Global Health: The Global Fund to Fight AIDS, TB and Malaria Is  
Responding to Challenges but Needs Better Information and	 
Documentation for Performance-Based Funding (10-JUN-05, 	 
GAO-05-639).							 
                                                                 
The Global Fund to Fight AIDS, Tuberculosis and 		 
Malaria--established as a private foundation in January 2002--is 
intended to rapidly disburse grants to recipients, including	 
governments and nongovernmental organizations. The Global Fund	 
has signed over 270 grant agreements and disbursed more than $1  
billion. Governments provide most of its funding; the United	 
States has provided almost one-third of the $3.7 billion the	 
Global Fund has received. In May 2003, the President signed	 
legislation directing the Comptroller General to monitor and	 
evaluate Global Fund-supported projects. GAO reviewed grants that
the Global Fund began disbursing before January 2004. This report
(1) describes the Global Fund's process for managing grants and  
disbursing funds, (2) identifies factors that have affected grant
performance, (3) reviews the basis and documentation of 	 
performance-based funding, and (4) notes recent refinements of	 
Global Fund processes.						 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-05-639 					        
    ACCNO:   A26333						        
  TITLE:     Global Health: The Global Fund to Fight AIDS, TB and     
Malaria Is Responding to Challenges but Needs Better Information 
and Documentation for Performance-Based Funding 		 
     DATE:   06/10/2005 
  SUBJECT:   Acquired immunodeficiency syndrome 		 
	     Funds management					 
	     Grant administration				 
	     Grant monitoring					 
	     Grants						 
	     Infectious diseases				 
	     Internal controls					 
	     International organizations			 
	     Policy evaluation					 
	     Tuberculosis					 
	     Malaria						 
	     Global Fund to Fight AIDS, Tuberculosis,		 
	     and Malaria					 
                                                                 
	     Malaria						 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-05-639

                 United States Government Accountability Office

                     GAO Report to Congressional Committees

June 2005

GLOBAL HEALTH

 The Global Fund to Fight AIDS, TB and Malaria Is Responding to Challenges but
    Needs Better Information and Documentation for Performance-Based Funding

                                       a

GAO-05-639

[IMG]

June 2005

GLOBAL HEALTH

The Global Fund to Fight AIDS, TB and Malaria Is Responding to Challenges but
Needs Better Information and Documentation for Performance-Based Funding

  What GAO Found

Global Fund policy is to manage grants in a transparent and accountable
manner, disbursing funds to recipients based on their demonstrated
performance as measured against agreed-on targets. In implementing this
performance-based funding system, Global Fund officials are to
periodically assess whether the grant's principal recipient has made
sufficient progress to warrant its next disbursement. After 2 years, the
Global Fund is to determine whether to continue funding the grant for an
additional 3 years. In making an assessment, officials consider several
information sources, including the recipient's reports on its performance
and expenditures and an independent agent's verification of the
recipient's reports.

Recipient countries' capacity to implement grants has been an underlying
factor in grant performance, according to Global Fund and other
knowledgeable officials. These officials, as well as principal recipients,
also cited guidance, coordination, planning, and contracting and
procurement as factors associated with challenges or successes in grant
performance. For example, recipients in three countries reported that they
could not meet their targets because they had not received national
treatment guidelines. However, several grant recipients reported that,
under certain circumstances, Global Fund guidance allowed them to quickly
redirect funds, thereby enabling them to meet their targets.

GAO found problems associated with the information sources that the Global
Fund uses in making performance-based funding decisions. For example, the
limited monitoring and evaluation capabilities of many recipients raise
questions about the accuracy of their reporting. Moreover, the Global Fund
has not consistently documented its determinations that recipients'
performance warranted additional funding. For instance, the Global Fund's
documentation did not explain its decisions to disburse funds to some
recipients who reported that they had met few targets. Further, the Global
Fund does not track or publicly document denied disbursement requests.

The Global Fund is taking steps to address challenges to grant performance
and improve the overall management of grants, including

o  reorganizing and strengthening its staff;

o  	developing a risk assessment mechanism and early warning system to
identify poorly performing grants;

o  streamlining reporting and funding procedures;

o  working with partners to strengthen recipient capacity; and

o  	clarifying certain guidance for the country coordinating mechanism-the
entity in each country responsible for developing grant proposals,
nominating grant recipients, monitoring grant implementation, and advising
the Global Fund on the viability of grants for continued funding. However,
the Global Fund has not clearly defined the role of these entities in
overseeing grant implementation.

United States Government Accountability Office

Contents

Letter                                                                   1 
                                  Results in Brief                          3 
                                     Background                             5 
              Grant Management Process Has Many Participants and Occurs in 
                                     Two Phases                             8 
                     Several Factors Affected Grant Performance            18 
              Sources Informing Funding Decisions Have Limitations, and    
                        Decisions Are Not Clearly Documented               24 
            Global Fund Is Taking Several Steps to Refine Grant Management 
                                   and Performance                         29 
                                     Conclusions                           38 
                        Recommendations for Executive Action               39 
                         Agency Comments and Our Evaluation                39 

Appendixes

                          Appendix I: Appendix II: Appendix III: Appendix IV:

Appendix V:

Objectives, Scope, and Methodology 41

Methodologies Used to Compare Grants' Disbursements 46

Comments from the Global Fund 53

Joint Comments from the Departments of State and Health and Human
Services, and the U.S. Agency for International Development 58

GAO Contact and Staff Acknowledgments 61

Tables Table 1:

Table 2: Table 3:

Table 4:

Table 5:

Global Fund Grants by Disease, Region, and Recipient
Type 6
Selected Characteristics of 130 Global Fund Grants 46
Information Related to Disbursements for Different Types
of Grants, Defined by Principal Recipient Type, Disease
Type, and Grant Size 47
Regression Models Showing Effects of Various Factors on
Percentage of Grants Disbursed, Minus the Amount
Disbursed in First Disbursement 49
Percentages of Timely and Untimely Disbursements, by
Disbursement Stage 52

Figures	Figure 1: Global Fund Grant Management Process 9 Figure 2: Global
Fund Grant Life Cycle 14 Figure 3: Global Fund Grant Rating System for
Phase 2 16

Contents

                      Figure 4: Global Fund Milestones 18

AbbreviationsEUR

CCM country coordinating mechanism
HHS U.S. Department of Health and Human Services
HHS/CDC U.S. Centers for Disease Control and Prevention
HIV/AIDS human immunodeficiency virus/acquired immune deficiency

syndrome NGO nongovernmental organization PWC PricewaterhouseCoopers TB
tuberculosis UNAIDS Joint UN Programme on HIV/AIDS UNDP UN Development
Programme UNICEF United Nations Children's Fund UNOPS UN Office for
Project Services USAID U.S. Agency for International Development WHO World
Health Organization

This is a work of the U.S. government and is not subject to copyright
protection in the United States. It may be reproduced and distributed in
its entirety without further permission from GAO. However, because this
work may contain copyrighted images or other material, permission from the
copyright holder may be necessary if you wish to reproduce this material
separately.

A

United States Government Accountability Office Washington, D.C. 20548

June 10, 2005

The Honorable Richard G. Lugar
Chairman
The Honorable Joseph R. Biden, Jr.
Ranking Minority Member
Committee on Foreign Relations
United States Senate

The Honorable Henry J. Hyde
Chairman
The Honorable Tom Lantos
Ranking Minority Member
Committee on International Relations
House of Representatives

In January 2002, the Global Fund to Fight AIDS, Tuberculosis and Malaria
(the Global Fund) was established as a private foundation to increase
spending for the prevention and treatment of the three diseases. The
Global
Fund is intended to rapidly disburse grants to recipients such as
governments and nongovernmental organizations, based on recipients'
performance against agreed-on targets; the Global Fund's policy is to make
such disbursements in a transparent and accountable manner. The Global
Fund approved its first grants in April 2002; as of April 15, 2005, it had
signed 271 grant agreements with recipients in 123 countries and
territories
and had disbursed more than $1 billion. The Global Fund is a
public-private
partnership, governed by a board of directors representing governments,
nongovernmental organizations, private companies, and persons living
with the three diseases. The Global Fund receives most of its funding from
national governments; the United States has provided almost one-third of
the $3.7 billion that it has received, and the President's 2006 budget
requests an additional $300 million.

In May 2003, we reported that the Global Fund had made progress in
establishing governance structures and had developed comprehensive
systems for monitoring grants but that the monitoring systems faced
challenges.1 Also in May 2003, the President signed legislation directing
the
Comptroller General to monitor and evaluate projects supported by the

1GAO, Global Health: Global Fund to Fight AIDS, TB and Malaria Has
Advanced in Key Areas, but Difficult Challenges Remain, GAO-03-601
(Washington, D.C.: May 7, 2003).

Global Fund.2 Since then, the Global Fund has continued to refine its
processes. This report reflects our review of grants that the Global Fund
began disbursing before the beginning of 2004-that is, grants that have
had at least 1 year in which to demonstrate performance. In this report,
we (1) describe the Global Fund's process for managing grants and
disbursing funds; (2) identify factors that have affected grant
performance; (3) review the basis for, and documentation of, the Global
Fund's performance-based funding; and (4) describe the Global Fund's
recent refinements for managing grants and improving their performance.

To describe the Global Fund's process for managing grants and disbursing
funds, we reviewed Global Fund documents and interviewed officials from
the Global Fund headquarters in Geneva, Switzerland. To identify factors
affecting grant performance, we reviewed principal recipients' progress
reports for 38 grants that had a first disbursement on or before December
31, 2003-to allow for at least 1 year of performance-and met additional
criteria, and we categorized reasons given for deviation from performance
targets. Also, to determine whether certain grant characteristics were
factors associated with the percentage of funds disbursed for each grant,
we analyzed disbursement data for 130 grants that began disbursing funds
by the end of 2003. Further, we reviewed documents obtained from field
visits to four countries-Indonesia, Kenya, Thailand, and Zambia3-and
interviewed a wide variety of government, civil society, and bilateral and
multilateral development officials in those countries who were involved in
grant implementation or oversight. In addition, we interviewed officials
from the Global Fund and multilateral technical assistance agencies. To
review the basis for, and documentation of, the Global Fund's
performance-based funding, we examined Global Fund documents-including its
policy manual, recipients' progress reports, and documents assessing the
eligibility of 25 of 28 grants considered by the Global Fund as of the end
of March 2005 for continued funding after an initial 2 years-and spoke
with Global Fund staff. To describe the Global Fund's recent refinements
for managing grants and improving their performance, we reviewed
additional Global Fund documents and conducted follow-up interviews with
Global Fund and other officials. We

2U.S. Leadership Against HIV/AIDS, Tuberculosis, and Malaria Act of 2003,
(P.L. 108-25), section 202(f).

3These four countries received more than $36 million in committed funds
for several grants that covered more than one disease. Three of these
countries, Kenya, Thailand, and Zambia, also have grants that cover both
government and civil society recipients.

conducted our work from June 2004 through March 2005 in accordance with
generally accepted government auditing standards. (See app. I for further
details of our scope and methodology.)

Results in Brief 	The Global Fund's grant management process takes place
in two phases. Several entities participate in the management of each
grant: the country coordinating mechanism (CCM), representing
country-level stakeholders; the grant's principal recipient; the Global
Fund secretariat; the grant manager; the Global Fund board of directors;
and the local fund agent, the Global Fund's representative in each
recipient country, responsible for verifying grant recipients' financial
and program reporting. The Global Fund board generally approves 5-year
grants but approves a budget ceiling for only 2 years; the secretariat
then signs an initial contract with the recipient for 2 years-phase 1 of
the grant-during which it is to disburse funds based on regular reviews of
the recipient's performance. Toward the end of the 2-year period, the
Global Fund decides whether to renew the grant for up to an additional 3
years, known as phase 2. The Global Fund bases its disbursement and
renewal decisions on four sources of information: (1) periodic progress
reports submitted by the recipient; (2) recipient expenditure data; (3)
the local fund agent's verification of the recipient's progress and
spending; and (4) contextual information, including information about
mitigating circumstances that may affect the recipient's performance.

Several factors have influenced grant performance. According to Global
Fund officials and other knowledgeable entities, recipient countries'
capacity to implement grants has been an underlying factor. In addition,
principal recipients as well as Global Fund and development partner
officials frequently cited four factors-guidance, coordination, planning,
and contracting and procurement-associated with challenges or successes in
grant performance. For example, recipients in three countries reported
that they could not meet their targets because they had not received
approved national guidelines for treating the diseases. However, several
grant recipients reported that, under certain circumstances, Global Fund
guidance allowed them to quickly redirect funds, thereby enabling them to
meet their targets. Our analysis to determine whether certain grant
characteristics were factors associated with the percentage of a grant's
funds disbursed found no significant association between the type of grant
recipient, grant size, or disease targeted and the percentage disbursed.

We found problems with the sources of information that the Global Fund
uses to make periodic disbursement and grant renewal decisions and with
its documentation of those decisions. Regarding the information sources,
we noted that recipient progress reports vary in quality; their
expenditure data are often incomplete; local fund agents' assessments are
inconsistent, reflecting varying levels of expertise; and contextual
information for phase 1 decisions is not systematically collected.
Further, although the Global Fund's policy is to disburse funds based on
grant performance, the secretariat did not document how it determined that
grants' performance warranted additional disbursements for the 38 grants
that we reviewed. For example, the documentation did not show why the
secretariat disbursed funds to some recipients who reported that they had
met few of their targets. Moreover, the secretariat does not track denied
disbursement requests. In addition, for about one-third of the grants that
we reviewed, Global Fund staff did not document the reasons for their
recommendations regarding continued funding after an initial 2-year grant
period. Stakeholders raised additional concerns regarding the grant
renewal process; for example, a representative of a local fund agent
stated that this process may occur too early in the life of a grant to
adequately reflect the grant's progress.

The Global Fund is taking a number of steps to address challenges to grant
performance and improve the overall management of grants, including

o 	reorganizing and strengthening its staff by increasing the number of
grant managers, adding a new position focused on documenting disbursement
decisions and other grant milestones, creating a new unit devoted to
problem grants, and taking measures to improve the monitoring and
evaluation of grants and the reliability of the performance data that
recipients report;

o 	developing a risk assessment mechanism and early warning system to
identify poorly performing grants and more systematically alert grant
managers when they need to intervene;

o 	streamlining reporting and funding procedures by allowing most
recipients to report semiannually instead of quarterly and considering
new, more efficient funding mechanisms than the current round-based
approach;

o 	working with partners to strengthen recipient capacity, for example, by
collaborating with the World Health Organization's (WHO) Roll Back

Malaria Department in 2004 to reprogram existing grants to incorporate
new, more effective malaria treatments; and

o 	clarifying some guidance for the CCM, the body in each country
responsible for developing grant proposals, nominating grant recipients,
monitoring grant implementation, and advising the Global Fund on the
viability of grants for continued funding after 2 years. However, the
Global Fund has not clearly defined the CCMs' role in overseeing grant
implementation.

To improve the quality of the information on which the Global Fund bases
its funding decisions, and the secretariat's documentation of these
decisions, we are recommending that the U.S. Global AIDS Coordinator work
with the Global Fund's Board Chair and Executive Director to (1) complete
efforts to ensure that local fund agents have the necessary expertise to
evaluate recipients' grant performance data, (2) continue to work with
development partners to strengthen the quality and consistency of that
data by enhancing recipients' capacity for monitoring and evaluating their
financial and program-related activities, and (3) continue efforts to
clearly document the Global Fund's reasons for periodically disbursing
funds and renewing grant agreements.

In providing written comments on a draft of this report, the Departments
of State (State) and Health and Human Services (HHS), the U.S. Agency for
International Development (USAID), and the Global Fund's Executive
Director largely concurred with our conclusions. The Global Fund's
Executive Director agreed with our recommendations and the U.S. government
agencies did not comment on them in their joint letter. (See apps. III and
IV for a reprint of their comments.)

Background 	Human immunodeficiency virus/acquired immune deficiency
syndrome (HIV/AIDS), tuberculosis (TB), and malaria, are devastating
millions of individuals and families, thousands of communities, and dozens
of nations around the world according to the UN's WHO.

o 	HIV/AIDS, the retrovirus that causes AIDS, is usually transmitted (1)
sexually; (2) from mothers to children before or at birth or through
breastfeeding; or (3) through contact with contaminated blood, such as
through the use of contaminated hypodermic needles. In 2004, it led to
between 2.8 and 3.5 million deaths, most of them in sub-Saharan Africa,
which is home to more than 60 percent of people living with the virus.

The number of people infected with HIV has risen in every region of the
world, with the steepest increases occurring in East Asia, Eastern Europe,
and Central Asia. In China, HIV/AIDS is now found in all 31 provinces,
autonomous regions, and municipalities; and in India, as of 2003, 2.5 to
8.5 million people had been infected. In Eastern Europe and Central Asia,
the number of HIV-positive people has risen ninefold in less than 10
years.

o 	TB, a bacterial infection transmitted by inhalation of airborne
organisms, ranks just behind HIV/AIDS as the leading infectious cause of
adult mortality, each year killing up to 2 million people, mostly between
the ages of 15 and 54 years. It is the most common killer of people whose
immune systems are compromised by HIV.

o 	Malaria, caused by a parasite, is transmitted in human populations
through the bite of infected mosquitoes. The disease kills more than one
million people per year, mostly young African children.

The Global Fund, established as a private foundation in Switzerland in
2002, was created as a partnership between governments, civil society, the
private sector and affected communities to increase resources to fight the
three diseases. As shown in table 1, 45 percent of the Global Fund's 271
grants, as of April 15, 2005, were focused on HIV/AIDS; 45 percent went to
recipients in sub-Saharan Africa; and 59 percent went to government
recipients.

Table 1: Global Fund Grants by Disease, Region, and Recipient Type

Number of Percentage of Categories grants totala

                                 Disease focus

                             HIV/AIDS                  122                 45 
                              Malaria                   71                 26 
                         Tuberculosis                   67                 25 
                           Integrated                   11                  4 

Region

                                Sub-Saharan Africa      121                45 
                         East Asia and the Pacific      38                 14 
                   Latin America and the Caribbean      36                 13 
                   Eastern Europe and Central Asia      31                 11 

(Continued From Previous Page)

                                                  Number of     Percentage of 
                                   Categories        grants            totala 
                                   South Asia            24 
             North Africa and the Middle East            21 

                               Type of recipient

Government: ministry of health 97

Multilateral organization: UN Development
Programme 51

Civil society: nongovernmental organization 45

Government: otherb 42

Government: ministry of finance 21

Civil society: private sector 8

Civil society: faith-based organization 4

Multilateral organization: otherc 3

Total: government 160

Total: civil society 57

Total: multilateral organizations 54

                                 Total 271 100

Source: GAO analysis of Global Fund data.

aTotal may not sum to 100 because of rounding.

bIncludes national AIDS commissions and multicountry government bodies
such as the Secretariat of the Pacific Community, which encompasses a
number of islands in the Pacific.

cIncludes one other UN organization other than UN Development Programme
and two other multilateral organizations.

In March 2005, the Global Fund reported that across all grants, it had

o  provided antiretroviral treatment to 130,000 people with AIDS;

o  tested more than one million people voluntarily for HIV;

o 	supported 385,000 TB patients with directly observed short-course
therapy;4

o 	given more than 300,000 people new, more effective treatments for
malaria; and

4TB patients are typically prescribed a regimen of antibiotics for 4 to 6
months. If patients fail to take their pills regularly, they may develop
resistant strains of the disease, which are much more difficult to treat
and require additional, expensive drugs.

o 	supplied more than 1.35 million families with insecticide-treated
mosquito nets.

The Global Fund's key principles are to (1) operate as a financial
instrument, not an implementing entity; (2) make available and leverage
additional resources; (3) support programs that evolve from national plans
and priorities; (4) operate in a balanced manner with respect to
geographic regions, diseases, and health-care interventions; (5) pursue an
integrated and balanced approach to prevention, treatment, care and
support; (6) evaluate proposals through an independent review process; and
(7) operate in a transparent and accountable manner and employ a
simplified, rapid, and innovative grant-making process.

Grant Management Process Has Many Participants and Occurs in Two Phases

Numerous entities participate in the Global Fund's processes for managing
grants. The Global Fund manages its grants in two phases, generally over a
5-year period. During phase 1, the Global Fund signs a 2-year grant
agreement with the principal recipient and periodically reviews
recipients' performance to determine whether to disburse additional funds.
Near the end of phase 1, the board reviews the grant's progress to
determine whether to renew the grant for an additional 3 years; if the
board approves continued funding, the grant enters phase 2. The Global
Fund board approved the first round of grants in April 2002 and approved
33 grants to enter phase 2 as of April 25, 2005.

Numerous Entities Are The following entities participate in the Global
Fund's grant management Involved in Managing process (see fig. 1).

Grants

Figure 1: Global Fund Grant Management Process

Source: GAO, based on information from the Global Fund.

o A country coordinating mechanism (CCM) representing country-level
stakeholders submits grant proposals to the Global Fund and nominates a
principal recipient to be responsible for implementing the grant.
According to the Global Fund, the CCM should be made up of high-level host
government representatives, representatives of nongovernmental
organizations (NGO), multilateral and bilateral donors, the private
sector, and individuals living with HIV/AIDS, TB, or malaria. CCMs are to
develop and forward grant proposals to the Global Fund, monitor grant
implementation, and advise the Global Fund on the viability of grants for
continued funding after 2 years.

o 	The principal recipient is a local entity nominated by the CCM that
signs an agreement with the Global Fund to implement a grant in a
recipient country. There may be multiple public and private principal

recipients for a single grant. The principal recipient is responsible for
overseeing the activities of any subrecipients implementing grant
activities and for distributing grant money to them.

o 	The secretariat is responsible for the Global Fund's day-to-day
operations, including managing the grant proposal process; overseeing and
managing grant implementation to ensure financial and programmatic
accountability; and acting as a liaison between grant recipients and
bilateral, multilateral, and nongovernmental partners to ensure that
activities at the country level receive necessary technical assistance and
are well coordinated. As of April 15, 2005, the secretariat had 165 staff.
Within the secretariat, the fund portfolio manager, or grant manager, is
responsible for reviewing grant progress and deciding whether to disburse
additional funds to the principal recipient.

o 	The secretariat reports to the Global Fund's board of directors. The
23-member board is responsible for overall governance of the Global Fund
and approval of grants. The board includes 19 voting representatives of
donor and recipient governments, NGOs, the private sector (including
businesses and foundations), and affected communities. Key international
development partners, including WHO, the Joint UN Programme on HIV/AIDS
(UNAIDS), and the World Bank, participate as nonvoting members. The World
Bank also serves as the Global Fund's trustee.

o 	The local fund agent is the Global Fund's representative in each
recipient country and is responsible for financial and program oversight
of grant recipients. This oversight role includes an assessment of
recipients prior to their receiving money from the Global Fund. To date,
the Global Fund has contracted with the following entities to serve as
local fund agents: four private firms, KPMG, PricewaterhouseCoopers (PWC),
Chemonics International, Inc., and Deloitte Emerging Markets; one private
foundation that was formerly a public corporation, Crown Agents; the Swiss
Tropical Institute; and two multilateral entities, the World Bank and the
UN Office for Project Services (UNOPS). PWC and KPMG serve as the local
fund agents in 91 of the 110 countries for which the Global Fund has
contracted local fund agents.

Global Fund Implements After the board approves a proposal submitted by a
CCM and vetted by an

Grants in Two Phases	independent, multinational technical review panel,
typically for a 5-year grant, the secretariat signs a 2-year grant
agreement with the principal

recipient. This initial 2-year period represents phase 1 of the grant; if
the board approves continued funding, the grant enters phase 2.

Phase 1 	For grants approved in 2002 and 2003, the local fund agent
conducted, or contracted with other entities to conduct, assessments of
the recipient's capacity to (1) manage, evaluate, and report on program
activities; (2) manage and account for funds, including disbursing to
subrecipients; and (3) procure goods and services and maintain a reliable
supply of drugs and other commodities. The local fund agent initially
conducted these assessments after the signing of the grant agreement but
now conducts them before the Global Fund signs an agreement with a
principal recipient. After the local fund agent determines that the
results of its assessments are satisfactory, the Global Fund instructs the
World Bank to disburse the first tranche of funds to the principal
recipient.

According to its policy, the Global Fund disburses subsequent tranches
based on performance to ensure that investments are made where impact in
alleviating the burden of HIV/AIDS, TB, and malaria can be achieved.
During the grant period, the portfolio managers are to link disbursements
to periodic demonstrations of program progress and financial
accountability. The grant agreements initially specified that principal
recipients would report their progress and request additional
disbursements on a quarterly basis. In July 2004, the secretariat changed
the default reporting/disbursement request cycle to every 6 months.5 As of
April 15, 2005, about 20 percent of the Global Fund's grants were on a
6-month schedule. According to the secretariat, some grant recipients may
choose to remain on a quarterly schedule or the secretariat may decide,
based on a grant's risk profile, to disburse only one quarter at a time.

According to secretariat officials, grant managers use four sources of
information to determine whether to disburse additional funds to grant
recipients.6

5Some board members, including the United States, disagreed with this
change.

6Secretariat officials stated that they apply this information with
varying weight on a case-by-case basis, that partial achievement of the
quarterly program goals may be acceptable, and that no single information
source is critical to the approval of continued funding. For example, if
other mitigating circumstances exist, such as the risk of treatment
interruption, the secretariat may accept partial achievement of program
goals.

o 	Recipient progress reports. Principal recipients submit progress
reports on meeting designated targets along with requests for further
funding at the end of each disbursement period. If program results or
expenses differ significantly from plans attached to the grant agreement,
the principal recipient is to explain the reasons for these deviations and
may also provide an overview of other program results achieved, potential
issues and lessons learned, as well as any planned changes in the program
and budget. The recipient forwards its progress report and disbursement
request to the Global Fund secretariat through the local fund agent.

o 	Recipient expenditure data. The progress reports contain cash-flow
information. Principal recipients are to outline expenditures for the
previous disbursement period, comparing amounts budgeted for grant
activities with amounts spent. Recipients are then to reconcile
expenditures and provide a current cash balance. Budgets may vary from
initial projections, owing to cost savings, additional expenditures, or
currency fluctuations.

o 	Local fund agent assessments. The local fund agent reviews and
validates the information in the progress update, performs ad-hoc
verifications of program performance and financial accountability, and
advises the Global Fund on the next disbursement. Local fund agents are to
highlight achievements and potential problems to support their advice and
may identify performance gaps to be addressed. Representatives of one
local fund agent, which covers grants in 29 countries, said that they base
their disbursement recommendations on two considerations: (1) the Global
Fund's level of risk in making additional disbursements to a recipient
that uses funds ineffectively and (2) the immediate effect of withholding
disbursement on program implementation, including the delivery of
disease-mitigating services. These representatives said that, overall,
they strive to tie the progress update to projected results in the grant
agreement.

o 	Contextual information. The secretariat also uses additional
information relevant to interpreting grant progress, such as news of civil
unrest, political disturbance, allegations of corruption, conflict, major
currency crisis, change of principal recipient, and natural disasters. A
secretariat official said that the secretariat did not document
requirements for such information for phase-1 decisions but did allow
grant managers to consider any information that would adversely affect
grant implementation in their decisions to disburse. This information is

typically obtained through informal communications with grant recipients,
bilateral and multilateral donors, or other development partners,
according to secretariat officials.

The principal recipient is to provide the CCM with copies of its
disbursement requests and progress reports, and CCM members may comment on
the progress of implementation based on their local knowledge and
experience, either through the local fund agent or directly to the
secretariat. If the secretariat decides to approve the disbursement
request, it may specify the level of disbursement and actions that the
principal recipient must take. The secretariat then instructs the World
Bank to make the disbursement. The secretariat may also decide not to
approve the disbursement request.

Phase 2 	When a grant reaches its sixteenth month, the Global Fund invites
the CCM to submit a request for continued funding for the period following
the initial 2 years. The Global Fund refers to this period as phase 2 of
the grant (see fig. 2).

Figure 2: Global Fund Grant Life Cycle

            Source: GAO, based on information from the Global Fund.

The CCM is to submit its request to the Global Fund by month 18, and the
secretariat is to evaluate the CCM's request using the four sources of
information described earlier.7 Based on its assessment of this
information, as informed by its professional judgment, the secretariat
gives the grant one of four scores, as shown in figure 3. It then provides
its assessment and recommendation-called a grant scorecard-to the board
regarding approval of the request, and the board decides on the request by
month 20. If the board approves the request, the principal recipient and
the Global Fund negotiate and sign a grant agreement extension over the
next 2

7Under phase 2, contextual considerations include financial and program
management issues; CCM functioning; major changes in the political or
technical environment in which the program operates; data quality; and any
indication that the grant is not advancing the Global Fund's operating
principles of promoting broad and inclusive partnerships, sustainability,
and national ownership. The secretariat reports these considerations in a
standard format for each grant.

months. At month 22, the Global Fund instructs the World Bank to make the
first phase-2 disbursement.

Figure 3: Global Fund Grant Rating System for Phase 2

            Source: GAO, based on information from the Global Fund.

Note: The grant's performance rating is combined with contextual
considerations to yield a decision as to whether the grant should be
continued under phase 2. For example, a performance rating of A combined
with no or minor contextual issues would likely result in a "go" decision,
whereas the same performance rating combined with more significant
contextual issues could result in a "conditional go," revised go," or even
a "no go" decision. Similarly, a lower performance rating could result in
any of the four possible decisions, depending on contextual
considerations. (This explanation is based on information provided to us
by the Global Fund in April 2005. The Global Fund's previously available
guidance indicated a simpler decision-making process.)

The secretariat sends its recommended scores to the board members, who
vote on the recommendations via e-mail.8 A "go" decision means that the
Global Fund approves proceeding to phase 2. "Conditional go" means that
the Global Fund approves proceeding to phase 2 after the principal
recipient undertakes specific actions within the time frame specified.
"Revised go" means that the principal recipient must reprogram the grant
and substantially revise the targets and budgets for phase 2. "No go"
means that the Global Fund does not approve the grant's proceeding to
phase 2.9 Currently, recipients denied further funding ("no go") cannot
formally appeal the board's decision. However, a board subcommittee may
consider a formal appeal process. If the "no go" decision affects patients
on lifelong treatment, the principal recipient may be eligible to receive
funding to sustain treatment for 2 more years.

As figure 4 shows, the Global Fund board approved the first round of grant
proposals in April 2002. The second, third, and fourth rounds were
approved in January 2003, October 2003, and June 2004, respectively. The
board is expected to approve a fifth round of proposals in September 2005.
As of April 25, 2005, the secretariat had reviewed 36 grants that became
eligible for continued funding under phase 2. The board approved 20
grants, conditionally approved 13, denied 1, and is still considering 2.
The board will continue to evaluate grants for phase 2 on a rolling basis
as they become eligible.

8Four objections from board members to a recommendation trigger an
additional vote, and four further objections elevate the decision to the
next formal board meeting, at which the decision is final.

9When the secretariat recommends a "revised go" or "no go" decision, the
board may refer the proposal back to the CCM, which must then deliver a
revised request for continued funding to the secretariat. The secretariat
then passes the request to the technical review panel, which provides a
recommendation to the board. If the panel recommends to discontinue
funding, the grant is terminated; if it recommends to continue funding,
the board votes by e-mail on whether to continue funding.

Figure 4: Global Fund Milestones

            Source: GAO, based on information from the Global Fund.

Several Factors Affected Grant Performance

According to Global Fund officials and other knowledgeable entities,
recipient countries' capacity to implement grants was an underlying factor
in grant performance. In addition, principal recipients for the 38
grants10 as well as Global Fund and development partner officials
frequently cited four factors associated with challenges or successes in
grant performance: (1) guidance, (2) coordination, (3) planning, and (4)
contracting and procurement. We found no significant association between
the type of principal recipient, grant size, or disease targeted and the
percentage of a grant's funds disbursed to the principal recipient.

10Our initial scope included 45 grants, but 7 of these grants lacked
progress reports or disbursement requests during the period of our review.
We found 75 progress reports or disbursement requests associated with the
38 grants, which represent 29 percent of the 130 grants that had received
a first disbursement as of December 31, 2003.

Limited Capacity in Recipient Countries Affected Performance, but Some
Governments Have Strengthened Health-Sector Capacity

Global Fund and development partner officials cited limited capacity in
recipient countries as an underlying factor that can negatively affect
grant performance.11 Global Fund grant managers said that in many cases,
grants experienced early delays because of weaknesses in recipients'
financial, procurement, and monitoring and evaluation systems. For
example, Indonesia's local fund agent found the principal recipient's
management and financial plans to be insufficient and asked the recipient
to rework them seven times before the local fund agent recommended grant
disbursements. Also in Indonesia, TB spending increased fivefold, greatly
straining capacity, particularly for monitoring and evaluating activities
at the district level. In Kenya, a lack of designated, adequately trained
staff at the principal recipient (the ministry of finance) and immediate
subrecipients (the ministry of health and the National AIDS Control
Council) slowed disbursements from the principal recipient and from the
immediate subrecipients to implementing organizations. Ethiopia has been
slow in implementing its first three grants, particularly those for TB and
malaria, owing to lack of monitoring and reporting capacity within the
ministry of health, delays in recruiting staff to manage financial
systems, slow decision-making processes, delays in starting the
procurement process, and cumbersome procurement procedures.

Despite limited overall capacity in recipient countries, we found
instances where recipient governments had worked with development partners
to strengthen capacity in the health sector, thus facilitating grant
performance. For example, according to the Indonesian government and WHO
officials, the Global Fund grant in Indonesia is building on a strong
foundation, using the country's 5-year strategic plan for TB, a joint
effort of the Indonesian government and WHO. Between 2000 and 2003, the
Dutch government helped train TB "soldiers" in Indonesian provinces, which
improved outreach and case-detection efforts under the Global Fund grant;
Indonesia's ministry of health had already established mechanisms to
quickly disburse funds to districts. In addition, the Zambian government
has worked with donors and other development partners to strengthen its
health sector financing mechanisms. As a result, donors, including the
Global Fund, contribute directly to an existing mechanism, the pooled
health sector "basket," and use the health sector donor group overseeing
these funds to monitor and evaluate grant progress in meeting targets.

11Country capacity refers to a country's human, scientific, technological,
organizational, institutional, and resource capabilities.

Zambia's health sector also has mechanisms in place to quickly channel
funds to the country's more than 70 districts. In Mongolia, the local fund
agent reported that the principal recipient and subrecipients had adequate
financial management systems in place to account for funds and that the
principal recipient could immediately start implementing the program with
little, if any, technical assistance. Some countries and grant recipients
are also seeking to strengthen their capacity through Global Fund grants,
according to the Global Fund and principal recipients.

Lack of Clear Guidance Impeded Performance, but in Some Cases Guidance
Helped Recipients Meet Targets

Grant recipients frequently reported that a lack of guidance from the
recipient country's government or the Global Fund caused them to fall
short of grant targets. For example, recipients in three countries
reported that they could not meet their targets because they had not
received approved national treatment guidelines. Indonesia's ministry of
health did not have guidelines ready for the voluntary counseling and
testing component of its HIV grant, delaying distribution of information
to the provinces. Senegal's ministry of health, another principal
recipient, did not have treatment plans needed for implementing its
malaria grant, preventing the principal recipient from receiving
antimalarial medication. In addition, some stakeholders reported that
guidance from the Global Fund was lacking or unclear or that they
encountered difficulties with Global Fund grant policies. For example, in
at least one instance, U.S. government officials reported that spending
delays in Kenya resulted from unclear guidance from the Global Fund
regarding altering programs to allow the use of newer, more effective but
expensive malaria drugs. The Global Fund recognized that procedures for
early grants were unclear and that this lack of clarity caused program
delays. Further, WHO officials and at least one recipient voiced concerns
over lack of flexibility when recipients sought to modify grant
activities. For example, one subrecipient in Thailand expressed concern
that it could not use Global Fund money to build or maintain a shelter for
HIV-positive women because this type of activity was not written into the
grant. Grant recipients also said that continued staff turnover in the
Global Fund's grant management teams made it difficult to receive clear,
consistent guidance. For example, recipients in Thailand said that they
had worked with four different grant managers over the life of their
grants and that this turnover had complicated communication.

However, several grant recipients reported that, under certain
circumstances, Global Fund guidance allowed them to quickly redirect funds
to meet existing targets. For example, the principal recipient in
Indonesia cited grant flexibility as a factor positively affecting
performance

in both its TB and HIV/AIDS round-1 grants, because this flexibility
allowed it to adjust its funding priorities in line with its targets.
Similarly, in Thailand, one subrecipient stated that the Global Fund
allowed it to change training modules to meet educational needs,
contributing to success.

Poor Coordination Slowed Grant Implementation, but Good Coordination
Facilitated It

The Global Fund secretariat reported that, in some cases, poor
coordination negatively affected grant implementation. For example, in
Ghana, internal rivalries between ministry of health units with different
responsibilities in the program are slowing implementation. In Senegal,
the Global Fund reported that the principal recipient did not meet its
target for coordinating and developing partnerships to promote
community-based programs for combating malaria.

However, effective coordination between grant recipients and local
community groups or development partners sometimes contributed to
recipients' meeting or exceeding their goals. Zambia's HIV/AIDS grant
exceeded its targets for training and provision of services because of
development partner, NGO, and private sector contributions. Similarly, in
Kenya, one NGO principal recipient leveraged the activities of other
groups providing HIV care kits. Another recipient in Kenya exceeded its
targets for condom distribution by working with local intermediaries to
increase demand by approaching new types of clients, such as shoe shiners,
open vehicle cleaners, security officers, staff at petrol stations, and
young men at salons. In Indonesia, a grant subrecipient was able to
provide treatment to a larger number of TB patients by partnering with
private physicians, because a significant number of patients sought
treatment at private clinics.

Planning Difficulties MadeIt Hard to Meet Targets, but Adept Planning Made
Them Attainable

Planning difficulties affected some recipients' ability to meet grant
targets. Recipients reported that they sometimes did not achieve targets
for a variety of reasons, including not budgeting sufficient time or money
to complete targets, or scheduling activities for the wrong time period.
One recipient in Zambia underestimated the time needed to analyze baseline
data on constituent needs prior to the planned distribution of educational
materials on malaria prevention to 1,000 households. The recipient
eventually printed the materials but did not reach the targeted households
within the planned time frame. In Sri Lanka, a malaria grant recipient
underestimated the cost of establishing a community center and had to
redesign its program plan to remain within the grant budget. According to
the progress report, the principal recipient established new targets to
use

the funds originally budgeted to build the center, delaying grant
implementation. Further, a recipient in Kenya did not conduct 3,000
planned community education skits aimed at preventing HIV during one
disbursement period and attributed the shortfall to a conflict with annual
school examinations. The Global Fund recognized that recipients'
difficulty in setting targets for the initial grants derived in part from
the fact that it was developing procedures and guidelines at the same time
that it was approving and signing round-1 and round-2 grants.

Conversely, in some cases, adept planning positively affected grant
performance. In Indonesia, several grant recipients reported that
effective planning for TB treatment allowed various districts to complete
work plans early in the grant, in turn allowing the provinces that oversee
those districts to meet their target of developing budgets on time. In
Haiti, one principal recipient exceeded its targets by planning activities
around World AIDS Day, increasing the demand for, and the principal
recipient's provision of, AIDS-related services such as condom
distribution.

Delays in Contracting and Procurement Hindered Grant Performance, While
Efficient Procurement Helped

Recipients frequently reported that contracting delays with subrecipients,
vendors, or other service providers caused them to miss quarterly targets.
For example, UNDP, a principal recipient in Haiti, was unable to hold a
planned HIV conference because of delays in signing a contract with a
subrecipient. Delays in selecting and reaching contracts with
subrecipients caused the Argentine grant to start slowly, the Global Fund
secretariat reported. In Thailand, the ministry of public health recipient
could not establish TB treatment services because of a subrecipient delay
in selecting a site and contract. Grant recipients and the Global Fund
secretariat also cited procurement delays as reasons for missing quarterly
targets. For example, recipients of malaria grants in Tanzania and Zambia
reported that they did not distribute the targeted number of bed nets due
to lengthy government procurement processes. In addition, during our visit
to Zambia, we found that local spending restrictions also affected
recipients' ability to meet and report on targets. A district health
director explained that spending restrictions delayed her purchase of a
new hard drive for her office's computer, which slowed the district's
grant activities and reports to

the principal recipient.12 In Kenya, we found that the limited capacity of
the Kenyan health ministry's procurement agency and the ministry's
reluctance to contract with outside procurement experts led to delays and,
as a result, to gaps in the supply of HIV test kits, which bilateral
donors had to fill. In Ghana, according to the Global Fund secretariat,
the government's slow, bureaucratic procurement processes caused delays
that contributed to the grant's poor performance in reaching people with
HIV/AIDS and opportunistic infections.

However, the Global Fund secretariat reported that some principal
recipients' efficient procurement helped them meet their targets. For
example, a principal recipient in Madagascar managed procurement
exceptionally well throughout the grant and, as a result, exceeded its
targets for distributing bed nets. The Global Fund disbursed this grant's
phase-2 funding early, because the recipient had implemented the program
rapidly and was therefore able to use the additional funds. Another
recipient in Madagascar consistently met targets, and its disbursements to
subrecipients accelerated. The Global Fund also reported that, after
initially strengthening its capacity, the principal recipient in Moldova
made substantial progress with procurement activities, thereby lowering
treatment costs per patient and realizing significant savings due to lower
acquisition costs.

No Significant Association To determine whether certain grant
characteristics were factors associated Found between Grant with the
percentage of funds disbursed,13 we analyzed 130 grants with first
Characteristics and disbursements on or before December 31, 2003. We found
no significant

association between the type of principal recipient, grant size, or
diseasePercentage of Funds targeted and the percentage of a grant's funds
disbursed, after taking into Disbursed

12The District Health Director reported that the hard drive crashed
because of a power outage and that the frequency of such outages could
result in the need to replace the hard drive repeatedly. The Director of
the Central Board of Health acknowledged this problem and stated that
spending ceilings have since been raised at the provincial level but not
at the district level.

13We analyzed the percentage of funds disbursed after the initial
disbursement.

account the time elapsed since the first disbursement.14 (See app. II for
details of our analysis.) For example, the Global Fund disbursed a smaller
percentage of grants to government recipients than to recipients in the
private sector and faith-based organizations, but these differences do not
incorporate other factors such as grant size or time elapsed since the
first disbursement. We also considered whether disbursements were made in
a timely manner, that is, within 135 days (a 90-day quarter plus a 45-day
grace period allowed by the Global Fund for reporting). Overall, we found
that 35 percent of the disbursements were made within 135 days and that
later disbursements were more timely than earlier ones. The number of
timely disbursements was too small at any given disbursement stage to
determine whether timeliness varied according to recipient or disease
type, or grant size.

Sources Informing Funding Decisions Have Limitations, and Decisions Are
Not Clearly Documented

We noted problems associated with the four information sources that the
secretariat draws on for periodic disbursement and renewal decisions. In
addition, although the Global Fund's stated policy is to disburse funds
based on performance and to operate in a transparent and accountable
manner, we found that the secretariat did not document its reasons for
periodic disbursement decisions during phase 1. Similarly, some of the
secretariat's recommendations regarding grant renewals for phase 2 have
not been fully documented, and stakeholders have raised additional
concerns regarding the timing of the phase-2 renewal process, dated
information, low grant expenditure, and potential politicization of
disbursement decisions.

Information Sources for We found the following problems associated with
the sources of Disbursement and Renewal information that the secretariat
uses in making periodic disbursement Decisions Have Limitations decisions
during phase 1 and determining whether to renew grants during

phase 2.

14Using a different methodology, the Global Fund concluded that NGO
recipients had a slightly higher disbursement rate than government
recipients [see The Global Fund to Fight AIDS, Tuberculosis and Malaria,
Investing in the Future: The Global Fund at Three Years (Geneva,
Switzerland: 2005)]. The Global Fund's methodology differed from ours in
several ways. For example, the Global Fund's analysis assumed that the
time elapsed since the first disbursement on the residual amount disbursed
had a linear effect, while we estimated the effect of time directly and
found it to be nonlinear. In addition, the Global Fund did not
simultaneously estimate the effects of time, principal recipient type,
grant size, and disease targeted, as we did.

o 	Recipient progress reports vary in quality. Some reports do not explain
why recipients missed targets, and the limited monitoring and evaluation
capabilities of many recipients raise questions about the accuracy of
their reporting. Secretariat officials acknowledged that guidance for
planning program activities, setting indicators, and monitoring and
evaluating progress was not available when initial grants were signed.
However, Global Fund secretariat and other officials have raised questions
about the ability of principal recipients to discharge their
responsibility for reviewing and monitoring the activities of
subrecipients to which they disburse funds. According to the Global Fund
official in charge of grant operations, many early grant proposals were
overly ambitious and hurriedly assembled; he said more recent proposals
were more realistic and better designed. UNAIDS officials also stated that
when principal recipients' progress updates show poor performance, it is
not always clear whether grants are underperforming or recipients are
failing to effectively report performance. For example, when a progress
update shows failure to achieve targets, the principal recipient and
subrecipients may have actually completed the activities but not
understood how to record them.

o 	Recipient expenditure data are incomplete. Recipients' cash-flow
reports do not include data on expenditures below the level of the
principal recipient. In addition, principal recipients may not always
document their disbursal of money to subrecipients. Moreover, Global Fund
and other officials have questioned whether some principal recipients have
the expertise needed to monitor subrecipients' expenditures. Further,
secretariat officials stated that although the achievement of program
targets and cash flow are closely linked, recipients' expenditures do not
necessarily indicate that they are meeting their targets. The officials
stated that utilizing this source of information is essential to guard
against treatment interruptions or irreparable harm to struggling programs
that are not yet viable but show strong potential.

o 	Local fund agent assessments are inconsistent. According to Global Fund
secretariat officials and others, the ability of local fund agents to
effectively verify program activities varies widely. A secretariat-
commissioned assessment reported that the current local fund agent system
does not provide grant managers with a sufficient level of risk

assurance for continued funding.15 The study, as well as Global Fund and
development partner officials, reported that although most local fund
agents are competent to assess and verify financial accountability, they
often lack the knowledge and experience needed to assess and verify
recipients' performance-specifically, recipients' ability to meet program
targets, monitor and evaluate progress, and procure and manage drugs and
other medical supplies. The study also stated that local fund agents'
assessments of financial and program-related capacity and verifications of
activities are limited and rarely include site visits to implementing
subrecipients.

o 	Contextual information is systematically collected for phase 2 but not
for phase 1. To better understand why recipients received phase-1
disbursements when they did not meet many of their performance targets, we
requested full disbursement dossiers from the secretariat; however, the
dossiers contained very little contextual information supporting the
disbursement decisions. The contextual information provided was often in
the form of hand-written notes or e-mail correspondence that had been
collected ad hoc. Secretariat officials acknowledged that while they
collect contextual information through detailed questions on the
scorecards for phase-2 decisions, they have no systematic method for
collecting such information for phase-1 decisions. Although the Global
Fund considers contextual information in its funding decisions, it does
not document the extent to which it uses such information.

Little Documentation Provided to Support Phase 1 Disbursement Decisions

Although the files for the 38 grants we reviewed contained information on
progress toward targets and cash flow, they contained little or no
documentation explaining why the Global Fund approved the disbursements.16
Overall, for the 38 grants we reviewed, we determined that recipients met,
on average, 50 percent of their targets; partially met 21

15Global Fund to Fight AIDS, Tuberculosis and Malaria: Review of the
Roles, Functions and Performance of Local Fund Agents, Final Report,
August 2004, Stein-Erik Kruse, Centre for Health and Social Development
(HeSo), and Jens Claussen, Nordic Consulting Group (NCG). This report has
not been shared with the board.

16The Global Fund provided 51 local fund agent reports associated with 24
of the 38 grants. Thirty-five of the local fund agent reports provided no
information on why funds should be released.

percent; and failed to meet 24 percent.17 For 6 percent of the targets,
the information in the progress reports was insufficient to determine
whether the target had been met, partially met, or not met. In some of
these cases, the Global Fund disbursed funds to recipients even though
they reported that they had met few or none of their targets. For example:

o 	The principal recipient for Sri Lanka's second malaria grant received
disbursements for its third and fourth quarters, although it had submitted
two progress updates showing that it met only 2 of its 14 targets for the
third quarter and 4 of 13 targets for the fourth quarter. The secretariat
provided no written information explaining its approval of the
third-quarter disbursement and provided only a one-sentence declaration of
agreement regarding the fourth-quarter disbursement. In both cases, the
local fund agent had recommended that the recipient receive less than the
amount requested, citing cash-flow considerations but not mentioning
performance against targets. In each case, the secretariat disbursed the
amount that the local fund agent recommended.

o 	The principal recipient for Thailand's TB grant received its second
disbursement although it had met only 1 of 29 performance targets. The
secretariat approved the full amount requested, stating that the recipient
had not requested sufficient funds in its previous disbursement request,
although the grant manager did not provide documentation to validate this
assessment. The local fund agent had noted the grant's poor performance
and, acknowledging the grant's low cash reserves, suggested a disbursement
of 25 percent of the recipient's request.

Further, the Global Fund secretariat does not systematically track denied
disbursement requests or publicly document denials. Secretariat officials

17We reviewed 75 progress reports with a total of 1,125 targets associated
with the 38 grants. For 116 (6 percent when the percentages for each
report are averaged) of these targets, the information available was not
adequate to determine whether or to what extent the target was met. For
the rest, we used the Global Fund's system for ranking target attainment.
The Global Fund considers a numeric target-for example, the number of
health workers trained-to be met if the recipient achieves over 80 percent
of it; partially met if the recipient achieves between 30 and 80 percent
of it; and not met if the recipient achieves less than 30 percent of it.
We placed partially met nonnumeric targets, for example, developing a
monitoring and evaluation plan, in a separate category and added the
targets in that category to the partially met numeric targets to arrive at
the 21 percent figure for partially met targets. The percentages add up to
101 percent due to rounding. See appendix I for a more detailed
description of our methodology.

acknowledged that they currently have no mechanism for tracking or
documenting these instances. According to these officials, the denial may
eventually be documented in a memorandum on the grant's disbursement
request history once a disbursement is approved or, if the grant is
ultimately canceled without further disbursement, in a grant-closing
memorandum.

Some Renewal Recommendations Were Not Clearly Documented

According to grant management officials, the secretariat is to
unequivocally demonstrate satisfactory performance of all grants
recommended to the board for continued funding under phase 2. However, we
found that the secretariat did not always clearly explain the overall
score it assigned each grant when it recommended the grant for continued
or conditional funding. Although a substantial part of the score is to be
based on recipients' performance against agreed-on targets (e.g., the
number of people to be reached by disease mitigation services), the final
score can also reflect grant managers' professional judgment, contextual
information from multilateral and bilateral donors, and past disbursement
rate data. Secretariat officials said that decisions based on these
information sources should be documented when an overall score does not
seem to reflect recipients' achievement of individual targets. However, we
did not find such documentation in the grant scorecards for 8 of 25 early
grants that the Global Fund has considered for continued funding after an
initial 2-year period. The secretariat gave 3 of the grants an overall
score of B2 yet recommended "conditional go," which corresponds to a B1
score. For another grant, the secretariat gave a B1 score for three
indicators, two of which concern the number of people reached by
treatment, care, or other disease mitigation services, yet made an overall
recommendation of "go," which corresponds to an A score. Such
discrepancies between scores and recommendations are significant, because
the recommendations determine the levels of action that recipients are to
undertake before receiving phase-2 funding. Seven of the scorecards also
raised concerns about the quality of recipients' data and their monitoring
and evaluation capabilities. Of the 25 grants, the Global Fund decided to
cancel one, and the secretariat's scorecard clearly explained the reasons
for recommending that the board cancel the grant.

Stakeholders Also Raised According to the Global Fund, the phase-2 renewal
process is a critical Concerns Regarding checkpoint to ensure that grants
show results and financial accountability. Phase-2 Renewal Process
However, some stakeholders raised concerns about the process that the

Global Fund used to review the first set of grants eligible for renewal.
For

example, a representative of a local fund agent stated that this process
may occur too early in the life of a grant and that progress may be better
evaluated when a grant approaches the 3-year mark. Further, officials
representing a Global Fund board member stated that data provided to the
board during the first round of renewal decisions did not contain
expenditure data. These officials stated that when they sought expenditure
data (i.e., amounts spent by grant recipients on program activities) on
the Global Fund's Web site in March 2005, the most recent information for
their grants of concern had been posted in June 2004. Subsequent data
submitted to the board for phase-2 renewal decisions contained expenditure
information. In one case, a recipient applying for phase-2 funding and
recommended by the secretariat for continued funding had received more
than 75 percent of its 2-year grant amount yet had transferred only 12
percent of this money to subrecipients for program activities. These
officials also raised concerns over the potential for the politicization
of board decisions because the board had returned three "no go"
recommendations to the secretariat for further consideration after some
recipients and NGOs lobbied board members.

Global Fund Is Taking Several Steps to Refine Grant Management and
Performance

The Global Fund's secretariat is launching a range of initiatives to
address challenges to grant performance and improve the overall management
of grants. Systemwide, the secretariat is (1) reorganizing and
strengthening its units, (2) developing a risk assessment mechanism and
early warning system, (3) streamlining reporting and funding procedures,
(4) working with partners to strengthen recipient capacity, and (5)
clarifying guidance for CCMs. However, the board has not clearly defined
the CCMs' role in overseeing grant implementation. The Global Fund has
also responded to country-specific challenges in Kenya and Ukraine.

Secretariat Is Reorganizing To improve grant management and documentation
of funding decisions and Increasing Its Staff to and to better support
underperforming grants, the Global Fund took the Better Manage Grants
following actions in 2004:

o 	Reorganized the secretariat's operations unit and increased the number
of staff from 118 to 165.18 For example, it added eight grant manager
positions and established regional teams, each with a team leader, so that
more than one grant manager is responsible for a set of grants in the
countries within a regional team. To better document periodic disbursement
decisions, the secretariat added a new position, known as a program
officer, to its grant management structure. The secretariat is currently
recruiting program officers for each regional team, who are to be
responsible for documenting disbursement and other decisions and keeping
track of grant milestones. Further, secretariat officials said that the
Global Fund is planning to recruit additional grant management staff to
conduct increased day-to-day recipient monitoring and assistance. The
program officers and the additional grant management staff accounted for
most of the increase in staff at the secretariat between 2004 and 2005,
according to a Global Fund administrative official.

o 	Created the Operational, Partnerships and Country Support Unit to focus
on problem grants. According to secretariat officials, this unit-which
also includes new positions to liaise with development and technical
assistance partners, local fund agents, and CCMs-will enable the
secretariat to address grant performance issues before they become serious
problems and will thereby better manage risk exposure. For example, the
unit could mobilize intervention by high-level recipient government
officials, solicit technical assistance from partners, or engage the
United Nations Children's Fund (UNICEF) to procure health commodities
until the recipient government can set up a viable procurement system.

o 	Strengthened its strategic information and evaluation unit to improve
monitoring and evaluation, data reliability, and quality assurance.

o 	To enhance the quality and consistency of the data that recipients
report, in June 2004 the secretariat issued a monitoring and evaluation
"toolkit" developed in cooperation with other donors and development
assistance partners.19 This toolkit guides grant recipients to select
consistent indicators to measure progress toward

18According to the Global Fund secretariat, its staffing review indicates
that it should be able to manage the volume of grants foreseen by 2007
with no more than 200 staff.

19Monitoring and Evaluation Toolkit, HIV/AIDS, Tuberculosis and Malaria,
June 2004.

key program goals, such as the number of people with AIDS who were reached
with drug treatment or the number of people given insecticide-treated bed
nets to prevent malaria. The secretariat has also required attachments to
each grant agreement that outline program indicators and the specific
activities that enable recipients to meet these indicators and overall
program goals. According to Global Fund officials, progress will more
easily and consistently be measured when all grants have aligned their
indicators and activities to this toolkit. Grant managers are currently
working to accomplish this goal with the recipients they cover.20
According to the Global Fund, these developments have been important in
harmonizing monitoring and evaluation approaches among partners at
national and international levels and will help simplify country-level
reporting to multiple donors by ensuring the use of a common set of
indicators to measure interventions. Partners provided training on the
toolkit in 2004. According to the secretariat, training is to continue in
2005. Recipients we met with in Thailand confirmed that they had received
the toolkit. However, they said that it was not in their native language
and therefore was not useful.21

o 	In March 2004, the board approved establishing a Technical Evaluation
Review Group with members from UNAIDS, WHO, and other partners to develop
a system for assessing and ensuring data reliability. The group first met
in September 2004. According to Global Fund officials, these efforts will
result in more systematic reporting and analysis by recipients, the Global
Fund, and partners and, consequently, in better comparisons of grants.

o 	To strengthen strategic information for monitoring grant performance,
in fall 2004, the secretariat created a "data warehouse" that contains
information from recipients' progress reports and disbursement requests,
donors, CCMs, and local fund agent assessments. Secretariat staff use the
database to prepare "scorecards" that rank grants for the phase-2 renewal
process.

20Because grant agreements from the earlier rounds were signed prior to
the launch of the toolkit, the strategic information team worked with
grant managers to draft or revise monitoring and evaluation plans and add
these plans to grant agreements and ensure the application of the
toolkit's standard monitoring and evaluation framework.

21The toolkit was published in the official UN languages of Arabic,
Chinese, English, French, Russian, and Spanish.

Secretariat Is Implementing a Risk Assessment and Early Warning System to
Identify Potential Nonperforming Grants

The secretariat has devised a risk-assessment model and early warning
system to identify poorly performing grants and to more systematically
alert grant managers when they need to intervene. Because the Global Fund
disburses grants to recipients in countries with varying levels of
economic development and capacity, its risk-assessment model will
incorporate grant size and performance as well as country development and
corruption indicators.22 By tracking key events in the context of grant
and country risk, the grant portfolio managers can determine whether
recipients have missed important milestones. The early warning system is
to generate reports using indicators-for example, time elapsed between
disbursements-to flag problems and trigger possible interventions. The
system will also incorporate contextual information from country-based
partners. When the system identifies slow-moving grants, staff from the
secretariat's Operational, Partnerships and Country Support unit will be
able to assess and follow up with the appropriate level of intervention.
For example, if a grant recipient in a high-risk country does not submit a
progress report and disbursement request at the expected time, the system
will alert staff that follow-up is needed.

Although the system has not been fully implemented, secretariat officials
said that their recent intervention in Tanzania exemplifies the way the
system should work. The Tanzania malaria grant was not demonstrating
progress after 1 year, as measured by the amount of funds disbursed
compared with the amount that the secretariat expected to disburse. After
following up with the principal recipient, secretariat staff realized that
political infighting-rather than technical limitations-were inhibiting
progress of the malaria program: competing groups were vying for control
of grant funds and uncertain of how to procure and distribute bed nets to
vulnerable groups, such as pregnant women or women with young children.
The government decided to give vouchers to members of vulnerable groups to
enable them to purchase bed nets at a lower price;23 however, the ministry
of health did not print or distribute the vouchers or specify where they
should be distributed. The Director of the Operational, Partnerships and
Country Support Unit traveled to Tanzania and met with

22The Global Fund secretariat has assessed several sources for the
information on country development and corruption indicators and plans to
use readily available information.

23The government decided to use a voucher system instead of distributing
the bed nets free of charge in order to develop a local bed-net industry
and thereby help to make the supply of bed nets more sustainable.

development partners and high-ranking host government officials to
encourage the government to take action. The Global Fund brought in
UNICEF, a key development partner in Tanzania, to work with the
government's malaria advisor as well as experts from the Swiss Tropical
Medicine Institute to resolve the problems and get the program back on
track.

Secretariat Is Streamlining Procedures

In response to concerns that grantee reporting requirements are difficult
and time consuming for recipients, grant managers, and local fund agents,
the secretariat instituted a new policy that changes the default for
reporting from quarterly to every 6 months. In addition, the secretariat
is considering new, more streamlined funding mechanisms than the current
round-based approach. However, the board has not endorsed these changes,
and some board members, including the United States, are opposed to them
at this time.

To decrease the administrative burden on grantees and to bring its
practice more in line with other donor agencies, the secretariat
instituted a semiannual reporting policy in July 2004. Some recipients
still report quarterly, such as those implementing grants that the
secretariat identified as high risk-for example, in countries with limited
human resource capacity-while others have the option of using quarterly
disbursements to meet their needs-for example, as a hedge against currency
fluctuations. However, this policy change did not require board approval,
and some board members, including the United States, do not support it.

Although the Global Fund strives to be a funding mechanism that seamlessly
fits into many country programs by providing additional funding where
needed, it recognizes that its current practice of financing grants
through rounds can disrupt countries' planning and time lines and strains
recipient capacity. In addition, some associated with the Global Fund said
that rounds might lead CCMs and recipients to concentrate their energy on
developing new proposals rather than implementing existing grants and that
repeated rounds add greatly to the secretariat's workload. A document
submitted to the board by the secretariat stated that although the
roundbased grant approval system worked well for launching the Global Fund

and identifying countries that submitted strong proposals,24 this system
forced recipients to adapt their planning cycles to those of the Global
Fund (rather than building on preexisting planning cycles), encouraged the
submission of smaller proposals, and left a considerable amount of time
between proposal submission and approval. This document presented several
options for the board, such as creating two continuous funding streams-one
for governments and another for civil society recipients. For example,
government applicants could submit their national strategic plans for the
coming years, highlighting financing gaps and facilitating integration of
Global Fund financing with existing planning and budgeting systems, such
as sectorwide approaches. According to the document, this approach would
create incentives for CCMs to improve and accelerate the disbursement of
funds and would ease the secretariat's workload, allowing secretariat
staff to spend more time managing grants and less time negotiating grant
agreements. The board has not set time frames for further discussing this
issue. According to U.S. board members, the board has not yet fully
discussed or approved these changes, and a majority of board members
oppose them at this time.

Global Fund Is Working Because most grant performance problems are
associated with limited with Partners to Improve capacity at the country
level, where the Global Fund has no presence and Technical Capacity at the
plays no part in program implementation, the Global Fund relies on its

technical partners to provide technical expertise to grant recipients.25
Country Level Although the partners we spoke with expressed their strong
support for the Global Fund, they also voiced concern that they have not
received

24At the request of the executive director, secretariat staff prepared a
discussion paper, outlining some of the challenges facing the Global Fund
and including several options for changing the core aspects of the current
business model, and introduced it to the eighth board meeting in June
2004. According to Global Fund officials, the paper was updated and
presented at the ninth board meeting in November 2004 and discussed at a
retreat after the official meetings. However, the board has not formally
endorsed the paper.

25According to the Global Fund, it has intensified dialogue on
coordination, information exchange, and assistance in program countries
with British, Canadian, French, German, Swedish, and U.S. bilateral
partners.

additional resources to provide the technical support that grant
recipients have requested.26

The Global Fund and partners reported that partners provided essential
support that strengthened recipients' capacity to prepare applications for
Global Fund financing and helped address the underlying problems that
affected grant performance. For example:

o 	UNAIDS, a key technical partner, has added about 30 monitoring and
evaluation officers in various countries who are available to support CCMs
in preparing grant performance reports for phase-2 renewals. UNAIDS has
also intensified its capacity-building support at the country level.

o 	Several WHO departments have provided critical technical support. For
example, WHO's Stop TB unit supported 50 countries when they developed
their applications for Global Fund financing. The Global TB Drug Facility
worked with recipients in eight countries to identify and resolve
procurement and supply management bottlenecks. WHO's HIV/AIDS Department
helped to develop comprehensive technical support plans for accelerating
the scale-up of antiretroviral therapy and prevention services in 15 to 20
countries.27 In addition, according to the Global Fund, it collaborated
closely with WHO's Roll Back Malaria Department in 2004 to incorporate
into existing grants new, more effective malaria treatments that use
artemisinin-based combination therapy.

o 	USAID and the U.S. Centers for Disease Control and Prevention (HHS/CDC)
are assisting grantees in a number of countries. For example, USAID is
supporting TB grants in numerous ways, including providing training on
procuring and managing medical supplies, addressing country-level
financial management constraints, and conducting human resource
assessments to determine existing capacity

26UNAIDS allocated additional resources to help support harmonized
monitoring and evaluation at the country level, which will benefit Global
Fund grantees in these countries. In commenting on a draft of this report,
U.S. government officials noted that grant proposals can and, in many
cases, should, contain a line item for improving technical capacity at the
country level.

27According to WHO, its "3 by 5" initiative sets the global target to
provide three million people living with HIV/AIDS in developing and middle
income countries with life-prolonging antiretroviral treatment by the end
of 2005.

needs. In another instance, HHS/CDC is assisting one grantee in reporting
and monitoring activities and revising project funds to improve grant
implementation. HHS/CDC has also coordinated the implementation and
monitoring of activities under another country's TB grant, participating
in supervisory visits to districts to assess their progress and compiling
and submitting quarterly reports for the TB grant to the ministry of
health.

UNAIDS and WHO officials in Geneva and in the field expressed strong
support for the Global Fund but consistently raised concerns about their
organizations' ability to respond to increasing numbers of requests from
grant recipients for help in addressing issues underlying performance
problems. For example, although UNAIDS recently added about 30 monitoring
and evaluation officers in its country and regional offices, officials
said that the agency's resources are being stretched thin and that it
cannot provide assistance to all Global Fund grant recipients. Likewise,
WHO officials said that its regional and country staff are dedicated to
providing technical assistance, but because WHO is not funded to support
Global Fund grants it is often unable to respond to all recipients'
requests for help. According to officials from WHO's HIV/AIDS, Stop TB,
and Roll Back Malaria departments, the Global Fund works under the
assumption that UN agencies have a mandate to provide technical
assistance. However, unless it gets more money from its member countries
for this purpose, WHO does not have the resources to keep up with the
massive increase in need for technical assistance owing to Global Fund
grants. In addition, WHO officials pointed out that the Global Fund does
not encourage recipients in African countries to take advantage of WHO's
Global Drug Facility to procure quality-assured TB drugs at the cheapest
prices available; instead, the Global Fund encourages competition and
reliance on local industry.

Global Fund Has Clarified Guidance and Taken Steps to Strengthen CCMs, but
Some Issues Remain

To strengthen accountability in recipient countries, the board has
clarified some roles and responsibilities for the CCMs. The board has
stated that CCMs are responsible for overseeing grant implementation and
are therefore to play an important role in deciding whether grants should
be renewed for phase-2 funding. To enhance and clarify CCM functioning,
the secretariat in March 2005, convened regional workshops in Zambia and
India on CCM best practices. In addition, to improve communication between
the Global Fund and CCMs, the secretariat is compiling contact information
for all CCM members. This information will enable it to communicate
directly with the members instead of relying on the CCM

chairperson to disseminate information. Secretariat officials acknowledged
that no formal studies conclusively demonstrate a link between CCM
functioning and grant performance. However, the Global Fund's March 2005
report stated that many of the (then) 27 grants eligible for phase-2
funding benefited from several factors, including full levels of
participation by CCM members in that body. Further, the report stated that
low levels of participation and involvement by CCM members were a key
factor in poor performance.28 Secretariat officials stated that they plan
to initiate a study at the end of 2005 to systematically investigate links
between CCM functioning and grant performance, given that a number of
additional grants will then have neared the 2-year mark and gone through
the phase-2 decision process.

In response to findings from several earlier studies commissioned by the
Global Fund on CCM functioning, in November 2004, the board agreed on
specific requirements for CCMs.29 However, it has not clearly defined
CCMs' role in monitoring grant implementation.30 In April 2005, the board
directed CCMs to develop tools and procedures for overseeing grants,
stating generally that these tools and procedures "should include but need
not be limited to" a list of five activities such as recording key
oversight actions and developing a work plan that "could include" site
visits. The board noted that because CCMs vary from country to country,
these guidelines can be adapted and their application paced as needed.
According to secretariat staff, the board has not reached consensus

28In one case, the secretariat cited poor CCM governance (acute
coordination problems and frequent turnover in the CCM chair) as a factor
in its recommendation to discontinue funding.

29Specifically, the board voted to require all CCMs to have a written plan
to mitigate against conflict of interest when principal recipients and CCM
chairs or vice chairs are from the same entity; establish and maintain a
transparent, documented process to solicit and review proposal-related
submissions, nominate principal recipient(s) and oversee grant
implementation, and ensure input from a broad range of stakeholders in
proposal development and grant oversight; show evidence of membership of
people living with or affected by the three diseases; and have NGO
representatives selected by their own sectors based on a documented,
transparent process.

30The board has also considered the funding of basic CCM operations. Some
board members prefer the Indonesia model, where a key bilateral partner
funds the CCM secretariat. Others prefer the Thailand model, where the
host government houses the CCM secretariat in a ministry and funds
administrative positions. At its tenth meeting in April 2005, the board
requested the secretariat to work with multilateral and private partners
to support CCM operations. In countries where partner support of CCMs is
not available, the board approved limited use of grant funds to support
CCMs.

regarding CCMs' oversight role because some members want clear, specific
requirements for CCMs while others prefer the more general guidelines. In
addition, in 2004, the board agreed on a checklist for measuring CCM
performance that focuses mostly on the makeup of the CCMs, participation
and communication among members, and governance and management. However,
the checklist did not include parameters for measuring the effectiveness
of CCMs in overseeing grant performance. Participants at the Zambia
workshop recommended that the secretariat develop more specific guidelines
defining the oversight role of the CCM.

Global Fund Intervened in at Least Two Countries to Address Grant
Performance

The Global Fund secretariat intervened in at least two countries in
response to grant performance problems. For example, in Kenya, the
secretariat intervened in 2004 at the request of donors and board members
to encourage the principal recipient to hold regular meetings with
subrecipients and designate staff to administer and monitor the grants.
The secretariat also intervened in Kenya to improve coordination by
facilitating new CCM procedures, such as designating multiple
minute-takers to ensure the accuracy of the minutes and making sure that
minutes are circulated promptly. According to one CCM member, two
additional people now take notes at each meeting; however, the minutes are
not being circulated in advance of the next meeting. In commenting on a
draft of this report, U.S. government officials said that, despite these
interventions, problems persist. For example, they said that CCM meetings
in Kenya are too infrequent and poorly prepared; decisions are made
outside of the meetings; and the minutes are often inaccurate.

In Ukraine, the Global Fund suspended three HIV/AIDS grants in January
2004 after investigating irregularities in the principal recipients'
procurements that development partners had brought to its attention a
month earlier. The secretariat had also found that after nearly 12 months
of a 24-month program, the recipients had spent less than 4 percent of the
total 2-year amount for the three grants. The Global Fund had disbursed a
total of $7.1 million to the principal recipients, from whom it obtained
$6.3 million in reimbursements. In March 2004, the secretariat signed an
agreement with a new principal recipient to continue the HIV/AIDS
mitigation activities specified in the original grants; in addition, it
transferred $300,000 to this entity to avoid interrupting ongoing
programs.31

31The new principal recipient, an international NGO, had been a
subrecipient under one of the original grants.

Conclusions	The Global Fund's mandate reflects inherent tensions. On the
one hand, the Global Fund is to function solely as a funding entity with
no implementing role and to encourage recipient country bodies such as the
CCM to be responsible for implementing and overseeing grants. On the other
hand, it is to disburse funds rapidly while also ensuring that recipients
are able to account for expenditures and produce measurable results in
addressing the three diseases.

In seeking to balance these tensions and further improve its performance,
the Global Fund has revised-and continues to revise-its processes. Some
systemwide changes require board approval or will take time to fully
implement, whereas others can be implemented relatively quickly. Capacity
in recipient countries, guidance, coordination, planning, and contracting
and procurement are pivotal to grant performance and therefore merit
continued attention. However, local fund agents' frequent lack of
expertise in assessing these factors, and many recipients' limited
monitoring and evaluation capabilities, raise questions about the accuracy
and completeness of the information that the secretariat uses to make its
periodic disbursement and funding renewal decisions. In addition, despite
recent improvements, the Global Fund's lack of consistent, clear, and
convincing documentation of its funding decisions may hamper its ability
to justify these decisions to donors and other stakeholders, in accordance
with its principles of transparency and accountability. To ensure that all
funding decisions are clearly based on grant performance and reliable
data, it is critical that the Global Fund resolve these issues in a timely
manner.

Recommendations for Executive Action

To improve the quality of the information on which the Global Fund bases
its funding decisions and the documentation explaining these decisions, we
recommend that the U.S. Global AIDS Coordinator work with the Global
Fund's Board Chair and Executive Director to take the following three
actions:

o 	complete efforts to ensure that local fund agents have the necessary
expertise to evaluate performance data on disease mitigation that
recipients submit,

o 	continue to work with development partners to strengthen the quality
and consistency of that data by enhancing recipients' capacity for
monitoring and evaluating their financial and program-related activities,
and

o 	continue efforts to clearly document the Global Fund's reasons for
periodically disbursing funds and renewing grant agreements.

Agency Comments and Our Evaluation

We requested comments on a draft of this report from the Executive
Director of the Global Fund, the Secretaries of State and HHS, and the
Administrator of USAID, or their designees. We received formal comments
from the Global Fund as well as a combined formal response from State,
HHS, and USAID (see apps. III and IV). The Global Fund concurred with the
report's conclusion and recommendations and noted steps it is taking to
improve documentation of grant performance such as organizing regional
training of principal recipient staff to improve the quality of their
reporting; defining universal and detailed performance indicators for each
grant to more systematically track performance; and tailoring grant
oversight and terms of reference for local fund agents based on grant
risk. State, HHS, and USAID largely concurred with the report's
conclusions but did not comment on the recommendations in their formal
response. Both the Global Fund and the U.S. agencies also submitted
informal, technical comments, which we have incorporated into the report
as appropriate.

We are sending copies of this report to the Global Fund Executive
Director, the U.S. Global AIDS Coordinator, the Secretary of HHS, the
Administrator of USAID, and interested congressional committees. Copies of
this report will also be made available to other interested parties on
request. In addition, this report will be made available at no charge on
the GAO Web site at http://www.gao.gov.

If you or your staff have any questions about this report, please contact
me at (202) 512-3149. Contact points for our Offices of Congressional
Relations and Public Affairs may be found on the last page of this report.
GAO staff who made major contributions to this report are listed in
appendix V.

David Gootnick, Director International Affairs and Trade

Appendix I

                       Objectives, Scope, and Methodology

In May 2003, the President signed a law directing the Comptroller General
to monitor and evaluate projects supported by the Global Fund.1 This
report reflects our review of grants that the Global Fund began disbursing
before the beginning of 2004-that is, grants that have had at least 1 year
to perform. In this report, we (1) describe the Global Fund's process for
managing grants and disbursing funds, (2) identify factors that have
affected grant performance, (3) review the basis for, and documentation
of, the Global Fund's performance-based funding, and (4) describe the
Global Fund's recent refinements for managing grants and improving their
performance.

Methodology for Describing To describe the Global Fund's process for
managing grants and disbursing the Global Fund's Process funds, we
reviewed Global Fund documents, including The Global Fund for Managing
Grants and Operational Policy Manual and related guidance documents; A
Force for

Change: The Global Fund at 30 Months; The Global Fund to Fight AIDS,

Disbursing Funds 	Tuberculosis and Malaria: Annual Report 2002/2003; and
Investing in the Future: The Global Fund at Three Years. We also
interviewed Global Fund officials in Washington, D.C., and in Geneva,
Switzerland.

Methodology for Identifying Factors That Affected Grant Performance

To identify factors affecting grant performance, we conducted three types
of analysis. First, we selected 13 countries that had grants with a first
disbursement on or before December 31, 2003, to allow for at least 1 year
of performance, and that had grants covering more than one principal
recipient. In addition, all but 4 of these countries had grants covering
more than one disease. We reviewed Global Fund dossiers for 38 grants to
recipients in these countries and categorized reasons given for deviation
from performance targets. (Our initial scope included the 45 grants to
these countries, but for 7 of the grants there were no disbursement
requests available during the period of our review.) We found 75 progress
reports/disbursement requests and 51 local fund agent assessments
associated with 24 of the 38 grants. These 38 grants represent 29 percent
of the 130 grants that had received a first disbursement by the end of
2003.

Starting with the grant's second disbursement, we included all
disbursement requests from each grant that were available on the Global

1U.S. Leadership Against HIV/AIDS, Tuberculosis, and Malaria Act of 2003,
(P.L. 108-25), Section 202(f).

Appendix IEURObjectives, Scope, and MethodologyEUR

Fund Web site as of November, 2005, and a few that we received
subsequently from the Global Fund. We requested full disbursement dossiers
from the Global Fund. These dossiers contained principal recipients'
progress reports and cash-flow/expenditure data, local fund agents'
reviews of the recipients' information and recommendations about further
disbursements, and, in most cases, additional documents such as
correspondence between the Global Fund secretariat and the principal
recipient. Using this information, we coded reasons given for deviation
from grantees' agreed-upon performance targets into 1 of about 30
categories. We grouped this information into 5 major categories- resources
or capacity; coordination; programmatic problems, needs, or changes;
procurement; and factors beyond recipients' control. Within these
categories, we developed specific subcategories such as guidance,
decisions or plans not made, done, or available; signing of contracts or
agreements delayed or not done; and limited trained human resources.

As in any exercise of this type, the categories developed can vary when
produced by different analysts. To address this issue, two GAO analysts
reviewed a sample of the progress reports and independently proposed
categories, separately identifying major factors and then agreeing on a
common set of subcategories. We refined these subcategories during the
coding exercise that followed. We then analyzed the reasons for deviations
from all of the recipients' progress reports and placed them into one of
the subcategories. When information in the progress reports was
insufficient to determine how to code a reason, we consulted the local
fund agents' reports. We tallied the counts in each subcategory and
identified the subcategories mentioned in the greatest number of grants.

The information in the disbursement requests varied in detail and quality.
The two analysts, together with a methodologist, therefore discussed and
documented categorization criteria and procedures throughout the analysis,
and the methodologist reviewed the entire analysis as a final step. As a
validity check on our document analysis and to identify frequently cited
factors that affected grant performance, we compared the information in
the subcategories mentioned in the most grants to information available
from our fieldwork (see below) and determined that both sources of
information reported similar findings.

In addition, we reviewed the 75 progress reports/disbursement requests
associated with these 38 grants and tallied the total number of targets in
each request. We ranked each target using the same numeric rating system

Appendix IEURObjectives, Scope, and MethodologyEUR

the Global Fund uses for phase 2 (see fig. 3).2 Because many of the total
1,125 targets were nonnumeric (e.g., developing a monitoring and
evaluation plan), we did the following: (1) if the principal recipient
clearly met the target, we ranked the target as met ("meeting or exceeding
expectations") and included it with the numeric targets that fell into
that category; (2) if the principal recipient clearly showed no progress
toward meeting the target, we ranked it as not met ("unacceptable") and
included it with the numeric targets in that category; and (3) if the
principal recipient partially met the target, we gave it a separate
ranking-"partially met nonnumeric target," and characterized it as
partially met, along with the partially met numeric targets. To arrive at
percentages for the targets in each category, we first calculated the
percentage for each progress report/disbursement request and then averaged
the percentages for the category from all the reports. Because the number
of targets in each report varied greatly, we averaged the percentages
rather than the numbers of targets to ensure that each report was given
equal weight. We excluded from our calculations those few targets for
which the information available was not adequate to determine whether or
to what extent the target was met.

Second, we reviewed documents obtained from field visits to four
countries-Indonesia, Kenya, Thailand, and Zambia-and interviewed a wide
variety of government, civil society, and bilateral and multilateral
development officials in these countries involved in grant implementation
or oversight. All four of these countries received more than $36 million
in committed funds for several grants that covered more than one disease,
and three of them (Kenya, Thailand, and Zambia) also have grants that
cover both government and civil society recipients. In addition, we
interviewed officials from the Global Fund, World Health Organization
(WHO), and the Joint UN Programme on HIV/AIDS (UNAIDS) in Washington,
D.C., and Geneva, Switzerland.

Finally, to determine whether the percentage of funds disbursed for each
grant (after the first disbursement) and the timeliness of the
disbursements were associated with grant characteristics such as type of
principal recipient, grant size, or disease targeted, we analyzed 130
grants with first

2The Global Fund also distinguishes between types of targets such as
number of persons reached with services versus number of service centers
established or strengthened, or number of persons trained to deliver
services. In our analysis of the targets, we did not make this
distinction.

               Appendix IEURObjectives, Scope, and MethodologyEUR

disbursements on or before December 31, 2003. (See app. II for a more
detailed discussion of this methodology.) To assess the reliability of the
Global Fund's data, we (1) posed a set of standard data reliability
questions to knowledgeable agency officials, (2) performed basic
electronic reasonableness tests, and (3) interviewed officials about a few
small anomalies that we found during our analysis. We found only one minor
limitation, namely that disbursement dates were not reported for less than
5 percent of the disbursements. Based on our assessment, we determined
that the data were sufficiently reliable to generate descriptive
statistics about the program, and to be used for advanced statistical
modeling work.

Methodology for Reviewing the Basis for, and Documentation of, the Global
Fund's Performance-Based Funding

To review the basis for, and documentation of, the Global Fund's
performance-based funding, we examined Global Fund documents- including
The Global Fund Operational Policy Manual and related guidance documents,
the dossiers for the 38 grants that had a first disbursement on or before
December 31, 2003, and documents supporting Global Fund decisions to
continue or discontinue funding 25 of 28 grants that had reached their
phase-2 renewal point and been reviewed by the secretariat as of March 31,
2005.3 We also analyzed local fund agents' assessments to determine how
often grant managers documented disbursement decisions. In addition, we
interviewed Global Fund officials in Washington, D.C., and in Geneva,
Switzerland, and officials from the Departments of State and Health and
Human Services (HHS), and the U.S. Agency for International Development
(USAID).

Methodology for Describing the Global Fund's Recent Refinements for
Managing Grants and Improving Their Performance

To describe the Global Fund's recent refinements for managing grants and
improving their performance, we reviewed Global Fund documents including
The Global Fund Operational Policy Manual and related guidance documents
and organization charts, and job descriptions for the positions of local
fund agent officer, country coordinating mechanism (CCM) coordinator,
program officer, and fund portfolio manager (grant manager). We also
examined Global Fund papers, including the Discussion Paper on the Core
Business Model for a Mature Global Fund; Update on New Measures of
Performance and Early Warning System; Update on the

3As of that date, the secretariat had provided us with documentation for
25 grants, which went to recipients in Argentina, Benin, Burundi, China,
Cuba, Ghana, Haiti, Honduras, India, Laos, Madagascar, Moldova, Mongolia,
Morocco, Panama, Rwanda, Senegal, Tajikistan, and Tanzania/Zanzibar.

Appendix IEURObjectives, Scope, and MethodologyEUR

Global Fund Information Management Platform; Revised Guidelines on the
Purpose, Structure and Composition of Country Coordinating Mechanisms and
Requirements for Grant Eligibility; and Performance Standards and
Indicators for CCM Monitoring. In addition, we reviewed the report
Investing in the Future: The Global Fund at Three Years and the Monitoring
and Evaluation Toolkit for HIV/AIDS, Tuberculosis, and Malaria. We also
reviewed documents from a March 2005 CCM workshop conducted in Zambia.
Further, we reviewed documents obtained during fieldwork in Kenya,
conducted follow-up correspondence with CCM members in Kenya, and reviewed
Global Fund documents concerning grants to Ukraine. Additionally, we
interviewed officials from the Global Fund, the Departments of State and
HHS, USAID, UNAIDS, and WHO.

We conducted our work from June 2004 through March 2005, in accordance
with generally accepted government auditing standards.

Appendix II

Methodologies Used to Compare Grants' Disbursements

This appendix provides descriptive information related to the 130 grants
that had received their first disbursements from the Global Fund on or
before December 31, 2003, and the results of analyses we undertook to
determine whether some types of grants had disbursed a larger percentage
of their 2-year funds than others and to estimate the number of
disbursements that were made in a timely fashion. Disbursements refer to
those from the Global Fund to the principal recipient, not from the
principal recipient to subrecipients. Data were current as of February 4,
2005.

Grant Characteristics Table 2 shows selected characteristics of the 130
grants we reviewed.

Table 2: Selected Characteristics of 130 Global Fund Grants

Category Number Percentagea

                            Principal recipient type

Civil society (private sector and faith-based
organizations) 20

Civil society (NGO) 7

Ministry of finance 11

Ministry of health 52

Government entities other than ministries of health or finance 20

Multilateral organizationb 20

                                  Disease type

HIV/AIDS 70

Malaria 32

Tuberculosis 28

                                  Grant sizec

                                 <$2 million              25             19.2 
                               $2-$5 million              43             33.1 
                              $5-$10 million              27             20.8 
                                >$10 million              35             26.9 
                     Number of disbursements                 
                                           1              10              7.7 
                                           2              34             26.2 
                                           3              34             26.2 

        Appendix II Methodologies Used to Compare Grants' Disbursements

(Continued From Previous Page)

                      Category               Number               Percentagea 
                             4                   34 
                             5                    6 
                             6                    8 
                             7                    4 

                         Percentage of grant disbursed

<50 40

70-74 46

75-99 27

100 17

                                Total 130 100.0

Source: GAO analysis of Global Fund data.

aPercentages may not add to 100.0 due to rounding.

bThe UN Development Program and other UN organizations.

cGrant size refers to the 2-year, or phase 1, amount. The grants ranged in
size from roughly $221,000 to more than $55 million.

No Significant Differences Table 3 shows the performance of different
grant types with respect to between Grant Type and receiving
disbursements. Some of these characteristics varied by type of Percentage
of Grant grant, although many of the differences were not significant.

Disbursed

Table 3: Information Related to Disbursements for Different Types of
Grants, Defined by Principal Recipient Type, Disease Type, and Grant Size

Grant category

Number of observations

Average days since first disbursementa

                        Average number of disbursements

                                          Average percent of grants disbursed

Average percentage of grants disbursed (of grants remaining after first
disbursement)b

                                            Percent of grants fully disbursed

Principal recipient type

Civil society (private                            
      sector and faith                               
           based)           20  575.70  3.95  71.94                65.96 10.0 
     Civil society (NGO)     7  560.86  3.29  76.39                72.23 42.9 
     Ministry of finance    11  463.73  2.09  56.84                 45.46 9.1 
     Ministry of health     52  557.58  3.35  62.52                 50.59 7.7 
                               Page 47               GAO-05-639 Global Health 

        Appendix II Methodologies Used to Compare Grants' Disbursements

                         (Continued From Previous Page)

                                                                Average        
                                                                percentage     
                           Average days                 Average of grants      
                                                                disbursed      
                                  since    Average      percent (of grants       Percent 
                                           number            of remaining             of 
                Number of         first            of              after first    grants 
                                                         grants                    fully 
Grant     observations disbursementa disbursements disbursed disbursement)b disbursed 
  category                                                                     
Other               20        548.85          2.60     64.86          50.33 
 government                                                                    
Multilateral                                                                   
organization           20        579.70          3.55     68.67          62.92 

Disease type

HIV/AIDS 70 548.93 3.19 64.65 51.52

Malaria 32 559.59 3.06 65.11 58.18

Tuberculosis 28 563.36 3.61 68.28 62.22

Grant size

       <$2 million         25       590.20      3.24      80.89      59.27    
      $2-$5 million        43       549.37      3.33      64.93      58.90    
      $5-$10 million       27       535.26      3.30      58.16      50.16    
       >$10 million        35       550.74      3.11      61.03      53.07    
          Total            130      554.66      3.25      65.54      55.53    

Source: GAO analysis of Global Fund data.

aA larger number of days since the first disbursement implies an earlier
first disbursement.

bThe percentage of grants disbursed (of grants remaining after first
disbursement) was calculated based on 128 grants, rather than 130, because
two grants were completely disbursed after a single disbursement.

Ministries of finance, on average, made a smaller number of disbursements
and disbursed a lower percentage of their grants, although they also had
made their first disbursements later and therefore had less time to make
disbursements. Similarly, larger grants disbursed lower percentages of
their grant amounts than smaller grants; but again, differences in the
time elapsed make it difficult to know whether these differences reflect
anything more than the time they had to make disbursements. Differences in
disbursements, percentages disbursed, and average days since the first
disbursement were insignificant across grants dedicated to the different
types of diseases.

Methodologies Used 	To determine whether the differences in the percentage
disbursed varied by type of grant, we used ordinary regression techniques.
The Global Fund also analyzed grant disbursements, reporting in March 2005
that disbursements are indicative of performance. Our analysis differed
from the Global Fund's in that we looked at the percentage of the 2-year
grant amount disbursed since the first disbursement, whereas the Global
Fund

Appendix II Methodologies Used to Compare Grants' Disbursements

looked at the percentage that was disbursed relative to the percentage
that was expected to be disbursed since the first disbursement. Because
the actual effect of time turns out to be nonlinear-meaning that although
time elapsed since the first disbursement has a significant effect on the
percentage disbursed, that effect decreases over time-we estimated the
effect of time directly before estimating differences in the percentage
disbursed across different types of grants.1

We fit bivariate regression models (models 1-4 in table 4) to estimate and
test the significance of the gross effects of time since first
disbursement, principal recipient type, disease type, and grant size2 (or
the effects of each of these factors, ignoring all others) and a
multivariate regression model (model 5) to estimate the net effects of
each (or the effects of each after controlling for the effects of the
others). Table 4 shows these results.

Table 4: Regression Models Showing Effects of Various Factors on
Percentage of Grants Disbursed, Minus the Amount Disbursed in First
Disbursement

                                     Model

1234

                          Time (in days) .785a .830 a

              Time-squared -.001a -.001a Principal recipient type

                              Civil NGO        6.27                     12.20 
                    Ministry of finance       -20.50        
                     Ministry of health       -15.37b                  -10.67 
                       Other government       -15.63                   -10.09 
                           Multilateral        -3.04                     -.24 

1Further, the expected disbursement, as the Global Fund calculates it, is
a straight-line function of the time since the first disbursement rather
than a "stair step" function reflecting the discrete points in time at
which each disbursement was made.

2In our regression models, we categorize grant size rather than treat it
as a continuous variable. We did this in part to simplify the
interpretation of results, since preliminary analyses (not shown) revealed
that the effect of grant size was not linear or monotonic. For example,
while grants of $2 to $5 million disbursed a lower percentage of the
residual grant amount than grants of less than $2 million, and grants of
$5 to $10 million disbursed a lower percentage of the residual grant
amount than grants of $2 to $5 million, grants greater than $10 million
did not disburse a lower percentage of the residual amount than grants of
$5 to $10 million.

Appendix II Methodologies Used to Compare Grants' Disbursements

(Continued From Previous Page)

                                  Model 12345

Disease type

                               Malaria 6.65 6.36

                            Tuberculosis 10.70 8.88

Grant size

                $2-$5 million                                 -0.37      9.45 
               $5-$10 million                                 -9.10      2.25 
                 >$10 million                                 -6.20      1.85 
                     Constant    -194.44  65.96    51.52    59.27     -209.50 
                  F-statistic      12.87   1.34     1.06       0.49      2.80 
                 Significance      <.001      .25      .35      .69      .002 
           Variance explained      17.1%   5.2%     1.7%       1.2%     22.6% 

Source: GAO analysis of Global Fund data.

Note: The reference (and omitted) categories for the principal recipient
type, disease type, and grant size variables are other civil society
grants, HIV/AIDS grants, and grants for less than $2 million,
respectively.

aDenotes effects which are significant at the .05 level.

bDenotes effects which are significant at the .10 level.

Model 1 in table 4 shows the effect that time, or days between the first
disbursement and February 2, 2005, has on the percentage of the grant
disbursed. The significant time-squared term indicates that the effect of
time is nonlinear. This nonlinearity makes the interpretation of the time
coefficients somewhat less straightforward, but the positive time
coefficients in table 4 indicate, not surprisingly, that grants that have
had more time to make disbursements have disbursed a larger percentage of
the 2-year grant amount remaining after the first disbursement. The
negative coefficients associated with the squared term means that over
time, time is less of a factor, or that the difference in the percentage
disbursed between 100 days and 200 days is greater than the difference
between 300 and 400 days. The sizable F-statistic at the base of the
column for model 1 attests to the significance of the effect of time, and
this nonlinear effect of time explains 17 percent of the variation in the
amount disbursed of the total remaining after the first disbursement.

Models 2, 3, and 4 estimate the gross effects of principal recipient type,
disease type, and grant size on the percentages disbursed, after
subtracting the amount of the first disbursement. These differences are
estimated using dummy variables to indicate the differences between the
grant categories named in the table and the omitted referent category
(civil society for

        Appendix II Methodologies Used to Compare Grants' Disbursements

principal recipient type, HIV/AIDS for disease type, and less than $2
million for grant size). The constants for each of these models reflect
the percentages disbursed for grants in the referent categories, and the
coefficients indicate the differences between the percentages for the
categories in the table and the percentages for the referent categories.
The overall percentage disbursed was 21 points lower for ministry of
finance grants, and roughly 15 points lower for ministry of health grants
and other government grants, than for civil society grants. These
differences reflect the results of ignoring, rather than controlling for,
differences in time since first disbursement, grant size, and disease
targeted.

Model 5 estimates all of these effects simultaneously and, as such,
provides us with net effect estimates, or estimates of each effect,
controlling for the others. It shows that the time each grant has had
since its first disbursement is the principal determinant of the amount
disbursed. Grants that made their first disbursement earlier disbursed
larger amounts of their remaining 2-year awards. After these effects were
controlled for, the differences between principal recipient types in the
percentages disbursed became smaller than they appeared before controls;
and the only difference, which had appeared marginally significant before
controls, became insignificant afterward. We found no significant
differences by disease type or grant size when looking at either the gross
or net effects.

Timeliness of We also looked at the extent to which disbursements were
made in a timely

Disbursements	fashion (i.e., in 135 days or less). As table 5 shows, 35
percent of all disbursements were timely, or within 135 days, and the
extent to which disbursements were timely was greater for later
disbursements than for earlier disbursements. The number of timely
disbursements is too small at most stages for us to determine whether
timeliness varies across grant types.

Appendix II Methodologies Used to Compare Grants' Disbursements

Table 5: Percentages of Timely and Untimely Disbursements, by Disbursement
Stage

                           Timeliness of disbursement

                 Disbursement                                  
                        stage     Untimely          Timely              Total 
                          2nd       94.0%            6.0%              100.0% 
                                            109       7                   116 
                          3rd       58.0%           42.0%              100.0% 
                                             47       34       
                          4th       39.2%           60.8%              100.0% 
                                             20       31       
                          5th       11.8%           88.2%              100.0% 
                                              2       15       
                          6th       41.7%           58.33%             100.0% 
                                              5       7        
                          7th       0.0%            100.0%             100.0% 
                                              0       4        
                        Total       65.1%           34.9%              100.0% 
                                            183       98                  281 

                   Source: GAO analysis of Global Fund data.

                                  Appendix III

                         Comments from the Global Fund

Appendix IIIEURComments from the Global FundEUR

Appendix IIIEURComments from the Global FundEUR

Appendix IIIEURComments from the Global FundEUR

Appendix IIIEURComments from the Global FundEUR

Appendix IV

Joint Comments from the Departments of State and Health and Human
Services, and the U.S. Agency for International Development

Appendix IV Joint Comments from the Departments of State and Health and
Human Services, and the U.S. Agency for International Development Appendix
IV Joint Comments from the Departments of State and Health and Human
Services, and the U.S. Agency for International Development

Appendix V

                     GAO Contact and Staff Acknowledgments

GAO Contact David Gootnick (202) 512-3149

Staff 	In addition to the person named above, Candace Carpenter, Martin de
Alteriis, David Dornisch, Cheryl Goodman, Kay Halpern, Bruce Kutnick,

Acknowledgments	Reid Lowe, Jerome Sandau, Douglas Sloane, Julie Trinder,
and Tom Zingale made key contributions to this report.

GAO's Mission	The Government Accountability Office, the audit, evaluation
and investigative arm of Congress, exists to support Congress in meeting
its constitutional responsibilities and to help improve the performance
and accountability of the federal government for the American people. GAO
examines the use of public funds; evaluates federal programs and policies;
and provides analyses, recommendations, and other assistance to help
Congress make informed oversight, policy, and funding decisions. GAO's
commitment to good government is reflected in its core values of
accountability, integrity, and reliability.

Obtaining Copies of The fastest and easiest way to obtain copies of GAO
documents at no cost

is through GAO's Web site (www.gao.gov). Each weekday, GAO postsGAO
Reports and newly released reports, testimony, and correspondence on its
Web site. To Testimony have GAO e-mail you a list of newly posted products
every afternoon, go to

www.gao.gov and select "Subscribe to Updates."

Order by Mail or Phone	The first copy of each printed report is free.
Additional copies are $2 each. A check or money order should be made out
to the Superintendent of Documents. GAO also accepts VISA and Mastercard.
Orders for 100 or more copies mailed to a single address are discounted 25
percent. Orders should be sent to:

U.S. Government Accountability Office 441 G Street NW, Room LM Washington,
D.C. 20548

To order by Phone:	Voice: (202) 512-6000 TDD: (202) 512-2537 Fax: (202)
512-6061

To Report Fraud, Contact:
Waste, and Abuse in Web site: www.gao.gov/fraudnet/fraudnet.htm

E-mail: [email protected] Programs Automated answering system: (800)
424-5454 or (202) 512-7470

Congressional	Gloria Jarmon, Managing Director, [email protected] (202)
512-4400 U.S. Government Accountability Office, 441 G Street NW, Room 7125

Relations Washington, D.C. 20548

Public Affairs	Paul Anderson, Managing Director, [email protected] (202)
512-4800 U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, D.C. 20548
*** End of document. ***