Expression(s) of Poverty
The Impact of Measuring
Poverty in Ghana in the Context
of Development Goals
Lianne Schmidt Research Master International Development Studies University of Amsterdam Lianne.schmidt@xs4all.nl +31 6 28721973 wordcount: 35.511 August 9, 2019 Supervisor: Dr. N.R.M. Pouw Second Reader: Dr. C.L. Vegelin Image: painting pictures of Ghana (Accra, March 3 2019, by L. Schmidt)
Foreword
To me, this thesis is more than just the final outcome of my studies. It is a project that has required a lot of determination and patience, which proved quite challenging in their combination. This thesis has come to represent my professional and personal journey over the past few years, which I feel very proud of.
First and foremost, I would like thank my supervisor, Dr. Nicky Pouw, for her guidance and trust. This research project greatly benefitted from her constructive feedback, which always provided new perspectives to my thinking. Thank you for challenging me in such encouraging and supportive ways; this has meant a lot to me.
I cannot think of a better second reader of this thesis than Dr. Courtney Vegelin. Her teaching inspired me to follow the path of International Development Studies. I would also like to thank Eva van der Sleen, my study advisor. Her kindness and encouragement provided me with the trust I needed to take time to heal. In addition, her support motivated me to take the leap and apply for the UN World Data Forum. My participation in the conference was made possible by a grant from the Amsterdam University Fund, for which I am very grateful.
This thesis relies heavily on the inspiration provided by Prof. Dr. Daniel Mügge. His work on the political economy of macroeconomic measurement has been incredibly influential in shaping my own academic thinking. I always looked forward to our reading group sessions and feel grateful for the insights it provided.
When I left for my fieldwork, I could not have imagined the kindness I would be greeted with in Ghana. I would like to thank Joshua Kobla Adzakpa in particular for welcoming me and sharing his time, opinions and connections with me. More in general, I am thankful to each and every research participant for generously sharing their ideas. I still feel touched by the (com)passion that drives much of their work.
I miss my home away from home in Accra where I was welcomed with the most delightful
akwaabe (welcome) every day. I have been incredibly lucky to meet Damiana Salm and Jordan
Fletcher, both extraordinary women and dear friends. This research project would not have been the same without their thoughtful reflections, laughter and support.
Of course, I am grateful for the support of my friends and family. Querine Hoejenbos and Caroline Heuschen need to be mentioned for always being there, not just in the context of this thesis. A special word of thanks goes to my yoga teacher Sasha van Aalst, for consistently providing gentle nourishment to my early morning practice. The clarity I (try to) cultivate on the mat has been an important foundation for writing this thesis.
To all of them and to you, the reader of this thesis:
Medaase. Thank you.
Lianne Schmidt Amsterdam, August 9, 2019
Abstract
The Sustainable Development Goals (SDGs) provide an ambitious agenda; not only for poverty reduction, but also for its measurement. This is based on the assumption that data will help to achieve the SDGs and the goal of eradicating poverty more specifically. Nevertheless, current literature does not address the impact of measuring poverty as part of the Development Goals. This research addresses that knowledge gap by examining the impact of consumption-based poverty measurement in Ghana. It does so by explicitly contextualizing the case study of Ghana within the international discourse on data for the SDGs. For this reason, this research builds on fieldwork undertaken at both the UN World Data Forum 2018 as well as research in Ghana itself. It integrates the results of semi-structured interviews with key stakeholders, participatory observation, document analysis and descriptive quantitative analysis of existing poverty data.
The results of this research show that consumption-based poverty data strongly affects how people think about poverty. Even though the height of the poverty line is set too low to correspond with stakeholders’ notions about what it means to live a life free from poverty, the headcount ratio powerfully shapes ideas about poverty reduction progress. The Development Goals have contributed to this impact of the poverty headcount ratio by featuring it as a primary indicator for the eradication of poverty. The discursive strength of the first Millennium Development Goal causes people to be familiar with Ghana’s poverty reduction success, but much less with the recent slowdown thereof. Most surprisingly, this research shows that this poverty data is hardly used for policymaking to address this slowdown and reach the SDGs. This can be explained by a mismatch between the aggregate levels of poverty measurement and the district level where most policymaking takes place in Ghana. Stakeholders mostly use poverty data to demonstrate accountability and affirm their legitimacy, which has become a necessary strategy to acquire sufficient funding for poverty reduction programmes. This has increased the emphasis on poverty data itself, rather than the ways it can be used to reach the SDGs.
This research concludes by introducing a much-needed conceptual framework for studying the impact of indicators in the field of international development studies. In addition, it provides specific recommendations for making the forms, uses and regulatory environment of poverty measurement more inclusive. Most importantly, poverty measurement should provide better insight in who is left behind through different forms of data disaggregation at the district level. Ultimately, this would enable the increased use of such data for evidence-based policymaking to contribute to eradicating poverty and reaching the SDGs, both in Ghana and beyond.
Index 1. Introduction 1 1.1 Background 1 1.2 Problem Statement 1 1.3 Research Approach 2 1.4 Thesis outline 2 2. Theoretical framework 3 2.1 Introduction 3
2.2 The Use of Indicators in International Development 3 2.3 (Inter)National Actors involved in Measuring Development 3
2.4 Measuring the Goal of Poverty Reduction 5
2.5 Measuring the Measurement Impact 6
2.6 The Power of Indicators as Goals 8
2.7 The Challenge of Expressing the Poverty Goal on a Country Level 9
2.8 Conclusion 10
3. Research Methodology 11
3.1 Introduction 11
3.2 Ontological and Epistemological Considerations 11
3.3 Research Questions 12
3.4 Conceptual Scheme 14
3.5 Operationalization 15
3.6 Methods 16
3.6.1 Unit(s) of analysis 16
3.6.2 Mixed Methods Research Strategy, Case Study Design
& Process Tracing 17
3.6.3 Qualitative Data Collection and Analysis 19 3.6.3.1 Participatory Observation at the UN World Data Forum 19
3.6.3.2 Document Analysis 19
3.6.3.3 Semi-structured Interviews in Ghana 20 3.6.3.4 Participatory Observation in Accra, Ghana 21 3.6.4 Quantitative Data Collection and Analysis 22
3.6.5 Ethics 22
3.7 Limitations 23
3.8 Conclusion 24
4. Research Context 25
4.1 Introduction 25
4.2 UN World Data Forum 25
4.3 Accra, Ghana 25
4.4 Development and Poverty Reduction in Ghana 27
4.4.1 A brief History of Ghana’s Development and
4.4.2 Measuring Poverty Reduction in Ghana 28
4.5 Conclusion 29
5. Global Data Forces at the World Data Forum 30
5.1 Introduction 30
5.2 The Global Discourse on implementing the Data Revolution 30
5.3 The First Contradiction: Supply and Demand 32
5.4 The Second Contradiction: Harmonization and Disaggregation 33 5.5 The Third Contradiction: Data as a Goal and data for the SDGs 34
5.6 Conclusion & Implications 35
6. Process Tracing Ghana’s Statistical Development 37
6.1 Introduction 37
6.2 A brief history of Ghana’s National Statistical System: from Centralization to
Decentralization 37
6.3 From Decentralization to Planning: the first Ghana Statistical Development Plan 38 6.4 From Planning to Realization: towards the second NSDS 39
6.5 Evaluating Ghana’s Statistical Capacity 40
6.6 Conclusion 42
7. Measuring and Understanding Poverty in Ghana 44
7.1 Introduction 44
7.2 Poverty Measurement in Ghana 45
7.3 Poverty Reduction Trends in Ghana 47
7.4 Understanding Poverty through Official Statistics 50 7.5 Understanding Poverty through different kinds of Data 52
7.5.1 Decomposing the Headcount Ratio: Disaggregation by Region
and other Dimensions 52
7.5.2 Decomposing the Headcount Ratio: Vulnerability 55 7.5.3 Understanding Poverty through Qualitative Data 55
7.6 Conclusion 56
8. Poverty Data in Governance 58
8.1 Introduction 58
8.2 The (inter)national Governance Context of Poverty Data Impact 59
8.3 Governance Effects of Poverty Data 60
8.3.1 Outline 60
8.3.2 Poverty Data for Policymaking, Signalling and Targeting 60 8.3.3 Poverty Data for Monitoring, Evaluation and Accountability 61 8.4 Knowledge and Governance Impact Interactions in the Field of Poverty 62
8.5 Measurement Impact on Poverty 63
8.5 Conclusion 65
9. Governing Data for Poverty Reduction and Inclusive Development 67
9.2 Governing Poverty Data as part of the Voluntary National Review 68 9.3 Governing Poverty Data through the Setup of a Quality Assurance Framework 70 9.4 The Challenges of Creating Inclusive Indicators: the Case of the
MDCP Indicator 72
9.5 Criteria for making Forms and Uses of Poverty Data more Inclusive 74
9.6 Conclusion 75
10. Conclusion 77
10.1 Answering the Main Research Question 77
10.2 Theoretical Reflection & Contribution: introducing a Framework for Studying
Expression of Development through Indicators 79
10.3 Policy Recommendations 82
10.4 Reflections on Methodology 83
10.5 Recommendations for Further Research 85
11. Bibliography 87
Appendix 1: Selection of relevant Development Goals 91 Appendix 2: Unit(s) of Analysis in relation to Research Questions and Methodology 92 Appendix 3: List of UN World Data Forum Conference Sessions attended 93 Appendix 4: List of Documents included in the Document Analysis 94
Appendix 5: Interview Guide Example 96
List of tables
Table 1: Operationalization of Main Concepts Table 2: Relevant Sustainable Development Goals Table 3: Millennium Development Goal 1.1
Table 4: Units of Analysis in relation to Research Questions and Methodology\ Table 5: List of UN WDF Conference Sessions Attended
Table 6: Overview of Qualitative Data Sources
List of figures
Figure 1: Conceptual Framework Figure 2: Research Design
Figure 3: Development of National Statistical System and Statistical Capacity in Ghana Figure 4: Poverty Reduction in Ghana
Figure 5: Inequality and Poverty Depth in Ghana
Figure 6: Proportion of Population living in Poverty disaggregated by Region
Figure 7: Framework for the Study of the Expression of Development through Indicators
List of images
Image 1: Letter by Pobee, M.A. A. (2018) to the United Nations to the United Nations to announce Ghana’s participation in the Voluntary National Review of the High Level Political Forum on Sustainable Development
Image 2: Ghana & Accra
List of Abbreviations
AGM Annual General Meeting CSO Civil Society Organization
e-GDDS Enhanced General Data Dissemination System
FAO Food and Agricultural Organization of the United Nations GSGDA Ghana Shared Growth and Development Agenda
GH1 Interview code; refers to interview code as explained in GHC old Ghanaian cedi
GHS new/current Ghanaian cedi GLSS Ghana Living Standard Survey
GPRS Growth and Poverty Reduction Strategy Paper GSS Ghana Statistical Service
HLG-PCCB High-level Group for Partnership, Coordination and Capacity-Building HLPF High Level Political Forum
IAEG-SDGs Inter-Agency and Expert Group on SDG Indicators IMF International Monetary Fund
IOs International Organizations
ISSER Institute for Statistical, Social and Economic Research LEAP Livelihood Empowerment Against Poverty programme MDCP Multidimensional Child Poverty
MDGs Millennium Development Goals
MDG1 The first Millennium Development Goal of Eradicating Extreme Poverty and Hunger, often used to refer to Target 1.A: Halving, between 1990 and 2015, the proportion of people whose income is less than $1.25 a day
NGO Non-Governmental Organization
NDPC National Development and Planning Commission NSDS National Strategy for the Development of Statistics NSS National Statistical System
PARIS21 Partnership in Statistics for Development in the 21st Century
PPME Policy Planning, Monitoring and Evaluation
RSPIR Research, Statistics, Information and Public Relations SDGs Sustainable Development Goals
SDG1 The first Sustainable Development Goal of No Poverty
SQ Sub-Question
UK United Kingdom
UN United Nations
UNSD United Nations Statistics Division
UN ESCAP United Nations Economic and Social Commission for Asia and the Pacific
US United States
UNSD United Nations Statistics Division VNR Voluntary National Review
Image 1: Letter by Pobee, M.A. A. (2018) to the United Nations to the United Nations to announce Ghana’s participation in the Voluntary National Review of the High Level Political Forum on Sustainable Development.
1
Never again should it be possible to say “we didn’t know”. No one should be invisible. This is the world we want – a world that counts. (The United Nations Secretary-General’s Independent Expert Advisory Group on a Data Revolution for Sustainable Development, 2014: 3)
1. Introduction
1.1 Background
On 11 July 2018, Martha Ama A. Pobee formally announced the participation of Ghana in the National Voluntary Review of the High Level Political Forum on Sustainable Development. This marked the start of an ambitious endeavour to measure and report national progress of Ghana on the Sustainable Development Goals (SDGs) in 2019.
The origin of this initiative can be traced back to 2015, when the international community set an ambitious agenda with the adoption of the SDGs; not just for development, but also for its measurement. With as much as 17 goals, 169 targets and 230 indicators, the SDGs seemingly not only aim to leave nobody behind, but also to leave little development unmeasured. Data on those living in deprivation is deemed to be crucial for effective evidence-based policymaking in order to lift people out of poverty (IAEG-SDGs, 2016). The first target of ending extreme poverty by 2030 has therefore become closely connected to the last target on statistical capacity building in developing countries (appx. 1, p. 91).
In this context, a lack of data in itself is increasingly regarded as a form of deprivation that should be ended (Serajuddin, Uematsu, Wieser, Yoshida, & Dabalen, 2015). Indeed, many developing countries are currently not well equipped to meet the statistical demands of the SDGs. Therefore the United Nations Secretary-General’s Independent Expert Advisory Group (2014: 2) has called for a “data revolution” involving more and better data on all relevant dimensions of development. This desire for more data is firmly rooted in the belief that better data will lead to better decisions and increased accountability, and is therefore crucial to achieving the SDGs and eradicating extreme poverty. However, the countries facing the greatest challenges towards reaching this goal, are often characterized by low statistical capacity (Jerven, 2013). Although the new poverty reduction agendas increased the demand for data, it did not present clear strategies for how to respond to it.
1.2 Problem Statement
Little research has been done on the actual impact and local adaptation of these statistical demands on a country level, which could reach much further than is often recognized. Although statistics are often regarded as representations of reality, they are rarely acknowledged for the social, economic and political realities that statistics have become (Jerven, 2013). Viewed from this perspective, the measurement of poverty not only reflects, but also comes to define development by determining what counts as poverty and progress as part of Development Goals. In addition, little is known about the assumed impact of poverty data on policies that contribute to reaching the goal of eradicating poverty.
2
These knowledge and governance effects of the use of statistics are often neglected in existing research, despite its potential effects on poverty itself. This process by which statistics arrange elements to characterize objects, and thereby ultimately potentially affect the object measured itself, has been called ‘expression’ by Didier (2008). In the light of the goal to eradicate extreme poverty, it clearly illustrates the relevance of this issue in relation to (research on) the implementation of the SDGs, and Ghana’s upcoming Voluntary National Review (VNR) more specifically. Therefore, this research aims to address this matter by answering the following research question:
How does the expression of poverty through consumption-based indicators relate to how poverty is understood and governed in Ghana in the context of Development Goals, and to its relation with progress on the goal to eradicate (extreme) poverty more specifically?
This research question uses the concept of expression to include both knowledge and governance effects of poverty measurement, as well as their relation to poverty itself. Importantly, it explicitly does so in the context of Development Goals, which have featured consumption-based poverty indicators as part of the first goal of eradicating poverty.
1.3 Research Approach
This research approaches the topic of poverty measurement from a unique perspective by specifically focussing on its impact on the understanding and governance of poverty. In order to answer this main research question, a mixed methods strategy will be used, integrating qualitative data collection and analysis as well as quantitative secondary data analysis. This research mainly builds on the data gathered at two fieldwork locations. First, the UN World Data Forum was visited, which was held from October 22-24, 2018 in Dubai. Participatory observation at this conference created a thorough understanding of the international discourse on the SDG data demands and their implementation. This way, the case study of Ghana could be explicitly positioned within this broader international context. Fieldwork in Ghana took place between January 3 until March 21, 2019, during which semi-structured interviews were held with key stakeholders. In addition, relevant meetings in relation to e.g. the Voluntary National Review process were observed to provide additional insight in ongoing processes. This approach thereby ensured the embedding of this research within current developments in Ghana, which contributes to the relevance of the results.
1.4 Thesis Outline
The following chapter provides the theoretical foundation to this research, by offering insight into relevant existing literature and theories. Chapter 3 (p. 11) describes the methodology that guided this research, including sub-questions and the research methods used to answer them. The subsequent chapters present the results of this research in line with the structure of the sub-questions. Ultimately, this leads up to a conclusion which provides an answer to the main research question, as well as a reflection on the research itself.
3
2. Theoretical Framework
2.1 Introduction
This theoretical framework presents an overview of the most important concepts and theories in relation to the measurement of poverty and its impact in the context of the SDGs. First, it describes the general role and impact of poverty indicators, followed by a description of actors relevant to measuring development. The measurement of poverty is explored more in-depth, after which its potential ways of impact are described. In doing so, this theoretical framework connects insights from anthropology, political economy, sociology, economic history and international development studies. Finally, existing research on measurement challenges experienced by developing countries is evaluated. Together, this provides a comprehensive overview of academic literature and theories that are relevant to the research topic.
2.2 The Use of Indicators in International Development
Although the terminology of the “data revolution” suggests the possible relevance of many qualitative and quantitative sorts of data, the term is mainly used by the UNs IEAG-SDGs to refer to statistics (2014). This is in itself an important sign of the growing prevalence of numbers in international development. In this field, statistics have become important instruments to measure performance, assess development progress and inform and evaluate policymaking (Fioramonti, 2014; Ward, 2004). For these purposes, development is measured according to a wide range of indicators, which can be defined as “statistical measures that are used to consolidate complex data into a simple number or rank that is meaningful to policymakers and the public” (Merry, 2011: 86).
According to Freistein (2015), the use of these indicators is not only motivated by their functional use, but also by their social and political relevance. Numbers have become a means to claim expertise and establish authority within a certain field, and to reaffirm the legitimacy for acting within it. This not only applies to international organizations (IOs) and national governments, but also to NGOs and private actors, which (increasingly) make use of statistics. Numbers have become important means of communication, both within and between organizations as well as towards the general public. In this context, statistics are believed to enlarge transparency and accountability (IAEG-SDGs, 2014). With regards to the SDGs, it is believed that the public availability of data will help to hold governments accountable for their contribution towards reaching the goals (ibid.). This explains the importance of the measurement of poverty reduction both on a national level and in an international context.
2.3 (Inter)national Actors involved in Measuring Development
As previously described, measuring development has become increasingly important for individual governments well as IOs. This is reflected by increased efforts to organize statistical activities and to improve statistical capacity, both on national and international levels.
4
On a country-level, the organization of the production and distribution of statistics is also known as the National Statistical System (Starr, 1983: 8) This system often includes a national statistical office that is in charge of the periodic collection, analysis and dissemination of data. This is far from just a technical exercise; Starr (1983) emphasizes its social, political, legal and financial dimensions that appear to be interrelated. How the statistical work is organized and to what extent it is independent from the state (the social dimension) is often determined by legislation (the legal dimension) and the extent to which it is dependent on foreign financial flows (the financial dimension). In addition, political priorities can also influence what is measured (the political dimension). All these dimensions share a historical component, in the sense that what is measured and in which ways is often path-dependent and shaped by history (Mügge, 2015: 418). This component is especially relevant to developing countries, in which statistical activities were often heavily shaped during the colonial period (Jerven, 2013).
With the increasing demand for data (in particular in the context of the SDGs), measuring development has also become increasingly demanding. Especially for developing countries, meeting the statistical demands presented by Development Goals has proved to be challenging and resource-intensive (Jerven, 2013). For these reasons, international organizations have been increasingly stimulating the production of statistics. Relevant actors include the United Nations Statistics Division (UNSD), the World Bank and the related PARIS21 initiative, and the IMF.
Historically, the United Nations have played a vital role in the creation of an internationally agreed upon statistical system, that serves as a data collection framework (Ward, 2004). Nowadays the UN Statistical Division (UNSD) coordinates the statistical activities of the UN; it aims to advance the global statistical system by supporting National Statistical Systems, improving methodologies and through international cooperation and coordination (ibid.). It also facilitates the review process for progress on the SDGs by offering a global SDG indicators database and functioning as the Secretariat of the earlier mentioned Inter-agency and Expert Group on SDG indicators (IAEG-SDGs, 2014). The UNSD also coordinates the organization of the UN World Data Forum and is formally headed by the UN Statistical Commission.
In addition to the UNSD, the World Bank is also an important actor relevant to this field of study. It is of course well known for its research and programmes in the field of poverty reduction, but its Development Data Group also has an important international role in the international coordination of statistics (Ward, 2004). In addition, the Bank has been offering financial and technical assistance for statistical capacity building to developing countries since the late 1990s (Clegg, 2010). The World Bank is also well known for bringing together development data from all over the world in accessible databanks. In addition, it also produces a relevant Statistical Capacity Indicator, which indicates the capacity of individual states to meet statistical standards (datatopics.worldbank.org/statisticalcapacity). Moreover, the World Bank had an important role in the creation of the Partnership in Statistics for Development in the 21st Century (PARIS21), which
5 “promotes the better use and production of statistics throughout the developing world” (paris21.org/about). One of the most important ways in which PARIS21 pursues this goal is by encouraging and assisting developing countries – including Ghana – with the design, implementation and monitoring of a National Strategy for the Development of Statistics (NSDS).
The IMF and its Statistics Department more specifically, is mostly relevant in this context for its Data Standard’s initiatives. These involve standards and strategies for the evaluation, monitoring and improvement of National Statistical System (Ward, 2004). The lowest level of this system is known as the Enhanced General Data Dissemination System (e-GDDS). This is aimed at countries with relatively low statistical capacity, including Ghana. Of course, the relevance of the IMF (and the World Bank) to Ghana and this research topic exceed the scope of statistical capacity building, and also include financial assistance and related Poverty Reduction Strategy Papers (as described by McKay, Pirttilä & Tarp, 2016).
Concluding, the production and dissemination of development statistics is not just a national affair, but is also directly related to international organizations through standard-setting, coordination and funding. It is in this context where the measurement of poverty takes place.
2.4 Measuring the Goal of Poverty Reduction
The relevance of measuring poverty reduction and inequality is illustrated by its prominence in both the Millennium Development Goals as well as the Sustainable Development Goals, where the eradication of extreme poverty features as the first goal (appx. 1, p. 91). Although the MDGs and SDGs as a whole represent a multidimensional approach to poverty, in line with its multidimensional nature, the first goals define poverty primarily through its economic dimension. In this context, the main indicator used to assess poverty reduction is the poverty headcount ratio based on the international poverty line. This headcount ratio is measured as the proportion of population living on less than $1.25 per day, making it a measure of absolute poverty. By taking into account country-specific levels of purchasing power (PPP), inter-country comparison is enabled, which is regarded as one of the main strengths of the indicator (Freistein, 2016).
Criticism focuses mostly on technical issues – the calculation of purchase power parity, the extrapolation from limited data as raised by Pogge and Reddy (2010) – while the act of measuring poverty itself is generally not disputed (Freistein, 2016). Nevertheless, current measurement and understandings of poverty offers limited insight in the persistence of poverty over time, according to Hickey and Bracking (2005: 851). They convincingly argue that the deeply political nature of poverty is not adequately addressed by using “narrow” measures. In the past, more qualitative dimensions of poverty have been taken into account through participatory poverty assessments such as the well-known ‘Voices of the Poor’ (Narayan, Patel, Schafft, Rademacher & Koch-Schulte, 2000). Although this World Bank-led project represents a growing recognition of the relevance of subjective experiences of the poor, such resource-intensive assessments of poverty have not been structurally
6
integrated into its primary poverty measurement methods, nor into the Development Goals. Instead, both the MDGs and the SDGs stick to ‘objective’ measures of poverty such as the previously described poverty headcount ratio, which can be more easily collected and compared. The SDGs make this measurement more country-specific by focussing on national poverty lines. Whereas the MDGs also included a separate indicator to assess the depth of poverty (the poverty gap), this indicator has not been included as a SDG indicator. Instead, the SDGs focus more on data disaggregation of the population living below the poverty line, by for instance sex, age and geographical location (appx. 1, p. 91).
All these poverty reduction figures are based on household consumption survey data. In order to measure the reduction of poverty, several data points are needed to track progress over time. As part of the earlier mentioned e-GDDS, the IMF recommends to update poverty statistics at least every 3 to 5 year interval (Serajuddin et al., 2015). Although the availability of such data has significantly expanded over the last 15 years, still substantial data gaps persist. In Sub-Saharan Africa, 70% of the countries lack adequate data on poverty; for as much as 20 countries it is even impossible to estimate poverty reduction (ibid.). Because household surveys are very resource intensive, the availability, periodicity and quality of poverty data tends to be weaker for poorer countries (Jerven, 2013). Even data that is available data is often plagued by validity and reliability problems. Paradoxically, the countries most of interest based on the phenomenon measured thus are the same countries of which such data is of limited availability and quality. Ward (2014: 101) explains that gaps in data are not value neutral; the can often be explained by political, conceptual or budgetary reasons. Despite – or instead because of – these gaps and validity problems, the use of this data can nevertheless have important effects.
2.5 Measuring the Measurement Impact
Although the SDGs mainly emphasize the importance of measurement in relation to monitoring and evidence-based policy, literature on the so-called ‘politics of numbers’ (Alonso & Starr, 1983) shows that the impact of measurement reaches much further. Indicators have become important lenses through which we see and understand society (ibid: 3). In line with this observation, Mügge (2015) argues that indicators can be viewed as powerful ideas that define what counts. Such definitions in the form of indicators often become widely accepted. This kind of institutionalization (for instance as part of the SDGs) makes the indicators more powerful, also in terms of their political impact. In this context, Merry (2011) makes a useful distinction between the knowledge and governance effects of indicator use, which appear to be closely related.
The knowledge effects of the use of indicators include the construction of what is often regarded as a discourse about truth (Camargo, 2009). Numbers seem to present an objective reality – for instance, that of poverty reduction. However, at the same time, indicators themselves contain (sometimes contested) assumptions and conventions about this reality, creating a paradoxical duality
7 as part of which measurement is expected to reflect reality as much as it is shaped by ideas about it (Camargo, 2009; Fioramonti, 2014). The construction of an indicator requires the selection and quantification of the single or multiple most important features of a phenomenon; a process which is influenced by social, economic, cultural and political factors and assumptions that are not readily recognizable in the indicator itself (Jerven, 2013: 11). In this context, the conceptualization of poverty as living on less than $1.25 per day became very powerful as an easy to comprehend, money-metric indicator (Freistein, 2016). The resulting technical conceptual categories (in this case: extreme poor and non-poor) thereby seemingly become universally applicable, irrespective of cultural or historical contexts; a process that Porter (cited by Broome and Quirk, 2015b: 828) calls reification. Hickey (2008) adds that the measurement of poverty in individual terms leads it to be viewed as an apolitical phenomenon, even though he argues that poverty and its reduction are deeply political. This argument relates to the work of Ferguson (1994), who shows how notions of (under)development become institutionalized through the work of the IMF and the World Bank. Based on the data used by these institutions, countries are categorized and labelled (e.g. based on income and poverty levels), thereby creating a powerful and seemingly a-political discourse about their development. The measurement of development and poverty thus not only reflects, but also influences how these phenomena are understood, with important governance implications. Ferguson (1994: 276) shows how the created development discourse provides the foundation for policy interventions, which tend to fail when they do not take into account local political realities that come to be obscured by this discourse. It illustrates how the knowledge and governance effects are closely related: through the institutionalization of increasingly indicator-based perceptions of development by international organizations, development becomes known and governable in ways that not necessarily align with local political-historical realities. The governance effects of indicators can be further broken down in parts in order to understand the scope of their impact in relation to poverty.
First of all, the use of indicators makes the unit that is measured – in this case the poor – legible by the one who is measuring (Espeland & Stevens, 2009). This is often the state, thereby rendering its (poor) population governable. On an international level, it also makes a country as a whole legible by IOs in ways that enable comparison with other countries (Freistein, 2013), for instance through the SDGs. It thereby enables (self)regulation of behaviour in the form of poverty reduction policy. This is an example of how poverty indicators have become important sources and instruments of power: the use of poverty indicators is linked to authority and the mandate to act to reduce poverty (Clegg, 2010). According to Freistein (2013: 373) as well as Ruggeri Laderchi, Saith, and Stewart (2003), this not only affects power relations, but also the possible responses to poverty reduction resulting from it. Preferably, the intended effects of any policy should be understandable and measurable in line with the definition of the object of this policy in the form of indicators – in this case, the poverty headcount ratio. The effects of the indicator used thereby affect the policies and
8
methods of implementation that impact the phenomenon measured, as expressed through those same statistics.
Concluding, statistics can have important effects on the phenomena they describe; which is reason to question if these numbers could be seen as ‘performative’ of development. In line with Sommerfeldt, Caine and Molzahn’s description of performativity (2014), this would imply that poverty could be seen as being ‘performed’ through its repetitive iteration through numbers. Although this can indeed impact how people and institutions behave, Didier (2008) convincingly argues that this does not mean that indicators construct phenomena completely from scratch, as the term performativity might suggest. Nevertheless, indicators both describe and potentially have an (indirect) effect on what they describe in doing so. For these reasons, Didier (2008: 307) has named this phenomenon ‘expression’, which indicates the process whereby statistics express (relational) arrangements of elements as characteristic of a certain object. This term is particularly useful to emphasize the (indirect) effects of statistics on the objects they describe, whilst also acknowledging these objects themselves are not created out of thin air through their statistical expression. Still, Didier argues that statistics do have the power to transform the object described by it, ultimately resulting in a circular effect.
In line with this reasoning, poverty can be seen as being expressed through statistics. Based on the previously described theories, it can be argued that the expression of phenomena such as poverty through statistics may have important knowledge and governance effects, which might ultimately impact that wat is measured – poverty – itself. However, this process – and its knowledge and governance effects – are often neglected in existing literature and empirical research, despite its high relevance as part of a framework that seeks to transform development through goalsetting.
2.6 The Power of Indicators as Goals
The goal of eradicating extreme poverty can be seen as an example of what is described by Broome and Quirk (2015a: 815) as “global benchmarking”. This involves “the development of comparative metrics of performance”, as part of which complex phenomena are translated into (seemingly neutral) numerical values, of which poverty measurement is regarded as an important example (ibid.). Benchmarking is regarded as a specifically transnational practice. In the case of the SDGs, the quality of the development outcomes of individual nation states is assessed based on the framework that was agreed upon by the international community represented in the UN. In this context, the use of poverty indicators gives rise to knowledge and governance effects that follow specifically from its use in a goalsetting framework.
Through the use of poverty indicators, global expectations about development and progress become normalized (Freistein, 2016: 373), making different countries comparable based on a certain understanding of poverty as articulated through Development Goals. Although this has been regarded as one of the main strengths of the MDGs, it is also the foundation on which the MDGs have been
9 criticized for their effects. According to Vandemoortele (2011; 2009), the quantitative targets have furthered a one-dimensional understanding of development – a knowledge effect – which has promoted one-size-fits all approaches to development – the governance effect. According to Fioramonti (2014), this could misguide development policy (and the support for it) on the long term, contrary to the intended effect of Development Goals.
Still, its measurability also explains part of the success of the MDGs, since the coupling of goals with numbers contributed to the salience and urgency of poverty reduction (Freistein, 2016). However, Vandemoortele (2011: 1) argues that as a consequence, “statistics have been abused to fabricate evidence of success” of MDG progress. This leads to the conclusion that ultimately, the reinforcement of the supremacy of numbers through development goals (as described by Fioramonti, 2014) could be regarded as a main strength and weakness of Development Goals, presenting great unsolved challenges for the implementation of SDGs on a country level.
2.7 The Challenge of Expressing the Poverty Goal on a Country Level
In line with the call for the data revolution, the challenge most emphasized is that of data availability in developing countries (UN-IEAG, 2014). Many developing countries were already struggling to meet the statistical demands of the MDGs; according to Jerven (2013: 6), this resulted in perverse effects, with statistical capacity being diverted to monitoring targets at the cost of local priorities. In the light of the increased statistical demands of SDGs, concerns about increasing inequality in data availability between countries have been voiced repeatedly (IAEG-SDGs, 2014).
This issue has been empirically examined through the Post-2015 data Test – a cooperation between several institutes that have assessed data quality in various countries in light of the demands of the SDGs (Kindornay, Bha-acharya, & Higgins, 2016). Although Ghana has not been included in this study, their findings show some general observations that bear relevance to this study. Research shows that most of the examined countries (including Tanzania, Senegal, Sierra Leone and Bangladesh) do provide data on poverty-related indicators, but this data is often not collected frequently enough to assess poverty reduction in a timely and reliable way. Current possibilities for data disaggregation (by for instance sex, age, region) are still limited (ibid., 2016: 33). The results illustrate the central importance of the work of the National Statistical Office, the legal framework under which it operates, the coordination among all actors involved in data collection and use and the availability of financial and human resources for data production. Krätke and Byiers (2014) draw a similar conclusion; they argue that for data to make a real difference, political economy factors should be considered during its production and use, including country priorities, institutional settings and actor incentives. Most importantly, the authors stress the relevance of national priorities, and using data to support these in the context of the SDGs.
This shows that ultimately, the use of data in the context of the SDGs is by no means just a technical issue – the local implementation and adaption of indicators and targets is shaped by political
10
and economic priorities (Jerven, 2013: 84), as well as the incentives of various actors within the country-specific institutional environment (Krätke and Byiers, 2014). These factors not only shape the expression of poverty through indicators, but also its governance and knowledge effects, which ultimately impact if and how progress is made towards the goal of eradicating extreme poverty.
2.8 Conclusion
Even though little empirical research has been done on the impact of development indicators, the integration of insights from existing literature does provide a useful theoretical framework for approaching the study of this topic.
Although poverty is widely acknowledged to be a multidimensional concept, it is mostly measured based in terms of what a person is able to spend per day as part of both MDG1 and SDG1. Distinguishing between knowledge and governance effects allows for a more detailed understanding of the impact of these indicators. Importantly, the concept of expression (as described by Didier, 2008) show how these measurement effects can ultimately impact the phenomenon measured – poverty – itself. The SDGs appear relevant to this expression in providing a unique framework for goals and evaluation, which also bring about ambitious data demands. Meeting these demands proves challenging for developing countries such as Ghana, which experience limitations in terms of statistical capacity and data availability.
Nevertheless, existing research does not explore the measurement impact of development indicators in such contexts, despite its relevance to meeting the SDGs. Therefore, this research will address the gap in existing literature by taking into account the aforementioned factors in the approach of the study topic, as described in the following methodology chapter.
11
3. Research Methodology
3.1 Introduction
This chapter presents the methodology that has guided this research. It does so by first describing the ontological and epistemological considerations, which are relevant to the approach of the research topic. Subsequently, the main research question is divided in several sub-questions, which together have guided this research. The relation between these research questions and the most important concepts is summarized through a conceptual scheme, after which each concept is also operationalized separately. The methods section includes an elaborate description of the different methods used as part of the mixed-methods strategy. Both qualitative and quantitative data collection and analysis are discussed, as well as the integration of these different sources of data, for instance through process tracing. Lastly, ethical considerations and limitations of this study will be discussed.
3.2 Ontological and Epistemological Considerations
As appears from the main research question, this research examines the knowledge and governance impact of the use of poverty statistics through the concept of expression. Instead of approaching statistics as a means for doing research, its role in thinking about poverty and governing itself is the object (or unit of analysis) of this research. The foundation of such an analysis about the meaning and impact of poverty indicators is provided by a poststructuralist epistemology.
Whereas poststructuralism mostly focuses on critically analysing the meanings conveyed by language (Wylie, 2006), this can also be extended to include statistics as a form of communication. Because similarly to how words can be viewed as acquiring meaning through differential relations (such as poor/non-poor), statistics have become particularly powerful tools in demarcating such difference when it comes to poverty, especially in the context of the SDGs. This “argumentative strength of statistics” (Camargo 2009: 6) is the reason that statistics can be regarded as a “discourse about the truth” (ibid.), which can be critically analysed from a poststructuralist perspective.
In addition, poststructuralism offers a critical perspective that allows for the deconstruction of how poverty is expressed through statistics. Such a deconstructive approach to statistics critically analyses the assumptions implicit in the use of statistics and its impact on how poverty becomes ‘known’, in line with the research (sub)questions. In this context, it is acknowledged that statistics not just describe phenomena such as poverty, but have the power to impact the phenomena described by it (Didier, 2008). This research thereby also takes a constructivist ontological approach (as described by Bryman, 2008: 19), regarding social phenomena - including both poverty and statistics - and their meaning as continuously socially (re)constructed. It is this (re)construction of poverty through statistics and its knowledge and governance effects that is central to this research, which therefore also inform the following sub-questions, research strategy and design.
12
3.3 Research Questions
Although the relevance of the impact of development indicators is increasingly recognized, little empirical research has been done within this field. Research tends to focus on indicators as global norms on a conceptual level, rather than their impact in expressing development in local contexts such as that of Ghana. More specifically, little research has been done on the distinct knowledge and governance effects of development indicators in the context of goalsetting as part of Development Goals. Moreover, the interrelation of the country-specific ways of expressing development through statistics with its knowledge and governance effects remains unexplored, despite its high relevance to the implementation of the SDGs and the progress on the goal of eradicating poverty more specifically. Especially in the light of its poverty reduction efforts and its upcoming participation in the SDGs’ Voluntary National Review, Ghana provides an interesting case study to explore these issues in-depth. For these reasons, this research has been guided by the following main research question:
How does the expression of poverty through consumption-based indicators relate to how poverty is understood and governed in Ghana in the context of Development Goals, and to its relation with progress on the goal to eradicate (extreme) poverty more specifically?
This question features the concept of expression as described by Didier (2008) to focus on the process through which statistics not only describe, but also characterize and potentially affect what is measured (poverty) through intermediary factors. The question is country-specific, in the sense that this research will take into account relevant political-economic and institutional factors specific to the case of Ghana and its National Statistical System more specifically. This country-specific expression of poverty is likely to be affected by challenges in statistical capacity and quality, of which the effects have been taken into account. The research question specifically focuses on how this all relates to the knowledge and governance effects of the use of poverty indicators, and thereby ultimately to poverty reduction itself in the context of the SDGs.
Importantly, the contribution of this specific research thereby lies in its focus on the impact of poverty measurement, rather than on the technicalities of measurement itself. Although those two issues are definitely related, the technical underpinnings of poverty measurement have been explored within existing literature quite elaborately (e.g. Pogge & Reddy, 2010). Instead, this research focuses on the impact of such measurement (in Ghana), which has not been covered by existing literature. It thereby aims to make a unique contribution to the broader debate on measuring development, in particular in the light of its increasing importance in the context of the SDGs.
This is done based on a series of sub-questions (SQs) that build upon each other to provide an answer to the main research question:
1) How can the current international discourse on measuring development and poverty as part of the SDGs be characterized, in which context the expression of poverty in Ghana takes place?
13 2) How has the direct institutional context in which poverty statistics are produced (the National
Statistical System) in Ghana developed since the late 1990s?
3) How have extreme poverty itself and its measurement evolved in Ghana since the late 1990s, based on the various indicators that are used to assess poverty by the national government and international organizations?
4) What are the knowledge effects of the use of poverty data based on quantitative indicators? a. how do different stakeholders perceive and define extreme poverty (reduction) in the
light of its measurement as part of Development Goals?
5) What are the governance effects of the use of poverty data based on quantitative indicators? a) (how) do different stakeholders perceive the relevance and make use of poverty data
during different phases of decision and policymaking processes?
6) How do different stakeholders perceive specific (statistical) challenges and opportunities in measuring poverty in the context of Development Goals, and how are these related to the observed knowledge and governance effects?
7) How are the specific knowledge and governance effects of the use of poverty measurements observed related to each other as well as to the specific country context of Ghana in which these arise?
8) How do these observed effects relate to national poverty reduction policies and results in Ghana, in particular in the context of the Development Goal to eradicate extreme poverty? 9) Based on these observations, how can the forms and uses of poverty measurement be made
more inclusive in the light of the Sustainable Development Goals?
In answering these research questions, specific attention is given to the specific country context of Ghana, for instance with regards to specific institutional regulations (such as national legislation), the organization of the National Statistical System, relevant political developments, and other country-specific factors relevant to the research topic. Within this context, the relevant stakeholders referred to in the questions (and identified by the IEAG-SDGs on the Data Revolution, 2014) include a selection of national statisticians, staff members working for IOs, government officials, relevant donors and staff members of civil society organizations and NGOs that make use of poverty statistics. This provides further insight in the differences and commonalities in terms of perceptions, experiences and incentives for the use of poverty statistics for different actors on multiple levels (global/national/local), which are also taken into account in the subsequent sub-questions. Together, the different research questions build upon each other sequentially, ultimately leading up to an answer to the main research question.
14
3.4 Conceptual scheme
The figure below provides a schematic overview of the main concepts and their interrelation. The numbers used refer to the various sub-questions outlined in the previous section, and illustrate how answering the different sub-questions ultimately results in a comprehensive understanding of the research topic.
15
3.5 Operationalization
The table below provides an overview of the operationalization of the main concepts relevant to the research questions, based on the theoretical framework.
Concept Dimensions Variables Indicators
Extreme poverty Economic/income-related, expressed
as…
Universal extreme income-poverty [in line with SDGs],
expressed as…
Proportion of population below the international poverty line, by sex, age, employment status and geographical location Country-specific (extreme)
income-poverty [in line with SDGs], expressed as…
Proportion of population living below the national poverty line(s), by sex and age Proportion of men, women and children of all ages living in poverty in all its dimensions according to national definitions
Depth of income-poverty [in line with MDGs], expressed
as…
Poverty gap ratio, based on both universal and national poverty lines
Non-income related This research will primarily focus on extreme poverty defined through a lack of income, which is of course closely related to non-economic forms of poverty. Although these will be taken into account in relation to income poverty and how stakeholders perceive poverty more in general, these will not be the primary focus of this research in itself.
Expression of extreme poverty Process of measurement Relational arrangement of elements, characterization of replicable objects
Poor/non-poor categories and their basis, both technical basis as well as underlying rationale Production; adding similar
items together
Counting, (dis)aggregated poverty data Reporting Bringing out characteristics of
the object described through indicators
Presentation of poverty data in numbers, graphs, graphics, maps
Interpretation/description added to the presentation of poverty data
Extent to which this is explicitly linked to Development Goals
Transformative potential
Through knowledge and governance effects, separately operationalized below
Knowledge effects
Understanding Defining poverty Extent to which the definition of poverty is similar/shaped by or limited to what is measured
Extent to which poverty measurements are seen as reality, representations of it or constructed based on assumptions Reification Extent to which poverty is understood as universal vs. contextualized
Governance effects
Power/Authority Legibility The possibility of organizations or actors to use or request data as a basis for governance Mandate to act The expected and experienced role of
organizations or actors in addressing poverty (in relation to the use of data)
Benchmarking National benchmarking National assessment of poverty levels and progress over time
Global benchmarking International, comparative assessment of poverty levels and progress over time Policy Use of poverty indicators & effects in different phases:
Agenda setting & issue recognition Policymaking & design
Execution & targeting Monitoring and evaluation
16
Table 1: Operationalization main concepts
In the light of the main research question, the concept of extreme poverty has been operationalized in line with SDG goal 1 on this topic (see appendix 1, p. 91). Although the SDG indicators do not specifically focus on the depth of poverty, the poverty gap has been included as an additional relevant poverty indicator, as it used to be included to measure poverty as part of the MDGs.
In addition, the concepts of knowledge and governance effects have been operationalized into dimensions, variables and indicators based on the theoretical framework. Lastly, the relevant characteristics and dimensions of the National Statistical System have been specified, to take into account the specific country-context, without limiting this to one specific (disciplinary) approach.
3.6 Methods
3.6.1 Unit(s) of analysis
The unit of analysis of this research directly follows from the main research question. Therefore, the main unit of analysis of this study is the country-specific expression of extreme poverty in Ghana. However, as appears from the foregoing conceptual scheme and operationalization, this is a very broad concept with multiple dimensions. As a consequence, it is a concept that cannot be measured directly or easily. Therefore, in order to study this unit of analysis, the research strategy entails multiple methods focusing on a specific aspect of this broad concept, as described in the following method sections. Together, the combination of methods has provided a comprehensive understanding of the main unit of analysis. Appendix 2 (p. 92) provides a schematic overview of the unit(s) of analysis in relation to each sub-question and the different research methods.
1 datatopics.worldbank.org/statisticalcapacity/SCIdashboard.aspx National Statistical System; context in which expression occurs
Organizational Responsibilities of different actors
Position, work and experiences of National Statistical Office
Involvement of IOs in statistical capacity building
Cooperation between different stakeholders Statistical capacity/quality IMF Enhanced General Data Dissemination
System; results/experiences
World Bank Statistical Capacity Indicator1
Experienced capacity to meet SDG/poverty measurement demands
Legal Legal framework Rules and regulations relevant to statistics, their impact and how these are experienced Political (in)dependence from political
priorities
Experienced political involvement in statistical priorities
Social Independence of statisticians Experienced independence in statistical work Financial Stability of government
funding
Financial flows related to Ghana’s statistical service
Dependence on foreign donors
17
3.6.2 Mixed Methods Research Strategy, Case Study Design & Process Tracing
In order to answer the research questions described, a mixed methods research strategy has been used for this case study. In doing so, this research has drawn inspiration from what Merry (2011:85) calls “an ethnography of indicators”, which involves
examining the history of the creation of an indicator and its underlying theory, observing expert group meetings and international discussions where the terms of the indicator are debated and defined, interviewing expert statisticians and other experts about the meaning and the process of producing indicators, observing data collection processes, and examining the ways indicators affect decision-making and public perceptions. (2011: 85)
Merry (2011: 85) furthermore adds that critical dimensions include analysing the sources of information used and “the forms of cooperation and resistance by countries and NGOs in the contest over who counts and what information counts”, thereby providing useful methodological guidelines for deconstructing the impact of indicators. Importantly, the emphasis of this research is not on the indicators as such, but on their use and the effects of this use in Ghana in the context of the SDGs. Therefore, the research strategy and design have been designed to support this research focus, and align with the different research questions more specifically. In order to do so, this research has mixed quantitative and qualitative research methods in ways that are informed by the approach of process tracing.
Process tracing involves “the analysis of evidence on processes, sequences and conjunctures of events within a case” (Bennett & Checkel, 2014: 7). It can be used to generate hypotheses and theories about causal mechanisms, based on a particular case. Because of this study’s emphasis on knowledge and governance effects of poverty measurement, process tracing presents a useful approach to studying and situating these effects in the specific context of Ghana. In doing so, process tracing allows for the distinction between ideational and materialist explanations in line with Jacobs (2014). Idealist explanations focus on the impact of the content of cognitive structures on actors choices (ibid.), thereby allowing for the study of the knowledge and governance effects relevant to this research. On the other hand, considering material explanations (which focus on the impact of variation in material parameters) is helpful to take into account the variation of poverty (as expressed through numbers) itself as well. In addition, it is important to study the impact of critical events in relation to these processes, such as the establishment of the SDGs. Lastly, Jacobs (2014: 42) stresses the importance of considering the ways strategic incentives for actors are generated within the specific context of the case. Altogether, these methodological guidelines provide a useful framework for the study of the expression of poverty in Ghana through a combination of methods.
Both qualitative and quantitative methods are used to answer the research questions.
Appendix 2 (p. 92) provides a comprehensive overview of what methods are used to answer each research question. Most of the sub-questions require qualitative research methods to understand and deconstruct the meaning and impact of poverty indicators, as well as the specific local context in
18
which these arise. For this purpose, the main methods to be used are participatory observation, document analysis, descriptive quantitative analysis and semi-structured interviewing. Participatory observation primarily took place at the United Nations World Data Forum, whereas qualitative interviews were held during fieldwork in Ghana. Before fieldwork, document analysis and quantitative analysis have been used to answer the sub-question (SQ3) on the development of poverty in Ghana over time. This way, the analysis of secondary data has provided more insight in possible material explanations as part of the process tracing approach. In addition, the findings have informed the qualitative research instruments used during semi-structured interviewing, which focus more on the tracing of ideational processes. Several sub-questions required the combination of these quantitative and qualitative methods to explore the impact of the use of poverty indicators through integrative analysis. This research has therefore taken a pragmatic approach to mixing methods in order to answer the research questions, as advocated by e.g. Johnson and Onwuegbuzie (2004). This combination of multiple qualitative methods allows for triangulation, thereby adding to the reliability and validity of the findings (Bryman, 2008: 608).
Interestingly, this means this study on the impact of the use of quantitative indicators has used those quantitative indicators itself as a means of doing so. Although this might seem paradoxical, it has contributed to a more comprehensive understanding of the research topic and has allowed for triangulation. Still, it is important to note that the use of quantitative data as part of this research never assumes that this data objectively represents (or in fact, is) reality, as is sometimes the case within (more positivist) paradigms of quantitative research (Morgan, 2007). Instead, the quantitative research methods data have been used to explore and assess how the use of quantitative data describes, constructs and possibly impacts realities of poverty. Ultimately, the integration of the results of the quantitative and qualitative analysis makes it possible to answer the main research question.
The figure below provides an overview of the research design, of which the various components are described more extensively in the following sections.
