Analyzing the influence of institutions on health policy development in Uganda: A case study of the decision to abolish user fees

Abstract

Background

During the 2001 election campaign, President Yoweri Museveni announced he was abolishing user fees for health services in Uganda. No analysis has been carried out to explain how he was able to initiate such an important policy decision without encountering any immediate barriers.

Objective

To explain this outcome through in-depth policy analysis driven by the application of key analytical frameworks.

Methods

An explanatory case study informed by analytical frameworks from the institutionalism literature was undertaken. Multiple data sources were used including: academic literature, key government documents, grey literature, and a variety of print media.

Results

According to the analytical frameworks employed, several formal institutional constraints existed that would have reduced the prospects for the abolition of user fees. However, prevalent informal institutions such as “Big Man” presidentialism and clientelism that were both ‘competing’ and ‘complementary’ can be used to explain the policy outcome. The analysis suggests that these factors trumped the impact of more formal institutional structures in the Ugandan context.

Conclusion

Consideration should be given to the interactions between formal and informal institutions in the analysis of health policy processes in Uganda, as they provide a more nuanced understanding of how each set of factors influence policy outcomes.

Introduction

For decades, many low- and middle-income countries have had difficulties financing their public health systems, as raising revenue through taxation is often difficult in economies characterized by a large informal sector and low growth ^1,2. Thus in the late 1980s, many countries were encouraged to implement user fees policies as part of broader health sector reforms ^3,4. In the years that followed, there was considerable debate about the benefits and harms of these policies, particularly in Africa ^5–10. However, despite the proposed advantages of user fees^11–14, in practice the results were mixed ^2,8. They were often associated with increased administrative costs, and were found to reduce access to essential services for the poorest and most vulnerable members of the population ². They also increased the likelihood that expenditures became catastrophic, causing households to forego consumption of other essential goods and services while pushing some into poverty¹⁵. Today, the discussions continue, and although it is generally agreed that the removal of fees can help to improve access to health services for the poor, it is also acknowledged that user fees can be beneficial if implemented correctly in appropriate contexts². Furthermore, immediate and hasty removal has been discouraged, due to its potential to create negative shocks in the health care system¹⁰.

Amidst this debate, the Ugandan government officially introduced user fees for all public health services in 1993 as part of their broader health reforms ^16,17. Eight years later during the 2001 presidential election campaign, President Yoweri Museveni announced the immediate abolishment of user fees 10 days before Ugandan citizens went to the polls, and went on to win a second term in office¹⁸.

The reasons behind Museveni's decision are unclear, and hypotheses range from suggesting it was a reaction to a report which outlined the burden user fees place on the poor in Uganda ^1,19, to asserting it was the result of political commitment within the country ²⁰. However, the decision's proximity to the election and its potential to secure voters' support also imply that it was a campaign strategy^20,21. Indeed, the fact that a Government report submitted to the International Monetary Fund the same week does not mention this policy direction (and actually suggests that user fees were to remain), supports this explanation²². What cannot be refuted, however, is that this was a significant decision that unfolded rapidly and carried with it significant consequences for the health system¹⁵.

Currently, in-depth policy analyses employing available analytic frameworks are sparse in low- and middle-income countries²³, and do not appear to have been undertaken for this particular case in Uganda. With an aim to address this gap in understanding ,this paper uses a “mechanism-centred” institutionalist approach to policy analysis²⁴, as a way to identify the variables that help explain how President Museveni was able to successfully adopt a policy abolishing user fees in the days leading up to the 2001 election. The first section outlines the methodology we employed to carry out this study, including a detailed discussion of the conceptual approach and analytic framework used. Next, we briefly describe the policy subsystem in Uganda to identify the key actors engaged in the policy development process in the country. In section three, we present the results of our analysis, and show that there were several formal institutional factors that were in place to constrain the President's ability to introduce the policy decision under consideration. However, we also show that the influence of dominant ‘competing’ and ‘complementary’ informal institutions in Uganda likely trumped these constraints, enabling Yoweri Museveni to abolish user fees with haste. In the final section, we conclude by suggesting the need for more analyses of the policy process in Uganda using appropriate frameworks, in order to further understand the interactions between formal and informal institutional structures in determining policy outcomes.

Methods

Case study approach

This study used an explanatory qualitative case study approach^25,26, and employed a conceptual framework that guided the process of identifying what must be addressed during data collection and analysis. The case was defined as ‘the influence of institutional factors in Uganda and their impact on the decision to abolish user fees in 2001’. This case was chosen due to the contention surrounding the decision to abolish fees, and the current lack of in-depth policy analyses performed to illuminate the factors that enabled this decision. An important strength of this methodology is that it calls for the use of multiple sources of data to inform the analysis, which enhances the potential for generating comprehensive accounts of the case, and reduces the chance that interpretation of the data will be misleading²⁶. Further, as Yin²⁵ outlines, the case study approach is highly suited towards undertaking analyses of complex health care services and systems.

Conceptual approach and analytic framework Conceptual approach

The conceptual approach used to guide data collection and analysis is rooted in Douglas North's classic assertion that institutions form “the rules of the game in a society” and are “any form of constraint that human beings devise to shape human interaction”²⁷. This broad definition of institutions has allowed for significant variation in how these constraints are conceptualized within the field of political science and policy analysis, which is illustrated in the debate surrounding the three branches of new institutionalism^28,29.

The historical institutionalist framework is often promoted as flexible enough to incorporate both the rational actor-centred calculus of individual actors emphasized by rational choice institutionalists, and the culturally embedded norms that characterize the thinking of sociological institutionalists^29,30. Further, due to the clarity and empirical tractability of the hypotheses commonly employed by historical institutionalists, it is this school that is most often argued to constitute the best analytical point of departure^24,30,31. We decided to adopt a similar “mechanism centred” logic²⁴, to guide our analysis, whereby the explicit hypotheses commonly associated with historical institutionalism were used as an overarching analytical framework. However, we acknowledged that the informal, culturally rooted, sociological institutions have often been shown to matter in African political arenas^32–34. Unfortunately, their informal nature makes them harder to study empirically, which is often cited as a key challenge^29,31,33. Thus, we have complemented our explicit historical institutional hypotheses with theoretical insights drawn from the literature on informal institutions in Africa to ensure the development of a comprehensive account^32,33,35.

Analytical framework

The framework used to guide the identification of formal institutional mechanisms that shape prospects for reform are divided into two parts, and is based on a similar approach used to undertake analyses of health care reform in other contexts^36,37 . The first part is referred to as “government structures”, and suggests that broad “big bang” reforms are more likely in unitary states, parliamentary systems where power is more concentrated in the executive, electoral systems that support the emergence of majority parties, and in systems with a history of party discipline^38,39. This allows the government to consolidate authority, increasing their capacity to successfully make decisions that have a large impact on the entire system. The number of “veto points” in such systems is minimized, which reduces the number of potential constraints on a given policy decision^39,40. The second part of the analytical framework used to guide analysis of formal institutional influences is based on the concept of path dependence, policy feedback, and policy legacies^29,41–43. Here, policy processes are viewed within their historical context, where past policies influence policy development and implementation through the resource/ incentive and interpretive effects they create, which shape the prospects for future reform^41,43. In short, resource/incentive effects refer to how past policies influence administrative capacities, give particular interests privileges by way of financial benefits, and establish access to both policy authority and sources of finance. Interpretive effects, on the other hand, describe how learning from the effects of past policies informs future policy making, and how easily these effects can be traced back to particular government decisions and actors. The overall result of these feedback mechanisms is that they create boundaries which become engrained and thus determine the new “rules of the game” for policy development and implementation⁴¹. The constellations of actors and institutions created by these feedback effects become more resilient to change as time progresses because they are subject to increasing returns, thereby locking in policy prospects which are “path dependent” ⁴².

Our complementary framework is rooted in years of scholarship focused on African political processes, which has shown the importance of informal institutions^33,44,45. We based our approach on what Goran Hyden has called the “economy of affection”, where reciprocity, trust, and social exchange underpin political processes³³. Three distinct, but highly interrelated, informal institutions are common within this political sphere:

1. “Big Man” presidentialism;

2. corruption; and

3. clientelism³².

Taken together, these concepts posit that power is concentrated in an executive who is able to control key societal resources, and who uses this control to establish a form of neopatrimonial rule in which patrons and clients engage in reciprocal exchanges to achieve their desired (and often self-interested) goals^32,33. This is particularly salient in light of the fact that state-building efforts in Africa after colonial rule have been penetrated to a large extent by traditional values and community interests³³. The result is a weak and soft state, which often enhances the role given to informal institutions like neopatrimonial rule³⁴. A recent report has suggested that this is indeed the case in Uganda⁴⁶.

We also adopted a taxonomy developed to tie the two types of institutions together³⁵. In the context of the often ineffective formal institutions in Uganda^46,47, observed outcomes that converge suggest that informal institutions were “substitutive”, whereas outcomes that diverge suggest that informal institutions were “competing”. If formal institutional rules are viewed to be effective, then convergent outcomes suggest that informal institutions were “complementary”, whereas divergent outcomes are described as being the result of “accommodating” informal institutions. This logic was used throughout the analysis to link formal with informal institutional concepts.

Data sources

Data collection was driven by the analytical frameworks outlined above, and the propositions outlined in them were used to establish which relationships among the key variables should be explored^25,26. Following this, data sources were purposively sampled based on their relevance to key themes and variables outlined in each framework. Multiple sources were consulted including: academic literature, government documents, grey literature, and a variety of print media gathered through searches of the Lexis Nexus online database.

Data analytic procedures

Data analysis was proposition-driven and employed a pattern matching strategy in which data was compared and linked to the relationships between key themes and categories proposed in the analytic frameworks^25,26. Documents, articles, and media sources were first reviewed individually, and case study notes were taken that highlighted the data's relation to the key thematic categories²⁶. In subsequent reviews of the documents, case study notes were refined and indexed based on category, then grouped to compile collections of relevant instances which related to each of the framework categories. The categorized and aggregated notes were then reviewed as a whole, and relationships between key themes and categories were compared against those proposed in the frameworks. Perceived gaps in understanding of the case were also noted and used to steer further data collection and analysis, which continued simultaneously until it was felt that a robust and comprehensive account of the case had been developed and that it had addressed each relationship implicated by the theoretical lenses used to guide the study²⁵. Finally, observations were assessed based on the potential for other interpretations emerging from the various data sources, until it was felt that these interpretations converged. This method of triangulation was employed to ensure any assertions made were firmly grounded in the data and to verify repeatability of the observations.

Results

Defining the subsystem

The development of policy in low and middle income countries is complex, and the influence of international and domestic actors who are often members of broad global policy networks must be considered at each stage in the policy cycle^48–50. The subsystem that has formed around health policy issues and user fees in Uganda reflects this. Domestically, the President and his administration (including the Ministry of Health), Members of Parliament (MPs), District Health Management Teams (DHMTs), and health workers are often identified as key actors^{1,4,18–20,51–52}. Internationally, the two main actors in this subsystem are the World Bank and International Monetary Fund, who have been integral to health policy development since the late 1980s (ibid). Although the general public is frequently mentioned and directly affected by health policies, policymaking systems in this country and in those that are similar in the region are generally comprised of the insulated upper classes that have direct access and contact with international players^53,54. It is this diverse and extensive subsystem that is the focus of subsequent sections of this analysis.

Institutional explanations of the policy outcome Government structures

When analyzing the prospects for the abolition of user fees in Uganda, it becomes clear that government structures at the time of the decision did not appear conducive to supporting the quick enactment of a decision to abolish user fees. Instead, they reduce the concentration of power held by the administration through the creation of more veto points. Firstly, the government in Uganda began a process of decentralization in 1987, devolving authority to districts as part of broader structural adjustment policies. The Local Governments Act (1997) reinforced this further, legally granting locally elected bodies control over many health care decisions^19,55. Devolution fragmented governance in the country, moving it away from a unitary system. Second, the Constitution developed in 1995 separated the legislature and the executive, giving the Ugandan Parliament an independent role in keeping the administration accountable for decisions⁴⁷. This structure reduces the capacity of the executive to consolidate authority centrally. Further, in 2001, Ugandan parliament functioned without political parties, so party discipline does not exist. MPs were not elected not for their political affiliation with a specific party or platform, but for their own character often linked to how they are able to serve their constituency or district⁴⁷. Overall, the authority of local governments to run their own district health systems, the enhanced role of the Ugandan Parliament as a separate body from the executive, and the lack of parties and party discipline suggests that a “big bang” reform like the one under analysis would have faced many potential veto points. In theory, these government structures should have reduced the capacity of President Museveni to make autonomous and unchecked policies. Therefore his ability to successfully abolish fees in 2001 appears to contradict this view of how government structures impact policy outcomes.

It is within these divergent outcomes that both “Big Man” presidentialism and clientelism likely played important roles in political development, interacting as ‘competing’ informal institutions that downplayed the influence of veto points^32,35In the “economy of affection”, big men who control resources (in this case the President), are able to use their comparative advantage to garner support from those who see continued patronage as a way to achieve personal political goals³³. Thus, it is conceivable that there would be extensive support for the President's decision amongst potential veto players, in return for reciprocation which would enable them to appease the communities they represented. Indeed, it is the fusion of community interests and formal government processes that have often been identified as key drivers in contemporary African politics³³. The lack of a party system at the time, in addition to the concentration of power in the big man provides no incentives for district managers or MPs to be in opposition. Instead it is in their own self-interest to enter into a reciprocal relationship with the only person in the country likely able to provide them with access to the resources they need to achieve their goals. In other words, being in opposition would likely lead to a losing outcome for anyone in a position to veto a decision³³. Therefore, it appears that the government structures that could have stymied President Museveni's decision through the creation of veto points were likely undermined because of the influence of informal institutions, which in this case appear to have enhanced his ability to make the decision to abolish user fees.

Policy legacies and feedback

In Uganda, Museveni's policy to abolish user fees also appears to have faced many constraints created by policy legacies, particularly those left by the 1993 National Health Policy which had effects on actors in the subsystem both internationally and domestically⁵⁶. As the war years ravaged Uganda's economy, reliance on external sources of development aid increased. Although the World Bank was willing to dedicate funds for the health system, by 1993 they would only do so if they had significant authority over the formulation of domestic policy⁵¹. With no real options the GoU conceded, and a user fees policy was put in place at the beginning of the 3-year National Health Plan in 1993—the result of a World Bank conditionality^16,56. This policy decision served to officially transfer administrative capacities for policy development to external donors and technical experts (namely those housed within the World Bank and the International Monetary Fund), thereby decreasing domestic policymaking capacities^57,58. This policy also resulted in the development of new interpretations of the health policy process which affected the government elites in Uganda, as they learned that it was necessary to surrender authority in order to access much needed aid, while viewing external actors as dominant in health policy decisions in the country (ibid).

In the context of the devolved system, the user fees policy was also influential in increasing administrative capacities for local governments in the country. Locally retained fees reduced DHMTs' reliance on central transfers, thereby increasing their autonomy over local financial decisions, which empowered them to take a greater role in the governance and delivery of health care^{19,55,59–60} . Additionally, as DHMTs gained access to a new source of finance to pay for human resources and system maintenance, many managers came to rely on this new revenue stream, which became an essential element in enabling them to run an effective district health service⁶¹ . This policy also had a similar effect on health professionals working within these district systems, who relied on this source of finance to ensure adequate levels of pay^20,52. Thus, as a result of learning from the user fees policy, new interpretations were developed among these actors which influenced them to view the policy as one which conferred positive benefits onto the health system- namely improvement in quality of services and positive changes in health staff motivation^4,52.

Overall, through resource/ incentive effects, the 1993 user fees policy increased capacities for health care administration and decision making for both international donors and DHMTs while decreasing the capacity of domestic government elites. New sources of finance were transferred to DHMTs and health professionals, who now had a budgetary source that they eventually came to rely upon as a way to maintain salaries and cover operating costs of the district health system. Policy learning created “interpretive effects”, as the central government no longer perceived itself to be developers of domestic health policy strategies but instead as agents for external policy driven by donors. Local governments learned that the user fees were an essential tool that enabled them to run a quality local service. Therefore, the policy legacies created by the user fees policy in 1993 would suggest that the swift abolition of user fees based on a decision by the President would have required a significant restructuring of administrative capacities, shifting of the resource distributions, and would have challenged the world view of several key actors who developed new interpretations of the health care system as a result of this past policy. For those actors set to lose from this decision (namely district managers and health professionals), it would also be easily traceable to one man—the President. Thus, the feedback from this past policy would serve to constrain the prospects for President Museveni's decision to abolish fees. Overall, these findings suggest that the observed outcome in 2001 also appears to be at odds with the constraints imposed by formal institutions in the form of policy legacies.

This divergent outcome provides another example to suggest that informal institutions were “competing”. News coverage from the country in the time since the decision confirms that the resource/incentive and interpretive effects created in 1993 still exist. The DHMTs, hospitals, and health workers continually express the negative effects of Museveni's campaign decision, elaborating on how the absence of fees has resulted in a public system that is no longer able to function due to the incompatibilities with the reality of district health systems^57–66. Here, the influence of “Big Man” presidentialism with its concentration of unchecked authority prevails. Despite the constraints presented by the decision's incompatibility with the policy legacy left by the 1993 user fee introduction, and the fact that the apparent difficulties caused by this decision are easily visible and traced back to the President, this informal institution may have allowed Museveni to make decisions without considering the financial implications or potential opposition from these key interests to pursue his own political interests. This apparently irrational decision is also consistent with what Hyden³³ described as the “policy deficit” (pg. 116), whereby the culturally accepted institution of “Big Man” rule allows leaders to make purely self-interested and political decisions, without considering the on-the-ground economic repercussions.

Overcoming the influence of past policies on government elites and international donors was likely the result of a change in a formal institution, albeit one that allowed informal institutions to work in a “complementary” way for President Museveni. A key development in the year before the election was the introduction of the Memorandum of Understanding for a Sector Wide Approach (SWAp) in Uganda, which set in motion a shift in the way donor funds were allocated within the country⁶⁷. Instead of loans specifically targeted to sectors, funds were pooled centrally, no longer earmarked, and therefore less prone to sector conditionalities⁴.

Although this decision in and of itself would not have increased domestic elite capacity to formulate health sector policies, it may have subdued the interpretive effects resulting from the 1993 policy—namely that decision making authority for the health sector had to be given to external donors if they were to receive financing for the health sector. The shocked reaction shown by the World Bank and International Monetary Fund after the decision was made to abandon fees suggests that this interpretive effect may have still been present. However, the SWAp MOU in 2000 allowed the President to challenge the dominant view of policymaking in the country, and muted the impact of the precedent set in 1993^67,68. Therefore, the MOU served to formally insulate the President from constraints traditionally imposed on the policy process by external donors, while allowing him to use his increased control over public sector resources to make a policy decision that was likely important in garnering support in his bid to re-election—another action consistent with descriptions of Africa's “economy of affection”³³.

Discussion

The decision to abolish user fees in Uganda in March 2001 has generated a variety of conflicting explanations as to why this policy was implemented. Although the exact motivations driving President Museveni's choice are difficult to verify empirically, there are analytical tools available which can aid in attempts to deduce how such a significant reform decision was enabled in this particular context. By undertaking what Thelen²⁴, referred to as a “mechanism-centred”, and what Fritz Scharpf³⁰ has described as an “interaction-consequential” approach to institutional analysis, this study utilized the explicit hypotheses derived from historical institutionalism and complemented these with insights derived from the role of informal institutions in Africa . The results show how the success of Yoweri Museveni's decision to abolish user fees for health services in Uganda in March 2001 is at odds with current analytical frameworks that highlight the influence of formal institutional constraints imposed by government structures and policy legacies in the development and implementation of policy. However, when complementing these historical institutionalist approaches with insights taken from the literature on the influence of sociologically constructed informal institutions in Africa, a more comprehensive picture emerges as to how these informal institutions interact and trump the more formal structures^32,35.. President Yoweri Museveni was able to enact his chosen policy decision quickly in the run up to the election, due to the enabling characteristics embodied in the three pillars of neopatrimonial rule: “Big Man” presidentialism, clientelism, and corruption. These informal institutions were classified here as “competing”, due to their divergent outcomes with what formalized institutional structures would dictate in the situation. However, the introduction of what has been called a successful SWAp⁶⁷, also provided an example of “complementary” informal institutions, whereby the intended outcome of the formalized agreement with external donors provided the President with greater capacity to take advantage of the informal rules that govern the “economy of affection” in Ugandan politics. Therefore, this study helps to confirm what both Helmke and Levitsky³⁵ and Bratton³² have suggested in the past—that it is essential to take into consideration both the informal and formal institutional influences as they interact, to determine policy outcomes in low-and middle-income countries.

Conclusion

Overall, given the current paucity of rigorous analytical studies undertaken to explain the policy process in low -and middle-income countries²⁴, this paper should be viewed as a step toward filling this gap that will hopefully inform similar work in the future. These case study results offer important insights to guide the selection of institutional frameworks for explaining health policy processes in comparable contexts. They will also hopefully inform further refinement of health policy analysis methodology for adoption by future researchers in studying policy making in Africa.