Date for Development in Africa
How can civil society use open budget data to
advocate for improved public services within the
Ugandan political context?
Name: Elma Jenkins
Student number: 10859578
E-mail: elma.jenkins@googlemail.com
Course: MSc International development studies, UvA Date: 25 November 2015
Word count: 26,821
Cover photo: Budget champion volunteers receiving training from the Advocates Coalition in Kampala on how to extract budget figures from Uganda’s open data web portal.
1
Master’s Thesis
First Supervisor
Dr. Linnet Taylor
Marie Curie Research Fellow, Governance and Inclusive Development University of Amsterdam Nieuwe Achtergracht 166 1018 WV Amsterdam The Netherlands E-mail:l.e.m.taylor@uva.nl http://www.uva.nl/over-de-uva/organisatie/medewerkers/content/t/a/l.e.m.taylor/l.e.m.taylor.html Second Supervisor Dr. Nicky Pouw
Assistant Professor, International Development Studies Graduate School of Social Sciences
Nieuwe Achtergracht 166 1018 WV Amsterdam The Netherlands E-Mail: n.r.m.pouw@uva.nl http://home.medewerker.uva.nl/n.r.m.pouw/ Second Reader Dr. Sean Higgins
Lecturer,International Development Studies Graduate School of Social Sciences
Nieuwe Achtergracht 166 1018 WV Amsterdam The Netherlands
E-mail: S.Higgins@uva.nl
2
Acknowledgements
Thank you to the wonderful people I had the honour of speaking with while undertaking this research, including but not limited to; ACODE, their budget champions and staff who were kind enough to let me travel with them and introducing me to some truly wonderful people; CUWEDA for your excellently planning, hospitality and dedication; Development Initiative, for setting me on the journey; Development Research and Training for the insightful
discussion giving me new angles for questioning; Bugisu NGO Forum for helping me connect to the youth and patiently translating; All the African Centre for Media Excellence journalists for being an inspiration, asking tough questions and setting the curve; Overseas
Development Initiative for giving me insights into your project and to all the individuals working on promoting truth patriotically and standing up for what they believe in. Finally to the UvA supervisors for guiding my questioning and patiently reminding me, again, to re-read.
3
Abstract
The debate on whether open data can be used for effective advocacy has many conflicting viewpoints. Different scholars see both an empowering effect through the increase of social accountability with evidence-based advocacy or the breakdown of trust in society as ICT-mediated information increases anonymity and creates passive transparency. This thesis project aims to add to this debate by assessing how civil society groups use, disseminate and apply a recently launched transparency initiative for open budget data in Uganda. The analysis further explores the landscape of budget data use; how transparent budget data is; how open data impacts on government accountability to public services and what impact the political context has on advocacy. Based on fieldwork that was carried out over 10 weeks in two urban centres and one rural area of Uganda, the analysis makes use of a mixed methods qualitative data collection approach and uses the capabilities approach as a lens with micro frameworks of surveillance, transparency and change mechanisms.
Overall, it is argued that civil society uses open data to increase bottom-up transparency through improved monitoring and evaluation work. However, impacts are currently only visible at the sub-national level and do not always lead to accountability. The challenges of scaling this up are limited by a weak legal framework, civil society branding as opposition in the pre-election climate and increasing regulation over CSOs in Uganda. The thesis
concludes, while that are strong theoretical and empirical links between transparency and advocacy, In Uganda it has been shown through comparisons with international standards that the open budget initiative in Uganda displays an opaque level of transparency and consequently a weak level of advocacy. It is suggested that the best way for civil society in Uganda to use digital data is by building local government capacity and awareness of open data requirements, especially to improve local hard copy displays in order to re-affirm local governments’ responsibility to provide public services and to include the media more. The research also emphasises the important role infomediaries will play as ICT-mediated open data increases, information can be presented out of context creating a false image of non-biased data and ultimately disempowering users.
Key words: Monitoring and evaluation, open data, budget transparency, civil society advocacy, social accountability, advocacy
4
Inhoud
Acknowledgements ... 2 Abstract ... 3 List of figures ... 6 Abbreviations ... 7 1. Introduction ... 8 2. Theoretical framework ... 12 2.1 Introduction ... 122.2 Openness and accountability ... 12
2.2.1 Open data and transparency ... 13
2.2.2 Open budget data ... 13
2.3 Advocacy and accountability ... 14
2.3.1 Civil society advocacy... 15
2.4 Data for advocacy ... 15
2.4.1 Surveillance ... 16
2.5 Governments and transparency ... 16
2.5.1 Criticisms of open data ... 17
2.6 Budget data for advocacy ... 17
2.7 Conceptual scheme ... 18
2.8 Conclusion ... 19
3 Research Methodology ... 20
3.1 Introduction ... 20
3.2 Research question and sub-questions ... 20
3.3 Epistemology ... 20 3.4 Methodology ... 21 3.5 Sampling ... 21 3.6 Methods ... 22 3.6.1 primary data ... 22 3.6.2 Secondary data ... 23 3.6.3 Units of analysis ... 23 3.7 Data analysis ... 23 3.8 Data validity ... 24 3.9 Limitations ... 24 4 Conclusion ... 24 4. Country context ... 25 4.1 Historical context ... 25
4.2 Economic and social development context ... 26
4.3 Political context ... 28
4.4 Civil society... 29
4.5 Technological context ... 30
5
4.6 The national budget and the data revolution ... 31
5. Conclusion ... 31
5 Empirical analysis ... 33
5.1 The landscape of data use ... 33
5.1.1 Responses to digitisation ... 33
5.1.2 Dissemination and engagement ... 38
5.1.3 Processes and products of advocacy ... 44
5.2 How transparent is Uganda’s budget? ... 46
5.2.1 What data do CSOs want? ... 46
5.2.2 Access issues ... 48
5.2.3 Overcoming the gaps ... 49
5.3 Conclusion ... 50
6. Empirical analysis ... 50
6.1 Open data impacts through advocacy ... 52
6.1.1 Impacts at the national level ... 52
6.1.2 Sub-national level impacts ... 55
6.1.3 Empowerment impacts ... 57
6.2 Political context of Uganda ... 59
6.2.1 The legal framework in Uganda ... 59
6.2.2 Safety using budget data ... 60
6.2.3 Citizens self-censorship ... 62
6.3 Conclusion ... 65
7. Conclusion ... 66
7.1 Open data landscape ... 66
7.2 What is budget data? ... 66
7.3 Impacts on and through advocacy practices ... 66
7.4 Political context ... 67 7.5 Conclusion ... 67 7.2 Reflections... 68 7.2.1 Methodological ... 68 7.2.2 Theoretical ... 68 7.3 Recommendations ... 69 7.3.1 For policy ... 69
7.3.2 For future research ... 70
Bibliography ... 71
Annex one - Operationalization of major concepts ... 77
Annex two – Overview of respondents ... 81
6
List of figures
Figure 1 - DI Uganda data portal (DI, 2015) ... 10
Figure 2- Conceptual scheme ... 19
Figure 3 - WB Uganda statistics ... 26
Figure 4 - Infrastructure map showing connectivity coverage in Uganda ... 27
Figure 5 – Uganda’s score in the civil society index (2009) ... 29
Figure 6 - Newspaper print data ... 34
Figure 7 - Uganda's governance structure ... 35
Figure 8 - Five stars of data engagement ... 37
Figure 9 - Tally of digital comments ... 38
Figure 10 - picture of local hard copy budget display ... 40
Figure 11 - Radio talk show host taking a call, Mbale ... 43
Figure 12 – Main types of budget information used for analysis (Renzio & Simson, 2013)... 44
Figure 13 - Tally of CSO identified budegt data gaps ... 47
Figure 14 - Participatory timeline, Nebbi ... 58
7
Abbreviations
ATI: Access to Information
ACODE: Advocate Coalition for Development and Environment BC: Budget Champion
BM: Black Monday
CBO: Community Based Organisation
CIPESA: Centre for ICT policy for Eastern and Southern Africa CSO: Civil Society Organisations
DI: Development Initiative
DRT Development Research and Training FOI: Freedom of Information
IO: International Organisations MDG: Millennium development goals M&E: Monitoring and evaluation MoF: ministry of Finance
NGO: Non-Governmental organisation NRM: National Resistance Movement
OECD: Organisation for Economic and Cultural Development ODI: Overseas Development Initiative
OGP: Open Government Partnerships RCT: Randomised Control Trials UDN: Uganda Debt network UvA: University of Amsterdam UN: United Nations
8
1. Introduction
In the post 2015 development agenda Ban-Ki-Moon, president of the UN, has called for a data revolution. In a UN report published in 2014 he argues, “never again should it be possible to say ‘we didn’t know’. No one should be invisible.” (IEAG, 2014: p3). The data revolution calls for the strengthening of national capacities through initiatives such as data labs, building Open Governance Partnerships (OGP) and open data initiatives to hold governments accountable and sustain development (UN, 2014: 3). While the amount of data in the world is increasing, estimates say 90% of today’s data has been created in the last two years, it will take some time for civil society, governments and academic groups to adapt to these new knowledge flows (UN-IEAG, 2014, p. 5). The UN aims to make poverty and its causes more visible with data (UN-IEAG, 2014). Particularly in African countries which are data poor as little is known about the reduction of poverty (Devarajan, 2011). The increasing interest in the data revolution is largely absent of much discussion around what the purpose of the increased digitisation is. Therefore, many organisations struggle with the use and application of the new digital data leading some to suggest that the
practice of open data is racing ahead of empirical or theoretical thinking (Fox, 2015, p. 346). This research aims to contribute to the debate on how data, which has been described as a tool for positive change in the world (Taylor, 2014), can be used for development purposes. Open data, as one facet of the data revolution, cuts across many data sets and has seen a dramatic increase globally (Davies, Tim, 2010, p. 8). In developing countries the data revolution is creating new funding streams for International Organisations (IOs) and Non-governmental Organisations (NGOs). Most noticeably the Bill and Melinda Gates foundation has committed 3.5 million (USD) to “inform discussion around calls for a data revolution and work with partners to build a roadmap and support improvements in development data” (QWIDS, 2013).
The definition of open data is data which “must be accessible, machine-readable (i.e. in a form where you can manipulate it with digital tools), and licensed to permit re-use, rather than restricted by copyright or intellectual property rights” (Davies T. , n.d, p. 20). Tim Berners-Lee developed a 5 star deployment scheme for open data. It outlines that data must be more than just online, it must also be available in a downloadable format such as Excel or CSV which makes it easier to locate and link to other data sets (Berners-Lee, 2012). This thesis will focus on open budget data as a tool to create social good.
Every country has a national budget which fulfills an important role in distributing revenue, resources and common goods fairly in a nation state. Budget data can be useful to many groups in society including businesses and advocacy groups. Budget data at the sub-national level, when looked at over time, can also show how collective national resources are being
9
allocated and any potential misuse revealing corruption or levels of incompetence in administrative structures.
Uganda has the highest levels of corruption in East Africa scoring 26/100 in the
Transparency International rankings with 86% of people admitting to paying a bribe (TI, 2014). Corruption negatively impacts on the quality of public services people receive from health care centres to education and infrastructure and held Uganda back from achieving the MDG targets. Open budget data is being explored as a way to increase transparency and engage civil society across the developing world in order to fight corruption (Renzio &
Masud, 2011). In Kenya the BudgITwebsite brings the national budget in a readable form
online while also promoting its importance as a document for improving social services.
There is also the World Bank initiative ‘Boost,’a budget tool which “collects and compiles
detailed data on public expenditures from national treasury systems and presents it in a simple user-friendly format” (Kheyfets, Mastruzzi, Merotto, & Sondergaa, 2011, p. 1). These initiatives have come about with the aim of making budget more open and accountable. The Government of Uganda has recently launched a web portal to make government budget data more transparent, consequently this is the first time sub-national budget data is
available in a widely accessible form (MoF, 2015). This initiative comes with a wave of other technological changes in Uganda such as improved mobile and internet connectivity which will increase access in society. These technological changes are happening against a
backdrop of increased pressure from international donors on developing countries to become more transparent, such as the IMFs 2015 Enhanced General Data Dissemination Systems (E-GDDS), which contains a growing focus on micro data (IMF, 2015). The web portal has also created interest from International Organisations (IOs) like the Development initiative (DI), an NGO focussed on using data to reduce poverty. The DI are trying to
improve transparency by taking digital data off the budget web portal and applying new PDF reading software in order to extract data quicker. Figure 1 shows an image of the DI budget data visualization map now available online.
10
Figure 1 - DI Uganda data portal (DI, 2015)
This research will seek to answer how civil society groups are using the new digital budget data for advocacy, with the aim of improving public services, within the political context of Uganda. It will focus on three areas of civil society operationalised as: radio, print media and NGOs in three geographically distinct areas of Uganda: Kampala, Mbale and Nebbi.
The UN report also recognises the risk of data inequality which creates a gap between the “data haves and the data have-nots” (UN-IEAG, 2014, p. 6). Others described this as a ‘digital divide’ linking it not only to lack of internet access but also a lack of usage and benefit from data (Fuchs & Horak, 2008). National statistics are a first step in overcoming this divide and Uganda recently completed its tenth census to date. This will help ground any future development plans, if used correctly. In much the same way that the use of new demographics in history created the branch of Social History with authors such as Howard Zinn (1980) writing new perspectives on military history, the data revolution could usher in new perspectives through open data and information sharing (Kitchin, 2014).
By looking at how civil society groups in Uganda are using digital budget data this research can fill a knowledge gap on how useful a data revolution is in a country like Uganda, by asking if and how civil groups are using the data within an authoritarian rule.
Outline of the study
Chapter 2 will discuss the literary and theoretical framework which will unpack the main concepts; civil society advocacy, open budget data and political context, to help analyse the empirical data. A conceptual scheme will further develop the concepts inter-relationships. Chapter 3 will outline the research methodology (qualitative), epistemology, research
11
design and introduce the units of analysis. Chapter 4 introduces the country context followed by the empirical chapters. Chapter 5 will cover sub-question 1 and2 – the
landscape of budget data and levels of transparency in budget data. Chapter 6 will cover sub questions 4 &5 - the direct impacts on public services and how Ugandan political institutions affect advocacy. The thesis concludes by answering the main research question and
reflecting on the theoretical framework, conceptual model and methodology of the study (Chapter 7). Lastly, recommendations for further research and policies are given.
12
2. Theoretical framework
2.1 Introduction
This chapter will build a theoretical framework with which to guide the data analysis. Starting with the concept of openness this section will build on Amartya Sen’s Capabilities Framework to include open data by using the concept of accountability (2.2). It will then introduce a micro theory of scales of transparency and accountability (2.2.1) and how these relate to the concept of open budget advocacy (2.3). However, open data can be two sided and the controlling aspect of open data in Uganda will also be explored (2.5). The thesis will also apply a new framework of surveillance to further explain how monitoring and
evaluation (M&E)1 is emerging as a tool for accountability. Finally the chapter will introduce two change mechanisms in order to explore how public services are improved at the micro level (2.6).
2.2 Openness and accountability
Openness as a social concept is gaining momentum with an increase of global surveys such as the Open Budget Index (OBI) and the Open Knowledge Survey promoting greater global open data standards. Open data has been described as a “spread sheet revolution” (IBP, n.d) and can include everything from government data to aid data. Openness has historical foundations, in 1914 American lawyer Brandeis said “sunlight is the most powerful of all disinfectants” referring to the powerful effect publicity can have as a measure against illicit behaviour such as gathering vast amounts of wealth or levying unfair taxes (Brandais and the history of transparency, 2009).
Development thinker Amartya Sen argues in his Capabilities approach (1999) that individual freedoms are crucial for inclusive development, capturing aspects of individual agency in poverty eradication. He distinguishes between two types of freedoms: substantive – which cover basic physical needs and instrumental (political, social, economic, transparency and protective freedoms) which are more complex (Amartya, 1999, p. 10). This thesis will focus on citizen’s freedom to information which includes guarantees of disclosure and access to evidence creating a trusting relationship between citizens and administrators. Expanding citizen’s freedoms, Sen argues, moves development beyond the narrow definition of GDP growth, which Sen classifies as a “means” rather than an “end” to development. Sen supports openness with his ground-breaking work on the causes of famine in India. He explained that famines in Ethiopia and Bengal were the cause of a lack of instrumental freedom and concluded that open societies experienced less famines (Sen, 1981). One way to view the Sen’s development approach is to view development as accountability
1
Monitoring and evaluation has a large literature define M&E as a tool for NGOs and INGOs to measure the success of projects (Chapman & Wameyo, 2001). This thesis is using the term differently to refer to the practice of M&E by civil society on their own governments which has a smaller but growing literature (Björkman & Svensson, 2009) and which will be explored further throughout the thesis.
13
(AccountAbility, n.d). The concept of openness is important to Sen’s Capabilities framework because the process of accountability requires openness in a globalising world.
Openness is theorised to create good governance through inclusive development by helping governments respond to citizens demands for accountability and justice (Soros, 2000). Soros, another prominent thinker on openness, has written about the importance of “a society that holds itself open to improvements” (Soros, 2000, p. 52) which involves opening institutions and governments. Soros established the Open Societies foundation in 1979 with the aim advocating for peoples freedom to participate in civic life and ultimately challenge top down governance structures.
2.2.1 Open data and transparency
Davies (2012) has built on Berners-less’s definition of open data by outlining some economic, political and organisational changes necessary to make data truly open.
He defines five starts of data engagement with the aim of “generating social and economic benefits […..] to hold state institutions to account” (Davies T. , 2012, p. 1) which will be used as an analytical tool in this thesis to measure the engagement level of Uganda’s open data. Fox (2007) has theorised that open data, as a form of transparency, can be placed on a scale between clear and opaque. An opaque level hides how institutions really behave while clear transparency reveals reliable data on performance, responsibilities and funding behind institution (Fox, 2007, p. 666). The scales of difference are due to the malleable nature of the term transparency, which explains why open data does not always lead to
accountability. The challenge of closing low accountability gaps has been theorised with the use of a ‘virtuous circle’ of accountability which can be initiated through a sandwich strategy of state and society working together (Fox, 2015, p. 347). This solution emphasise the
importance of feedback loops, where data users and data producers communicate with each other preferably in real time and aligns with Janssen et al’s thinking that the benefits of data openness are more complex than often credited (Janssen, Charalabidis, & Zuiderv, 2012).
2.2.2 Open budget data
Making budget data open and freely available is one of the key aims of the open data
movement because of its importance to development (Heald, 2012). Yet there are still many challenges to overcome especially around measuring budget transparency and scaling up open data to include participation and public engagement (Renzio & Masud, Measuring and Promoting Budget Transparency:, 2011). Open budget data then becomes a mechanism for positive change by increasing citizens participation and accountability to public service (Gray, 2015). Top down open budget initiatives require further research to explore how these potentials can be fully realised. The dangers of applying a broad brush stroke of accountability through transparency, without taking into account individual country contexts also requires further research.
14
2.3 Advocacy and accountability
Advocacy has various meanings which are dependent on the context. The Open Budget Survey (OBS) has defined budget advocacy as “a strategic approach to influence
governments’ budget choices, aimed at achieving clear and specific outcomes e.g., healthier people, less poverty, or improved governance (IBP, Orientation to budget advocacy, n.d). This global survey is run by the International Budget Partnership (IBP) an international network which started in India to promote the use of budget data. The IBP advocates to improve the internal development of a county and describes budget advocacy as evidence based advocacy.
The concepts of accountability and transparency are often coupled together by “Civil-society campaigners around the world who have incorporated the right to know into both their strategies and their tactics, with the hope that transparency will empower efforts to change the behaviour of powerful institutions by holding them accountable in the glare of the public eye” (Fox, 2015, p. 663) . As such, accountability is emerging as a key strategy for development by rooting advocacy in knowledge and information (Houtzager, 2012). This builds on the above definition of advocacy to include social accountability as an advocacy approach and supports Gray’s call to make data citizen readable (Gray, 2015: 4).
Accountability encompasses many activities from monitoring of public and private assets, access and dissemination of public information, redress of complaints and grievances to citizen participation (Fox, 2015). Joshi & Houtzager (2012) argue that current trends in accountability have come to focus on widgets or mechanisms such as web portals or scorecards. These however are low hanging fruit in advocacy practices and do very little to tackle low data engagement rates (Houtzager, 2012, p. 153). Creating an environment of listening and debate by implementing a watchdog dynamic alongside the widget is suggested which also helps to adopt a longer term view of accountability.
CSOs, however, are often faced with difficulties in enforcing accountability measures and have to resort to ‘soft’ accountability (Fox, 2007). This includes naming and shaming as well as building a dialogue through one-on-one training in contrast to applying real pressure on public officials with sanctions, or what Fox (2007) terms ‘hard’ accountability. CSO advocacy is therefore often focussed on strengthening data transparency over accountability at national level.
In order to find out why transparency does not always lead to accountability, it is important to study the format and function of open data carefully (Heald, 2012). It has been argued that open data policies can actually weaken effective transparency with misleading communication (Heald, 2012). For example, a country may score highly on an open data index and but this may not always result in accountability because of low uptake in civil society. Heald (2012) argues that we need to be realistic about what open data can achieve.
15
It cannot provide answers “to profound ideological and practical questions concerning the scope of the state” (2012: 47) but transparency of public expenditure can improve empirical evidence with which to inform citizens views.
2.3.1 Civil society advocacy
Civil society is seen as an engine for social change (Pearce & Howell, 2001) which has led to initiatives such as the World Bank Global Partnership for Social Accountability (GPSA); a grant giving organisation which funds social accountability initiatives. The complexities of transporting European concepts of civil society into an African context need to be carefully considered (Obadare, 2014). Because normative assumptions of civil society – that civil society actors will always be for the benefit and empowerment of the poor – can reflect European concepts of democracy (Pearce & Howell, 2001, p. 59). It can be said that civil society in Uganda is active in representing the views of the poor, especially online (HRW, 2011).
A growing strategy of civil society advocacy in developing countries is through independent monitoring and evaluation (M&E) of government programmes (Gildemyn, 2014). The application of M&E by civil society moves it beyond the traditional realm of INGO project impact measurement into a more participatory mechanism. However, Gildemyn argues M&E has limitations for tackling the unequal power relations between CSOs and service providers. Studies in Ghana and Uganda have shown that M&E impacts behaviour at the sub-national level. Bjorkman and Svensson (2009) conducted Random Control Trials (RCTs) of community M&E on primary health care centres in two villages in Uganda. They found the behaviour of health care staff at a local level did change with increased engagement by citizens in their local health centres. In addition there was increased health care utilization and improvements to healthcare outcomes (Björkman & Svensson, 2009). The research suggest the influence of M&E is limited by location but that it can be effective as a mechanism for social change at a local level.
2.4 Data for advocacy
Data is increasingly being used by civil society advocates online and offline. Gray (2015) has conducted a study using digital methods in order to map the online landscape of open data use globally. He categorised four growing areas of interest around open budget data. The most prominent (65%) were to improve the visualization of data (DI, 2015). Three other categories which were much smaller in scope but worth mentioning here are: projects encouraging citizen budget monitoring by actively equipping civil society with information to perform monitoring and evaluation; projects for journalists to investigate money flows such as farm subsidies and finally data for advocacy where budget data is built into an advocacy campaign.
16
2.4.1 Surveillance
A new framework of surveillance can be applied to open data to further clarify how open data can be used as a tool for social accountability. Surveillance is defined as a “focused, systematic and routine attention to personal detail for the purpose of influence,
management, protection or direction” (Lyon, 2007, p. 14). There is a tendency to view surveillance negatively in our growing digital age, but this does not always have to be the case. As Lyon points out, surveillance contains elements of care and control, which can be used by a range of people. An example is the crowdsourcing website in India which allows citizens to report bribery and corruption online (Janaagraha, 2015).
Monitoring and evaluation (M&E) practices by civil society can be theorized as a form of bottom up surveillance. Surveillance is never a solo activity, instead it is one in which the watcher and the watched know they are part of. As a mechanism for advocacy, M&E enables the poor to scrutinize authorities. Therefore, in such bottom up forms of
surveillance, data can contain elements of care. Furthermore, as a community activity M&E itself helps ensure inclusive development by involving citizens in the decisions which affect their lives.
2.5 Governments and transparency
Governments are the largest providers of data reflected in the recent growth of open government data websites in America and the UK. Despite this, government bodies often adopt a simplistic approach to data sharing, viewing open data as an ends rather than a means to development. Janssen et al (2012) identify several open data myths commonly held by governments such as: disclosure will automatically benefit society, all data is required to be open and open data automatically results in open governments (Janssen, Charalabidis, & Zuiderv, 2012). In addition, much government open data is actually voluntarily given making it important to study how data is made available as this can communicate a message about the motivations and objectives behind government open data policies. These could be to do with investments in democratic accountability, public policy influence and equity but it could also be about social control or gaining quick results with the plethora of open data indexes emerging, rather than real structural change which Sen theorises as necessary for development (Susha, Zuiderwijk, Janssen, & Gronlund, 2015). Government open data theories can be related to Africa by considering how data can be used as a tool for control, as discussed by Lyon (2007). Political leaders in Africa typically exert control through the use of state resources to create patronage structures as a way of retaining power (Tangri & Mwenda, 2005). This is expressed in Uganda through the control of data as well. The theoretical framework of the log frame can be applied here to explain how budget data is controlled by authoritarian political institutions in Uganda to impede transparency and retain power. This is in contrast to Fox’s ‘virtuous circle’ (2015) which creates accountability through data feedback loops. The logical framework has many
17
working definitions but is typically characterised by a hierarchy of objectives beginning with prioritised inputs which lead to logical outputs (Bakewell & Garbutt, 2005, p. 2). The log frame is a formal procedure for project planning, often criticised for being in-flexible and exclusive as it is inherently results focussed (Bakewell & Garbutt, 2005, p. 11). It has been accused of limiting development thinking because it requires the prioritisation of change indicators. It is useful here to theorise how Ugandan political institutions retain control in the face of increasing demands for transparency. Planning government interventions with a narrow input-output work flow could potentially help authoritarian regimes, such as in Uganda, control data with a one-dimensional approach.
2.5.1 Criticisms of open data
There are also criticisms of the open data movement. Dawes and Meijer (2009) argue that opening data, especially administrative data of governments can be detrimental because it can reduce social trust, questioning open data’s ability to foster inclusiveness and generate public value. Furthermore government data is often not created with the intention of external use in mind which can result in bias, misunderstanding and misuse. Furthermore analysis of data by un-trained citizens can limit findings to low hanging fruit by missing complex trends and comparisons between data sets are limited if users do not know how the data was collected or how data quality has been assured. Data can also lose meaning outside of its context creating further engagement problems.
Government data is rarely designed for public use so a careful study of the character of data needs to be undertaken first. Meijer (2009) also challenges the impact of bottom up
surveillance by arguing that ICT mediated data actually creates anonymity further distancing the policy makers from reality (Meijer, 2009). These criticisms need to be take into
consideration when understanding how open data is being used by civil society in Uganda.
2.6 Budget data for advocacy
Global indexes show that the state of open budget data is low, especially in Africa where often budget data is only available at the national level (Renzio & Simson, 2013). Renzio and Simson (2013) have compiled a list, drawn from African NGO publications, showing the main uses of budget data and how it is analysed by African CSOs. These include drawing
comparisons across sector spending and measuring CSOs own budget priorities against the governments. This list will be used in the data analysis to explain data use in Uganda but will require further abstraction to understand how sub-national data is used in Uganda.
Swedburg and Hedstrom (1999) argue that much empirical and theoretical social science work neglects the task of identifying the change mechanisms which explain phenomenon in the world. Change mechanisms are causal patterns which aid analytical abstraction from the empirical data in order to define the relationship between phenomenon . (Hedström & Swedberg, 1996). Jetzek et.al have taken this abstraction and placed it within a discussion of open data. They outline two mechanisms which explain how open government data creates
18
value, these they name the transparency and participation mechanisms (Jetzek, Avital, & Bjor, 2014, p. 105).
Mechanisms can generate macro-level change through micro-level patterns which
reproduce contexts, actions and transformations. (Jetzek et.al., 2014: 103). This last point is important for understanding the impact of micro level collective action at the macro level. It can therefore be possible to view micro level causal mechanisms as building up to a larger picture of macro changes.
2.7 Conceptual scheme
Figure 2 is a scheme of the main concepts and their inter-relationships. The main theories which relate to my research question outline how open data creates hard accountability, as a form of advocacy, and clear transparency while also emphasising the difference in scales of these concepts. The main research questions has driven the focus of exploring the concept of openness. Theorists argue for an increase in feedback loops with data to strengthen social value and have overcome the challenge of low accountability by
identifying change mechanism. In comparison to the one dimensional open data modus of Ugandan political institutions which are intended to retain power. The conceptual scheme shows how these approaches lead to different outcomes on the scale and reflects the main question by asking how these two approaches, characterised by Lyon (2007) as either caring or controlling, influence each other.
19
Figure 2- Conceptual scheme
2.8 Conclusion
This chapter has drawn on literature related to open data, transparency, social
accountability, political transparency and civil society advocacy. It has drawn together theories with which to frame and develop the main concepts. Sen’s capabilities approach has enabled the development of advocacy into its above discussed parts – social
accountability and transparency. Using Lyon (2007) the concept of civil society advocacy has been developed to lead into a discussion of the emerging field of bottom up monitoring and evaluation. This chapter has also drawn on cross-continental writers in order to be sensitive to different cultural perspectives and to describe the Ugandan political context. Micro theorists will provide the empirical analysis with scales to assess the data against.
20
3 Research Methodology
3.1 Introduction
This chapter will introduce the sub questions (3.2), the epistemological departure points and the methodology and research design which follows on from this (3.3, 3.4). It will then describe how sampling was conducted and the units of analysis (3.5). The methods will be outlined under sub-sections of primary and secondary data again backed up with sound logical explanations (3.6). Section 3.7 will describe how the data will be analysed in order to answer my research questions. There is also an initial reflections on the limitations of the chosen methods (3.8). Annex one contains an operationalization table of the main research concepts which defined the indicators in order to inform the interview questions and help outline the scope of the study.
3.2 Research question and sub-questions
The main research question is divided into four sub-questions for further conceptual development.
Sub question 1:What is the landscape of open budget data use in Uganda? Sub question 2: How transparent is budget data?
Sub question 3: How has open budget data impacted governments accountability to public services?
Sub Question 4: How does the political context affect the advocacy practices of civil society groups?
3.3 Epistemology
Sociologists have argued the importance of identifying the philosophical belief system, or paradigm, behind research (Cresswell J. , 2009, p. 1). A paradigm is a set of basic beliefs or worldview which can guide the research in data collection and evaluation (Lincoln & Guba, 1994, p. 105). This research employs a subjective pragmatic epistemology, a theoretical departure point which rejects the positivist notion of an absolute reality. The approach seeks to move away from debates about the nature of reality and instead uses mixed data methods to focus on a research problem in order to find practical solutions. It draws on the
constructivist paradigm in that it acknowledges the construction of multiple realities
through individual interactions in society (Cresswell J. , 2009, p. 8). The scientific paradigm drives the nature of scientific questioning towards qualitative how questioning with an inductive line of enquiry which will drive the research question and sub-questions in order to create theory from empirical research. The epistemology will define the research
21
holds that realities are based on local and specific experiences so they can be both multiple in nature and shared (Guba & Lincoln, 1994, p. 110).
3.4 Methodology
The increasing emphasis on quantitative data for good governance in Africa ( (Delapalme, 2011) and the lack of data showing the qualitative impact of development spending drove the qualitative methodological focus for this thesis. This was in order to research how the stated impacts of data for development are being applied in Uganda. This research employs an exemplifying case study design, that is to research a typical (or exemplary) case over a rare or unique case in order to examine key social processes and which will allow for some generalizability. The research design acts as a framework for the collection and analysis of data. By selecting a typical case study the research can draw links between concepts easier. (Bryman, 2008, p. 56).
A case study approach allows researchers to fully explore the structure of the organisations being researched. In some ways this case study is also revelatory – digital budget data has only been available in the accessible online format since 2013 while the toll free call line has been live for less than one year. Two main qualitative data collection methods were
selected: semi-structured interviews and observations. The research also uses document analysis and a focus group method. These methods were chosen because they yield rich data during collection, necessary for the exploratory research design.
The research will use a qualitative mixed methods approach in order to triangulate findings. Mixed methods research means adopting a mixture of methods in the data collection which can be either mixed qualitative, quantitative or both (Sumner & Tribe, 2008).
3.5 Sampling
My initial respondent sample were selectively chosen, known as purposive sampling
(Bryman, 2008: 458) followed by a snowballing approach2. The samples were pre-select to
locate respondents who had come into contact with or previously used the web portal. Civil society was operationalised as NGOs, CBOs and the media in order to research a wide variety of civil society actors who drive advocacy work and have a presence on the ground. NGOs selections were further narrowed down to those who are working on improving public services. With the media this was improbable (journalists inform on a large range of subjects) so I selected media contacts who had been through data journalism training and had therefore been introduced to the web portal.
Respondents from three geographical areas (Kampala, Nebbi and Mbale) were choosen to ensure the findings were not limited to one geographical area. By spending some time in the
2
22
field scoping out the political angle and interest of data use within different media houses, the research was able to target a range of sample respondents here too.
3.6 Methods
3.6.1 primary data
Semi-structured Interviews: A total of 23 in depth interviews using semi-structure
interviews were conducted with open ended questions in order to keep the discussion on
track but also to allow the respondents space to drive the discussion3. As the most widely
employed qualitative method, interview data will provide the bulk of the rich data essential for qualitative research (Bryman, 2008: 436). Rambling in the conversation was steered towards answering the main research question and by focussing on the four sub-question areas it was possible to keep time and adapt the length of the interview without losing essential data. The research questionnaire was divided into roughly four sections, each corresponding to a sub question. As not every question was suitable for every interview it was necessary to carry a wide selection of relevant questions. For example questions suitable for a Community Based Organisation (CBO) would not be suitable for a radio presenter. Confirming statements were used after responses to verify the answers for the recording and follow up questions were noted down after reading through each transcribed interview. After checking if it was fine to do so, questions were e-mailed to the respondents though unfortunately few gave the time to respond back.
There were some issues of non-response to interview requests when in the field therefore not everyone in the target sample was interviewed. In these situations discretion was used when making direct approaches to organisations. For instance NGOs were directly
approached but government ministries, without formal correspondence, were not. I made sure to introduced myself, the UvA, the purpose of the research and to assure respondents verbally that their responses would not be used or shared outside of the research. For those who were interested I offered to share the finding of my research with them and several contacts are keen on maintaining a relationship for future work.
Observations: Observations allow behaviours to be directly observed rather than inferred
from secondary data (Bryman, 2008:254). During the observations I took an overt
participant observer approach (2008:257) in order to gain the perspective of a civil society organisation. This approach made note keeping and requesting translations easier. At these events I was able to gain contextual knowledge to the research through non-structured conversations and to network for more respondents. I found this method particularly useful for providing insight into how the civil society community collaborates to promote digital budget data for advocacy, an element which would be difficult to research otherwise. It also helped broadened the scope of the civil society as operationalized in the study to include
3
23
dimensions such as academics or government quangos which were beyond the scope of this study.
Participatory focus group: The research included a participatory method in order to
explore the central themes of the research question from the perspective of a CSO working at sub-national level (Bryman: 2008, 473). A participatory time line method was selected in order to map the experiences of using budget data on the ground. This method will help triangulate findings and include the voices of the people who need transparency in public service delivery the most, those in the poorer rural communities. This method is preferable over a focus group due to the language barrier and in order to make the research more interactive.
3.6.2 Secondary data
Secondary data collected from the field includes government correspondence, photos and media coverage as well as online comments extracted from the web portal. Secondary data was especially important for data on the media component who have a greater online presence. The data collected includes:
1. Quantified feedback date from the Budget.go.ug website
2. Photos made of public displays of budget information as well as from observations 3. Online and in print budget coverage of the last two years from blogs and news
articles
4. Official correspondence from Ministries regarding the use of open data.
3.6.3 Units of analysis
The units of analysis were NGO and CBO organisations as well as individuals from the media. The main focus was on groups who had come into contact with and used budget data in order to analyse how data is being used, with a focus on groups who are creating social value. These include NGO volunteers, advocacy groups, radio chat show hosts and
journalists from the media. This requires an abstraction beyond individual behaviour to look at collective actions in Ugandan civil society.
3.7 Data analysis
After recording and transcribing a process of open and axial coding was used to break down and organise the data for analysis (Bryman, 2008: 542). Open coding is useful for examining and categorising data which can then be grouped together in order to find the relationship between concepts and to locate core categories with which to build a middle-range theory (Bryman, 2008: 7). I will triangulate my findings with secondary sources and data across the range of CSOs represented to analyse the impacts of the data revolution on civil society.
24
3.8 Data validity
Bryman (2008) outlines three prominent criteria for the evaluation of research validity: reliability, replicability and validity. For qualitative research however some amendments have been made because of the difficulty in measurability. Four criteria for qualitative research have been theorised by Guba & Lincoln (1994). These are: credibility –
strengthened in this study with the mixed methods approach which provides triangulated findings. Transferability – which relates to generalizability. In this study it is recognised that there is a low generalizability, so the research will not attempt to make any generalising theories beyond the rural and urban contexts of Uganda. Dependability has been assured with careful transcription and record keeping of data and finally confirmability has been assured by conducting independent research throughout my time in the field.
3.9 Limitations
All respondents spoken to knew about the budget website however not all had used or downloaded data. Locating CBOs who had access to the website because the digital divide was a foreseeable challenge which may require reassessing the generalizability of the study results. This can be overcome by making sure my data analysis states clearly states which civil society group the findings can be generalised to.
4 Conclusion
In total the primary data collected amounted to 23 in-depth, semi structured interview, 3 participant observations and 1 participatory timeline. The secondary data amounted to one official correspondence letter, roughly 200 online comments from the web portal and 4 -5 news articles which use sub-national data. The research worked from a constructivist epistemological departure point which informed a qualitative methodology and the rich data methods used. The mixture of methods applied during the data collection was helpful in order to collect data from offline and online sources helping to establish if there are any useful applications of open data at the sub-national level and if any common techniques are emerging around its use. The reflective approach of the research allowed the researcher to judge which civil society organisations to conduct research on and interpret uses of open data, as suggested by the inductive line of enquiry. The case study approach required selecting atypical CSO groups as the main focus and to increase generalizability.
25
4. Country context
This chapter will begin with an overview of the historical context (4.1) followed by an
economic & social context of the country, comparing GDP and HDI indexes (4.2). Section 4.3 will outline the political changes President Yoweri Museveni has made and the institutions in place. A brief overview of what civil society looks like in Uganda (4.4) will be followed by the technological changes and the uneven growth of ICT. Looking closer at the topic of budget data sections 4.5 & 4.6 will cover historical changes to the budget and how open data has impacted on budget use.
4.1 Historical context
Uganda gained independence in October 1962 from British colonial rule. The British handed power over to the Uganda Legislative council (LEGCO) which went on to form the Ugandan parliament. This move instilled in Uganda the first democratic structures which would later be corroded by Oboto and his supporters (Mutibwa, 1992). The British had used the
Buganda tribe to rule the country through because of their strength, owning to their political structuring which centralised power through an hereditary leadership. The country is split into several distinct kingdoms following tribal identities. Rivalries
between the kingdoms were exasperated during and after colonialism following favouritism by the British. For instance, the British granted the Buganda tribe non-heredity land titles increasing tensions and laying the grounds for future tribal hatred as well as granting them greater freedoms through independence in the form of economic and political preference (Mutibwa, 1992).
Since independence democracy has alluded the country as two dictators have held power for over 17 years, between 1963 until the general elections in 1980. The colonial
administrative structures such as the national budget, were not altered in this time and a system of closed records and data continued in Uganda. The political structures which Oboto created and which allowed Amin to retain power were not challenged until Amin’s overthrowal when the National Resistance Movement (NRM) led by Yoweri Museveni demanded a return to democratic elections for the first time since the negotiation of independence (Mutibwa, 1992).
Elections were finally held in 1980 when Yoweri Museveni was elected president, and have been held every five years since with Museveni continuing to hold power. Economic reforms followed the political reforms because the county had seen economic stagnation and debt during Amin’s regime. These included reducing poverty, enhancing transparency, openness and accountability (Kuteesa, Magona, Maris, & Wokadala, 2006). However, the need to retain power by the ruling party has seen the corrosion of these reforms as spending priorities are re-directed and selective transparency is implemented. In the last election (2011) it was estimated the NRM spent Shs350 billion maintaining power and is set to spend more in the coming election (Matsiko, 2014).
26
4.2 Economic and social development context
East Africa generally performs poorly in comparison to west and south Africa in areas such as transportation and trade. Figure 3 below shows the percentages of Uganda’s GDP in comparison to the country’s debt and the tax revenues (WB, World Development Indicators, 2016). The Initial peak in 2006 can be explained by the discovery of oil, which dipped in 2011 due to inflation and has slowly been making a recovery since with a GDP growth of 5.9% in 2014. This macro-economic stability is driven by the agriculture and telecoms industries ( Alexis Rwabizambuga, 2015, p. 4). Government debt, however, remains consistently high reflecting the reliance on international aid which in 2013 was just over 1,000M official development assistance (ODA) from DAC countries (WB, World
Development Indicators, 2016).
27
Figure 4 shows the spread of mobile connectivity across Uganda, indicated by the purple
dashes, demonstrating the growth of the telecoms industry ( Ranganathan & Foster, 2015).
Figure 4 - Infrastructure map showing connectivity coverage in Uganda
The main challenges Uganda faces for infrastructure building are securing electricity supply, roads construction and the HIV/ AIDS endemic ( Alexis Rwabizambuga, 2015). Traditional economic sectors are dominated by exports of coffee, tobacco and fish. Uganda is a land locked country with a population of 37.58 Million (WB, World Development Indicators, 2016). The country has been following IMF policies of low inflation through control of interest rates in order to encourage international investment and GDP growth. In 2014 the budget deficit was 4.9% and growing from previous years. This widening budget deficit holds the country back from investing in human and infrastructural development. The country experienced international donor aid cuts in early 2014 after President Yoweri Museveni signed into law a bill imposing tough penalties for homosexuals. The move sees the country attempting to distance itself from donor influence.
Uganda has been rapidly urbanizing due to increasing international interest in the country, putting this into context however is important as currently only 18% of the population live in urban areas. Urbanisation is built up along the major trade routes following the now defunct regional railway networks (Sengendo, Lwasa, & Mukwaya, 2010, p. 268). The majority of the countries revenue is made and collected in the only registered city, Kampala, though there are several urban centres in the running to be given city status in the coming years (Jinja, Mbale, Arua, Mbarara and Gulu) which will rival the economic strength of Kampala in the future.
28
The district areas of Kampala, Nebbi and Mbale will be the focus of this research. Kampala is the fastest urbanizing city of Uganda with population increase of 5.6% / annum. Mbale is within the country average of 5% / annum growth rate (UBOS, 2014) while Nebbi in the north is predominantly rural, it goes against the average urbanising trend but this is typical of the northern districts. With rapid urbanization also comes rapid technological change, the African telecoms market is booming with higher speed internet connecting more
businesses. With the increased data, poverty can be mapped more closely to peoples lived experiences on the ground.
Looking at average figures, poverty would appear to have declined in the urbanized region of Kampala however Uganda comes in one-hundred and sixty-fourth on the HDI and growth in consumption is consistently low for the poorest income levels in society, demonstrating a rising income inequality gap (HDI, 2014, p. 39). The development of public services has also not matched the economic growth of the country as the World Bank estimates of 40-60% of urban accommodation are slum dwellings, corruption has a large part to play in this uneven development. In his five year development plan Yoweri Museveni has included building human capacity as a target to improve social and economic development.
4.3 Political context
The current political climate sees tensions building in the upcoming 2016 elections.
Museveni has been in power since 1986 and only recently ended the one party rule. He has been accused of a sham democracy in previous elections while there have been a number of human rights abuses and bribery cases linked to the 2011 election campaign (Izamo & Wilkerson, 2011, p. 65) and a host of laws have recently been passed concerning freedom of speech and ICT control. This suggests political control through violence by the ruling party. The continuation of corruption within the current regime is shown by the Transparency International rankings, especially in the kick-backs for public service contracts (WB, April 2004). Some argue that Museveni’s power is weakening due to corruption charges, for instance amendments to the anti-corruption bill were passed in parliament in order to tighten the law on corruption and bring political leaders in line with the law (Parliament, 2015). Internal attempts from the ministry to stamp out corruption have resulted in act such as the Leadership Code of 2002 (WB, April 2004).
Andrew Mwenda a Ugandan journalist reports on how Museveni continues to hold power and how international donors have been slow coming to realise that their support is propping up an authoritarian system. This quasi-authoritarian regime, once hailed as the economic success of Africa, until recently did not come under much international criticism due to reluctance to find fault with a president that has seen the longest reign of peace since taking power. In this way corrupt elite officials have continued to go uncharged (Mwenda, 2014). The strong leadership of Museveni was, arguably, necessary after the Amin years but now the international community is starting to take note of the slow
29
democratic reforms. There is decreasing trust in government bodies which are seen as corrupt, closed institutions, particularly the Presidency, Parliament and the Police (AfroBarometer, 2015).
Uganda has been following a decentralisation process since 1997 which involves making local governments in each district semi-autonomous in order for people to hold their local officials to account and have more say in local affairs (OECD, 2004). Despite this, some entities in Kampala are still dealt with directly through the central government office making them effectively personal entities of Museveni, another sign of his strong grip on the
country.
4.4 Civil society
Civil society in Uganda find themselves working in an increasingly regulated society. For example, the recent the passage of the NGO regulation bill, which some view as at attempt to curtail civil society, only allows NGOs to work with certification by the central
government. Figure 5 shows Ugandan civil society score as measured by CIVICUS, a civil society strengthening alliance. They conduct a global index based on four dimensions and 72 indicators.
Figure 5 – Uganda’s score in the civil society index (2009)
The structure dimension measures the size, strength and vibrancy of civil society which is strong in Uganda. The environment dimension measures the overall political, social,
economic, cultural and legal environment within which civil society exists which in Uganda is somewhat disabling.
Dimension Scores: Structure: 1.8 Environment: 1.4 Values: 1.9 Impact: 2.3
30
The value measures Ugandan CSOs practice and promotion of positive social values, which is moderate. Finally the impact dimension describes the extent to which civil society is active and successful in fulfilling several essential functions in Ugandan society and politics. Increasingly Ugandan CSOs have been taking part in international discussions which has led President Yoweri Museveni to be distrustful of CSOs with large or external funding (Monitor, 2012). CSOs influence on policy in Uganda has generally been decreasing (Robinson & Friedman, 2007) which has led some CSOs to take on a more outwardly vocal role through ICT.
4.5 Technological context
Across East Africa Uganda excels in open budget data. According to the Open Budget Index (OBI) Uganda scores 62/88 globally, the highest in East Africa and second to South Africa in the continent, although it ranks as the lowest within its ‘substantial openness’ category (OBI, 2015). The web portal launched in 2013 provides open data access to national and sub-national budget data with the aim of eventually building capacity of local governments so each district uploads their own budgets online. This transparency initiative sets Uganda above Kenya which comes second for open data in East Africa. With the completion of a fibre optic cable in to Kampala this spring (ItNews, 2015) Uganda will have increased global connectivity, matching the trend across Africa to connect the continent.
Uganda has a history of reforming open budget data since the 1980’s. For example the 1992 Mid-Term Expenditure reforms (MTER) which implemented budgetary fiscal discipline and aimed to reduce poverty. This led to newspapers regularly publishing the national
government releases and a reformed participatory budget process (Kuteesa, Magona, Maris, & Wokadala, 2006). In addition the MoF plans to launch a mobile app of the web portal along with other changes, such as improvements to the search engine, reflecting the increasing use of technology by the central government. Civil society also seem to be responding by using ICT technology to campaign more openly and effectively for their aims (HRW, 2011).
Regarding ICT uptake, only 2,2% of all households have a working laptop in their homes with large differences between rural and urban areas. Furthermore between 30 to 50% of people (UCC, 2015)have not accessed the internet at a cafe or other network point. Furthermore, literacy rates among the general populous are low so reading data is a challenge. Together these statistics demonstrates there is a digital divide in Uganda (Fuchs & Horak, 2008).
4.5 Participatory budget structure
Uganda has a strong written policy for a participatory budget structure which was created to allow feedback of local needs in to the budget. The Local Government Act of 1997, which bought into effect decentralisation also set out requirements for a participatory process to
31
create the sub-national budget in order to improve services to the poor (IFPRI, 2011).
Participation takes the form of local meetings in the smallest administrative unit, the village, which then feed into the larger units to create an inclusive budget. There have been
complaints however that power is concentrated at the sub-county and district level and that the participatory process is merely consultative (Kiwanuka, 2012). Furthermore the central government only send 15% of the total budget to the districts (DI, 2015) leading to low capacity among local leaders to implement development projects and poorly planned local services (IFPRI, 2011).
The primary form of income for the majority of Uganda’s 111 districts is in the form of development grants from the central government. To supplement this local governments are expected to collect local revenues through taxes and rents. Many however are reliant on the central government for money. The grants are dispersed at district level where local governments take charge of spending and developing local services and infrastructure. This further adds to the challenge of making local voices count in the participatory structure. Budget data can help bring power back to the participatory structures by empowering people through information. The data can help citizens see how resources are being spent locally and help them make informed decisions when planning future budgets.
4.6 The national budget and the data revolution
The national government considered budget date published in books as open data for a long time. Consequently budget data was expensive to reproduce which meant many media houses and NGO's alike could not access to them. There were many complaints from CSOs however that these were difficult to access ( Nandyona, 2014). Requesting information was a long process with low success rates, as the Development Research and Training (DRT) a Ugandan NGO who focus on policy research and analysis to reduce inequality have
documented (Anderson, 2015). Even now basic information such as school enrolment, fees and demographic data requires money or contacts to access. Between five and ten years ago the only information you could find about the budget would be from the budget speech or from internal contacts. For those who did not have internal contacts, gaining access was problematic (ibid). In this context CBOs have to work with an ill- informed public while media houses gain little experience writing or working with factual data sparking a call for a
data sensitive revolution in Africa as seen in the High-Level Conference on the Data
Revolution in Africa held March 2015 in Addis Ababa, Ethiopia (Adieno, 2015).
5. Conclusion
Overall Uganda’s recent consistent economic growth has seen the country attempt to move away from donor policies, but this growth is uneven and does not benefit the majority of society. Inequality holds CSOs back from participating and limits the effects of
32
dictatorships and cronyism. International donors are starting to wake up to the slow
reforms in politics and the increasing amount of laws in place which are limiting CSO efforts for reform and are shifting their aid from bi-lateral to multi-lateral which is creating tensions in the country between CSOs and the ruling party.
33
5 Empirical analysis
Chapter five will start by describing the landscape of budget data, how the portal came about and local/global responses to it. Section 5.1 will describe how CSOs are engaging citizens with budget data both online and offline and what advocacy products and processes have emerged with open data at the national and sub-national level. Section 5.2 will study how transparent the budget website it based on CSO data needs.
5.1 The landscape of data use
This section will describe and explain the landscape of digital budget data use: how civil society accesses, engages with and disseminates the budget as well as exploring the divisions of labour between the national and sub-national level.
5.1.1 Responses to digitisation
In 2013 the Ugandan Ministry of Finance (MoF) paired up with the Overseas Development Initiative (ODI) a UK think tank who work through partners in developing countries. They aim to inspire and influence policy through training, research and consultation in order to
reduce poverty. Together the ODI and MoF built Uganda's first open budget web portal4.
The website includes a search function down do the parish level listing budget planning figures and quarterly release dates. It includes a feedback section and a toll free line available in five languages. The website represents the first time sub-national level data is freely available in Uganda. To understand the landscape of digital data use, this section will look at how and why budget data are digitised online and offline through the lens of civil society responses. The section will conclude by drawing comparisons with international data engagement standards, five steps which open data practitioners can take in order to
improve engagement, as established by Davies (2012). Responses to national digital data
National budget data, to the districts level, have been printed every quarter in major
national newspapers for the past few years. Figure 6 gives an idea of the level of print detail. It includes central government budgeting figures, given as grants and split into sectors.
4
34
Figure 6 - Newspaper print data
CSOs such as the BUGISU NGO forum in Mbale, an NGO which wants to deepen democracy through participation, have been using budget data prior to the web portal and view print data as Uganda’s first open data step to create accountability around public funds,
“what now the ministry (of finance) started doing is that they would publish, each time they release money, they publish it in the paper so that everyone knows that money has come [laughs] there should not be any excuse for not implementing. Interview 16
Open budget data reforms in Uganda have been written about extensively (Tumusiime-Mutebile, 2010). However, it is important to bear in mind that national budget data represents only about 15% of the total country budget (DI, 2015) reminding us of the voluntary nature of open data in Uganda. There is low transparency in many other government departments, most prominently the defence and state house spending, demonstrated by the low response rates to Access to Information (ATI) requests. Between the 2008 and 2011 a total of ten out of thirty-three requests were granted (AFOIC, 2012). Responses to sub-national digital data
Sub-national budget data includes figures of the sub county budget data down to the village level. Figure 7 shows where this fits in Uganda’s governing structures. Since
decentralisation in 1997 each district plans their own sub-national budget and the money is received at the sub-county level.
