The changing donor landscape of health sector aid to Vietnam: A qualitative case study

Social Science & Medicine 132 (2015) 165e172

Contents lists available at ScienceDirect

Social Science & Medicine journal homepage: www.elsevier.com/locate/socscimed

The changing donor landscape of health sector aid to Vietnam: A qualitative case study Sarah Wood Pallas a, *, Thi Hai Oanh Khuat b, Quang Duong Le b, 1, Jennifer Prah Ruger a, 2 a b

Yale School of Public Health, New Haven, CT, USA Center for Supporting Community Development Initiatives (SCDI), Hanoi, Viet Nam

a r t i c l e i n f o

a b s t r a c t

Article history: Available online 14 March 2015

The study objective was to identify how donors and government agencies in Vietnam responded to donor proliferation in health sector aid between 1995 and 2012. Interviews were conducted with key informants from donor agencies, central government, and civil society in Hanoi in 2012 (n ¼ 34 interviews), identified through OECD Creditor Reporting System data, internet research, and snowball sampling. Interview transcripts were coded for key themes using the constant comparative method. Documentary materials were used in triangulation and validation of key informant accounts. The study identified a timeline of key events and key themes. The number of donors providing health sector aid to Vietnam increased sharply during the late 1990s and early 2000s, then leveled off and declined between 2008 and 2012. Reasons for donor entry included Vietnam's health needs, perceptions of health as less politically sensitive, and donor interests in facilitating market access. Reasons for donor withdrawal included Vietnam's achievement of middle-income status, the global financial crisis, and donors' shifting global priorities. Key themes included high competition among donors, strategic actions by government to increase its control over aid, and the multiplicity of government units involved with health sector aid. The study concludes that central government and donor agencies in Vietnam responded to donor proliferation in health sector aid by endorsing aid effectiveness policies but implementing these policies inconsistently in practice. Whereas previous literature has emphasized donor proliferation's transaction costs, this study finds that the benefits of a large number of less coordinated donors may outweigh the increased administrative costs under certain conditions. In Vietnam, these conditions included relatively high capacity within government, low government dependence on aid, and government interest in receiving diverse donor recommendations. Vietnam's experience of donor proliferation followed by donor withdrawal illustrates a trajectory that other countries may experience as they transition from low-to middle-income status. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Vietnam Development aid Donor proliferation Aid effectiveness Global health

1. Introduction Development assistance for health in low- and middle-income countries has increased fourfold since 1990, accompanied by a proliferation in the number of donor organizations delivering such aid (Murray et al., 2011; Ravishankar et al., 2009; McCoy et al.,

* Corresponding author. Department of Health Policy & Management, Yale School of Public Health, P.O. Box 208034, New Haven, CT 06520-8034, USA. E-mail address: [email protected] (S.W. Pallas). 1 Present address: Sustainable Health Development Center e VietHealth, Hanoi, Viet Nam. 2 Present address: Perelman School of Medicine and Leonard Davis Institute of Health Economics, University of Pennsylvania, Philadelphia, PA, USA. http://dx.doi.org/10.1016/j.socscimed.2015.03.027 0277-9536/© 2015 Elsevier Ltd. All rights reserved.

2009). Donor organizations providing health sector aid include government agencies (e.g., UK Department for International Development), international organizations (e.g., World Bank), foundations and charitable groups (e.g., Bill and Melinda Gates Foundation), corporations (e.g., GlaxoSmithKline), and publicprivate partnerships (e.g., GAVI Alliance). Although there is evidence that health sector aid has contributed to population health improvements in some settings (Levine and Kinder, 2004; Sachs, 2005), previous literature has also identified costs that may arise when aid is delivered through a large number of donors, including higher administrative costs for recipient country governments, duplication and fragmentation of donor efforts, loss of recipient country government staff to better paid positions in donor-funded projects, lower performance incentives for donors and recipients,

166

S.W. Pallas et al. / Social Science & Medicine 132 (2015) 165e172

and greater opportunities for corruption (Easterly, 2002; Organisation for Economic Co-Operation and Development [OECD], 2003; Rogerson, 2005; Acharya et al., 2006; Knack and Rahman, 2007; Djankov et al., 2009; Lawson, 2009). These findings from previous literature suggest that donor proliferation may be limiting the potential health impact of aid delivered to low- and middle-income countries. To address such concerns, a series of international policy agreements on aid effectiveness have been adopted over the past decade, with the 2005 Paris Declaration on Aid Effectiveness representing a prominent milestone (OECD, 2004, 2008, 2011a; United Nations Development Program [UNDP], 2002). These international agreements, endorsed by donors and recipient country governments, express a policy consensus around a core set of aid effectiveness principles: donor harmonization, donor alignment with recipient country systems, recipient country ownership of the development agenda, mutual accountability between donors and recipients, and managing aid for development results (Rogerson, 2005). These principles are intended to reduce the administrative costs of managing aid-funded activities for both donors and recipients, thereby permitting more time and money to be spent on intervention implementation and improving the translation of aid into results (OECD, 2003; Rogerson, 2005; Knack and Rahman, 2007; Vandeninden and Paul, 2012). Although aid coordination is not a new concept (Buse and Walt, 1996, 1997; Walt et al., 1999; Easterly, 2007), the current global landscape of health sector aid is distinctive in terms of both the extent of donor proliferation and the consensus around a set of aid effectiveness principles as the appropriate policy response. Given this new global health landscape, understanding how donors and recipient countries are responding to donor proliferation in health sector aid in specific country settings is important. There are limited cross-country quantitative data on this question, as the Organisation for Economic Co-Operation and Development (OECD) collects data on implementation of the Paris Declaration for aggregate development aid but not health sector aid specifically (OECD, 2010). Existing quantitative indicators also do not give insight into how or why aid effectiveness principles are adopted or resisted in different recipient countries. For example, donors and recipients may resist implementing aid effectiveness principles because aid coordination activities impose new costs (Rogerson, 2005; Lawson, 2009; Vandeninden and Paul, 2012), weaken the recipient government's bargaining position (Easterly, 2002; Rogerson, 2005; Acharya et al., 2006), or increase aid volatility (Fielding and Mavrotas, 2008). More evidence is needed about the factors that influence donor and recipient country government responses to donor proliferation in health sector aid, and why aid effectiveness principles are or are not applied in different settings, in order to maximize the population health impacts of aid. To this end, this paper contributes to the literature by presenting findings from a retrospective qualitative case study of the changing donor landscape in health sector aid to Vietnam. Vietnam had 29 donors providing health sector aid in 2008, an increase of 16 donors relative to 1995 (Supplementary Appendix Figure 1) and more than any of the other 155 countries that received health sector aid during this period, as measured by a count of net donors that reported to the OECD's Creditor Reporting System under general and basic health aid, population and reproductive health aid (including HIV/ AIDS), or water and sanitation aid (OECD, 2011b). Between 1995 and 2010, total aid commitments across these subsectors grew fivefold, accompanied by donor proliferation in each subsector: from eight to 25 donors in general and basic health, from four to 16 donors in population and reproductive health, and from eight to 18 donors in water and sanitation, with some donors providing aid across multiple subsectors (Supplementary Appendix Figures 1, 2).

During this period, Vietnam also developed from a low-income to a middle-income country, doubling its per capita income and demonstrating progress in multiple population health indicators (Supplementary Appendix Table 1) (World Bank, 2012; World Health Organization, 2012). The case of Vietnam is typical of other low- and middle-income countries in some regards e for example, its history of European colonialism, civil conflict and reconstruction, and low-income status at the start of the donor proliferation period e but distinctive in other ways, such as its political structure, attractiveness to foreign investors, and rapid economic growth rate. Examining Vietnam's experience in detail can illuminate how such contextual features shape donor and government responses to donor proliferation in health sector aid. Practitioners, policy makers, and researchers may find the study results useful in understanding factors that influence health sector aid allocations and in designing responses to donor proliferation in health sector aid. 2. Methods The study used a retrospective qualitative case study design, which was selected because it is well-suited to investigating complex social phenomena with multiple causes in real-world settings (George and Bennett, 2005; Yin, 2009; Creswell, 2007), such as donor proliferation in health sector aid, for which randomization for hypothesis testing is not possible. The case was defined as Vietnam's experience of donor proliferation in health sector aid between 1995 and 2012, covering the years before, during, and after the global health aid scale-up and major international aid effectiveness agreements. Data for the case study were collected from key informant interviews, documents from donor and recipient organizations, and published literature. Triangulation across documentary evidence and firsthand accounts in a case study design was employed to gain insight into how donors and recipients in Vietnam responded to donor proliferation in health sector aid, both publicly and privately, and why they responded as they did. The sampling frame for key informants included individuals with experience in Vietnam's central government, in foreign aid donor agencies, or in civil society groups involved in aid-funded health projects in Vietnam. Key informants were identified by mapping the organizations providing or receiving health sector aid in Vietnam between 1995 and 2012, as reported by donors to the OECD Creditor Reporting System (OECD, 2011b), and through referrals, i.e., snowball sampling (Cresswell, 2007; Rubin and Rubin, 1995). For each donor identified in the Creditor Reporting System as having provided health sector aid to Vietnam, internet research was conducted to map the specific projects funded with this aid, and to identify the offices and individuals (e.g., project officers, contracting officers, chiefs of party, unit directors) that had administered the aid on the donor and recipient sides. The authors (KTHO, LQD, SWP) contacted organizations, offices, and, where possible, individuals identified through the aid flow mapping by email and phone to invite their participation in the study. Given the retrospective nature of the study, some individuals involved with health sector aid projects in the past had moved on to other roles; whenever possible, such individuals were contacted at their new locations, or a referral was requested to another appropriate person with knowledge of the project. Key informants were also asked for referrals to others who might be knowledgeable about the study topic. All key informants were requested to participate in their personal capacities rather than as official spokespeople for any current or former employer. Purposive sampling was used to identify a diverse range of key informants knowledgeable about the case of donor proliferation in health sector aid to Vietnam,

S.W. Pallas et al. / Social Science & Medicine 132 (2015) 165e172

continuing until no new concepts emerged from successive interviews (i.e., until theoretical saturation was achieved) (Creswell, 2007; Strauss and Corbin, 1998). The resulting key informant sample was composed of individuals based in Hanoi with firsthand experience in designing, implementing, or overseeing health sector aid activities in Vietnam between 1995 and 2012 in a donor agency, Vietnam central government office, or non-governmental or private sector organization. The study protocol was reviewed by the Yale University Human Subjects Committee (protocol # 1203009835; exemption from further review was granted under 45 CFR 46.101(b) (2)). All key informants were provided with the study's informed consent form and written or verbal informed consent was obtained from all subjects prior to the interview. The informed consent procedure specified that key informants' identities would be kept confidential and would not be associated with their interview responses in publication of the study findings. Semi-structured interviews averaging one hour in length were conducted by the authors (SWP and LQD) in Hanoi between May and August 2012, with two interviews conducted by phone with informants no longer located in Vietnam. Interviews were conducted in English, with two interviews conducted in Vietnamese with simultaneous translation by one of the authors (LQD). The interviewers followed an interview guide with a standard series of open-ended primary and follow-up questions. Interview locations were chosen by key informants and included informants' offices, the study team's offices at the Center for Supporting Community Development Initiatives, s. If the informant agreed, the and public locations such as cafe interview was digitally recorded; otherwise, the authors took notes. No repeat interviews were conducted. All recordings and notes were transcribed and coded by two independent coders (SWP and a research assistant) for key themes using the constant comparative method (Creswell, 2007), with codes emerging from the transcripts added to an initial start list of codes drawn from the study's research questions and background literature. The start list of codes included the concepts of donor entry, donor exit, aid volume increases, aid volume decreases, donor size distribution, donor competition, recipient competition, recipient government control, donor control, donor coordination, recipient government prioritization of health, aid effectiveness, accountability, and aid package features; each start code included sub-codes for examples, causes, and consequences related to the concept. Differences between coders were resolved through discussion. Atlas.ti version 7 was used to manage the coded data. Documentary materials were also collected from donor and government agencies during key informant interviews and via online research to triangulate with key informant accounts in identifying key themes and the timeline of responses to donor proliferation in health sector aid to Vietnam. The analysis approach sought to answer the study question of how donors and recipients responded to donor proliferation in health sector aid to Vietnam between 1995 and 2012 by constructing a composite timeline of events grouped into thematic periods, and to answer the study question of why donors and recipients responded as they did by identifying key themes e i.e., repeated causal claims or accounts from key informants e that helped explain the trajectory of events. Key informants did not provide feedback on the transcripts or analysis results. The authors brought complementary backgrounds to the study. Two of the authors (KTHO and LQD) were Vietnamese with training in medicine and public health, prior experience working in government in Vietnam, and currently working in a civil society organization in Vietnam that received health sector aid. Two of the authors (SWP and JPR) were American with training in the social sciences and health policy, prior experience working with donor

167

organizations that provide health sector aid, and currently working as faculty or studying in a U.S. university. The study was conducted as part of the PhD dissertation research of SWP, advised by JPR, and supervised in the field by KTHO. These research team characteristics permitted diverse perspectives relevant to the phenomenon of donor proliferation in health sector aid to be considered in the study design and analysis, such as local and foreign, public health and political economy, policy and implementation, and academic and practitioner. The identities and affiliations of the authors may have also facilitated access to key informants by establishing credibility and accountability. 3. Results 3.1. Interviewee sample In total, 34 interviews were conducted with 41 key informants (some interviews included two informants), out of 48 interviews requested (response rate: 71%). Seven interviews were conducted with key informants with experience in Vietnam's central government agencies, 23 interviews were conducted with key informants with experience in donor agencies, and four interviews were conducted with key informants with experience in civil society or the private sector. Of the key informants, 54% were Vietnamese, including Vietnamese nationals working in international or foreign government aid agency offices in Hanoi. The larger proportion of key informants from donor agencies relative to government reflects the fact that many donor organizations interface with the same Vietnamese government units, as well as a slightly higher response rate for key informants from donor agencies relative to government (70% versus 63%). 3.2. Timeline of key events in donors' and recipients' responses to donor proliferation in health sector aid in Vietnam Early donor proliferation and efforts to coordinate aid. Although a few donors (such as Sweden and United Nations agencies) provided health aid to Vietnam in the 1970s and 1980s, the number of donors to Vietnam's health sector increased rapidly in the 1990s following Vietnam's Doi Moi economic reforms and the lifting of the U.S. trade embargo. Early health sector aid was focused in the areas of water and sanitation, population and reproductive health, and primary health care. Key informants described the 1990s as a time of rapid and chaotic growth in the number of donor-funded projects, with donors duplicating each others' efforts unknowingly. In the late 1990s, the number of health aid donors continued to increase even as the volume of aid commitments leveled off, especially for general and basic health and population and reproductive health aid (Supplementary Appendix Figures 1, 2). Key informants suggested that donors proliferated in the health sector because Vietnam had clear health needs, improving population health was a priority for Vietnam's government, and health was perceived as a less politically sensitive sector. Both the Government of Vietnam and donor organizations made some efforts towards aid coordination during this early phase of donor proliferation: the Ministry of Health (MOH) developed a directory of health sector donor activities and Sweden and the Netherlands sponsored feasibility studies on donor coordination mechanisms and began consultative meetings with the MOH, along with the World Bank and World Health Organization. During the late 1990s and early 2000s, donors and the MOH also considered adopting a Sector Wide Approach (SWAp) that would pool donor funds into a common account to support a sector-wide strategy. Key informants reported that the MOH decided not to pursue a SWAp because of capacity constraints in human resources and

168

S.W. Pallas et al. / Social Science & Medicine 132 (2015) 165e172

information systems, and anticipated difficulties in securing cooperation from other parts of government on which the health sector strategy results would depend, conclusions informed by donor-funded feasibility studies and MOH officials' visits to African countries that had adopted SWAps. Entry of high-volume disease-specific donors. Donor proliferation continued in the early 2000s with the arrival of the newly created Global Fund to Fight HIV/AIDS, Tuberculosis, and Malaria and U.S. President's Emergency Plan for AIDS Relief (PEPFAR), which increased funding for HIV/AIDS within Vietnam's health sector (Supplementary Appendix Figure 2b). Key informants perceived the Global Fund and PEPFAR as operating largely independently of each other and other donors, with Global Fund grants channeled primarily through disease-specific units within the Ministry of Health and PEPFAR funds channeled primarily through American contractors working at the provincial level. Key informants contrasted the increased opportunity for country ownership through the Global Fund e which required a Country Coordinating Mechanism composed of donor, government, and civil society representatives to jointly develop and oversee grants e with the more donor-driven approach of PEPFAR. Institutionalization of aid effectiveness policies and processes. During the 2000s, the Government of Vietnam participated actively in OECD-sponsored international aid effectiveness policy forums (OECD, 2003; Dodd and Olive, 2011). After signing the Paris Declaration on Aid Effectiveness (OECD, 2008), the Government of Vietnam issued the Hanoi Core Statement, which translated the Paris Declaration commitments into national-level targets that Vietnam's donors endorsed in June 2005 (Dodd and Olive, 2011). In 2006, Vietnam's Prime Minister signed Decree 131, a new law stipulating how aid was to be designed, negotiated, managed, and evaluated, which assigned primary responsibility for aid coordination to the Ministry of Planning and Investment (MPI) (Government of the Socialist Republic of Vietnam, 2006). In parallel to these general aid policy developments, aid coordination in Vietnam's health sector was also becoming more institutionalized. A formal Health Partnership Group (HPG) composed of donor, government and civil society representatives was established in 2004 and began meeting twice a year, supplemented by Technical Working Groups in specific health domains. A Joint Annual Health Review (JAHR) process was initiated in 2006e2007 with donor funding and a dedicated MOH team, resulting in annual publications of key population health indicators and in-depth analysis of particular health issues in Vietnam that fed into the Government of Vietnam's Five Year Plans. In 2008, the Ministry of Health developed a Statement of Intent that applied the principles of the Paris Declaration and Hanoi Core Statement to the specific sector of health, which was formally adopted by the Health Partnership Group in 2009 (Dodd and Olive, 2011). Achievement of middle-income status and donor withdrawal. In 2008, the Dutch government announced that it would be withdrawing its aid from Vietnam beginning in 2012 in anticipation of Vietnam attaining middle-income country status. The Swedish government followed with a similar announcement shortly thereafter. Key informants suggested that Vietnamese government officials were surprised by these announcements and did not initially believe that the Dutch and Swedish governments would withdraw aid as both governments were among Vietnam's earliest donors. In 2009 and 2010, the United Kingdom, Canada, Switzerland, and Norway all announced their intention to withdraw aid to Vietnam in the next few years. In 2010, Vietnam became a middle-income country according to World Bank standards. Other donors, such as Australia, Germany, and Korea, began reallocating their aid from health to other sectors in Vietnam during this period due to shifts in their headquarters'

priorities or because they perceived the health sector to have too many donors. Key informants suggested that the global financial crisis had also prompted some donors to withdraw or scale back operations in Vietnam during this period, such as the Rockefeller Foundation (due to endowment losses) and the Global Fund (which adopted a new policy limiting its giving to middle-income countries in the face of global resource mobilization shortfalls). Key informants reported that PEPFAR was planning to reduce its funding levels to Vietnam in future years, with the expectation that the Government of Vietnam would assume more HIV/AIDS program costs now that the country had attained middle-income status. Some key informants expressed concerns that HIV/AIDS activities for socially marginalized at-risk populations such as injection drug users, sex workers, or men who have sex with men would not be supported by the Government of Vietnam and would end if donors withdrew support. The scope of future government support for HIV/AIDS services, including through the national health insurance scheme, was an area of active policy debate at the time of data collection. Key informants also reported that aid-recipient civil society organizations were compensating for donor withdrawal by seeking alternative funding sources (e.g., local businesses and international foundations), developing revenue-generating services, and evolving their health-specific activities into platforms for service delivery in other sectors with ongoing aid flows. As a consequence of donor withdrawal, key informants anticipated a reduction in the number of civil society organizations active in the health sector and in these organizations' ability to engage in health policy discussions with government. 3.3. Key themes from donors' and recipients' responses to donor proliferation in health sector aid in Vietnam The study identified three key themes from donors' and recipients' responses to donor proliferation in health sector aid in Vietnam: (i) high levels of competition among donors, (ii) strategic actions by government to increase its control over aid relative to donors, and (iii) multiplicity of government units involved with health sector aid. High levels of competition among donors. Competition among donors was perceived by key informants as a routine feature of donor activity in Vietnam since the period of large-scale donor entry into the country in the mid-1990s, and was not perceived to be unique to health sector aid. Key informants described how donors competed with one another to gain visibility with key audiences in Vietnam and to take credit for successful health projects in order to report successes back to their headquarters: So AusAID did seem more political, but you know, if you build a bridge we can say, “This bridge was built with funds from the people of Australia.” But you put some money into developing a system for health insurance, then there's no way an Australian coming to Vietnam would see that. (Interview 10) And from the donor's side, yes, we wanted to report as well to our headquarters what we were able to do in the country. And there were times we tend to focus more on reporting on our own achievement in our single project instead of looking at how we contribute to the successfulness of, the success of the HIV prevention here in the country. (Interview 12) Bilateral donors also competed with one another for trade advantages for their domestic businesses seeking to enter the Vietnamese market. As one key informant described regarding the 1990s:

S.W. Pallas et al. / Social Science & Medicine 132 (2015) 165e172

I think at the same time there were still a lot of bilateral donors and a lot of them with their financial interests to try to sell equipment were still there. Japan, Korea, probably France, Spain, Italy, and when these countries have equipment that they've provided to various hospitals and you can see, it was mainly just to provide equipment free but paid from the foreign budget, like their foreign policy budget, but they paid to companies from their countries. (Interview 10) According to key informants, donor competition increased as Vietnam became a middle-income country: [W]hat happens when a country becomes a middle income … is that this country becomes an attractive potential business partner for developed countries. And of course the political objective in providing aid becomes influenced by this factor. And as a result there is more tendency for individual donors to profile themselves individually rather than achieving results jointly. (Interview 13) The need to secure future budgets from donors' headquarters and home country legislatures was identified as a root cause of donor competition. As one key informant stated: [M]ost Europeans and Australia and particularly America, … they're linked directly to the foreign affairs people, they have a lot of pressure from business interests to promote Danish business in Vietnam or whatever, so they have to justify their existence in commercial terms, in strategic and diplomatic terms, and in development outcome terms. And you cannot miss any of those targets or you'll get less funding. (Interview 6) Strategic action by government to increase control over aid relative to donors. Key informants reported that the Government of Vietnam acted strategically to enhance its control over health sector aid and to maximize aid funds in the context of donor proliferation. One government strategy mentioned was to resist coordination among donors. Key informants suggested a range of motivations for government resistance to donor coordination, including preserving diversity among donors and maximizing donor funding: But [government], it seems to me that they're quite into divide and conquer and having a diversity. So a couple of times donors have come together wanting to give consolidated feedback and they sort of said “No, we'd rather have 15 opinions.” You know, like, “It's interesting for us to get the diversity of opinions and we've the capacity to manage a whole, like, raft of donors.” When we went into a division of labor exercise, the feedback from government was basically like, “If the only objective is to reduce our transaction costs, we're not interested. We can, it's worthwhile for us to manage the transaction.” (Interview 4) [T]here is sometimes resistance in the government mechanisms to actually have coordination and sharing of information because it's in their own interest to have unilateral or bilateral approaches with each [donor] agency, because each [donor] agency wants to fund different things and they can sometimes double fund or have more control, I would say. (Interview 33) According to key informants, another government strategy was to reallocate donors across provinces within Vietnam. Key informants described scenarios in which a donor desired to provide its aid to a particular province but was encouraged to switch or expand provinces by government interlocutors:

169

Even when we said, “Okay, maybe Danang is rich enough, they have a lot of money, we don't need to include Danang, we should focus on the poor provinces,” the government would say, “No, for the, for political reasons we want to include all the provinces,” and they will all say party line because even the richer provinces still need system development, need capacity building, maybe they need much less on infrastructure. (Interview 5) MOH divide[s], we divide in the project in the group of provinces …. For example, ADB they can take about seven provinces in the north midlands or south midlands, and the World Bank in the north, the northwest, something like this, not separately. That means the MOH decide[s] about this, you see? (Interview 16) Key informants also suggested that the Government of Vietnam used new laws as a strategy to increase its control over development aid and to maximize benefits from aid flows. A frequently cited example among key informants was Decree 131, which required establishment of a separate Project Management Unit (PMU) for every donor-funded project within government even though one of the Paris Declaration indicators was a reduction in parallel PMUs (OECD, 2010). Key informants noted the tension between this requirement and the government's espoused support for aid effectiveness principles: [O]n the one hand they say Vietnam has signed up to harmonization and alignment and all that that brings with it, but on the other hand there is a Decree 131, which any new initiative needs a separate project management unit and should be managed separately and in parallel. So these project management units, again just being flippant, are sometimes a money spinner for different individuals or departments. (Interview 33) But even with Decree 131 we are forced, every project we have is forced to give you a Project Management Unit, which is top-up salaries, arrange things, staffing, et cetera … .It's salary subsidy because the reforms of public administration have not been strong enough … .And government whilst they, MPI, will sprout the Hanoi Core Statement, it's very much about, means “We make all decisions about where we spend your money.” (Interview 3) Key informants also indicated that donor proliferation and accompanying increases in health funding had given the government more leverage, enabling government to more effectively exploit donors' lack of local knowledge and need to disburse funds: In the sense, we [Vietnam] have too many resource[s] and donor[s] sometime[s] have to please the government to use the fund[s]. So, to what I myself observe is that in most of the meeting[s] between the government and the donors, the donors sometime come with a very strong point and argument, but after discussing with the government official, later on I see they are convinced by the government of how to use the fund[s] the way the government want[s]. Because, I guess, because they may not do the study carefully about the situation, so they have to rely on information of the government and the government always tr[ies] to argue their own way, like say[ing], “This is Vietnamese condition and situation and then your idea is good but [it] need[s] to be adjust[ed].” So, basically, by the end there is a kind-of consensus in favor of the government proposal. (Interview 2) Multiplicity of government units involved with health sector aid. Key informants suggested that donors' tendencies towards vertical programs in focal health domains, while valuable in

170

S.W. Pallas et al. / Social Science & Medicine 132 (2015) 165e172

addressing specific disease burdens and building needed capacity, had contributed to the creation of costly parallel government systems that duplicated or displaced those of the Ministry of Health (MOH). According to key informants, the Ministry of Health did not have a monopoly on interactions with health sector donors, many of whom engaged with other central government ministries and provincial governments. Although the Ministry of Health was the lead technical agency at the central government level, key informants noted that its decisions about aid were subject to approval by the Ministry of Planning and Investment (MPI) or Ministry of Finance (MOF): But if, even as the health sector … we want to organize by the SWAp [Sector-Wide Approach], [we] have to starting with MPI, not from MOH....If MOH work[s], nothing happens, you see. Because the funds is also put in the … MPI first and finally put in the MOF, and then put now to MOH. If two big ministries [MPI and MOF] they not agree, nothing will happen with the lower [MOH]. (Interview 16) [Informant A] For example, money from the World Bank or DfID program or PEPFAR or anything, sometime[s] the government want[s] to have the involvement of a number of key ministries in Vietnam including Ministry of Investment and Planning, because they're responsible for planning everything. This is kind of a planning mentality in Vietnam. So they have the very important voice in how the resource should be used … .The other key ministry involved is Ministry of Finance because they say, “Ok, this is your money but you give to us and then we …

not the MOH, and so that's the power of the purse is at the province. (Interview 20) Now I don't think the larger number of [donor] partners has been a big problem, partly because many of these partners don't go through MOH. They deal directly with these provinces, local, or they deal directly with people, like NGOs. Many just say “Hi”, inform the planning people that they're there, what they want to do it gets approved and then they're just on their own. (Interview 5) Provincial People's Committees control budgets for provincial development activities, providing an alternative source of funding for provincial Departments of Health to central MOH funds (Priwitzer, 2012). Key informants indicated that when donor funds were channeled through the MOH, this strengthened the Ministry's position relative to provincial governments: Some of the [people] in the health field in Vietnam are of the view that MOH actually has very little leverage and maybe only controls 10 percent of the funds. So its main leverage actually comes from the [donor] partners … And of course, all the other resources are, or to some extent, already locked up like salaries et cetera, so the provinces are keen to play along with the MOH because the MOH brings, in a sense, extra money [from donors] and in a true sense flexible money, not just for infrastructure but the workshops, giving incentives to staff, organizing trainings, stuff like that. (Interview 5)

[Informant B] … must use [it] … within the financial system of Vietnam.” (Interview 2) 4. Discussion Key informants reported that some donor organizations saw MPI or MOF rather than MOH as their main government interlocutors, especially for the initiation of new projects: I understand that the provincial hospitals, when they [are] looking for the financing of course they will approach both Ministry of Health and MPI. Because MPI is [the] ministry who actually, who is responsible for financing of the, because all these provincial general hospitals they are public sector, public hospitals. So MPI is the boss, is the institution, ministry who [is] looking for financing for the public sector. So they approach MPI. MPI will say, “That's okay, you have need so we'll looking for …” so they will propose to us [donor]. If we refuse I think, I guess that they will propose to Japan or other donors. (Interview 29) At the central level we work mainly with MPI … to be sure project[s] are in the government strategy … and if we think for [a] project it's necessary to have technical assistance we negotiate with the technical ministry and also with … MPI. But after that, as [the] loan agreement will be signed between [donor name] and Minister of Finance … .at the end of the process, Minister of Finance will exam[ine] if financial conditions are correct in terms of debt management and so on. (Interview 7) Key informants also described how some donors chose to bypass the Ministry of Health to work directly with provincial governments, including the elected Provincial People's Councils and their executive implementation bodies, the Provincial People's Committees: And the reason we chose a provincial level approach is because most of the health care spending decisions are made at provincial level, not at the MOH level, central level, but at the provincial level. And funding's usually coming from Provincial People's Committee,

Previous literature suggests that donors and recipient country governments should apply aid effectiveness principles of harmonization (i.e., coordination among donors), alignment (i.e., donors' adherence to recipient countries' priorities and administrative systems), and ownership (i.e., recipient country leadership and responsibility for development strategies) to mitigate the inefficient transaction costs of donor proliferation and increase the development benefits of aid; however, this case study found that donors and government officials responded to donor proliferation in Vietnam's health sector by endorsing aid effectiveness policies while maintaining practices that conflicted with aid effectiveness principles. 4.1. Harmonization Maintaining a diverse and competitive donor landscape served important interests of both donors and the Government of Vietnam. Donors in Vietnam's health sector were interested in certain types of benefits (e.g., visibility, attributable results, trade advantages) that would be potentially more difficult to obtain if their efforts were integrated with those of other donors in a harmonized approach. In addition, key informants reported that the Government of Vietnam wanted to maintain a diversity of ideological and technical perspectives from donors as it determined how to combine its traditional socialist policies with selected marketoriented reforms, and that preserving a large number of uncoordinated donors gave the government more power in individual donor negotiations. Key informants also suggested that the Government of Vietnam had higher human resource capacity compared to other low- and middle-income countries, making the transaction costs of donor proliferation relatively more manageable. Harmonization was also resisted by disease-specific government units that had become ‘donor darlings,’ which might have lost

S.W. Pallas et al. / Social Science & Medicine 132 (2015) 165e172

funds under a more coordinated donor approach that spread aid across a broader range of health domains. 4.2. Alignment The multiplicity of potential government contact points for health sector donors e including the Ministry of Planning and Investment, the Ministry of Health, disease-specific units, provincial health departments, and provincial People's Committees e increased the number of government units with interests in maintaining a fragmented donor landscape. The direct engagement of diverse government units with donors limited the Ministry of Health's control over health sector aid and may have lessened enthusiasm for donor harmonization and alignment, which would have empowered some government entities at the expense of others. For example, the Ministry of Health stood to potentially lose influence with donors if it encouraged alignment to existing government systems that privileged the Ministry of Planning and Investment and the Ministry of Finance. Key informants suggested that officials in Vietnam's Ministry of Planning and Investment would prefer to channel aid through normal government financial and personnel systems whereas the head of a disease program in Vietnam's Ministry of Health would prefer a stand-alone donorfunded project that provided the program with top-up salaries for staff, international trips, or enhanced visibility. For donors, channeling funds through routine Government of Vietnam systems also potentially reduced their ability to account for funds invested and attribute results to their investments for Vietnamese and domestic audiences. 4.3. Ownership The Government of Vietnam enjoyed a relatively strong bargaining position with donors due to its economic context. During the period of health sector donor proliferation in the 1990s and 2000s, Vietnam's economy was growing rapidly, presenting an attractive market for bilateral donors' domestic businesses and turning the Government of Vietnam into a viable borrower for donors who provided aid in the form of loans. Moreover, although Vietnam's volume of health sector aid increased between 1995 and 2012, the Government of Vietnam was not dependent on aid to finance its health sector budget. The Government of Vietnam was therefore better positioned to resist donor demands for harmonization and alignment and to impose conditions e such as for aid allocation to particular provinces or for the creation of Project Management Units e that served the government's institutional and political objectives. 4.4. Explaining the disjuncture between aid effectiveness principles and practice in Vietnam's health sector The findings from the case study indicate that donor agencies and government units in Vietnam were pursuing a range of political, economic, and institutional objectives in addition to maximizing the population health benefits from health sector aid. Harmonization, alignment, and ownership were perceived by donors and government officials as tools that could increase the efficiency and sustainability of aid-funded health programs, but these principles also constrained the achievement of other important objectives and were therefore applied selectively by both donors and government units. In addition, the interests of those donor and government officials who enacted aid effectiveness policies differed in some cases from the interests of those charged with implementation of aid-funded projects. The evidence from Vietnam thus highlights the importance of competing organization- and

171

individual-level objectives within donor agencies and government units in explaining the divergence between aid effectiveness policies and practice. The case study also illustrates that donor and government organizations and their objectives change over time, in turn influencing the evolution of responses to donor proliferation in health sector aid. In Vietnam, events external to the health sector, such as the global financial crisis, political transitions in bilateral donors' home country governments, shifts in donors' global priorities, and Vietnam's achievement of middle-income status, prompted some donors to withdraw from Vietnam's health sector and others to adopt more stand-alone project approaches in order to demonstrate results and increase visibility. Vietnam's experience of donor proliferation followed by donor withdrawal in health sector aid illustrates a trajectory that other countries may experience as they transition from low-income to middle-income status. In addition to these external events, turnover in donor and government staff over time may have contributed to changes in the degree of commitment to aid effectiveness policies. 4.5. Limitations The study findings should be interpreted in light of several limitations. First, key informants' responses may have been subject to recall bias or social desirability bias; triangulation across informants and comparison of informant narratives with documentary evidence where possible were used to help mitigate these potential biases. Second, some individuals invited to participate in the study declined or were unable to be interviewed during the data collection period; it is not possible to know what individuals who did not participate might have contributed to the study. Third, interviews were only conducted with key informants located in Vietnam's capital city of Hanoi and did not capture the views of informants working at sub-national levels or in donors' global headquarters. The study findings should therefore be understood as reflecting Vietnam's experiences with donor proliferation as perceived by those in the key informant sample, namely individuals with firsthand experience working on health sector aid in Hanoibased donor offices, central government agencies, or civil society organizations. 5. Conclusions Whereas previous literature has described the costs of donor proliferation and the expected benefits of aid effectiveness principles, this study of Vietnam identified a broader range of factors influencing how donors and government agencies respond to donor proliferation in health sector aid. These factors include the number of entry points for donor engagement with government, the financial need and human resource capacity of the recipient country government, the existence of competing organization-level and individual-level objectives within donor and government agencies, and the evolution of donor and government objectives over time in response to external events and internal staffing and policy transitions. Although this qualitative case study does not measure how responses to donor proliferation in health sector aid affected population health outcomes in Vietnam, it suggests that donors' and governments' political, economic, and institutional incentives may override considerations of population health benefits in determining aid allocations and responses to donor proliferation in the health sector. Future research should consider how the phenomenon of donor withdrawal from a country's health sector affects total funds available for population health, which health subsectors are prioritized, and the relative roles of government and non-governmental actors in health policymaking and

172

S.W. Pallas et al. / Social Science & Medicine 132 (2015) 165e172

service delivery. Acknowledgments SWP would like to acknowledge the feedback received from Achyuta Adhvaryu, Thad Dunning, Elizabeth H. Bradley, Ingrid Nembhard, Jeremy Shiffman, and participants in presentations about the study at the Yale Global Health Research in Progress Seminar, Georgia State University Institute of Public Health, and Yale Health Policy and Management Colloquium. The authors gratefully acknowledge the key informants who participated in interviews for the study, as well as Chhitij Bashyal for research assistance and the staff of the Center for Supporting Community Development Initiatives (SCDI) for logistical support provided during the study. Funding for the study was provided by the Wilbur G. Downs International Health Student Travel Fellowship awarded by the Yale University Schools of Medicine, Nursing, and Public Health and the Yale School of Medicine Office of Student Research. SWP also benefited from the support of a U.S. Agency for Healthcare Research and Quality T-32 training grant (#5T32HS017589). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.socscimed.2015.03.027. References Acharya, A., Fuzzo de Lima, A.T., Moore, M., 2006. Proliferation and fragmentation: transaction costs and the value of aid. J. Dev. Stud. 42 (1), 1e21. Buse, K., Walt, G., 1996. Aid coordination for health sector reform: a conceptual framework for analysis and assessment. Health Policy 38, 173e187. lange? coordinating external resources to the Buse, K., Walt, G., 1997. An unruly Me health sector: a review. Soc. Sci. Med. 45 (3), 449e463. Creswell, J., 2007. Qualitative Inquiry and Research Design: Choosing Among Five Approaches. Sage Publications, Inc., Thousand Oaks, CA. Djankov, S., Montalvo, J., Reynal-Querol, M., 2009. Aid with multiple personalities. J. Comp. Econ. 37, 217e229. Dodd, R., Olive, J.M., 2011. Player or referee? aid effectiveness and the governance of health policy development: lessons from Viet Nam. Glob. Public Health 6 (6), 606e620. Easterly, W., 2002. The cartel of good intentions: the problem of bureaucracy in foreign aid. J. Policy Reform 5 (4), 223e250. Easterly, W., 2007. Are aid agencies improving? Econ. Policy 22 (52), 633e678. Fielding, D., Mavrotas, G., 2008. Aid volatility and donor-recipient characteristics in ‘difficult partnership countries’. Economica 75, 481e494. George, A., Bennett, A., 2005. Case Studies and Theory Development in the Social Sciences. MIT Press, Cambridge, MA. Government of the Socialist Republic of Vietnam, 2006. Decree No. 131/2006/ND-CP Issued by the Government on 9 November, 2006 on the Management and

Utilization of Official Development Assistance. Government of the Socialist Republic of Vietnam, Hanoi. Knack, S., Rahman, A., 2007. Donor fragmentation and bureaucratic quality in aid recipients. J. Dev. Econ. 83, 176e197. Lawson, A., 2009. Evaluating the Transaction Costs of Implementing the Paris Declaration. Concept Paper Submitted by Fiscus Public Finance Consultants to the Secretariat for the Evaluation of the Paris Declaration. Fiscus Public Finance Consultants, Ltd., Oxford. Levine, R., Kinder, M., What Works Working Group, 2004. Millions Saved: Proven Successes in Global Health. Center for Global Development, Washington, DC. McCoy, D., Chand, S., Sridhar, D., 2009. Global health funding: how much, where it comes from, and where it goes. Health Policy Plan. 24, 407e417. Murray, C., Anderson, B., Burstein, R., Leach-Kemon, K., Schneider, M., Tardif, A., Zhang, R., 2011. Development assistance for health: trends and prospects. Lancet 378, 8e11. OECD, 2003. Harmonising Donor Practices for Effective Aid Delivery: Good Practice Papers. OECD, Paris. OECD, 2004. Joint Marrakech Memorandum. Managing for Development Results Second International Roundtable, Marrakech, Morocco, 4e5 February 2004. OECD, Paris. OECD, 2008. The Paris Declaration on Aid Effectiveness and the Accra Agenda for Action. OECD, Paris. OECD, 2010. Aid Effectiveness 2005e10: Progress in Implementing the Paris Declaration. OECD, Paris. OECD, 2011a. Busan Partnership for Effective Development Cooperation. Fourth High Level Forum on Aid Effectiveness, Busan, Republic of Korea, 29 November e 1 December 2011. OECD, Paris. OECD, 2011b. Creditor Reporting System Database. Retrieved from. www.oecd.org/ dac/stats (accessed 20.12.11.). Priwitzer, K., 2012. The Vietnamese Health Care System in Change: A Policy Network Analysis of a Southeast Asian Welfare Regime. Institute of Southeast Asian Studies, Singapore. Ravishankar, N., Gubbins, P., Cooley, R., Leach-Kemon, K., Michaud, C., Jamison, D., Murray, C., 2009. Financing of global health: tracking development assistance for health from 1990 to 2007. Lancet 373, 2113e2124. Rogerson, A., 2005. Aid harmonisation and alignment: bridging the gaps between reality and the Paris reform agenda. Dev. Policy Rev. 23 (5), 531e552. Rubin, H., Rubin, I., 1995. Qualitative Interviewing: The Art of Hearing Data. Sage Publications, Inc., Thousand Oaks, CA. Sachs, J., 2005. The End of Poverty: Economic Possibilities for Our Time. Penguin Press, New York, NY. Strauss, A., Corbin, J., 1998. Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory. Sage Publications, Inc., Thousand Oaks, CA. United Nations Development Program [UNDP], 2002. Monterrey Consensus on Financing for Development. Monterrey Conference on Financing for Development, Monterrey, Mexico, 18e22 March 2002. United Nations, New York, NY. Vandeninden, F., Paul, E., 2012. Foreign aid transaction costs: what are they and when are they minimised? Dev. Policy Rev. 30 (3), 283e304. Walt, G., Pavignani, E., Gilson, L., Buse, K., 1999. Health sector development: from aid coordination to resources management. Health Policy Plan. 14 (3), 207e218. World Bank, 2012. World Development Indicators Database. Retrieved from. http:// data.worldbank.org/data-catalog/world-development-indicators (accessed 01.03.12.). World Health Organization, 2012. Global Health Observatory Database. Retrieved from. http://apps.who.int/ghodata/ (accessed 01.03.12.). Yin, R., 2009. Case Study Research: Design and Methods, fourth ed. Sage Publications, Inc., Thousand Oaks, CA.