Rolando Avendano
Carolyn Culey
Charlotte Balitrand
Development Co-operation Report 2018
Joining Forces to Leave No One Behind
Development Co-operation Report
Chapter 5. Data and diagnostics to leave no one behind
Abstract
Ensuring progress on the leave no one behind commitment set out in the 2030 Agenda for Sustainable Development requires a new approach, one that is based on counting people, and takes into account the factors that contribute to their exclusion. This chapter examines the data challenges posed by the leave no one behind agenda and considers the potential of new and existing data sources to meet them. It makes the case for increased data disaggregation and assesses the tools available for country diagnostics and improving the targeting of resources to those at risk of being left behind. It highlights data success stories, considers the potential for scaling up new data initiatives, and urges greater investment in data and national statistical systems as prerequisites for meeting and monitoring the pledge to leave no one behind.
Rolando Avendano and Charlotte Balitrand from PARIS21 and Carolyn Culey from Development Initiatives.
Counting the invisible: The case for disaggregated data
The 2030 Agenda’s leave no one behind commitment presents a series of data challenges, starting with the need to ensure that everyone is counted. At present, many countries lack basic data such as comprehensive civil registration and vital statistics systems. Those living in poverty are most likely to be excluded; the poorest 20% of the global population currently account for 55% of unregistered births (Development Initiatives, 2017[1]). Governments cannot ensure that everyone is included in progress if they do not know they exist in the first place. Better data are fundamental for targeting resources at ending poverty and delivering better public services and goods for all.
Counting everyone as individuals means that national statistical systems will also need to pursue more sophisticated data disaggregation strategies. Existing statistics typically capture national averages, but they often mask disparities at the subnational, community and household levels. Implementing the Sustainable Development Goals (SDGs) requires a radically different approach. To this end, the data community has been charged with ensuring that data are disaggregated by income quintile, gender, geography, age and disability. For example, disaggregation of income is needed because, while poverty is multidimensional, income remains a key predictor of well-being and is the most widely used indicator of poverty. Data on daily income shows the 1.4 billion people who make up the poorest 20% earn on average USD 1.75 day, well below the USD 24.9 for the rest of the world. Globally and within most countries, the gap between the poorest 20% and the rest of the population is growing. Still, today’s measurement of progress is largely reliant on survey data, which suffer from several shortcomings in relation to the leave no one behind agenda (Box 5.1).
Box 5.1. Shortcomings of survey data in measuring leave no one behind
Many surveys do not record information on marginalised and excluded groups.
Countries without data are often those most affected by conflict and insecurity, where people are at risk of being left behind.
Data are often based on prevalence estimates and rely on national averages.
Poverty measurement is affected by multiple factors, including survey implementation and sampling frames.
Survey data are mostly collected at the household level, and risk ignoring intra-household disparities.
Having disaggregated data makes a difference in the way public policy is delivered. A lack of disaggregated data has prevented countries from putting leave no one behind into practice. In Nigeria, for example, it hampered the allocation of resources to prevent the surge in HIV infections. As a result, the government is now working on a survey regrouping 36 states to fill this gap.1 When available, disaggregated data can allow for more effective anti-poverty and inclusion policies, as Box 5.2 illustrates. The ability to create and implement effective policies strongly depends on the quality of these data.
Box 5.2. Examples of where disaggregated data have made a difference
Bolsa Familia, a conditional cash transfer programme in Brazil, has had a proven impact in reducing extreme poverty among 24 million families. An important element of its success has been the quality of disaggregated data on beneficiaries. Using per capita family income as declared in the Cadastro Único (Single Registry), the programme targets families living in poverty at the municipality level.1 Specific groups are also prioritised: quilombola families, indigenous families, families living off recycling and families in which there is child labour (Gazola-Hellmann, 2015[2])
Data availability can also improve public service delivery. In Lanet Umoja, Kenya, a community-generated data project revealed the existence of 4 500 more households than previous records indicated. The collected data helped 12 000 residents gain access to clean water for the first time (Developement Initiatives and Open Institute, 2016[3])
1. An estimate on the number of families living in poverty is established at the municipality level, based on data from the Demographic Census and the National Household Survey.
Delivering effective policies on leave no one behind requires greater investment in data. To reduce existing data gaps and improve the production of disaggregated data, national statistical systems need further investment and increased capacity. Today, only 56 of the 102 countries with statistical plans have secured adequate financing to implement those plans (UNDESA, 2018[4]). Scaling up investment in national data systems is an essential first step in directing policies and resources to those who need them most.
Strategies and tools for disaggregating data
Undertaking the leave no one behind agenda raises some important challenges in data disaggregation. At a minimum, data have to be disaggregated by income quintile, sex and gender, geography, age, and disability. Many of the standardised measurement tools currently in use must be rethought to capture unseen disparities and fundamental aspects of identity, such as intra-household inequalities in asset ownership by gender (Asian Development Bank, 2018[5]). And when gathering information on small groups, survey data may need to be collected from a disproportionally large sample to be trustworthy. Improving coverage and including a disaggregation perspective into current methods is therefore important.
Challenges for disaggregation: Income, sex and gender, geography, age, and disability
Disaggregating data according to the dimensions of income, sex and gender, geography, age, and disability is challenging and there is no one size fits all survey tool or data set as outlined in Table 5.1.
Table 5.1. Disaggregated data are missing from major global datasets
1. A World Bank interactive computational online tool to monitor global poverty. It provides harmonised data on poverty from different surveys.
2. A data collection programme providing policy makers with demographic and health information.
3. A UNICEF survey generating data on equity to track progress towards elimination of disparities and inequities.
Source: (Development Initiatives, 2017[1]) P20 Initiative: Baseline report, http://devinit.org/wp-content/uploads/2017/03/P20-Initiative-baseline-report.pdf
Geographic disaggregation is an essential element for understanding how policies are implemented and their distributional impact. However, there is no statistical harmonisation today for the definition of urban and rural concepts (UNDESA, 2017[6]). Still, as in the case of Colombia (Figure 5.1), geographic disaggregation highlights a far more complex picture when considering national and subnational (i.e. clustered household level) poverty rates, with poor households also present in the wealthiest areas.
Figure 5.1. Geographic distribution of the poorest 20% of the population in Colombia
Source: (Development Initiatives, 2017[1]), P20 Initiative: Baseline report, http://devinit.org/wp-content/uploads/2017/03/P20-Initiative-baseline-report.pdf.
The SDG agenda encourages the collection of gender-disaggregated data to monitor policies on inequality and access to decent work, as well as a diversity of sectors ranging from sanitation to financing, education and health (UN Women, 2016[7]). Exclusion and marginalisation go beyond gender, but most major surveys do not provide much information on gender identity or sexual orientation.
For age disaggregation, while data can be obtained from major surveys (e.g. demographic and health surveys, multiple indicator cluster surveys, living standards measurement surveys) and the data are usually publicly available, existing survey methodologies tend to focus on the 15-49 year-old age group. Data on age disaggregation, however, show that the very old and the very young are disproportionately represented among the poorest 20%.
Disability is referenced in five of the SDGs, seven of their targets and numerous indicators. A major challenge faced in collecting disability data is the way in which a person’s disability status is determined (Chapter 3 Chapter 11). More emphasis is being put today on identifying the impact of disability on people’s lives rather than their status. Sustained advocacy for more and better disability data has started to pay off; the UN Statistical Commission’s city group (The Washington Group on disability statistics) has developed a standard module to be applied in different contexts.
Understanding how these different dimensions of disaggregation can combine and reinforce each other is essential in designing policies to promote social and economic inclusion. Clear standards and classification criteria need to be developed for disaggregation dimensions such as age and geographic location following, for example, the standard developed for collecting data on disability.
Smart and alternative sources for disaggregating data
The surge in new data sources from digital technology is stimulating new thinking and methods to fill gaps in disaggregated data. Gradually, national statistical systems are adapting to the new technological environment, exploring the smart use of existing and new sources such as administrative data, citizen-generated data and geospatial data.
Administrative data hold potential for increased disaggregation at the individual level. A more systematic use of administrative registers could complement traditional sources (censuses, demographic and household surveys, etc.) with more disaggregated data. Other administrative sources, such as firm-level data, provide valuable, sometimes untapped disaggregated data. As a first step, a mapping of existing administrative registers could inform the future use of these data sources. Understanding legal barriers is also crucial, and advocacy is needed to mainstream the use of administrative records (Ploug, 2016[8]). Countries like Ecuador and Viet Nam see the opportunity of such registers to improve statistical production (Ploug, 2016[8]). Survey modules, such as DFID’s Guide to disaggregating programme data by disability, are also being integrated into administrative systems (DFID, 2015[9]).
When it comes to citizen-generated data, new tools have the potential to fill data gaps and improve micro-data on general living conditions. Pilots from DataShift (CIVICUS, 2017[10]) to the Humanitarian OpenStreetMap Team (GPSDD et al, 2016[11]) provide disaggregated data for the SDGs on gender, social justice inclusion, access to public services and the environment. Although some concerns exist on the coverage, comparability, capacity and sustainability of citizen generated data, its relevance and contribution to SDG is increasingly recognised (Chapter 6).
Geospatial data produced using the Global Positioning System (GPS), satellite imagery, remote sensing and cartography are used to improve population coverage and provide accurate geographic boundaries for field surveys and censuses. Organisations such as Open Data Cube are providing free open sources of geospatial data to facilitate the use and analysis of satellite imagery. Integrating geospatial data with census or survey data is needed to produce spatially disaggregated population estimates which can then be aggregated for national or administrative purposes. For this, quality control and supervision across different levels of national statistical offices (NSOs) are important.
Using disaggregated data for better policies and planning
While having and using disaggregated data is crucial for policies and investments to deliver on the pledge to leave no one behind,2 a recent assessment of demand for these data from national decision makers found that the demand was weak due to lack of specification of population groups (Serajuddin et al, from the World Bank, 2015[12]). Policy makers can make the most of the data revolution through improved data planning and by making links between use of data for policy design and its use for better targeting of policies.
Improving disaggregated data planning and production
Innovative tools that link available disaggregated data to policy formulation are needed. The goal of producing more granular data requires that national policy frameworks, and in particular national development plans, incorporate, systematically, the need to produce disaggregated data in order to deliver and evaluate the effectiveness of policies and programmes. National statistical and data strategies and national development plans are often designed separately despite the interlinkages and synergies between them. PARIS21’s (2018[13]) Advanced Data Planning Tool (ADAPT) could help reinforce these synergies (Figure 5.2).
Figure 5.2. The advanced data planning tool: Linking data and policies
Source: (PARIS21, 2018[13]), “Advanced Data Planning Tool (ADAPT)”, http://www.paris21.org/node/2905.
Compatible with methodologies such as the Generic Statistical Business Process Model and the Generic Activity Model for Statistical Organizations, the tool assesses gaps between data supply (e.g. data inventories) and data demand (e.g. demanded indicators) at different levels, including sectors and subnational policies. This and similar tools can be instrumental in identifying untapped data sources and incorporating disaggregated data into monitoring and evaluation frameworks.
From identifying needs to designing better policies
Identifying those who are excluded by their income or condition – and improving the targeting of policies to support them – are intertwined (Figure 5.3). Pre-existing data sources (e.g. household surveys or census information) or on-demand surveys, which could exclude some people, can inform policy making to leave no one behind. Indeed proxy means tests have been used extensively for policy interventions that aim to benefit the poorest people.3 At the same time, the proxy means test poses some challenges from a statistical perspective: they demand capacity and co-ordination; priority populations groups may be excluded from the sample; there is a risk of arbitrary selection of beneficiaries based on a limited group of variables; and a risk of overlooking progress or decline in household conditions over time.4 Still, while imperfect the method is considered to be effective when compared to other tools for identifying groups in need.
Figure 5.3. Examples of data tools for leave no one behind
Notes:
1. Acronyms: Advanced data planning tool (ADAPT); Microdata Cataloguing Tool (NADA) – International Household Survey Network (IHSN); Every Policy Is Connected to People, Planet and Prosperity (EPIC) – United Nations Economic and Social Commission for Asia and the Pacific (UNESCAP); Humanitarian OpenStreetMap Team (HOT); Swedish International Development Cooperation Agency (SIDA); Innovations for Poverty Action (IPA); United States Agency for International Development (USAID); European Union (EU).
Source: Authors
Better tools for identifying needs can bolster policy design, whereas better tools to identify target groups can improve policy delivery. Investing in both is necessary. Better targeting could also improve the monitoring of policies, align development agendas, and encourage accountability. A good identification system that informs existing targeting tools allows planners to better understand the trade-offs associated with the commitment to reach the furthest behind first, to determine those who are more excluded and more vulnerable, and to decide what policy support can be provided.
Integrating leave no one behind into national development plans
To be effective, monitoring of progress of the leave no one behind agenda should be compatible with national monitoring tools, in particular national development plans. While there is a consensus within the international and donor community on its importance, achieving full disaggregation across all social groups will be a complex and a long-term endeavour. However, an estimate of the degree to which disaggregated data are referred to in national development plans or equivalent national policy documents (Figure 5.4) shows that these data are only minimally incorporated in national policy frameworks. Systematically including a disaggregated perspective into national planning is clearly in the starting block.
Figure 5.4. Distribution of PARIS21 data disaggregation index scores among National Development Plans
Notes: Based on the PARIS21 Use of Statistics in Policy Making Index. The figure represents the distribution of PARIS21’s disaggregation score. This score provides an estimate of the use of disaggregation-related terms and concepts in 199 national development plans and poverty reduction strategy papers between 2001 and 2017.
Source: (Avendano et al, forthcoming[14]) “Proposing a use of statistics indicator in national development plans”.
The leave no one behind agenda goes hand-in-hand with assessing the distributional impact of policies. As more and better disaggregated data inform the implementation of national plans, they also inform of the impact of policies in different dimensions. In recent years, development co-operation providers have made efforts in exploring the effects of poverty reduction and inclusion policies by designing better quality country assessments. With the underlying idea that the poorest and most vulnerable populations should benefit most, country diagnostics help set policy and programme priorities and examine their distribution between different sectors of the population. From job quality among women to educational attainment of indigenous groups, disaggregated data can provide a subtler description of the sometimes heterogeneous effects of policy.
Box 5.3. Systematic Country Diagnostics: Identifying data gaps for a multidimensional approach to policy
Implemented in more than 90 countries, the World Bank’s Systematic Country Diagnostics (SCD) is a policy diagnostics tool to identify key constraints for a country to achieve its development objectives. The SCD puts strong emphasis on assessing data quality and identifying data gaps, which can be critical in providing a multidimensional assessment and formulating policy.
Good quality disaggregated data have been essential for the assessments and recommendations in the SCDs. In Bangladesh, the SCD focused on improving job quality and availability. Sex disaggregation data allowed it to examine how young women are rapidly shifting towards the manufacturing sector and young girls gaining access to education.
In Panama, information on ethnic disaggregation revealed that the country’s indigenous groups are reported to have the lowest levels of electricity coverage out of 12 Latin American countries. However, important data gaps on income disaggregation and educational attainment hindered a better assessment of these communities. In Uruguay, data disaggregation by age and occupational status allowed for the identification of unemployed youth and estimates of spending by the social protection system on this vulnerable population.
While disaggregation is not the only objective, the SCD’s data diagnostic tool provides support to the leave no one behind agenda by improving assessments of data gaps and linking them to the World Bank’s areas of intervention.
Sources: (Washington, 2015[15]); (The World Bank, 2015[16]) and (The World Bank, 2015[17])
Disaggregation: The challenge of working out where to start
Countries seeking to implement the leave no one behind agenda face problems in knowing where to start. The P20 initiative, designed by Development Initiatives, proposes that decision makers at all levels should focus their attention on the poorest 20% of people and investing in better data about these people. This poorest quintile includes everyone currently in, or vulnerable to, extreme poverty as well as those who by reasons of their identity (age, disability, belief, ethnicity, sexual orientation) are most vulnerable to poverty or exclusion. The P20 initiative uses three bellwethers drawn from the SDG framework and based on income, nutrition and civil registration. The state of the global P20 (Figure 5.5) confirms that the poorest 20% of the world’s population receive 1% of global income, account for 46% of new cases of stunting and have 55% of all unregistered births. The same approach can be applied at the national or subnational level (Chapter 8).
Figure 5.5. The state of the world’s poorest 20% - the P20
Source: (Development Initiatives, 2017[1]) P20 Initiative: Baseline report, http://devinit.org/wp-content/uploads/2017/03/P20-Initiative-baseline-report.pdf.
Scaling up data initiatives
There is still much work to be done in convincing governments of the importance of investing in more disaggregated data. And in the absence of solid evidence of how governments and development partners are addressing leave no one behind there is a risk that the pledge will become little more than a slogan or a hashtag. For example, from the 42 countries who submitted voluntary national reports in 2017, only 14 provided an indication of the availability of data to leave no one behind, and the majority (11) noted that additional disaggregated data is needed (CCIC, 2018[18]). From a data perspective, mainstreaming the leave no one behind agenda requires resources, partnerships, building capacity and securing political commitment.
Building national capabilities for data disaggregation
The leave no one behind agenda calls for rethinking the spectrum of skills and capacities needed to fully harness the benefits of the new data ecosystem (OECD, 2017[19]). Information technologies, the emergence of new data providers and users, and the increasing complexity of the data ecosystem point towards a new way in which national statistical systems can support evidence-based policy making. As a first step, upgrading technical and organisational capacities within national statistical systems to exploit existent and new data sources for data disaggregation will be mandatory. In this direction, improving co-ordination between NSOs with other agencies and the private sector is critical: more than 40% of NSOs report being interested in establishing partnerships to access geospatial data, but are lacking sufficient knowledge to pursue them (PARIS21, 2018[20]).
Improving NSO capacities for supervision and quality control, as in the case of geographic disaggregated data, will be important. National statistical systems will also be challenged to explore new and innovative financing mechanisms to complete the disaggregation agenda, engaging with a new range of actors. In the long term, other skills, including legal expertise and communication, will be essential to mainstream the outcomes of improving data disaggregation maps.
Principles for diverse and new data partnerships
As data ecosystems of producers and users become more complex and involve diverse players and data sources, the vision and principles of the Inclusive Data Charter and guide partnerships for more granular data about people. While there are technical and methodological challenges to improving data disaggregation, some of the largest barriers are political. Founding members of the Charter have drawn up detailed action plans for implementing it (Box 5.4).
Box 5.4. The Inclusive Data Charter principles
Principle One: All populations must be included in the data.
Principle Two: All data should, wherever possible, be disaggregated in order to accurately describe all populations.
Principle Three: Data should be drawn from all available sources.
Principle Four: Those responsible for the collection of data and production of statistics must be accountable.
Principle Five: Human and technical capacity to collect, analyse and use disaggregated data must be improved, including through adequate and sustainable financing.
Source: (GPSDD, 2018[21]), “Inclusive Data Charter”, http://www.data4sdgs.org/sites/default/files/2018-08/IDC_onepager_Final.pdf.
Assessing and managing trade-offs related to costs, coverage, privacy
Recent studies have focused on the estimated cost of producing data for the SDGs, rather than considering the specific costs of having data to leave no one behind. The annual funding gap for monitoring the SDGs is estimated to be USD 200 million per year (GPSDD et al, 2016[11]) and (OECD, 2017[19]). In 2016, aid commitments for statistics reached USD 623 million, an increase of 6% compared to the last two-year average (PARIS21, 2018[22]). However, current estimates are largely based on expanding existing methodologies to measure SDG progress, with many indicators relying on prevalence data. Leaving no one behind requires a new approach to costing, with an emphasis on counting people, and on collecting data that are sufficiently disaggregated to identify groups and individuals who are marginalised. Currently, no cost estimates are available, and it is likely the costs for disaggregation will be higher than current estimates.
In particular, making data anonymous and guaranteeing coverage when producing disaggregated data is central, and can be costly. While digital technologies can be an effective way to collect certain forms of disaggregated data, the need to anonymise those data prior to publication may incur significant costs (Johnson et al, 2017[23]). In the case of geospatial data, the provision of maps from satellite imagery relies on having spatial data infrastructure, which remains too costly for many developing countries.
There are also potential risks and trade-off between increasing data disaggregation and ensuring data confidentiality at a time when big data, artificial intelligence and algorithms are increasingly used for policies and marketing by private sector (Rieland, 2018[24]). While disaggregation by ethnicity, personal situation and location may be an essential part of a national statistical office’s remit, legislative and regulatory frameworks as well as systems for storing and anonymising date need to protect citizens rights to privacy and guarantee that disaggregated data are not used for discriminatory purposes.
Conclusion
In their efforts to leave no one behind and deliver the 2030 Agenda, the data and statistics community will need to address two key challenges with urgency: counting everyone and producing data that go beyond national averages and reflect peoples living conditions so that development policies can tackle the multidimensional causes of poverty and vulnerability.
Going forward the statistical community, governments and development partners should focus on:
Expanding the use of untapped data sources (e.g. administrative data) and involve new actors (e.g. citizen-generated data, geospatial data providers) to provide a more accurate picture of who is excluded.
Integrating leave no one behind principles into the tools used to design and implement national development policies.
Strengthening national statistical systems through a new and holistic approach to capacity development (PARIS21, 2018[25]). National statistical systems require comprehensive sets of individual skills and organisational practices, recognising leadership, management and communication skills as effective vehicles for strengthening data systems and responding to their challenges.
Tailoring national SDG results framework to the pledge to leave no one behind with providers of development co-operation aligning to these results frameworks.
Stepping up global co-ordination, investment and understanding of what it will take to meet growing demand for disaggregated data.5 The planned data disaggregation guidelines to be shared at the 50th session of the United Nations Statistical Commission in March 2019 should provide useful reference and basis for mobilising the necessary financing to have and use the right data about individuals and there needs and to monitor progress for the furthest behind.
References
ADB (2018), Measuring asset ownership and entrepreneurship from a gender perspective: methodology and results of pilot surveys in Georgia, Mongolia and the Philippines, Asian Development Bank, http://dx.doi.org/10.22617/TCS189212-2. [5]
Avendano et al (forthcoming), Proposing a use of statistics indicator in national development plans, PARIS21, http://paris21.org/sites/default/files/inline-files/Proposing%20a%20Use%20of%20Statistics%20indicator%20for%20National%20Development%20Plans%20PARIS21%20draft.pdf. [14]
CCIC (2018), Progressing national SDGs implementation: An independent assessment of the voluntary national review reports submitted to the United Nations High-level Political Forum on Sustainable Development, Canadian Council for International Co-operation, Ottawa, Ontario, https://ccic.ca/wp-content/uploads/2018/06/ES-Eng.pdf. [18]
CIVICUS (2017), The data shift, Civicus, http://civicus.org/thedatashift/#av_section_2. [10]
Developement Initiatives and Open Institute (2016), Using community-generated data to deliver and track the Sustainable Development Goals at the local level: A case study from Lanet Umoja, Kenya Case Studies, Development Initiatives, http://devinit.org/wp-content/uploads/2016/10/Using-community-generated-data-to-deliver-and-track-the-Sustainable-Development-Goals-at-the-local-level_DI_Case-study.pdf. [3]
Development Initiatives (2017), P20 Initiative: Baseline report, Development Initiatives, http://devinit.org/wp-content/uploads/2017/03/P20-Initiative-baseline-report.pdf. [1]
DFID (2015), DFID’s guide to disaggregating programme data by disability, DFID, https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/530605/DFID_s_guide_to_disaggregating_programme_data_by_disability.pdf. [9]
Gazola-Hellmann, A. (2015), How does Bolsa Familia work? Best practices in the implementation of conditional cash transfer programs in Latin America and the Caribbean, Inter-American Development Bank, Washington, DC, ht, https://publications.iadb.org/bitstream/handle/11319/7210/How_does_Bolsa_Familia_Work.pdf?sequence=5&isAllowed=y. [2]
GPSDD (2018), Inclusive Data Charter, GPSDD, http://www.data4sdgs.org/sites/default/files/2018-08/IDC_onepager_Final.pdf. [21]
GPSDD et al (2016), Open mapping for the SDGs: A practical guide to launching and growing open mapping initiatives at the national and local levels, GPSDD, http://www.data4sdgs.org/resources/open-mapping-sdgs. [11]
ILO and Development Pathways (2017), “Exclusion by design: An assessment of the effectiveness of the proxy means test poverty targeting mechanism”, International Labour Office and Development Pathways, https://www.social-protection.org/gimi/gess/RessourcePDF.action?ressource.ressourceId=54248. [26]
Johnson et al (2017), “The cost(s) of geospatial open data”, Transactions in GIS, Vol. 21/3, pp. pp. 434-445, https://doi.org/10.1111/tgis.12283. [23]
OECD (2017), Development Cooperation Report 2017: Data for Development, Organisation for Economic Co-Operation and Development, https://doi.org/10.1787/20747721. [19]
PARIS21 (2018), Advanced Data Planning Tool (ADAPT), PARIS21, http://www.paris21.org/node/2905. [13]
PARIS21 (2018), Capacity Development 4.0 (CD4.0), PARIS21, Paris, http://www.paris21.org/capacity-development-40. [25]
PARIS21 (2018), Draft report on responses to the joint survey on new approaches to capacity development and future priorities, PARIS21, http://www.paris21.org/results-cd40-task-team. [20]
PARIS21 (2018), Partner Report of Support to Statistics 2018, PARIS21, http://www.paris21.org/press2018. [22]
Ploug, N. (2016), Improving data disaggregation by a wider use of administrative registers in data production, Danemarks statistics, http://unstats.un.org/sdgs/files/meetings/egm-data-dissaggre. [8]
Rieland (2018), Artificial intelligence is now used to predict crime. But is it biased?”, Smithsonian.com, http://www.smithsonianmag.com/innovation/artificial-intelligence-is-now-used-predict-crime-is-it-biased-180968337. [24]
Serajuddin et al, from the World Bank (2015), “Data deprivation: Another deprivation to end”, Policy research working papers, No. 7252, The World Bank, Washington, DC, http://openknowledge.worldbank.org/handle/10986/21867. [12]
The World Bank (2015), “Uruguay : Systematic Country Diagnostic”, The World Bank, http://dx.doi.org/License: CC BY 3.0 IGO. [17]
The World Bank (2015), Bangledesh: More and better jobs to accelerate shared growth and end extreme poverty, The World Bank, http://openknowledge.worldbank.org/handle/10986/23101. [16]
Washington, D. (ed.) (2015), Panama: Locking in success, World Bank, http://openknowledge.worldbank.org/handle/10986/22035. [15]
UN Women (2016), Making every woman and girl count: supporting the monitoring and implementation of the SDGs through better production and use of gender statistics, UN Woman, https://kampania17celow.pl/wp-content/uploads/2017/07/making-every-woman-and-girl-count.pdf. [7]
UNDESA (2018), The Sustainable Development Goals Report 2018, UNDESA, http://unstats.un.org/sdgs/files/report/2018/TheSustainableDevelopmentGoalsReport2018.pdf. [4]
UNDESA (2017), Data Disaggregation and the SDGs: An Overview, UNDESA, http://ggim.un.org/meetings/2017-4th_Mtg_IAEG-SDG-NY/documents/Session_3_Benjamin_Rae.pdf. [6]
Notes
← 1. The Institute of Human Virology accessible at: www.ihv.org/news/2018/Institute-of-Human-Virology-IHV-Will-Undertake-Largest-HIV-Survey-Ever-Conducted-in-a-Single-Country.html.
← 2. The IAEG-SDGs is intensifying its work for a more common framework and operationalisation on data disaggregation. The IAEG-SDGs group defines disaggregation as the breakdown of observations within a common branch of a hierarchy to a more detailed level to that at which detailed observations are taken.
← 3. The method provides a reliable income estimate based on household characteristics using different variables such as “type of house wall” or “livestock ownership” to estimate household wealth. To confirm accuracy of proxies, multiple regression models are used as a way to evaluate relationships between the variables and the level of household welfare, both derived from household surveys.
← 4. In Indonesia, as of 2017, the proxy means test survey had not been repeated in the past four years, while in Mexico it had not been repeated in more than ten years (ILO and Development Pathways, 2017[26]).
← 5. The IAEG-SDGs data disaggregation work stream is engaging with custodian agencies to establish collaboration mechanisms.