Combating COVID-19: Data everywhere but not the kind we need

By Julia Schmidt, Policy Analyst, Archita Misra, Policy Analyst and Johannes Jütting, Executive Head, Partnership in Development for the 21st Century (PARIS21)

This blog is part of a series on tackling COVID-19 in developing countries. Visit the OECD dedicated page to access the OECD’s data, analysis and recommendations on the health, economic, financial and societal impacts of COVID-19 worldwide.

statistics-covid-19-shutterstoc-1688596069Earlier this year at the Munich Security Conference, World Health Organisation Director-General Tedros Adhanom Ghebreyesus said, “We’re not just fighting an epidemic; we’re fighting an infodemic”. He was referring to the excessive amount of information surrounding the COVID-19 pandemic. Data dashboards, aggregators and charts of all types have formed the basis of much of what we know about the pandemic, lending a veneer of legitimacy to often contradictory or competing claims. While it is true that on some levels we have never had so much data, it may not be the data we need for sustained policy response and recovery. This is especially true among least-developed countries, where looming data gaps, even in foundational statistics, persist and may seriously undermine the ability of governments to develop effective COVID-19 response and recovery measures.

What your data dashboard isn’t telling you

In a short period of time, a surge of data actors of all kinds—analysts, aggregators, modellers, visualizers—have filled public consciousness with dashboards, graphs and numbers on a variety of aspects of the pandemic. But while data might be available in quantity, quality data is often lacking. At times data can come from dubious  origins, leading one to question which data can be trusted.

In many developing countries, “foundational data” (basic data collected by governments upon which we can build data dashboards, models and analysis) are often unavailable. According to the WHO, only about half the deaths that occur worldwide are registered with information on cause of death. Shockingly, 81 countries (all the low-income and two-thirds of the lower-middle-income) either do not register deaths at all or have serious data quality issues.

With gaps in even the most fundamental statistics, it is no wonder that many countries are finding it difficult to address questions on the socio-economic impact of COVID-19. How many wage workers are experiencing income shocks? How many rural and urban children can no longer attend school? What is the effect on small and medium enterprises and informal markets?

Evidence is growing that COVID-19 amplifies inequities.  Women, immigrants, racial and ethnic groups are impacted differently. Yet the data needed to inform targeted policies are missing.

COVID-19 is squeezing capacity of national statistical systems

As countries adopt stringent containment measures, COVID-19 has created a dual shock that affects data demand and supply. This disrupts the data value chain for official statistics (see Figure 1). National statistical offices (NSOs) are finding it more difficult to do their job. The full or partial closure of workplaces has forced statistical offices and other institutions to shift to teleworking arrangements, often without the adequate technical infrastructure. National Statistical Systems that rely on technical assistance and capacity development programmes supported by development partners are likely to be the most hit, with budget decreases foreseen in several African statistical authorities.

Figure 1PARIS21-2020

Source: PARIS21, 2020

In response, most statistical offices (especially those with lowest capacity) have interrupted their data production and postponed field-based data collection. The World Bank’s global assessment of 122 national statistical offices is a stark reality check. About 65 percent are either fully or partly closed, 90 percent have instructed staff to work from home, and 96 percent have fully or partially stopped face-to-face data collection. Even where operations are ongoing, disruptions have severely limited the ability of two thirds of  the offices to produce essential statistics and meet international reporting requirements. For those that had planned census activities in 2020, preparatory activities were impacted in 55 percent of cases. Fieldwork had to be postponed to later in 2020 in 40 percent of cases, and to 2021 and beyond for another 16 percent.

A pandemic like COVID-19 also calls on national statistical offices to produce more timely and frequent data due to volatility in the economy and society, also rendering standard imputation methods inadequate. These interruptions in data collection are affecting access to key economic indicators related to prices, employment, international trade, finance and economic growth worldwide.

Three urgent changes for more trustworthy data

Many governments and development partners are now realising that they have been partially flying blind when designing their policies and aid packages. Now is the time to change course and accelerate the production of trusted foundational data. Three concrete areas of action could support this transition.

  • Support systems over silos

Support to data and statistics goes mostly to sectors, and not enough to systems support. While this is understandable, given the sectoral focus of many international organisations, it makes it difficult to design and implement the multidimensional recovery and social protection policies that are now required. It also becomes harder to make connections between intersectional issues, such as, for example, those affecting migrant women who have lost their income while being forced to return to their home villages.

Besides supporting the ad-hoc surveys that are needed, development partners should also invest in foundational national data, such as administrative data and civil registration systems, and institutional capacity development of national statistical systems. This would also ensure a consistent standard for data production over time, something that alternative data providers and unofficial sources cannot guarantee. This is essential, bearing in mind that the next big crisis will have different characteristics, involving different data requirements.

  • Build and maintain trust through data governance

Harnessing the potential of data from social media, contact tracing apps, satellites or mobiles phones to tackle crises such as COVID-19 can play an important role. However, if data access and use is not regulated, the health crisis may become a trust crisis. Therefore, development partners should also support institutional mechanisms and data governance frameworks to access, share and process data with accountability. As The GovLab co-founder Stefaan Verhulst reminded us, governance frameworks need to involve data protection and privacy rights, open data policies, trusted data re-use agreements and transparency requirements and safeguards. The current rush to set up contracts with private data providers in some countries might possibly solve some short-term data demands but backfire in the long run.

  • Back promises with funding

 Official development assistance from wealthy countries to support data and statistics in developing countries is only around half of what it needs to be to ensure that they have adequate data to support sustainable development and measure SDG progress. Often, donor support to statistics goes to issues with more visible impacts, like building schools or hospitals. While data systems support the success of development objectives across sectors, policy or funding focused on data systems are rare.

Getting the data right, as the current COVID-19 pandemic shows us, has life and death implications. Informed public debate needs reliable, trustworthy data on wide-ranging issues from small business bankruptcies, to gender inequality and poverty. This calls for empowering national statistical systems that produce the foundational data we need for our recovery and response. Some of our COVID-19 response funds should go to strengthening national statistical systems in developing countries – to trace the consequences of the crisis and determine the impact of fiscal stimulus used.

Most importantly, every single data point on death or recovery represents a human life. We owe it to the dignity of each person, regardless of the country that they live in, to portray that number accurately.

One thought on “Combating COVID-19: Data everywhere but not the kind we need

Comments are closed.