The Sustainable Development Goals (SDGs) were set up to support sustainable health, tackle poverty and enhance peace and prosperity for present and future generations, at all levels, locally to globally.¹ The implication is that the SDGs can be achieved through governmental and non-governmental interventions, supported by data and statistics to ensure that they are on course to deliver those goals and targets. The rise of an evidence-based policy paradigm and the idea of managing by results has led aid agencies and international policy-makers to place statistical measurement at the heart of monitoring and evaluation of official development assistance.²

Statistical development in Africa has attracted the interest of international policy-makers as well as regional and national bodies across the continent. There is a growing sense that statistics should be the backbone of sound policy decisions.³ There are now increasingly persistent calls for African policies to be driven by evidence-led research, turning away from gut feelings or ideological-driven agendas as nations embark upon the process of achieving the SDGs. Poor statistics hurt African governments’ ability to make good policy decisions; reversing this requires the collection of sound data and its effective use in addressing the issues of transparency and accountability.

Reliable statistics provide the evidence needed to assess solutions to socio-economic problems facing Africa. For example, no government can build schools without prior knowledge of the numbers of children likely to be enrolled. Similarly, no government can claim to have reduced crime rates unless it can compare statistics on current crime rates to those of previous years. A country needs to know what crops it grows well, and where, if it is to prevent famine and malnutrition in children. Donors can only know whether their assistance is changing lives if they have access to quality data, stored securely and readily accessible for decision-making purposes. Development programmes should produce measurable results, and developmental decisions should be informed by the analysis and interpretation of data by government and/or educational agencies. In short, statistics constitute the barometer for measuring whether governments are making progress in addressing the concerns of their populations.

In this context, in 2014, the United Nations launched an appeal for a ‘data revolution’ prior to launching the SDGs.⁴ The concept of a data revolution highlighted the need for reliable statistics to address the widening gap between developed and developing countries in terms of access to and use of information. Statistics enable the state to address crucial issues affecting the lives of its citizens. They are a means by which citizens potentially can hold governments and their policies accountable.

In recent years, the state of statistics in Africa has been subjected to intense criticism. The major grounds have been, first, that development data produced by African regimes tends to be fabricated in order to reflect well on the regime and that collecting verifiable data inside closely guarded societies is virtually impossible.⁵ Second, it is argued that in any case, statistics gathered in Africa are often flawed and do not present the true situation on the ground.⁶ These criticisms tend to focus largely on export–import data and the economic sector of Africa,⁷ while pointing to the collapse of African statistical agencies. The concerns, which have also have been expressed by the World Bank, are strong indications of African statistical agencies’ inability to generate reliable and comparable data needed to evaluate the continent’s progress. Indeed, very few of the statistics produced in Africa, and especially sub-Saharan Africa, are sufficiently reliable to use, and critics contend that virtually all are guesstimates.⁸ This situation has persisted for so long that critics suggest that securing reliable data in and for Africa is an unachievable goal.

This chapter, however, takes the view that rather than engage in sheer scepticism, it is more helpful to appreciate the progress made by African states and to focus on the way forward in achieving and monitoring governance in relation to the SDGs.⁹ We present a situation analysis and review the state of statistics across Africa in relation to the SDGs. The main issue here is that Africa offers a contrasting picture. On the one hand, some valuable statistical data exists, even though they are rarely used to plan and implement policies. On the other, many of the statistics that are gathered and published may not be of much help in addressing issues that matter to international and national development agencies. Moreover, there is a lack of technical capacity to analyse these data and make it available to the public as a basis for determining basic needs at the sub-regional or sub-county levels.

We argue that this gap must be addressed if African countries are to tackle real issues facing their populations with a view to achieving critical SDG targets. In this respect, statistics in Africa cannot be understood in isolation from the social conditions in which they are produced, processed and managed. Part of the reason that the available statistical data is underused has to do with the conditions for accessing and storing it in a continent that is traditionally more reliant on paper than on electronic media.¹⁰ Thus, in considering the state of statistics in Africa and in analysing the challenges that statisticians and data collectors face, attention needs to be given to the socio-economic and political conditions in which the information is collected, processed, stored, managed and used.¹¹

We explore these issues in four sections. The first defines ‘data’ and ‘statistics’, terms that are often mistaken for one another but that have distinct meanings. The second offers an overview of statistical censuses as carried out in Africa. To produce this, we used a Strengths, Weaknesses, Opportunities and Threats (SWOT) model to understand the issues associated with producing good quality statistical data in Africa. The third focuses on these key challenges in relation to gathering reliable statistics in Africa. The last section suggests some of the ways that these challenges can be overcome, especially in relation to addressing the schism between demand for and the supply of data.

Defining the terms statistics and data

The starting point for discussing statistics and data in the African context is to consider how these terms should be defined. At its most basic level, data is information about a subject of interest (for example heights in a population), which can come in different forms.¹² Data can be quantitative (numerical) or qualitative (descriptive), for example, the answers to interview questions. The important point about data is that if they are to be of value to policy-makers, they need to reflect what they need to know. For example, data on the number of children of a specific age attending school in a region of the Democratic Republic of Congo will be a good indicator of the number of schools to be built for potential class intakes.

When a census is conducted of children of school-going age in a given region, it is possible, based on the data, to make statements directly about the population, for example, its average age. Generally, we can only estimate a particular characteristic or variable in the entire population; it is practically impossible to collect information on many issues at once. Rather, we can take a representative sample of the population we want to know about. We also want the sample to reflect the diversity of that population (for example, boys and girls, ethnic origin, disability cases). There are techniques for ensuring this as well as for random sampling.¹³

Descriptive statistics are a mathematical tool for analysing and organising data about a given state of affairs in a summary form.¹⁴ They illustrate different characteristics of a particular sample or population, making it possible to present the data in a meaningful way; statistics obtained from a sample of the population can be used to make inferences about the characteristics of the population. For instance, it can be helpful to an international developer or national policy-maker to know the mean, or average, of a particular variable. This is calculated by adding up the value of all the numbers reflecting a particular variable and then dividing that sum by the total of all the numbers.

Descriptive statistics are also about the spread or variability of a dataset, i.e. how much the data clusters around the mean, or whether the values are widely dispersed. The standard deviation is a measure used to quantify the amount of variation or dispersion of a set of data values. The way a standard deviation is calculated for data from a population is different from the way it is calculated for a sample. Data that has been statistically processed using measures including standard deviation, frequency, mode, range and interquartile range, are often referred to as statistical data.¹⁵ It is the raw information from which statistics are created. Statistics cannot exist without data, but it is possible to have data without statistics.

For data and statistics to remain relevant to policy-makers in Africa and beyond, the results should be interpreted in such a way that is easy for the decision-makers to understand. This is critical, not least because they need to have confidence in what statisticians recommend. Effective communication between statisticians and decision-makers is fundamental to using statistical data to decide whether and how to plan an intervention.

This discussion about data and statistics raises two fundamental questions. First, what sort of statistics are available about Africa? Second, to what extent do the existing statistical data provide a reliable foundation upon which to base policy decisions? To address these questions, we first turn to the nature and issues facing one of the most crucial aspects of data in Africa, namely census data.

Census data

The availability of a significant amount of cross-sectional census data in Africa makes it possible, theoretically, to monitor and explore the state of a country’s current development across many aspects of social life (e.g. health, economics, education and science). They provide a means by which public policies can be continuously evaluated. This has led some scholars to make an association between good data and good governance.¹⁶ However, while census data have often been collected across sub-Saharan Africa, they tend not to have been collected regularly. We cannot compare statistics gathered at irregular intervals, often of several years, given the existence of many gaps and inconsistencies in data collection. Policy-makers find it difficult to use the data effectively to assist in implementing policies.¹⁷

Investigations using census data are likely to yield meaningful results because census data are a complete enumeration of all individuals in a country at a given time, allowing a meaningful understanding of progress, including monitoring the SDGs, which is crucial for implementing interventions. A population census is the result of the process of collecting, compiling, evaluating, analysing and publishing or otherwise disseminating demographic, economic and social data pertaining to all people in a country, or in a well-defined part of a country, at a specified time.¹⁸ Censuses play a vital role in developing the official statistics needed to assist state agencies, businesses, other organisations or the public in planning, decision-making, monitoring or assessing policies.

Census data tend to be collected in such a way that the identity of the respondents is protected and that the data are relevant, accurate, reliable, timely, objective and comprehensive. Generally, such data are compiled, reported and documented in a scientific and transparent manner and disseminated impartially. Moreover, they tend to be collected in accordance with national and international standards and classifications that are appropriate for distribution by gender, disability, region and similar socio-economic features.¹⁹ A census is a complex and costly enterprise, especially in terms of careful planning and mobilisation of people and resources,²⁰ particularly so because all inhabited areas must be visited to provide a fair coverage of the entire population.

The majority of countries in sub-Saharan Africa conducted their first population censuses in the 1970s as a result of the African Census Programme (ACP), which was established by the United Nations Economic Commission for Africa. The ACP provided significant technical and financial assistance, received through the United Nations Population Fund, which enabled many countries to conduct censuses, especially in the 1980s and 1990s. By the 1990s, sub-Saharan Africa had assembled an impressive volume of population data, and national statistical institutes had developed expertise in collecting, processing and analysing these data.

In recent years, however, without ACP support, serious financial difficulties have prevented the organisation of population censuses. This has resulted in increased intervals between censuses or in a lack of censuses, as illustrated in Table 2.1. This, in turn, has reduced the quality and volume of statistics available to governments for planning and formulating policy and for efforts to monitor the MDGs and, later, the SDGs.

Apart from the lack of funding for censuses initiatives, the intervals in census data collection in African countries have also been due to wars, political instability, economic crises, and inadequate policies and leadership.²¹ Unless satisfactory solutions are found, many sub-Saharan African countries will find it impossible to use reliable statistics to monitor SDGs. Strategies are needed to address this issue. Effective awareness-raising, good organisation, rigorous planning of operations and optimum use of new technologies will make censuses less difficult to finance and more likely to produce reliable statistics for monitoring and achieving the SDGs.²²

In countries where population and housing censuses are carried out, the data represent a significant source of information on health matters (for example, immunisation and family planning) regardless of how inadequate and incomplete vital registration programmes may be. This is why it is crucial for governments in sub-Saharan Africa to ensure that full censuses are carried out if they are to meet the growing demand for statistical information at the national and subnational levels and to support the SDGs.

Post-apartheid South Africa is among the few African countries that have made progress in this area. Conducting its first population census in 1996, South Africa subsequently carried out censuses in 2001 and 2011, largely to compensate for the unreliable, uneven statistics produced during the apartheid regime, as illustrated by the overall figures produced under the Native Laws Amendment Act, the Areas Amendment Bill and the Group Areas Act 1950, all of which grossly underestimated the overall figures of urban and city residents.²³

Kenya has also worked to correct past distortions. In 2006, the government passed Law No. 4 of the Statistics Act, making it mandatory for the state to carry out regular censuses for every ten years on the basis of a printed questionnaire.²⁴ This is part of the broader context for the Kenyan population and housing census that aimed to deliver on the country’s vision: ‘Counting our People for Implementation of Vision 2030’. By generating information at all administrative levels, the Kenyan government has sought to provide a sound basis for assessing policies relating to its population. Despite this ambition, however, Kenya has not collected any fresh data in keeping with its current population growth.²⁵

The Democratic Republic of Congo illustrates an extreme situation. Relevant statistics do exist, but they date back to collection efforts by the National Institute of Statistics in 1984. Since independence in 1960, the DRC has had a turbulent history, and the earlier statistics have become obsolete; any projections that might be drawn from them will no longer be helpful in planning interventions. The idea of a second census was mooted and planned for July 2011 but did not materialise, even though a decree on its organisation was signed in August 2009.²⁶ It would be particularly valuable to have a census in this post-conflict country, which is heavily in debt, with a very poor population, despite its huge reservoir of mineral resources. The results of an up-to-date, well-conducted census would allow economic and social planning based on reliable statistical data that could contribute to the reconstruction of the country and help build capacity in the National Institute of Statistics. It could provide the nation, as well as international organisations, with reliable data for monitoring the SDGs.²⁷

Having considered issues relating to census data, we now need to consider other statistical activities at the regional and sub-continental levels.

Statistical activities in Africa

The Mo Ibrahim Foundation has examined the main activities undertaken by statisticians in Africa and considered the crucial significance of data for policy-making and service delivery. A report released in 2016 by the Mo Ibrahim Foundation on Africa’s data revolution noted that there has been progress in the quantity of data being collected over the past ten years, especially in household surveys and population censuses. It noted, for instance, that:

However, the report also noted that there continued to be challenges in the frequency and the quality of data produced. For instance:

The most readily available statistical data in Africa have been collected by western-based institutions. These data are important for sectors responsible for budgeting and planning where no other reliable data exists.³⁰ Local-level data are sparse in sub-Saharan statistical systems,³¹ and the available local data often do not provide the information needed for realistic planning. Decision-makers at international, national and local levels need data that are disaggregated down to the lowest level of administration. The ability to disaggregate data (breaking them down into sub-population, district, locality, and so on) enables policy-makers to plan appropriate programmes, determining which evidence-based interventions are most appropriate and deciding where they are most needed.

For instance, small sample surveys do not provide enough information to allow a health service in a given African country to determine precise locations where resources need to be allocated. ‘Services are delivered through local authorities who need intelligence on their local communities to know how best to serve the people. Counting people to make people count is what the Data Revolution is about.’³² Disaggregating data can show where aggregate data are masking discrepancies. For example, by looking at disaggregated data for smaller sub-populations, a national policy-maker or an international developer can recognise whether outcomes vary by sub-population and whether strong results by some sub-populations are masking poorer results by others.

Data collected at irregular intervals tend to be of uneven quality, as we illustrate in Table 2.2, which is based on information compiled from USAID-sponsored Demographic and Health Surveys³³ (DHS) and the UNICEF Multiple Indicators Clusters Surveys³⁴ (MICS).

Over the past 20 years, official statistics in Africa have suffered stagnation and obsolescence due to a progressive lack of sufficient human and financial resources allocated by governments. Beginning in the 1980s, when policy-making in the continent began to be dominated by structural adjustment programmes,³⁵ the main effect has been the steady reduction or curtailment of budgets for data collection and statistical analysis over decades.³⁶

The number of experts involved in data collection has grown over this period, but there has been an absence of basic controls for the reliability and quality of data.³⁷ Big data is an example of information produced outside official controls that should describe the process of drawing together disparate datasets to offer new insights into a population.³⁸ One of the challenges of this new trend is that there is no guarantee that disparate datasets are unbiased or will remain relevant to the sectors that are of interest to international and national policy-makers. Moreover, it is unclear how the rapidly growing pools of data generated through data digitisation and algorithms match publicly held databases, for instance those related to control of diseases in any given country. Another concern is whether big data is to be seen as an outright challenge to the credibility of African statistical institutes or even as a gradual takeover of the growing market for statistical information.

At the same time, there is evidence that some African governments are beginning to use data in new ways. Moving beyond doubts expressed about technical expertise in national data collection and statistics,³⁹ and beyond long-standing patterns of corruption and misfeasance, countries such as South Africa, Mozambique and Namibia have begun leveraging big data to improve efficiency and effectiveness of government in policy areas such as citizen security, taxation and smart cities. Other countries in sub-Saharan Africa have been slower in appreciating the role of data and statistics as an aid to government in discharging its obligations towards its citizens⁴⁰ and embarking on the road to prosperity.⁴¹

A detailed and richer picture of the state of statistical data in Africa is captured through use of the SWOT technique as described below.

SWOT analysis

A situation analysis aimed at understanding the emerging internal strengths, internal weaknesses, external opportunities and external threats for managing statistics across the continent should help explain the factors that must be addressed if realistic solutions are to be developed.

As Table 2.3 shows, Africa’s strengths in the field of statistics include emerging new frameworks, regulations and action plans geared towards enhancing statistical development. After a significant decline in the quality of work by national statistics offices in Africa from the 1970s, the Addis Ababa Plan of Action for Statistical Development in Africa was officially adopted in May 1990 by the United Nations Economic Commission for Africa, Conference of Ministers for Economic Development and Planning. The Plan, which was at the cutting-edge of statistical advocacy, promoted evidence-based development. This trend has been enhanced by the fact that most African countries have joined the global effort to reduce poverty by supporting the achievement of the SDGs, including sustained and equitable economic growth, which is in line with the objectives of previous pan-African initiatives. The regular exchange of knowledge and experience, as well as agreed good practices among national statistical experts in Africa, has also helped.

These strengths, if anything, are overshadowed by weaknesses and threats for statistical reliability. As the SWOT model suggests, despite positive trends across Africa, the picture is mixed. We recognise the significant efforts by international agencies to ensure that Africa has reliable data. For instance, recent improvements in Nigeria’s national accounts compilation methodology have effectively doubled the estimated Nigerian GDP, making Nigeria the leading economic power in sub-Saharan Africa, even before South Africa. However, this reinforces the argument: it shows real improvement in Nigerian statistics, but nevertheless, the actual data on the country’s macroeconomics remains questionable.

In recent years there have been improvements in statistics in Africa in relation to the continuous development of both the Demographic and Health Surveys (DHS) and the Multiple Indicators Clusters Surveys (MICS). However, even here significant challenges remain. As has been discussed and shown in Tables 2.1 and 2.2, data are gathered irregularly, and some sources have not been updated in recent years. In some cases, key information is lacking about a particular issue for sub-regions or for counties. Where census data have not been collected consistently, their quality and timeliness can be compromised, especially when there are delays in collecting and releasing data. Timeliness is a serious problem. Where there is a long interval between data collection and reporting, by the time the data are released, the circumstances that led to a demand for them may have changed, making them redundant, irrelevant and difficult to compare. This encourages a culture where policy-makers and the media tend to rely on data estimated on the basis of gut feelings. Many of the DHS and MICS depict situations at the national level that mask the reality at the subnational or regional levels. These gaps, and their impact on the quality and timeliness of data, must be tackled if DHS and MICS sources are genuinely to become major pillars for national and international development programmes in Africa.

Overall, the challenges for statistical development in Africa remain enormous, and the scarcity of capacity to tackle them is pronounced. The shortcomings are largely due to the lack of significant investment in obtaining data needed for planning and implementing policies. National statistical institutes in most sub-Saharan countries have limited human and financial resources, and often there are insufficient skills among staff responsible for data collection and management. For instance, a statistical department may have only two or three competent statisticians and demographers, while the majority of staff lack appropriate training.

There are also technical issues, especially when data are in paper form and liable to destruction or where there is no digital preservation strategy in place. Even where appropriate technology exists, there are the additional challenges of power cuts, poor equipment and low bandwidth, which compound the difficulties of accessing and sharing data. All of this significantly impacts the data that are collected. For instance, data on older women, youth and agricultural activities remain a challenge.⁴² Despite the persuasive UN language (‘nobody left behind’), major investment is needed if all areas of statistical concerns are to be adequately addressed.

The result is that Africa has inadequate statistical information on key development indicators, such as environmental/climate change, gender, governance, HIV/AIDS control. Often, laws relating to statistics are not enforced, and plans for improvement are not implemented, leaving national governments without reliable information about the populations they should be serving. As has been noted, data from various sources (census, surveys, civil registration) tend not to be disaggregated to the community level, which is where interventions are needed.⁴³ Census population data are readily available, and over 80 per cent of African countries have conducted a census in the last ten years. However, there are two significant challenges: first, how can the coverage and quality of the census be reliably assessed, and second, how can we ensure that data are disaggregated to the lowest level of administration where they are needed for planning purposes.

Surveys in Africa have tended to be driven by international organisations, and there is an assumption that well-structured data exist to inform policy planning and implementation. It is true that many African countries have conducted at least one or two household surveys over the last three to five years. But, as World Bank staff have noted, because only 33 per cent of African countries have conducted more than two poverty-related surveys in the past ten years, comparative data are not readily available, and civil registration, which represents the only credible way to count people on an ongoing basis and thus to produce useable vital statistics, has not been adequately resourced.⁴⁴

Another major statistical challenge for sub-Saharan Africa is that much of the data that exist are buried in files across government ministries and agencies. As one scholar has indicated, ‘sometimes, sourcing of this administrative data is made very difficult due to administrative bottlenecks’.⁴⁵ End users are often faced with secrecy and confidentiality issues. Significant data may be buried in an obscure special report, in published documents held only in a few libraries or on administrative files. Sometimes staff employed in government-run National Statistics Institutes are simply not aware of the detailed statistics contained in these documents. Weaknesses in records systems often make it difficult to find relevant information on government files. Even if the relevant source is identified, the data may not be arranged according to time series, which makes the users’ task very daunting. Annual reports and other documents often contain figures for the most recent years only.

Other constraints arise from the social conditions under which the data are collected and recorded. These can include misconceptions about why the statistics are being collected, uncooperative attitudes of participants from whom data are being collected, statistical illiteracy, ineffective statistical legislation and lack of dedication by enumerators.⁴⁶ These challenges are often the result of political influence over how statistics are collected and presented.

Political influence is a challenge for statistics in Africa.⁴⁷ Too often, policies are driven by political views rather than by empirical analysis.⁴⁸ Even where data are available, they may be ignored by African policy-makers, especially where they threaten the personal views or political needs of leaders. Often, data captured through national and international efforts and their potential value are not known to the public. The prevailing culture of secrecy, the fear of misuse of the information and the perceived need for confidentiality add to the difficulties of accessing and using the data.

By way of summary, the issues presenting a challenge in Africa can be categorised as those of a technical nature related to the data and statistics themselves and those associated with the overall management infrastructure (for example, laws and policies, human and financial resources).

Technical data-related issues include:

Infrastructural management-related issues include:

These weaknesses have contributed to the argument that African statistics are a tragedy. The statistical narratives emerging from the continent ‘tell us less than we would like to think (they do) about income, poverty and growth in the region’.⁴⁹ This concern has been echoed by the World Bank’s chief economist for Africa, Shantayanan Devarajan, who has argued that because of the state of statistics in Africa: ‘We cannot be sure whether there is growth or that poverty is declining’.⁵⁰

Overcoming the challenges

We have discussed some of the challenges of gathering reliable data and statistics in Africa. Let us now consider some of the ways that the challenges can be overcome. There is a need for a pan-African approach to transforming the current situation. By 2009, a number of African institutions, including the Economic Commission for Africa, the African Union and the African Development Bank, had developed a Strategy for the Harmonisation of Statistics. The strategy, which was adopted that year by African heads of state, remains a valuable way forward. It is a continent-wide effort to produce and disseminate harmonised statistics to inform African development at the local and regional levels.⁵¹

The strategy provides a framework for defining policies and good practices for developing, producing and using statistics and is intended to guide the process of harmonising concepts and definitions, adapting international or peer-agreed good practices, such as quality frameworks, and using common methodologies for producing and disseminating statistics to facilitate comparisons of statistics of African Union member states across time through coordination and collaboration of national, regional and international stakeholders.

Advocates of the strategy believe that statistical services in regional institutions⁵² need to be strengthened in order to generate timely, reliable and harmonised statistical information to support political, economic, social and cultural integration across the continent. Working with individual African governments, the strategy offers the potential to tackle the challenges of producing, analysing and disseminating the quality information needed to inform decision-making and evaluate the results of policies. However, to achieve this, it will be necessary to develop statistical capabilities within a reasonable timeframe. Establishing such capabilities will require human, material and financial resources; mobilising these resources will require significant and constant support from all development partners.

The strategy proposes that national statistical systems should be structured to:

With sufficient resource allocation in the coming years and a significant dose of political goodwill, Africa can address the overall poor quality and inconsistent nature of basic statistics in countries throughout the continent and be in a position to monitor the implementation of the SDGs by 2030. However, the goals can be achieved only if the situation becomes a priority for governments and development agencies in coming years.

Conclusion

This chapter has aimed to provide a frank assessment of the current state of statistics in and for Africa. We have noted that statistical development has a chequered past, owing largely to the social and political conditions under which statistics are produced. We have explored the current state of statistics in selected countries, investigated key challenges and considered ways in which those challenges can be overcome. The focus has been on both internationally generated statistics, such as DHS and MICS, and national efforts to collect and manage statistics.

By exploring the strengths, weaknesses, opportunities and threats facing statistical development in Africa, we have considered the major challenges that need to be overcome if African countries are to achieve the SDGs. We have noted, in particular, the importance of ensuring that national statistical institutions are properly resourced and have the autonomy needed to resist pressure to give undue credit to a particular leader. We also noted the importance of making statistics a national priority and building the capacity of existing institutions so that Africa can generate accurate, timely, relevant, accessible and unbiased data. Implementing these changes will require that governments act in partnership with civil society groups and international agencies to support statistical priorities.