant and negative
influence on the 1997 HIV/AIDS rate indicating that wealth has a negative
influence on the HIV/AIDS rate as more money can bring about more
86
education and health care. In this model, however, GDP p/c is not
significant.
• The Percent of men in the military in 1997 in estimations I and II has an
insignificant and negative influence on the 1997 HIV/AIDS rate. This result
is puzzling as the percentage of men in the military was predicted to have a
positive influence on the HIV/AIDS as men in the military are often far from
home and engage in sex with multiple, irregular partners. In any case, the
result is insignificant.
• The effect of the Gender Parity Index of adult literacy in estimations I and II
is positive but insignificant. The positive sign indicates that the more
unequal literacy is between men and women, the less the women will know
about the disease and hence, be able to protect themselves.
• The coefficient of Gender inequality in estimations I and II has a positive and
significant relationship with HIV/AIDS indicating that the less rights a
woman has, the more likely she will not be able to protect herself and be
more vulnerable to the disease.
Though only 22 of 89 countries in this dissertation had a multisectoral
HIV/AIDS program in place in 1997, this re-estimation reveals early evidence of
their success.

87
Models 2a, 2b, 2c: Determinants of HIV/AIDS 2005 and 1995
Model 2a examines the relationship between a group of political, societal,
and economic variables including the variable for multisectoral program with the
HIV/AIDS rate in 2005. Model 2b provides the results from the two-stage
estimation of the same model and, Model 2c repeats Model 2a with variables from
1995 instead of 2005. The purpose of Models 2a, 2b, and 2c is to identify those
variables that have the strongest relationship with the HIV/AIDS rate. These results
will be presented in estimations I, II, and III of Model 2a.
Estimations IV, V, and VI of Model 2a represents three institutional
variables: rule of law, government effectiveness, and decentralization. These
institutional variables are interacted with multisectoral program variable to see if
strong political institutions along with a multisectoral program work together to curb
the 2005 HIV/AIDS rate (this is not done in Model 2c because institutional variables
for 1995 are not available). Below are algebraic forms of Models 2a, 2b, and 2c:
2a] HIV/AIDS rate 2005 = α + β
1
M-S program + β
2
Gini + β
3
Muslim percentage
+ β
5
Male population (15-30) + β
6
Rule of Law*M-S program + β
7
Government
Effectiveness*M-S program + β
8
Decentralization*M-S program + β
9
GDP P/C +
β
10
Literacy Rate + β
11
Telephones + β
12
AIDS rate of neighbor + γ
1
Civil War +
γ
2
Legal prostitution + ε

88
2b] HIV/AIDS Rate 2005 = α + β
1
M-S program + β
2
Gini + β
3
Muslim percentage
+ β
5
Male population (15-30) + β
6
Rule of Law*M-S program + β
7
Government
Effectiveness*M-S program + β
8
Decentralization*M-S program + β
9
GDP P/C +
β
10
Literacy Rate + β
11
Telephones + β
12
AIDS rate of neighbor + γ
1
Civil War +
γ
2
Legal prostitution + ε

Instrumental Variables for M-S program: New Political Regime, Africa

2c] HIV/AIDS Rate 1995 = α + β
1
M-S program1995 + β
2
Gini + β
3
Muslim
percentage + β
4
Male population (15-30)1995 + β
5
GDP P/C1995 + β
6
Literacy
Rate1995 + β
7
Telephones1995 + β
8
AIDS rate of neighbor1995 + γ
1
Civil War +
γ
2
Legal prostitution + ε

89

Table 6: Results Model 2a-Dependent Variable: 2005 HIV/AIDS Rate

Model 2a
Variable I II III IV V VI
Constant -0.204002 -0.225880 -0.245507
(0.134092) (0.290462) (0.340808)
M-S Program -0.142722* -0.165138** -0.163504*
(0.076159) (0.082679) (0.094555)
Gini 0.617133** 0.658479** 0.483732
(0.270015) (0.299794) (0.405079)
Civil War 0.075059 0.084659 0.101688
(0.053881) (0.058460) (0.066929)
Legal Prost. 0.053746 0.055294 0.087875
(0.055244) (0.057427) (0.068323)
Percent Muslim 0.096858 0.118225 0.175929*
(0.074207) (0.081812) (0.099110)
Male Pop (15-
30) 0.269130 0.313918
(1.344333) (1.533629)
GDP P/C -0.010605 -0.046299
(0.067268) (0.114987)
Literacy Rate 0.151864
(0.233064)
No. of phones -0.018460
(0.526899)
AIDS rt. Nbr. 0.633441
(0.533042)
Rule of
Law*MS -0.20446
(0.166971)
Gov’t. Eff.*MS -0.018623
(0.039265)
Decent*MS -0.011595
(0.010125)

R-Squared 0.133699 0.147988 0.176924
Observations 89 85 76

*Significant at 10% level, **Significant at 5% level, ***Significant at 1% l
90
Results
Results from Model 2a indicate that having a multisectoral HIV/AIDS
program over the time period 1995 to 2005 led to a negative and significant
influence on the 2005 HIV/AIDS rate over three executions of the model.
Specifically, the results for all variables are as follows:
• The Gini coefficient in estimations I and II has a significant and positive
influence on the 2005 HIV/AIDS rate indicating that income inequality raises
the HIV/AIDS rate. This may reflect that the greater is income inequality for
any given level of GDP per capita; a larger percentage of the population will
be left without access to either important information and/or treatment. (Gini
is positive but not significant in estimation III).
• A Civil War any time over 1995 to 2005 is positive in all models indicating
the a civil war worsens HIV/AIDS due to forced sex and sex with irregular
partners. In all estimations of the model, however, the result is insignificant.
• Legalized prostitution was predicted to have a negative relationship with the
HIV/AIDS rate as safe and monitored intercourse should help to curb further
spread of the disease. In all cases, however, the estimated effect turns out to
be positive but not significant. Perhaps monitoring needs to be improved.
• The Muslim percentage of the population in estimations I and II and III has a
positive influence on the 2005 HIV/AIDS rate. The variable is only
significant in estimation III. This result is opposite to what was found in
Model 1 (the variable had a negative influence on HIV/AIDS indicating that
91
the rules and morals instilled when practicing the religion help in curbing
HIV/AIDS). Perhaps these rules and morals aren’t always adhered to and so
practicing the religion has a positive influence on the disease.
• Male population (aged 15-30) in estimations II and III has a positive
influence on the 2005 HIV/AIDS rate indicating that being a male at a highly
sexually active age increases the vulnerability to the disease. This result,
however, is insignificant.
• GDP per capita 2005 in estimations II and III has an insignificant, negative
influence on the 2005 HIV/AIDS rate. This indicates that income or wealth
has a negative but not statistically significant influence on the HIV/AIDS rate
since income may be associated with better access to information about the
disease and to health care.
• The Literacy Rate in estimation III has a positive but insignificant influence
on 2005 HIV/AIDS. The positive sign was something of a surprise possibly
relating that illiterate people may receive much useful information about
HIV/AIDS via local radio.
• The number of Telephone mainlines per 100 people in the population has a
negative but insignificant influence on the 2005 HIV/AIDS rate.
• The Neighboring country with the highest HIV/AIDS rate has a positive but
again insignificant influence on the 2005 HIV/AIDS rate. Close proximity to
high HIV/AIDS countries was believed to increase exposure and hence,
92
vulnerability to the disease. But, as noted, the result was not statistically
significant.
• All three interaction terms: rule of law * multisectoral program, government
effectiveness * multisectoral program, and decentralization * multisectoral
program have negative but not significant influences on the 2005 HIV/AIDS
rate. The purpose of including these interaction terms was to combine the
strong political determinants that are conducive to implementing effective
HIV/AIDS policies with an actual HIV/AIDS policy: the multisectoral
program. Together, the inclusion of these variables should have a negative
relationship with HIV/AIDS and they do but it is not a significant one.
Model 2a explored several political, societal, and economic variables and it
seems that a robust finding is that: the presence of a multisectoral HIV/AIDS
program has a negative and significant influence on the 2005 HIV/AIDS rate, the
main hypothesis for this dissertation. It is now important to look at the two-stage
model (Model 2b) to ensure that there is no reverse causality between the HIV/AIDS
rate and a multisectoral program. Instrumental variables chosen for this model are:
the regional dummy variable for Africa (since only Africa has multisectoral
programs which are funded by the World Bank), and the variable representing a new
political regime over the period 1995 to 2005. These two variables both have
positive and significant influences on a multisectoral program rate but should not be
expected to have a significant relationship with the 2005 HIV/AIDS rate. Two tests
of the instruments will also be executed to ensure their validity. The relevance test is
93
executed to determine that the correlation between the instruments and the
multisectoral program variable is not equal to zero. The exogeneity test is executed
to determine that the correlation between the instruments and the error from the
second stage result is equal to zero. This test requires the regression of the second
stage residuals on the instruments.
Table 7: Results Model 2b-Dependent Variable: 2005 HIV/AIDS Rate

Model 2b
Variable I II III
M-S Program -0.412839** -0.490707* -0.473605
(0.212860) (0.295036) (0.311532)
Gini 0.512794
(0.436418)
Civil War 0.109407* 0.111217 0.100506
(0.064291) (0.072344) (0.071208)
Legal Prostitution 0.123141* 0.134259* 0.090182
(0.065530) (0.071296) (0.073236)
Percent Muslim 0.148051* 0.163551 0.155265
(0.084574) (0.101461) (0.103189)
Male Pop. (15-30) 0.873978
(1.778549)
GDP P/C 0.032847 -0.082221
(0.093962) (0.102694)
Literacy Rate 0.127045 0.074039 0.065430
(0.117888) (0.251361) (0.255272)
No. of telephones -0.269969 0.167967
(0.499624) (0.560324)
AIDS rt. Neighbor 0.804623** 0.708849 0.391931
(0.416627) (0.531005) (0.606472)
R-Squared 0.018772 -0.020041 0.036721
Observations 79 76 76

*Significant at 10% level, **Significant at 5% level, ***Significant at 1% level
94
Table 8: Wald Test – Model 2b - Relevance Test for Instruments: Africa and
New Political Regime
Wald Test
Coefficient 1=0
Coefficient 2=0
Coefficient 3=0
Coefficient 4=0
Coefficient 5=0
Coefficient 6=0
Coefficient 7=0
Coefficient 8=0
Coefficient 9=0
Coefficient 10=0
Coefficient 11=0
Coefficient 12=0

F-statistic= 15.55061
Probability=0.000000

Table 9: Exogeneity Test - Model 2b - Exogeneity Test (Residuals) for
instruments: Africa and New Political Regime
Variable I
New Regime 0.072704
Africa 0.102810
Gini -0.672574
Civil War 0.038398
(0.062262)
Legal Prostitution -0.050820
Percent Muslim -0.084385
Male Population 0.635440
GDP P/C -0.018886
Literacy Rate 0.240036
No. of Telephones -0.553609
AIDS rate of Neighbor -0.153786
R2 0.129131
OBS 76
95
Results
In Model 2b, the variables representing a new political regime and the
dummy variable for African countries were again used as instruments for the
variable: multisectoral program. In this two stage model, those IVs as well as all
other variables from Model 2a were examined. The IVs representing a multisectoral
program have significant and negative influences on the 2005 HIV/AIDS rate in
estimations I and II. To ensure that the chosen instruments were valid, a relevance
test and an exogeneity test were preformed Conditions required for these tests were
met (F-statistic greater than 10 and coefficients on instruments equal zero) indicating
that instruments were valid. The two-stage model reconfirms that multisectoral
programs are useful for fighting HIV/AIDS and the result also eliminates the
possibility of reverse causality. The results for all variables are as follows:
• The Gini coefficient in estimation III has a positive influence on the 2005
HIV/AIDS rate indicating that income inequality worsens HIV/AIDS as a
disproportionate amount of assistance and information for the rich is likely to
leave the poor more vulnerable to the disease. This result, however, is not
significant.
• A Civil War any time over 1995 to 2005 has a positive effect on the
HIV/AIDS rate in all models possibly due to forced sex and sex with
irregular partners that tends to occur during civil wars. In estimation I of the
model, the result is also statistically significant.
96
• Legalized prostitution was predicted to have a negative relationship with the
HIV/AIDS rate as safe and monitored intercourse should help to curb further
spread of the disease. In all estimations, however, results are positive and
significant in estimations I and II. Perhaps indicating that legalizing
prostitution increases the demand for it.
• The Muslim percentage of the population in estimations I and II and III has a
positive influence on the 2005 HIV/AIDS rate, but is significant only in
estimation I. This result is opposite to what was found in Model 1 (the
variable had a negative influence on the 1997 HIV/AIDS rate indicating that
the rules and morals instilled when practicing the religion help in curbing
HIV/AIDS). Perhaps these rules and morals aren’t always adhered to and so
practicing the religion has a positive influence on the disease.
• Male population (aged 15-30) in estimation III has a positive influence on the
2005 HIV/AIDS rate indicating that being a male at a highly sexually active
age increases the vulnerability to the disease. This result, however, is
insignificant.
• GDP per capita 2005 has a positive influence on HIV/AIDS in estimation II
but in III has a negative one indicating that income may have mixed
influences on the HIV/AIDS rate. This could be due to the fact that while
many of the low-income countries in this sample have a high HIV/AIDS rate,
some of the highest rates are in the middle to low-income countries. In this
model, however, GDP p/c is not significant.
97
• The Literacy Rate in estimations I, II, and III have positive but insignificant
influences on 2005 HIV/AIDS. Though insignificant, the positive signs are
puzzling as literacy should lead to more education and awareness about
HIV/AIDS. In many developing countries, however, a great deal of the
illiterate population receives information about HIV/AIDS via local radio.
• The number of Telephone mainlines per 100 people in the population has a
negative influence on the 2005 HIV/AIDS rate in estimation II as technology
in a country is a good indicator of better access to information and health
care. There is a positive influence in estimation III however, revealing mixed
results. Both results, however, are insignificant.
• The Neighboring country with the highest HIV/AIDS rate has a positive
influence on the 2005 HIV/AIDS rate as close proximity to high HIV/AIDS
countries tends to increase exposure and hence, vulnerability to the disease.
The result is significant in estimation I of the model.

Model 2b explored several political, societal, and economic variables and it
seems that a robust finding is again that: the presence of a multisectoral HIV/AIDS
program has a negative and significant influence on the 2005 HIV/AIDS rate, the
main hypothesis for this dissertation. The next model (Model 2c) will examine all
the same political, societal, and economic variables but for the year 1995 to identify
differences over the time period.

98

Table 10: Results Model 2c-Dependent Variable: 1995 HIV/AIDS Rate

Model 2c
Variable I II III IV
Constant -0.009895 0.004878 -0.021665 -0.55517
(0.022494) (0.026455) (0.047246) (0.059719)
M-S Program -0.006364 -0.006333 -0.004833 -0.008359
(0.010976) (0.0011056) (0.012236) (0.014543)
Gini 0.115451** 0.100214* 0.123109** 0.168165**
(0.050530) (0.054310) (0.057242) (0.076917)
Civil War -0.008367 -0.007414 0.003700
(0.009785) (0.011060) (0.014039)
Legal Prostitution -0.014529 -0.015841* -0.018066* -0.014537
(0.009564) (0.009687) (0.010667) (0.012974)
Percent Muslim -0.024472* -0.029283** -0.029605** -0.022435
(0.013326) (0.014031) (0.015331) (0.019742)
Male Pop. (15-30) 0.108040 0.101923
(0.268635) (0.368707)
GDP P/C -1.17E-05** -1.09E-05* -1.37E-05** -1.71E-05**
(5.51E-06) (5.92E-06) (6.06E-06) (7.79E-06)
Literacy Rate 0.020170
(0.033302)
No. of telephones -0.070991 -0.089668
(0.077351) (0.131230)
AIDS rt. Neighbor 0.442485*** 0.424033*** 0.452947*** 0.614779***
(0.076695) (0.079569) (0.081979) (0.113554)
R-Squared 0.481800 0.490282 0.485262 0.601969
Observations 89 89 81 59

*Significant at 10% level, **Significant at 5% level, ***Significant at 1% level

99
Results
In Model 2c, the variable for multisectoral programs is negative but not
significant due to the lack of multisectoral HIV/AIDS programs that were in place in
1995. The results for all variables are as follows:
• The Gini coefficient in estimations I, II, III, and IV all have a positive and
significant influence on the 1995 HIV/AIDS rate indicating that income
inequality worsened HIV/AIDS as a disproportionate amount of assistance
and information for the rich is likely to leave the poor more vulnerable to the
disease.
• A Civil War any time over 1985-1995 is negative in estimations II and III
and positive in estimation IV indicating mixed results. None of the results,
however, are significant.
• Legalized prostitution has a negative relationship with the 1995 HIV/AIDS
rate as safe and monitored intercourse help to curb further spread of the
disease. In estimations II and III, results are significant.
• The Muslim percentage of the population in estimations I and II and III has a
negative and significant influence on the 1995 HIV/AIDS rate. This result is
similar to what was found in Model 1 (the variable had a negative influence
on the 1997 HIV/AIDS rate indicating that the rules and morals instilled
when practicing the religion help in curbing HIV/AIDS).
• Male population 1995 (aged 15-30) in estimations III and IV has a positive
influence on the 19955 HIV/AIDS rate indicating that being a male at a
100
highly sexually active age increases the vulnerability to the disease. This
result, however, is insignificant.
• GDP per capita 1995 in estimations I, II, III, and IV has a negative and
significant influence on 1995 HIV/AIDS indicating that wealth was useful in
the fight against HIV/AIDS.
• The Literacy Rate 1995 in estimation IV has a positive but insignificant
influence on 1995 HIV/AIDS. Though it is insignificant, the positive sign for
this result is puzzling as literacy should lead to more education and
awareness about HIV/AIDS. In many developing countries, however, a great
deal of the illiterate population receives information about HIV/AIDS via
local radio.
• The number of Telephone mainlines per 100 people in the population has a
negative influence on the 2005 HIV/AIDS rate in estimation II as technology
in a country is a good indicator of better access to information and health
care. There is a positive influence in estimations II and IV, however, both
results are insignificant.
• A Neighboring country with the highest HIV/AIDS rate in 1995 has a
positive influence on the 1995 HIV/AIDS rate as close proximity to high
HIV/AIDS countries tends to increase exposure and hence, vulnerability to
the disease. The result is significant in estimations I, II, III, and IV of the
model.
101
Results from model 2c reveal that as of 1995, income inequality and having a
neighbor country with a high HIV/AIDS rate contributed to a higher HIV/AIDS rate.
The result for income inequality is the same for 2005 and for 1997 (model 1)
indicating that income inequality remains a large contributor to worsening
HIV/AIDS rates. The variables, legalized prostitution and percentage Muslim had a
positive influence on HIV/AIDS in 2005 but a negative influence in 1995. While
these practices may have been effective in curbing the HIV/AIDS rate at an earlier
stage in the epidemic, it is clear that they were not as effective ten years later.

Models 3a and 3b: Change in HIV/AIDS rate between 1995 and 2005
Model 3a examines the relationship between the same several political,
societal, and economic variables from Models 2a, 2b, and 2c and the change in the
HIV/AIDS rate from 1995 to 2005. Two more variables are also included in Models
3a and 3b: the HIV/AIDS rate in 1995 to see how past rates affect the change over
time and the dummy variable for the World Bank MAP which began in 2000 (in the
middle of the change period). Model 3b provides the results from the two-stage
estimation of the same model. The purpose of Models 3a and 3b are to identify those
variables that have the strongest relationship with the change in the HIV/AIDS rate.
Below are algebraic forms of Models 3a and 3b:

102
3a] ΔHIV/AIDS Rate 1995-2005 = α + β
1
M-S program + β
2
Gini + β
3
Muslim
percentage + β
4
ΔMale population (15-30) + β
5
ΔGDP P/C + β
6
ΔLiteracy Rate +
β
7
ΔTelephones + β
8
ΔAIDS rate of neighbor + β
9
AIDS rt. 1995 + γ
1
Civil War +
γ
2
World Bank program + γ
3
Legal Prostitution + ε

3b] ΔHIV/AIDS Rate 1995-2005 = α + β
1
M-S program + β
2
Gini + β
3
Muslim
percentage + β
4
ΔMale population (15-30) + β
5
ΔGDP P/C + β
6
ΔLiteracy Rate +
β
7
ΔTelephones + β
8
ΔAIDS rate of neighbor + β
9
AIDS rt. 1995 + γ
1
Civil War +
γ
2
World Bank program + γ
3
Legal Prostitution + ε

Instrumental Variables for M-S program: New Political Regime, Africa

103
Table 11: Results Model 3a-Dependent Variable: ΔHIV/AIDS Rate (1995-2005)

Model 3a
Variable I II III
Constant 0.009776 0.013698 0.033707
(0.004663) (0.005639) (0.017628)
MS Program -0.002787
(0.009268)
GINI -0.036751
(0.035435)
Civil War -0.009832 -0.010106
(0.006542) (0.006756)
World Bank MAP -.019989*** -0.017354** -0.017321**
(0.006370) (0.007090) (0.007233)
Percent Muslim 0.028078*** 0.021700** 0.018544*
(0.008425) (0.009410) (0.009898)
Legal Prostitution -0.003189
(0.006609)
Δ Male Pop. (15-30) -0.045244** -0.027674 -0.025627
(0.017463) (0.020349) (0.020909)
ΔGDP Per Capita 0.081950** 0.072932* 0.060204
(0.42403) (0.043444) (0.060204)
Δ Literacy Rate -0.012307 -0.008988 -0.009182
(0.008281) (0.008454) (0.008764)
Δ Telephones 2.01E-05
(0.000294)
AIDS rate 1995 -0.026271 -0.019450
(0.058049) (0.061745)
Δ Neighbor AIDS rt. 0.001109 0.001114
(0.00915) (0.000935)
R-Squared 0.341884 0.372984 0.388660
Observations 79 79 79

*Significant at 10% level, **Significant at 5% level, ***Significant at 1% level

104
Results
Model 3a measures the change in the HIV/AIDS prevalence rate over the
period 1995-2005. In this model, the variable for multisectoral program is
insignificant but the variable for having the World Bank’s MAP leads to lower
HIV/AIDS rate over the time period in all three executions of the model. Once
again, there is an indication that multisectoral programs are effective in combating
HIV/AIDS. The results for all variables are as follows:
• The Gini coefficient in estimation III is negative which is unexpected as
income inequality in past models has worsened HIV/AIDS as a
disproportionate amount of assistance and information for the rich is likely to
leave the poor more vulnerable to the disease. In this estimation, however, it
is insignificant.
• A Civil War any time over 1995 to 2005 is negative in estimations II and III
and These results, however, are insignificant.
• Legalized prostitution has a negative relationship with the change in
HIV/AIDS in estimation III as safe and monitored intercourse help to curb
further spread of the disease. This result is insignificant.
• The Muslim percentage of the population in estimations I and II and III has a
positive and significant influence on the change in the HIV/AIDS rate. This
result seems to indicate that the rules and morals instilled when practicing the
religion are not helping to curb HIV/AIDS overtime.
105
• The change in Male population (aged 15-30) from 1995 to 2005 in
estimations I, II, and III has a negative influence on the change in HIV/AIDS
rate indicating that being a male at a highly sexually active age decreased the
vulnerability to the disease overtime. This result is significant in estimation I
of the model. Perhaps males have taken more steps to fight the disease
overtime.
• The change in GDP per capita from 1995 to 2005 in estimations I, II, and III
has a positive influence on the change in HIV/AIDS indicating that wealth
was not an important factor in fighting the disease over the time period. The
result is significant in estimations I and II.
• The change in the Literacy Rate from 1995 to 2005 in estimations I, II, and
III has a negative but insignificant influence on the change in HIV/AIDS. It
seems that over the ten-year period, being literate had a positive influence in
the fight against HIV/AIDS.
• The change in the number of Telephone mainlines per 100 people in the
population has a negative influence on the change in the HIV/AIDS rate in
estimation III as technology in a country is a good indicator of better access
to information and health care. However, this result is insignificant.
• The HIV/AIDS rate in 1995 has a negative but insignificant influence on the
change in the HIV/AIDS in estimations II and III indicating that past
HIV/AIDS leads to a lower HIV/AIDS rate overtime.
106
• A change in the HIV/AIDS rate of a high-AIDS Neighboring country has a
positive influence on the change in the HIV/AIDS rate as close proximity to
high HIV/AIDS countries tends to increase exposure and hence, vulnerability
to the disease. The result is insignificant in estimations I, II, and III of the
model.

Results from Model 3a reveal that being a member of the World Bank MAP
(which requires the establishment of a multisectoral HIV/AIDS program) was a
strong factor for combating HIV/AIDS over the period 1995 to 2005. Other
noteworthy results indicate that overtime, the practice of Islam leads to higher
HIV/AIDS rates and a higher GDP per capita does not necessarily aid in the fight on
the disease. To ensure that there is not reverse causality, a two stage least squares
model is executed below (Model 3b). The same IVs as in model 2b are used for the
multisectoral program variable: New Political Regime and Africa.

107

Table 12: Results Model 3b-Dependent Variable: ΔHIV/AIDS Rate (1995-2005)

Model 3b
Variable I II
M-S Program 0.025345** 0.022897
(0.010704) (0.021958)
Gini 0.004016
(0.024663)
Civil War -0.008350 -0.008646
(0.006802) (0.007110)
World Bank -0.017918*** -0.016294**
(0.006485) (0.007673)
Δ Male Pop. (15-30) -0.043152** -0.032984
(0.018677) (0.0221829)
Percent Muslim 0.026465*** 0.023464**
(0.009005) (0.009954)
Legal Prostitution -0.004559
(0.007099)
Δ GDP P/C 0.077013* 0.065696
(0.045834) (0.050954)
Δ Literacy Rate -0.013685 -0.011184
(0.009188) (0.009832)
Δ No. of telephones -3.66E-05
(0.000317)
AIDS rt. 1995 -0.002186
(0.068269)
Δ AIDS rt. Neighbor 0.001066
(0.000991)
R-Squared 0.268680 0.301223
Observations 79 79

*Significant at 10% level, **Significant at 5% level, ***Significant at 1% level

108

Table 13: Wald Test - Model 3b - Relevance Test for instruments: Africa and
New Political Regime
Wald Test
Coefficient 1=0
Coefficient 2=0
Coefficient 3=0
Coefficient 4=0
Coefficient 5=0
Coefficient 6=0
Coefficient 7=0
Coefficient 8=0
Coefficient 9=0
Coefficient 10=0
Coefficient 11=0
Coefficient 12=0
F-statistic= 18.95421
Probability=0.000000

109
Table 14: Exogeneity Test - Model 3b - Exogeneity Test (Residuals) for
instruments: Africa and New Political Regime

Results
Model 3b represents a two-stage test using the identical instrumental
variables (from Model 2b) for a multisectoral program: new political regime and
Africa. The results are puzzling as the multisectoral program variable has a positive
Variable I
New Regime 0.002103
Africa -9.04E-05
GINI 0.012032
Civil War -0.003629
World Bank MAP 0.003313
Legal Prostitution -0.005818
Percent Muslim -0.008624
AIDS 1995 -0.59962
Δ Male Population 0.010532
Δ GDP P/C 0.003354
Δ Literacy Rate 0.002194
Δ No. of Telephones 0.000117
Δ AIDS rate of Neighbor -0.000450
R2 0.049465
OBS 79
110
influence on the change in HIV/AIDS rate but the World Bank MAP multisectoral
program leads to a significantly lower HIV/AIDS rate in both estimations of the
model. This may be due to the fact that the World Bank MAP began in the middle of
the time period in 2000 and may have a strong influence on the actual change in the
HIV/AIDS rate. To ensure that the chosen instruments were valid, a relevance test
and an exogeneity test were preformed. Conditions required for these tests were met
(F-statistic greater than 10 and coefficients on instruments equal zero) indicating that
instruments were valid. The results for all variables in the two-stage model are as
follows:
• The Gini coefficient in estimation II has a positive influence on the change in
HIV/AIDS as income inequality translates into a disproportionate amount of
assistance and information for the rich and is likely to leave the poor more
vulnerable to the disease. In this estimation, however, it is insignificant.
• A Civil War any time over 1995 to 2005 is surprisingly negative in
estimations I and II indicating that civil war actually leads to less AIDS
overtime. Perhaps this is due to a lack of reporting or population. These
results, however, are insignificant.
• Legalized prostitution has a negative relationship with the change in
HIV/AIDS in estimations I and II as safe and monitored intercourse help to
curb further spread of the disease. This result, however, is insignificant.
• The Muslim percentage of the population in estimations I and II has a
positive and significant influence on the change in the HIV/AIDS rate. This
111
result seems to indicate that the rules and morals instilled when practicing the
religion are not helping to curb HIV/AIDS overtime.
• The change in Male population (aged 15-30) from 1995 to 2005 in
estimations I and II has a negative influence on the change in HIV/AIDS rate
indicating that being a male at a highly sexually active age decreased the
vulnerability to the disease overtime. This result is significant in estimation I
of the model. Perhaps males have taken more steps to fight the disease
overtime.
• The change in GDP per capita from 1995 to 2005 in estimations I and II has a
positive influence on the change in HIV/AIDS indicating that wealth was not
an important factor in fighting the disease over the time period. The result is
significant in estimation I.
• The change in the Literacy Rate from 1995 to 2005 in estimations I and II has
a negative but insignificant influence on the change in HIV/AIDS. It seems
that over the ten-year period, being literate had a positive influence in the
fight against HIV/AIDS.
• The change in the number of Telephone mainlines per 100 people in the
population has a negative influence on the change in the HIV/AIDS rate in
estimation II as technology in a country is a good indicator of better access to
information and health care. However, this result is insignificant.
112
• The HIV/AIDS rate in 1995 has a negative but insignificant influence on the
change in the HIV/AIDS in estimation II indicating that past HIV/AIDS leads
to a lower HIV/AIDS rate overtime.
• A change in the HIV/AIDS rate of a high-AIDS Neighboring country has a
positive influence on the change in the HIV/AIDS rate as close proximity to
high HIV/AIDS countries tends to increase exposure and hence, vulnerability
to the disease. The result is insignificant in estimation II of the model.

A strong result from Model 3b reveals that, again, being a member of the
World Bank MAP (which requires the establishment of a multisectoral HIV/AIDS
program) was a strong factor for combating HIV/AIDS over the period 1995 to 2005.
From all models executed thus far it is clear that multisectoral programs are effective
in combating HIV/AIDS.

Model 4: 2005 Expenditures on HIV/AIDS
Model 4 examines the relationship between several political, societal, and
economic variables and per capita spending on HIV/AIDS. Several new variables
are incorporated in this model as the dependent variable is no longer the HIV/AIDS
prevalence rate. New variables include Gross National Income (GNI) to see how
income influences spending, decentralization to see if many government voices can
effectively push for more spending, secondary school enrollment to see how a more
educated population (which typically earns a higher income) effects spending, a
113
dummy variable for if the health minister has a MD to see if a knowledgeable
government member can effectively push for more spending and finally, an
interaction term of the MAP along with the variable for multisectoral HIV/AIDS
program to see if there is a combined effect on spending on HIV/AIDS. The purpose
of Model 4 is to identify those variables that have a significant effect on spending on
HIV/AIDS. Below is the algebraic form of Model 4:
4] Expenditures on HIV/AIDS Rate 2005 = α + β
1
M-S program + β
2
GNI +
β
3
HIV/AIDS rate 1995 + β
5
Male population (15-30) + β
4
Decentralization +
β
6
School Enrollment + β
7
MAP*MS + γ
1
Civil War + γ
2
Healht Minister has an
MD + ε

114
Table 15: Results Model 4-Dependent Variable: 2005 Expenditures on AIDS
Model 4
Variable I II III IV
Constant 0.178888 -0.013522 1.449366 1.475469
(0.497075) (0.535998) (1.482108) (1.540815)
M-S Program 1.497075** 1.261325** 1.514862** 1.458460**
(0.583357) (0.609971) (0.603896) (0.618036)
Income 0.000517*** 0.000522*** 0.000593*** 0.000600***
(0.000153) (0.000153) (0.000157) (0.000168)
Civil War -0.436154 -0.432395 -0.408353 -0.450552
(0.409369) (0.411767) (0.413830) (0.429748)
AIDS rate
1995 6.538072* 7.003055** 8.285092** 7.988311**
(3.607833 (3.628738) (3.644769) (3.851298)
MD 0.503096
(0.537965)
Male Pop.
(15-30) -12.35356 -14.48587
(9.827541) (10.28538)
Decentralizati
on 0.079156 0.096377 0.089854
(0.066238) (0.065875) (0.066697)
School
Enrollment -1.004011 -1.127729
(0.934131 (0.995938)
WB*MS 0.377614
(0.932767)
R-Squared 0.188401 0.201684 0.264973 0.273784
Observations 89 88 86 86

*Significant at 10% level, **Significant at 5% level, ***Significant at 1% level

115
Results
Model 4 looks at the 2005 government expenditures on HIV/AIDS and
having a multisectoral HIV/AIDS program has a positive and significant influence
on HIV/AIDS in all four estimations of the model indicating that Multisectoral
programs are promote spending on fighting the disease. Specifically, the results for
all variables are as follows:
• GNI in estimations I, II, III, and IV has a significant and positive influence on
HIV/AIDS expenditures indicating that national income is important for
government spending on HIV/AIDS.
• A Civil War any time over 1995 to 2005 is negative in all four estimations
indicating the a civil war worsens spending on HIV/AIDS due to resources
being spent on the war instead of the disease. In all estimations of the model,
however, the result is insignificant.
• The HIV/AIDS rate in 1995 has a positive and significant influence on 2005
HIV/AIDS expenditures across all four estimations of the model. This
indicates that past HIV/AIDS leads to higher future spending on the disease.
In all estimations, results are significant. .
• If a government’s health minister is a Medical Doctor, there is a positive
influence on HIV/AIDS spending as predicted but the result is insignificant.
• Male population (aged 15-30) in estimations III and IV has a negative but
insignificant influence on 2005 HIV/AIDS expenditures. This seems to
indicate that with that not enough spending is taking place even though a
116
large population of males at a highly sexually active age increases the
vulnerability to the disease.
• Decentralization in estimations II, III, and IV has an insignificant but positive
influence on 2005 HIV/AIDS expenditures indicating that multiple
governmental voices have a positive influence on HIV/AIDS spending.
• Secondary school enrollment rates (2005) in estimations III and IV have a
negative and insignificant influence on 2005 HIV/AIDS. Though it is
insignificant, the negative sign for this result is puzzling as education should
lead to more income and spending on HIV/AIDS. In many developing
countries, however, enrollment rates often do not translate into actual
completion rates, making potential incomes less than may be expected.
• The interaction term of MAP with multisectoral program is positive in
estimation IV indicating that this combination of multisectoral activity does
lead to an increase on 2005 HIV/AIDS expenditures but this result is
insignificant.

Model 4 explored several political, societal, and economic variables and it
seems that a robust finding is that: the presence of a multisectoral HIV/AIDS
program has a positive and significant influence on 2005 expenditures on HIV/AIDS
rate.

117
Model 5: 2005 Knowledge of Condoms by sexually active women (aged 15-49)
Model 5 examines the relationship between several political, societal, and
economic variables and women’s (aged 15-49) knowledge of condoms. The purpose
of Model 5 is to identify those variables that have a significant effect on condom
awareness. Below is the algebraic form of the model:

5] Condom Knowledge 2005 = α + β
1
M-S program + β
2
Gini + β
3
Muslim
percentage + β
4
AIDS expenditures+ β
5
GDP P/C + β
6
Literacy Rate +
β
7
Telephones + β
8
AIDS rate of neighbor + γ
1
Legal prostitution + γ
2
World Bank
Program + ε
118

Table 16: Results Model 5 Dependent Variable: 2005 Female Condom
Knowledge
Model 5
Variable I II III
Constant 0.520325*** 0.621702** 0.446012*
(0.119074) (0.275519) (0.253584)
M-S Program 0.139863** 0.152942*
(0.072293) (0.083233)
Gini -0.453113 -0.369479
(0.319535) (0.310286)
World Bank
Program 0.018830
(0.065867)
Legal Prostitution 0.065404 0.050021
(0.053396) (0.053034)
Percent Muslim -0.175780** -0.217361** -0.179083**
(0.068657) (0.084530) (0.088426)
AIDS Expenditures 0.004130 -0.006903
(0.016771) (0.017092)
GDP P/C 0.058787 0.075139
(0.109477) (0.098770)
Literacy Rate 0.323876*** 0.313163* 0.381603**
(0.111900) (0.184853) (0.181436)
No. of telephones -0.279908 -0.443517
(0.591843) (0.575548)
AIDS rt. Neighbor 0.323876 0.406645 0.569597
(0.293843) (0.401070) (0.403883)
R-Squared 0.450515 0.445387 0.492876
Observations 47 47 47

*Significant at 10% level, **Significant at 5% level, ***Significant at 1% level

119
Results
Model 5 looks at condom awareness among females (between ages 15-49).
Results show that multisectoral HIV/AIDS programs in estimations I and III have a
positive and significant influence on condom awareness. In estimation II, the World
Bank MAP variable is included and it has a positive but insignificant influence on
condom awareness. These results indicate that multisectoral HIV/AIDS programs
are helpful in bringing about greater condom awareness for sexually active women.
It is important to note that data on female condom knowledge was only available for
47 out of the 89 countries in the database for this dissertation. Specifically, the
results for all variables are as follows:
• The Gini coefficient in estimations II and III has a negative and insignificant
influence on condom awareness indicating that income inequality worsens
knowledge as a disproportionate amount of assistance and information likely
goes to the richer portion of the population.
• Legalized prostitution is predicted to have a positive relationship with female
condom knowledge as sex-education for prostitutes (as in Thailand) is ideally
provided by the government. In all estimations, however, results are
insignificant.
• The Muslim percentage of the population in estimations I and II and III has a
negative and significant influence on female condom awareness as condoms
are often associated with promiscuity in the religion.
120
• 2005 HIV/AIDS expenditures have a positive influence on condom
awareness in estimation II and a negative influence in estimation III. These
mixed results are puzzling as a higher income should help to promote more
HIV/AIDS education. Both results, however, are insignificant.
• GDP per capita 2005 in estimations II and III has an insignificant but positive
influence on condom awareness indicating that wealth has a positive
influence on condom knowledge.
• The Literacy Rate in estimations I, II, and III has a positive and significant
influence on female condom awareness indicating that more education and
awareness is an important factor in knowledge about condoms.
• The number of Telephone mainlines per 100 people in the population has a
negative influence on condom awareness. This result is puzzling as
technology in a country is a good indicator of better access to information
and health care. This result, however, is insignificant.
• A Neighboring country with the highest HIV/AIDS rate has a positive
influence on condom awareness in all three executions of the model as close
proximity to high HIV/AIDS countries tends to increase exposure and
ideally, knowledge about preventing the disease. The result, however, is
insignificant.

121
Model 5 explored several political, societal, and economic variables and it
seems that a robust finding is that: the presence of a multisectoral HIV/AIDS
program has a positive and significant influence on female condom awareness.

Model 6: Percentage growth in GDP per capita from 1990 to 2005
Model 6 examines the relationship between several political, societal, and
economic variables and GDP per capita growth between 1990 and 2005. A variable
that represents average investment over the time period (as measured by fixed capital
formation) is included as well as lagged values (1995) of variables where data was
available. The purpose of Model 6 is to identify those variables that have a
significant effect on GDP growth. Below is the algebraic form of the model:

6] GDP Growth Rate 1995-2005 = α + β
1
HIV/AIDS rate 1995 +
β
2
InvestmentAverage + β
3
School Enrollment95 + β
4
Telephones95 +
β
5
Government Effectiveness + β
6
Decentralization + β
7
Gini + β
8
GNI95 + γ
1
Civil
War + ε

122
Table 17: Results Model 6-Dependent Variable: 1990-2005 GDP Growth Rate
Model 6
Variable I II III IV
Constant 0.017766*** 0.005566 0.028560* 0.018067
(0.003758) (0.008300) (0.015710) (0.017426)
AIDS rate 1995 -0.096012** -0.081742* -0.042168 -0.030835
(0049171) (0.048266) (0.049873) (0.050150)
Average Investment 0.054237* 0.073700** 0.083312**
(0.030715) (0.035342) (0.036576)
School Enrollment 95 -0.022939* -0.030659** -0.029450**
(0.012177) (0.012549) (0.012668)
Telephones 1995 0.111289** 0.10350** 0.129136**
(0.044974) (0.043303) (0.051881)
Civil War 0.002041 0.004084 0.002961
(0.005781) (0.005762) (0.005625)
Gov’t. Effectiveness 0.003376** 0.003600**
(0.001706) (0.001692)
Decentralization 0.000846 0.001344
(0.000880) (0.000892)
GINI -0.064981** -0.057255*
(0.028771) (0.032773)
Income 1995 -3.11E-07
(3.59E-06)
R-Squared 0.044170 0.181325 0.239907 0.263798
Observations 87 85 84 82

*Significant at 10% level, **Significant at 5% level, ***Significant at 1% level

123
Results
Model 6 examines economic growth over the period 1990-2005. The
lagged variable for HIV/AIDS 1995 has a negative and significant influence on
growth in estimations I and II of the model. This result confirms that the disease is
harmful to economic growth. Specifically, the results for all variables are as follows:
• Average investment has a positive and significant influence on GDP growth
in estimations II, III, and IV of the model as expected.
• Secondary school enrollment rates (1995) in estimations II, III and IV have a
significant but negative influence on GDP growth. The negative sign for this
result is puzzling as education should lead to more income and hence, GDP
growth. In many developing countries, however, enrollment rates often do
not translate into actual completion rates, making school enrollment rates less
productive.
• The number of Telephone mainlines per 100 people in the population in 1995
has a positive and significant influence on GDP growth in estimations II, III,
and IV of the model. This result indicates that infrastructure and in
particular, communication technology, in a country can contribute to
economic growth.
• A Civil War any time over 1995 to 2005 has a positive but insignificant
influence on GDP growth in estimations I, II, and IV of the model. The
positive sign of this coefficient is puzzling since wars usually drain a
country’s resources.
124
• The variable representing Government Effectiveness has a positive and
significant influence on GDP growth in estimations III and IV of the model.
This result indicates that a more effective government will take policy actions
that are conducive to GDP growth.
• Decentralization in estimations III and IV has a positive but insignificant
influence on GDP growth indicating that having multiple governmental
voices has a positive influence on economic growth.
• The Gini variable is negative and significant in estimations III and IV
indicating that income inequality is not conducive to GDP growth.
• Past GNI (1995) has a negative and insignificant influence in 1990-2005
GDP growth. This result is puzzling as past income should be positive
indicator of growth. Countries with incomes that are low and do not fluctuate
much however, may not see much growth overtime.

In Model 6, the economic effect from HIV/AIDS was revealed through the
negative influence that HIV/AIDS has on GDP growth. In Models, 1, 2a, 2b, 2c, 3a,
and 3b, political determinants of HIV/AIDS were revealed through such variables as
having a multisectoral program which is effective in fighting HIV/AIDS. Model 4
revealed that spending on HIV/AIDS is increased by the presence of multisectoral
programs and Model 5 revealed that multisectoral programs are helpful for women’s
knowledge of condoms. In the final model, Model 7, it is important to explore those
variables that are conducive to the existence of a multisectoral HIV/AIDS program.
125
Model 7: Existence of a multisectoral HIV/AIDS program by 2005
Model 7 examines the relationship between several political, societal,
regional, and economic variables and the existence of a multisectoral HIV/AIDS
program by the year 2005. Regional dummy variables for Africa, Asia, and Latin
America and the Caribbean are included to see if a particular continent is conducive
to having multisectoral programs. Institutional variables are included in this model
to examine the importance of strong governmental qualities on the existence of a
multisectoral program. The purpose of Model 7 is to identify those variables that
have a significant effect on the existence of multisectoral HIV/AIDS programs.
Below is the algebraic form of the model:

7] Has a Multisectoral Program as of 2005* = α + β
1
GNI95 + β
2
HIV/AIDS Rate
1995 + β
3
Government Effectiveness + β
4
rule of law + β
5
decentralization +
β
6
Male Population (15-30) + β
7
AIDS rate of neighbor + β
8
Gini + γ
1
Health
Minister has a MD + γ
2
Africa + γ
3
Asia+ γ
4
Latin America+ γ
5
New Political
Regime + ε

126
Table 18: Results Model 7-Dependent Variable: Has a Multisectoral Program

Model 7
Variable I II III IV V
Income 1995 0.252448* -.826966* -0.938679*
(0.139556) (0.483395) (0.51275)
AIDS 1995 1.336911 1.408615 -0.909305
(3.616794) (3.861834) (4.484286)
MD 0.111390
(0.474447)
Africa 0.676113*** 0.687617** 0.049665 1.264420* 0.864945
(0.257335) (0.349510) (0.369604) (0.730190) (0.861676)
Asia 0.856480** 0.883236** 1.346614* 1.206950
(0.400108) (0.467541) (0.784411) (0.821253)
Latin America 0.437889 1.483920* 1.249997
(0.491520) (0.817986) (0.878711)
Gov’t. Effective. 0.266273 0.868086 1.007784
(0.797583) (1.168869) (1.196594)
New Regime 0.688128*** 0.446368 0.699864* 0.593654
(0.241836) (0.324248) (0.395410) (0.414133)
Rule of Law -0.373798
(0.933839)
Decentralization 0.088552*
(0.053293)
Male Pop. 95 9.102239 8.234091
(7.909476) (8.124062)
AIDS N. 95 3.470616
(5.164216)
Gini 1.408969
(0.5307)
OBS 89 89 88 81 81

*Significant at 10% level, **Significant at 5% level, ***Significant at 1% level

127
Results
Model 7 is a Probit model that seeks to find what variables bring about the
existence of multisectoral HIV/AIDS programs. Specifically, the results for all
variables are as follows:
• Past GNI (1995) has a negative and significant influence on multisectoral
program existence in estimations III, IV, and V of the model growth
indicating that multisectoral programs exist more often in low-income
countries and may be deemed more necessary in such countries as well as be
more likely to receive World Bank funding.
• Past HIV/AIDS rate (1995) has a positive influence on multisectoral program
existence in estimations III and IV but a negative influence in estimation V.
All results are insignificant but it seems that just having a high past
HIV/AIDS rate does not necessarily mean that a multisectoral program will
be established.
• If a government’s health minister is a Medical Doctor, there is a positive
influence on multisectoral program existence in estimation V (as predicted)
but the effect is statistically insignificant.
• Being on the continent of Africa has a positive and significant influence on
multisectoral program existence in estimations I, II, and IV of the model
indicating that African countries are likely to have multisectoral HIV/AIDS
programs.
128
• Being on the continent of Asia has a positive and significant influence on
multisectoral program existence in estimations I, II, and IV of the model
indicating that Asian countries are likely to have multisectoral HIV/AIDS
programs.
• Being in the Latin America and the Caribbean region has a positive and
significant influence on multisectoral program existence in estimation IV of
the model indicating that LA and C countries are likely to have multisectoral
HIV/AIDS programs.
• The variable representing Government Effectiveness has a positive but
insignificant influence on multisectoral program existence in estimations II,
IV, and V of the model. This result indicates that a more effective
government will take policy actions that are conducive to multisectoral
program establishment.
• A new Political Regime has a positive and significant influence on
multisectoral program existence in estimations I and IV of the model. This
indicates that new leadership is conducive to multisectoral program
establishment.
• Rule of Law has a negative influence on multisectoral program existence in
estimation III indicating that adherence to the law is not important in
multisectoral program establishment. This result is, however, insignificant.
129
• Decentralization in estimation III has a positive and significant influence on
multisectoral program existence indicating that having multiple governmental
voices have a positive influence on multisectoral program establishment.
• Male population 1995 (aged 15-30) in estimations IV and V has a positive
but insignificant influence on multisectoral program existence. This seems to
indicate that having a large population of males at a highly sexually active
age has a positive influence on multisectoral program establishment.
• The Neighboring country with the highest HIV/AIDS rate (1995) has a
positive influence on multisectoral program existence in execution V of the
model as close proximity to high HIV/AIDS countries tends to increase
exposure and ideally, policies to prevent the disease. The result, however, is
insignificant.
• The Gini variable is positive and insignificant in estimation V indicating that
income inequality is not conducive to multisectoral program establishment.

In Model 7, the determinants of multisectoral program establishment are
examined. Significant results reveal that being in Africa, Asia, LA and C, as well as
being a decentralized, low-income country with a new political regime (over the
period 1995 to 2005) all have a positive influence on program existence. The seven
models presented above reveal many significant relationships between political,
societal, economic, and regional variables and variations on the HIV/AIDS rate.
Significant relationships were also found between these variables and female
130
condom awareness, economic growth, and the existence of a multisectoral
HIV/AIDS program. Two results that are especially important indicate that the
HIV/AIDS rate does have a negative effect on economic growth in low-income
countries and that multisectoral HIV/AIDS programs appear to be an important
political policy tool for reducing the HIV/AIDS rate. To further examine how
multisectoral HIV/AIDS programs can be effective in curbing the HIV/AIDS rate, a
difference-in-difference approach is presented below.

Difference-in-Difference Approach

To determine if multisectoral programs have been effective in helping to slow
HIV/AIDS prevalence rates, a difference-in-difference approach is now presented.
In this technique, three data points are used: the HIV/AIDS prevalence rate during
the first year of our data coverage, namely, 1995, the HIV/AIDS prevalence rate
when a multisectoral program was implemented (this differing by country) and, the
HIV/AIDS prevalence rate at the end of the data range: 2005. The differences in
these rates are used to estimate what happens to the HIV/AIDS rate when a
multisectoral program is introduced in the country. The difference of those two
values was then taken to obtain the actual “difference in the differences.” This
difference-in-difference value is shown in italics; a negative number revealing that
the multisectoral HIV/AIDS program was effective in bringing the HIV/AIDS rate
down. A positive difference-in-difference number suggests that the multisectoral
131
HIV/AIDS program was not effective in bringing down the HIV/AIDS rate. For
example, Botswana had a HIV/AIDS rate of 8% in 1995, a rate of 35.8% in 2000
when the multisectoral program was implemented and a HIV/AIDS rate of 24.1% in
2005. The difference in the rates between 2005 and 2000 (24.1-35.8) as well as the
difference in the rates between 2000 and 1995 (35.8-8) are calculated and then the
difference in those two rates is calculated.
These data points were available for only 31 countries in the sample; results
are listed in the following tables.

Table 19: Difference in the Differences - I

Angola Difference
With MS 0.037 0.039 -0.002
With no MS 0.039 0.055 -0.016
0.014

Benin Difference
With MS 0.019 0.02 -0.001
With no MS 0.02 0.0371 -0.0171
0.0161

Botswana Difference
With MS 0.241 0.358 -0.117
With no MS 0.358 0.08 0.278
-0.395

Burundi Difference
With MS 0.06 0.075 -0.015
With no MS 0.075 0.1132 -0.0382
0.0232

132
Table 20: Difference in the Differences – II

Cameroon Difference
With MS 0.054 0.055 -0.001
With no MS 0.055 0.047 0.008
-0.009

Congo Difference
With MS 0.053 0.049 0.004
With no MS 0.049 0.054 -0.005
0.009

Ghana Difference
With MS 0.023 0.022 0.001
With no MS 0.022 0.03 -0.008
0.009

Guinea Difference
With MS 0.015 0.032 -0.017
With no MS 0.032 0.0154 0.0166
-0.0336
Lesotho Difference
With MS 0.232 0.289 -0.057
With no MS 0.289 0.084 0.205
-0.262

Madagascar
With MS 0.005 0.003 0.002
With no MS 0.003 0.0015 0.0015
0.0005

Mali Difference
With MS 0.017 0.019 -0.002
With no MS 0.019 0.03 -0.011
0.009

Mozambique Difference
With MS 0.122 0.16 -0.038
With no MS 0.16 0.14 0.02
-0.058

133
Table 21: Difference in the Differences - III

Niger Difference
With MS 0.011 0.04 -0.029
With no MS 0.04 0.0135 0.0265
-0.0555

Nigeria Difference
With MS 0.039 0.058 -0.019
With no MS 0.058 0.019 0.039
-0.058
Senegal Difference
With MS 0.008 0.014 -0.006
With no MS 0.014 0.01 0.004
-0.01

Sierra Leone Difference
With MS 0.016 0.07 -0.054
With no MS 0.07 0.0299 0.0401
-0.0941

South Africa Difference
With MS 0.188 0.215 -0.027
With no MS 0.215 0.0324 0.1826
-0.2096

Swaziland Difference
With MS 0.334 0.386 -0.052
With no MS 0.386 0.07 0.316
-0.368
Tanzania Difference
With MS 0.065 0.088 -0.023
With no MS 0.088 0.094 -0.006
-0.017

Uganda Difference
With MS 0.041 0.15 -0.109
With no MS 0.15 0.075 0.075
-0.184

134
Table 22: Difference in the Differences – IV

Zambia Difference
With MS 0.165 0.169 -0.004
With no MS 0.169 0.283 -0.114
0.11

Zimbabwe Difference
With MS 0.201 0.337 -0.136
With no MS 0.337 0.2506 0.0864
-0.2224
Bangladesh Difference
With MS 0.0009 0.001 -0.0001
With no MS 0.001 0.0002 0.0008
-0.0009

Cambodia Difference
With MS 0.026 0.039 -0.013
With no MS 0.039 0.036 0.003
-0.016

Vietnam Difference
With MS 0.004 0.0024 0.0016
With no MS 0.0024 0.001 0.0014
0.0002

Brazil Difference
With MS 0.005 0.19 -0.185
With no MS 0.19 0.05 0.14
-0.325
Colombia Difference
With MS 0.006 0.007 -0.001
With no MS 0.007 0.0031 0.0039
-0.0049

DR Difference
With MS 0.011 0.017 -0.006
With no MS 0.017 0.028 -0.011
0.005

135

Table 23: Difference in the Differences - V

Kazakhstan Difference
With MS 0.001 0.002 -0.001
With no MS 0.002 0.0004 0.0016
-0.0026

Peru Difference
With MS 0.006 0.005 0.001
With no MS 0.005 0.0035 0.0015
-0.0005

Thailand Difference
With MS 0.015 0.037 -0.022
With no MS 0.037 0.024 0.013
-0.035
136
Discussion of Results

From the difference in difference equations it is evident that a multisectoral
program was effective in the reduction of HIV/AIDS rates in the majority of
countries where data was available. To check for significance, differences in
HIV/AIDS prevalence rates with and without the program were divided by the
number of years with and without the program respectively. For example, the
difference in Botswana’s HIV/AIDS rate between 2000 and 2005 is divided by six
for the number of years the multisectoral program was in place and the difference in
the rate between 1995 and 2000 is divided by five for the number of years before
multisectoral program implementation. A t-test was then executed to see if the two
rates were significantly different from each other: the time the multisectoral program
was in place and the time it was not. The t-test reveals a strong significant difference
in HIV/AIDS prevalence rates indicating that for these cases, the multisectoral
program was significantly effective in the decline of HIV/AIDS prevalence rates.
The results are present in the table below.

137
Table 24: Difference in Difference Results

Years with
program Years without program
Angola -0.000666667 -0.002
Benin -0.00025 -0.002442857
Botswana -0.0195 0.0556
Burundi -0.003 -0.006366667
Cameroon -0.000125 0.002666667
Congo 0.002 -0.000555556
Ghana 0.000166667 -0.0016
Guinea -0.00425 0.002371429
Lesotho -0.0285 0.022777778
Madagascar 0.0004 0.00025
Mali -0.001 -0.001222222
Mozambique -0.012666667 0.0025
Niger -0.00725 0.003785714
Nigeria -0.0038 0.0065
Senegal -0.0004 0.0008
Sierra Leone -0.0135 0.005728571
South Africa -0.0045 0.03652
Swaziland -0.0104 0.052666667
Tanzania -0.007666667 -0.00075
Uganda -0.007266667 0.015
Zambia -0.000571429 -0.0285
Zimbabwe -0.0272 0.041766667
Bangladesh -0.00001 0.0008
Cambodia -0.0026 0.0005
Vietnam 0.000177778 0.0007
Brazil -0.012333333 0.028
Colombia -0.0005 0.000433333
Dominican Republic -0.0012 -0.001833333
Kazakhstan -0.0002 0.000266667
Peru 0.00025 0.000214286
Thailand -0.001466667 0.0026

138
Table 25: t-Test: Two-Sample Assuming Equal Variances

t-Test: Two-
Sample Assuming
Equal Variances

with program without program
Mean -0.005413827 0.007650876
Variance 6.2318E-05 0.000321503
Observations 31 31
Pooled Variance 0.00019191
Hypothesized
Mean Difference 0
Df 60
t Stat -3.712924502
P(T<=t) one-tail 0.000225497
t Critical one-tail 1.670648544
P(T<=t) two-tail 0.000450993
t Critical two-tail 2.000297172

Conclusion

From regression results in Models 1 through 7 and the difference in
difference results, it seems evident that many political and societal determinants have
a significant effect on the HIV/AIDS rate. All of these determinants have been
reviewed above. Also reviewed above is the negative relationship between
HIV/AIDS and economic growth. Most important to this dissertation however, is the
significant finding that multisectoral HIV/AIDS programs seem to be an important
means of fighting HIV/AIDS.

139
Chapter 5: Concluding Remarks

Reflection on the Findings
The core premise of this dissertation has been that over the period 1995 to
2005 a country’s success in combating HIV/AIDS lies in the government’s ability to
implement a multisectoral HIV/AIDS program. A multisectoral program can be
effective because it includes representation from government agencies beyond just
the health ministry and this creates a broader base for successfully working with
NGOs, faith-based organizations, and local communities, in the fight against
HIV/AIDS. In this dissertation the relationship between the adoption of a
multisectoral HIV/AIDS program and a reduction in the prevalence of the disease
confirms that these programs have been effective in the majority of middle and low-
income countries included in this database. The importance of this research lies in
the urgent need to combat AIDS as a goal in itself, but also because actively
addressing the AIDS problem is key to sustaining economic growth in those
countries plagued by this epidemic.
The literature review conducted in this dissertation suggests that there are
negative economic effects from the disease, a main one being the fall in labor
productivity caused by illness and then death. Country simulations have also been
run and these too predict dramatic falls in future economic growth, labor
productivity, and savings due to the disease. The literature also pointed to several
societal determinants of HIV/AIDS, such as a lack of education and extreme gender
140
inequality. It was found in more than one study that higher education levels for
women lead to more knowledge about the disease. Similarly, in countries where
economic equality is higher for women, there is less exposure to risky circumstances
that may bring about the disease. The limited amount of literature on the political
determinants of HIV/AIDS revealed that an open and decentralized political
environment is conducive to an effective anti-AIDS campaign because policymakers
at sub-national levels can more effectively address the needs of vulnerable groups in
society. Finally, the meager literature on multisectoral HIV/AIDS programs was
reviewed, its main contribution being the provision of guidelines for program
implementation and in some cases suggestions for program improvement. For
example, it has been argued that successful multisectoral program implementation
requires flexibility as each country is unique, but a necessary condition for success is
the inclusion of the health sector.
Overall, my own research on those multisectoral HIV/AIDS programs in
place in the MAP countries reveals that the majority of developing countries have
experienced significant improvements in AIDS-awareness and prevention since
program inception. MAP funds have been used to establish a multisectoral approach
in many countries that had only health sector-based programs in the past and this
shift has produced vast improvements. Some of these advances include the
establishment of HIV-testing and treatment centers, a significant rise in condom
distribution, and an up-scaling of education on HIV/AIDS. My interviews and case
studies revealed that while no one blueprint will work in all countries, multisectoral
141
HIV/AIDS programs are an effective policy tool for combating the disease. For
example, multisectoral HIV/AIDS programs worked in Uganda because an
influential leader took up this cause by touring the country to promote awareness
about the disease from the national to community level. In Ethiopia, the
multisectoral approach allowed faith-based organizations to reach vulnerable groups
in society. In Thailand, public agencies and businesses worked together to establish
and monitor a safe sex-worker program. Though differences in all countries with
multisectoral programs are apparent, in their own way each of these programs proved
to be an effective policy tool for fighting the disease.
From research to date, this is the first study that provides a quantitative
analysis of multisectoral HIV/AIDS programs and evidence of their success. In this
dissertation, a variable for having a multisectoral HIV/AIDS program in place during
the years 1995-2005 was created for 89 middle and low-income countries in an effort
to provide quantitative evidence that these types of programs are effective in fighting
HIV/AIDS. Quantitative results, based on this 89-country sample, confirm that there
have been negative economic effects from the disease over time. Specifically, the
HIV/AIDS rate in 1995 had a negative influence on GDP per capita growth. From
the models that I ran on the political determinants of HIV/AIDS, the most significant
finding was that having a multisectoral program over the period (1995-2005)
negatively influences the 2005 HIV/AIDS rate.
Other significant findings revealed that societal determinants such as having a
neighbor country with a high HIV/AIDS rate has a positive influence on the disease
142
and economic variables such as income inequality also lead to higher HIV/AIDS
rates. When examining the change in the HIV/AIDS rate over time and government
expenditures on HIV/AIDS, the most significant finding, again, is the presence of a
multisectoral program over the ten year period. When examining females (ages 15-
49), the existence of a multisectoral HIV/AIDS program is the most significant,
positive determinant for the knowledge of condoms as a contraception method.
When examining those determinants of multisectoral program existence, having a
new political regime and being on the continent of Africa are the most significant
findings.
Difference in difference equations was executed to confirm the result that
multisectoral programs significantly matter for bringing down HIV/AIDS rates. The
difference in results indicates that a multisectoral program was effective in the
reduction of HIV/AIDS rates in the majority of countries where data was available.
The t-test for significance reveals a strong significant difference in HIV/AIDS
prevalence rates between the time periods in which the countries did and did not
have a multisectoral HIV/AIDS program. This result indicates that for these cases,
the multisectoral program was a significant contributor to the decline of HIV/AIDS
prevalence rates.
It has been shown that there are negative economic effects due to HIV/AIDS
and there are positive political determinants (such as a government-implemented
multisectoral program) that help to fight the disease. Most important to this
dissertation is the significant result that multisectoral HIV/AIDS programs are an
143
important policy choice for fighting HIV/AIDS. Ideally, these results can add to and
bolster the importance of policies that call for a multisectoral approach to fighting
HIV/AIDS. Multisectoral HIV/AIDS programs may be difficult to implement as
they are unique to each country. Nonetheless, evidence in this dissertation confirms
positive effects and offers strong endorsement for their implementation.

Future Research
Throughout this study, a major problem has been the lack of available data
regarding HIV/AIDS statistics, policies, and multisectoral programs, as well as
incomplete information on the political histories of the included countries. With new
and improved data, I hope to reconfirm the results obtained in the dissertation and
identify more significant determinants of the disease. I would also like to focus on
other developing regions to see if and how multisectoral programs work differently
in other areas. Some examples of what I would like to investigate: In Thailand’s
multisectoral program, there is an emphasis on contraception promotion; is this
policy tool specific to the country or is it common across the majority of Asian
countries? And if so, has this been more successful than programs in other regions
and why? Similarly, Ethiopia is a predominantly Christian (60.8% of the population)
country and its multisectoral program funds several faith-based organizations; should
this type of multisectoral program be applied in predominately Christian countries?
Again, complete country data is a problem when identifying similarities and
differences across regions.
144
I would also like to take a closer look at individual multisectoral programs
and identify what works best and how. This would involve a hands-on look at actual
multisectoral programs, including interviews with key program officials. Ideally,
different types of multisectoral programs, such as a more faith-based approach in
Ethiopia opposed to a more sex-worker focused approach in Thailand can be
identified. This provides a compelling comparison for studying those attributes that
work best and worst within different types of programs. Once identified, the weak
points in a given multisectoral HIV/AIDS program can be sufficiently improved.
Similarly, I would like to compare different types of implementation
strategies. In this dissertation, multisectoral HIV/AIDS programs include those
implemented by health ministers, presidents, and international organizations. I
would like to take a more in-depth look at each strategy and identify the differences
as well as the pros and cons of each. Again, this type of research would require more
data and interviews with program officials and other policymakers.

145
Bibliography
African Medical and Research Foundation. 2001. Inventory of Agencies with
HIV/AIDS Activities and HIV/AIDS Interventions in Uganda. New York:
Pfizer International Corporation.

AIDS Education Global Information System. 2008. Fact Sheets, San Juan
Capistrano, CA.

Ainsworth, Martha. 1998. Setting Government Priorities in Preventing HIV/AIDS
based on a World Bank Policy Report, Confronting AIDS: Public Priorities in
a Global Epidemic. New York: Oxford University Press.

Ainsworth, Martha and Waranya Teokul. 2000. Breaking the silence: setting realistic
priorities for AIDS control in less-developed countries. The Lancet 356 55-
60.

Allen, Tim and Suzette Heald. 2004. HIV/AIDS Policy in Africa: What has Worked
in Uganda and What has Failed in Botswana? Journal of International
Development 16 1141-1154.

Altman, Dennis. 1999. Globalization, Political Economy, and HIV/AIDS. Theory
and Society 28(4) 559-584.

Arndt, C. and J.D. Lewis. 2000. The Macro Implications of HIV/AIDS in South
Africa: A Preliminary Assessment. The South African Journal of Economics
68(5) 380-392.

“A Portrait in Red.” The Economist 13 Mar. 2008.

“A Testing Journey.” The Economist 13 Mar. 2008.

Avert Charity. 2003. HIV & AIDS Statistics for Africa, West Sussex, UK.

Ayvalikli, Didem, Brown, Jonathan C., and Nadeem Mohammad. 2004. Turning
Bureaucrats into Warriors: Preparing and Implementing Multi-Sector HIV-
AIDS programs in Africa. Washington D.C.: The International Bank for
Reconstruction and Development/the World Bank.

Baldani, Jeffrey, Bradfield, James, and Robert Turner. 1996. Mathematical
Economics. Orlando: The Dryden Press.

146

Barnett, Tony and Alan Whiteside. 2002. Aids in the Twenty-First Century. New
York: Palgrave Macmillan.

Barro, Robert J., and Jong-Wha Lee, 1993. International Comparisons of Educational
Attainment. NBER Working Paper #4349, April 1993.

Barro, Robert J. and Xavier Sala-i-Martin. 1999. Economic Growth. Cambridge: The
MIT Press.

Benatar, Solomon R. 2002. The HIV/AIDS Pandemic: A Sign of Instability in a
Complex Global System. Journal of Medicine and Philosophy 27(2) 163-177.

Bodiang, C.K. 2001. HIV/AIDS: The Multisectoral Approach a focus on Africa.
Prepared for the Swiss Agency for Development and Co-operation. Basel:
Swiss Centre for International Health Swiss Tropical Institute.

Bonnel, Rene. 2000. Economic Analysis of HIV/AIDS. AIDS Campaign Team for
Africa. Washington D.C.: The World Bank.

Boone, Catherine and Jake Bastell. 2001. Politics and AIDS in Africa: Research
agendas in political science and international relations. Africa Today 48(2) 2-
33.

Canning, David. 2006. The Economics of HIV/AIDS in Low-Income Countries: The
Case for Prevention. Journal of Economic Perspectives 20(3) 121-142.

Central Intelligence Agency. 2008. The WORLD Factbook, Washington D.C.

Collier, Paul. 2007. The Bottom Billion Why the Poorest Countries are Failing and
What can Be Done About it. New York: Oxford University Press.

Commonwealth Secretariat. Social Transformation Programmes Division. 2003.
Guidelines to Implementing a Multi-Sectoral Approach to HIV/AIDS in
Commonwealth Countries. London: Commonwealth Secretariat.

Demographic and Health Surveys. 2008. Knowledge of Contraceptive Methods,
Calverton, MD.

De Waal, Alex. 2006. Aids and Power: Why There is no Political Crisis – Yet. New
York: Zed Books Ltd.

Dickinson, Clare. 2006. The Politics of national HIV/AIDS responses: a synthesis of
literature. London: hlsp institute.
147

Dombo, M. 2004. Multisectoral and integrated strategies work best for mitigating the
impact of HIV/AIDS on children, households and communities: Lessons
from one of Uganda’s first multisectoral and integrated programs. Prepared
for the 15
th
International AIDS Conference, Bangkok, July 11-16, 2004.

Edwards, Jeffrey and Rashid B. Al-Hmoud. 2004. AIDS Mortality and Economic
Growth: A Cross-Country Analysis Using Income-Stratified Data.
Southwestern Journal of Economics 6.2.

Galaty, David, Gavian, Sarah, and Gilbert Kombe. 2006. Multisectoral HIV/AIDS
Approaches in Africa: How are they Evolving? Stuart Gillespie, ed., AIDS,
Poverty, and Hunger: Challenges and Responses. Washington D.C.:
International Food Policy Research Institute.

Gauri, Varun and Evan S. Lieberman. 2004. AIDS and the State: The Politics of
Government Responses to the Epidemic in Brazil and South Africa.
Downloaded from meetings of the American Political Science Association,
Chicago, September 2004.

Gauri, Varun and Evan S. Lieberman. 2006. Boundary Institutions and HIV/AIDS
Policy in Brazil and South Africa. Studies in Comparative International
Development 41(3) 47-73.

Glick, Peter J. and David E. Sahn. 2004. Determinants of HIV knowledge and
Behavior of Women in Madagascar: An Analysis Using Matched Household
and Community Data. Working Paper No. 168, Cornell Food and Nutrition
Policy Program.

Glick, Peter J. and David E. Sahn. 2007. Changes in HIV/AIDS Knowledge and
Testing Behavior in Africa: How Much and for Whom? Journal of
Population Economics 20(2) 383-422.

Glick, Peter J. and David E. Sahn. 2008. Are Africans Practicing Safer Sex?
Evidence from Demographic and Health Surveys for Eight Countries in J.
Strauss, ed., Economic Development and Cultural Change. Chicago: The
University of Chicago Press.

Gorgens-Albino, Marelize, Mohammad, Nadeem, Blankhart, David, and Oluwole
Odutolu. 2007. The Africa Multi-Country AIDS Program 2000-2006.
Washington D.C.: The International Bank for Reconstruction and
Development/The World Bank.

148
Gray, Clive and Malcolm McPherson. 2001. The Leadership Factor in African Policy
Reform and Growth. Economic Development and Cultural Change, 49(4)
707-740.

Gwatkin, D.R. and G. Deveshwar-Bahl. 2001. Inequalities in knowledge of
HIV/AIDS Prevention: An Overview of Socioeconomic and Gender
Differentials in Developing Countries. POPLINE Document Number:
284047.

Harman, Sophie. 2007. The World Bank: failing the Multi-Country AIDS Program.
Global Governance, 13(4) 485-492.

Hemrich, Gunter and Daphne Topouzis. 2000. Multi-Sector Responses to
HIV/AIDS: Constraints and Opportunities for Technical Co-operation.
Journal of International Development, 12 85-99.

HIV InSite. 2008. HIV/AIDS-Specific Country Profiles, San Francisco, CA.

Hollertz, V. 2001. Mae Chan Workshop on Integrated Community Mobilization
towards Effective Multisectoral HIV / AIDS Prevention and Care. Prepared
for the United Nations Development Programme [UNDP], South East Asia
HIV and Development Project, Bangkok, May 2001.

Hossain, Rajib. “AIDS: Epidemic Bell Ringing.” The Daily Star. 28, Oct. 2007.

Interagency Coalition on AIDS and Development. 2000. Political Commitment,
Governance and AIDS: A Discussion Paper. Ottawa: Interagency Coalition
on AIDS and Development.

Jones, Charles I. 1996. Human Capital, Ideas, and Economic Growth. Prepared for
the VIII Mondragone International Economic Seminar on Finance, research,
education and Growth, Rome, June 1996.

Jones, Charles I. 2002. Introduction to Economic Growth. New York: W.W. Norton
& Company, Inc.

Karim, S.S., Karim, Q.A., Soldan, K. and M. Zondi. 1995. Reducing the Risk of HIV
Infection among South African Sex Workers: Socioeconomic and Gender
Barriers. American Journal of Public Health 85 1521-1525.

Lieberman, Evan S. 2007. Ethnic Politics, Risk, and Policy-Making: A Cross-
National Statistical Analysis of Government Responses to HIV/AIDS.
Comparative Political Studies 40(12) 1407-1432.

149
Lule, Elizabeth. 2007. The Importance of a Multisectoral approach in addressing
HIV/AIDS. Prepared for the Africa Region HIV/AIDS Consultation on
Multisectoral Response, Rwanda, June 2007.

Mankiw, N., Romer, Davis, and David Weil. 1992. A contribution to the Empirics of
Economic Growth. Quarterly Journal of Economics 107 407-438.

Manning, Ryann. 2002. AIDS and Democracy: What do we Know? A Literature
Review. Prepared for AIDS and Democracy: Setting the Research Agenda,
Cape Town, April 22-23, 2003.

Mattes, Robert. 2003. Healthy Democracies? The Potential Impact of AIDS on
democracy in Southern Africa. Occasional Paper 71, University of Cape
Town, April, 2003.

McDonald, Scott and Jennifer Roberts. 2002. Growth and multiple forms of human
capital in an augmented Solow model: a panel data investigation. Economics
Letters 74(2) 271-276.

McDonald, Scott and Jennifer Roberts. 2006. AIDS and Economic Growth: A
Human Capital Approach. Journal of Development Economics 80(1) 228-
250.

Mwabu, Germano. 2008. Health Economics for Low-Income Countries in T. Paul
Schultz and John Strauss, eds., Handbook of Development Economics
Volume 4. Oxford: Elsevier.

National AIDS Commission. 2003. HIV/AIDS in Malawi: Estimates of the
prevalence of infection and the Implications. Silver Spring: Global Health
and Development Strategies Division, Social & Scientific Systems, Inc.

North, Douglass C. 1990. Institutions, Institutional Change and Economic
Performance. Cambridge: Cambridge University Press.

Nugent, Jeffrey B. 2002. Reexamination of Development Policies: Some Political
Economy Lessons of Success. Prepared for the Reexamination of
Development Policies: Market Institutions and Government Workshop, Jan.
29, 2002.

Oomman, N., Bernstein, M., and Steven Rosenzweig. 2007. Following the Funding
for HIV/AIDS: A Comparative Analysis of the Funding Practices of
PEPFAR, the Global Fund and World Bank MAP in Mozambique, Uganda
and Zambia. The Center for Global Development, Oct. 10, 2007.

150
Opuni et al. 2002. Current and Future Resources for HIV/AIDS. State of the Art:
AIDS and Economics. IAEN Policy Document.

Organization for Economic Cooperation and Development. 2008. Statistics Portal.
Paris, France.

Over, Mead. 1992. The Macroeconomic Impact of HIV/AIDS in Sub-Saharan Africa.
Population and Human Resources Department. Washington D.C.: The World
Bank.

Over, Mead. 1998. Coping with the Impact of AIDS based on a World Bank Policy
Report, Confronting AIDS: Public Priorities in a Global Epidemic. New
York: Oxford University Press.

Over, Mead. 1998. The Effects of Societal Variables on Urban Rates of HIV
Infection in Developing Countries in M. Ainsworth, Fransen, L., and M.
Over, eds., Confronting AIDS: Evidence from the Developing World.
Brussels: European Commission.

Over, Mead, Peter Heywood, Julian Gold, Indrani Gupta, Subhash Hira, and Elliot
Marseille. 2004. HIV/AIDS Treatment and Prevention in India Modeling the
Cost and Consequences. Human Development Network Health, Nutrition,
and Population Series. Washington, D.C: The International Bank for
Reconstruction and Development / The World Bank.

Patterson, Amy S. 2006. The Politics of Aids in Africa. Boulder: Lyne Reinner
Publishers, Inc.

Poku, Nana K and Alan Whiteside, eds. 2004. The Political Economy of AIDS in
Africa. Burlington: Ashgate Publishing Company.

Political Base. 2008. Legalization of Prostitution. Sausalito, California.

Prins, Gwyn. 2004. The Political Economies of AIDS. Prepared for the Centre for the
Study of AIDS, February 2, 2004.

Putzel, James. 2003. Institutionalizing an Emergency Response: HIV/AIDS and
Governance in Uganda and Senegal. A report submitted to the Department
for International Development. London, May, 2003.

Putzel, James. 2004. Governance and AIDS in Africa: Assessing the International
Community’s “Multisectoral Approach.” Prepared for the 2004 Annual
Meeting of the American Political Science Association, September 2 -
September 5, 2004.
151

Putzel, James. 2004. The Global Fight Against AIDS: How Adequate are the
National Commission? Journal of International Development 16 1129-1140.

Rodrik, Dani. 2003. In Search of Prosperity: Analytic Narratives on Economic
Growth. Princeton: Princeton University Press.

Rodrik, Dani. 2007. One Economics Many Recipes: Globalization, Institutions, and
Economic Growth. Princeton: Princeton University Press.

Sachs, Jeffrey D. 2005. The End of Poverty: Economic Possibilities for Our Time.
New York: Penguin Group.

“Sacking the wrong health minister.” The Economist 16 Aug. 2007.

Singer, Merrill. 1998. The Political Economy of Aids. Amityville: Baywood
Publishing Company, Inc.

Talbott, John R. 2007. Size Matters: The Number of Prostitutes and the Global
HIV/AIDS Pandemic. Africans Against AIDS, Inc. New York, New York.

Thurman, Sandra L. 2001. Joining Forces to Fight HIV and AIDS. The Washington
Quarterly 24 191-196.

Ukpolo, V. 2004. AIDS Epidemic and Economic Growth: Testing for Causality.
Journal of Asian and African Studies 39(3) 169 – 178.

UNAIDS: Joint United Nations Programme on HIV/AIDS. 1998, 2003, 2005, 2008.
Country Responses, Geneva, Switzerland.

UNDP: United Nations Development Programme. 2008. UNDP Country Offices and
other Programmes, New York, New York.

UNESCO: United Nations Educational, Scientific, and Cultural Organization. 2008.
Gender Parity Index, Quebec, Canada

UN: United Nations. 2001. General Assembly Special Session on HIV/AIDS, New
York, New York.

USAID: United States Agency for International Development. 2008. Country
Profiles, Washington D.C.

152
USAID: United States Agency for International Development. 2004. Annex 1.
USAID’s Experience Assessing and Mitigating Multisectoral Development
Impacts of HIV/AIDS, Washington D.C.

U.S. Department of State. 2008. Country Pages, Washington D.C.

Whiteside, Alan, Hickey, Alison, Ngcobo, Nkosinathi, and Jane Tomlinson. 2003.
What is driving the HIV/AIDS epidemic in Swaziland, and what more can we
do about it? Prepared for the NERCHA and UNAIDS, Mbabne and Durban,
April, 2003.

World Bank. 2008. Countries & Regions, Washington D.C.

World Bank. 2007. World Development Indicators, Washington D.C.

World Health Organization. 2008. Statistical Information System, Geneva,
Switzerland.

World Resources Institute. 2008. Earth Trends: Total Male Population, Washington
D.C.

Abstract (if available)

Abstract The proposed dissertation offers an explanation for the large differences in HIV/AIDS rates among 89 low and middle-income countries throughout the Sub Saharan African, Asian, and Latin American regions over a ten-year period (1995-2005). The HIV/AIDS rates in these countries vary widely and seemingly independently of economic wealth. One possible determinant of these differences is the presence and degree of development of strong multisectoral programs aimed at both prevention and cure of HIV/AIDS. The main hypothesis for this dissertation is: "A country's success in combating HIV/AIDS lies in the government's ability to implement an effective multisectoral program." This hypothesis is explored through quantitative models using data from the ten-year period (1995-2005). Results show that the presence of a multisectoral program over the ten-year period is associated with a significantly lower HIV/AIDS incidence rate by 2005. This effect is produced by controlling for various political, economic, societal, and institutional factors. Although there is some anecdotal evidence which suggests that multisectoral programs help to improve the HIV/AIDS problem in developing countries, there has been little if any empirical work done on this subject to date.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The role of local actors in international development: how incentive types affect foreign aid effectiveness

PDF

The political economy of Brazilian trade policy: domestic determinants, world and regional strategies

PDF

The political economy of R&D collaboration: micro- and macro-level implications

PDF

Essays on political economy of privatization

PDF

Strategic communication in the fight against HIV/AIDS in the Russian Federation

PDF

Inter-temporal allocation of human capital and economic performance

PDF

Essays on the economics of infectious diseases

PDF

Through targeted public relations awareness campaigns, community based organizations can raise the level of HIV/AIDS awareness among black women, thereby reducing the rate of HIV/AIDS infections ...

PDF

Essays in political economy and mechanism design

PDF

Financial crises and trade policy in developing countries

PDF

Political economy of private commercial arbitration in Colombia

PDF

Economic policy-making in a changing environment: the politics of success in Peru vis-à-vis China

PDF

The political economy of oil, terrorism and institutional development in Iran and its impact on the Middle East

PDF

An empirical analysis of the quality of primary education across countries and over time

PDF

The murky risk of trade protectionism in an interconnected and uncertain global economy: a state, industrial, and regional analysis

PDF

Regulatory burden: a systems analysis of weapons acquisition in the U.S. Department of Defense

PDF

An empirical characterisation of the Bombay Stock Exchange

PDF

The disease burden among aging people living with HIV/AIDS

PDF

Stage of readiness and learning styles -- a tailored HIV/AIDS prevention intervention for youth detained in the Los Angeles County Juvenile Justice system

PDF

The impact of trade liberlaization on firm performance in developing countries -- new evidence from Pakistani manufacturing sector

Asset Metadata

Creator Davis, Dollie (author)

Core Title Political determinants and economic effects of HIV/AIDS: a push for the multisectoral approach

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Political Economy

Publication Date 10/30/2008

Defense Date 07/15/2008

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Economic development,HIV/AIDS,multisectoral,OAI-PMH Harvest

Place Name Africa (continents), Asia (continents), Carribbean (region), South America (continents)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Wise, Carol (committee chair), Chi, Iris (committee member), Nugent, Jeffrey B. (committee member)

Creator Email dollieda@usc.edu,dolliesdavis@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m1724

Unique identifier UC1433759

Identifier etd-Davis-2422 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-121356 (legacy record id),usctheses-m1724 (legacy record id)

Legacy Identifier etd-Davis-2422.pdf

Dmrecord 121356

Document Type Dissertation

Rights Davis, Dollie

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu