Deploying data for better governance

Data-driven decision-making can transform public service delivery, addressing key questions on equity and transparency.

Data science has transformed our world, and AI continues to disrupt creatively, transforming work and workplaces. In the age of big data and open data, governments must begin to use data-driven tools for better public governance.

Big Data for Social Good teaches how to use big data, data science, and economics to address contemporary social issues. Big data can help develop scientifically effective policies, bringing the concept of ‘evidence-based policymaking’ closer to reality.

Big data can reveal how our neighbourhoods and schools influence our life outcomes, allowing us to improve possibilities for everyone.

Covid -19 disrupted everyday life and upended livelihoods even as governments scrambled to respond to the pandemic supposedly guided by evidence or the science, as they made decisions to impose lockdowns or open sectors or geographies.

In practice, making several of these decisions to shut or reopen schools, to requisition hospital beds and how many, to impose travel restrictions or quarantine requirements or ease them, posed the challenges of using and interpreting data to make decisions that impact entire geographies or populations.

It highlighted the importance of enabling policy-makers and public officials to learn how to use data-driven tools and provide training to simplify what might appear at first sight as a complex technical subject. In essence, highlighting the need to address the issue of equity and the bureaucracy.

Mainstreaming data science and analytics into government decision-making is not an easy task. It is often the case that many in government develop anxiety around inferential statistics or data analytics. The first step, therefore, is to dispel this sense of trepidation by leveraging their interest in solving real world public policy problems.

Introductory training for those government officials at the operational levels in domains that have a large citizen-public service interface such as public health, school education, agriculture, and criminal justice will help begin the process of systematic evidence evaluation. Data-driven decision-making in government is about public service delivery.

In business, if the application of a new algorithm helps grow profits, that becomes the standard for decision-making. In government, the imperatives of developing data-based use-case applications are far more nuanced, like fairness, equity and politics.

A business analyst can optimise for profits, but it takes a social scientist to optimise for these other bottom lines. Interpreted thus, I would suggest that data analytics in government is akin to policy planning with empathy. For an algorithm to effectively deliver a government resource or a public service effectively, equitably, and transparently, it must be developed defining one objective function alone: to maximise public welfare.

Addressing inequalities

Government departments must begin by combining a conceptual framework and a set of objective functions that circumscribe that domain. An important resource in this approach that can be transformational will be to draw on open access data sets and open-source use-case applications that can help introduce what might be described as ‘open data analytics’ that hold the potential to drive innovation in public service delivery on a government-wide basis.

Departments collect similar administrative data and follow similar service delivery processes on common issues of public interest such as malnutrition, school education, affordable housing, child welfare, and public health.

This common set of use cases and processes means that one agency can develop analytics and supporting materials, share the code, and enable other government departments to replicate the solution in their own domains. Karnataka, which already has a high degree of digitisation and e-governance, is well positioned to lead the open analytics movement.

A key challenge in public policy and practice is resolving the problem of structural inequality and casteism in resource allocation and hence, in advancing human development. It is axiomatic that caste and class determine who gets what and why, they also shape decisions like where we live, which schools our children go to, the access we have to healthcare and the quality, and the opportunities for upward mobility in our own communities.

You might ask if resource allocation has historically been driven by inequality, how might data-driven decision-making be any different. Political economy doubtless is a hard constraint. However, if we can demonstrate how inequality manifests through government data and frame the analysis to show how a particular intervention will exacerbate the negative impact across caste and class, it would give pause to the political decision-makers and would be a step forward in making a trade-off between political necessity and fairness.

At present, government officials at most levels learn a lot about policy and public service delivery but little about identifying at-risk populations and hence, programme interventions that constitute a necessary condition for providing the required social protection support.

Greater attention to and targeted capacity building in programme design is essential. Therefore, training government functionaries to consider data analysis-driven and algorithm-based approaches to public service delivery will be the next step in improving public governance.

The second challenge in public praxis is the imperative of equitable resource allocation of scarce public resources. Most governments work in resource-constrained environments where ‘Less is More’, and optimisation solutions can be game-changing. Data Science can help in this process and enable governments to allocate limited resources based on needs.

The COVID-19 pandemic showed how difficult and politically fraught the process of procurement can be in the backdrop of limited resources: from testing equipment, PPE kits, oxygen, and vaccines, every aspect was a challenge.

The supply/demand mismatch is perhaps the most common problem in government, affecting disaster response, provision of health services, and social protection services. Developing science-based data optimisation solutions can unlock new value here, but, again, only if the government begins to think critically and begins work on providing a framework for open analytics in public governance.

The states in India constitute the theatres of development praxis, and there is no greater use for data analytics than in public health, school education, agriculture and climate change, and skills and livelihoods. The states must, therefore, turn attention to building capacities for data-driven public governance.

(The writer is the Director, School of Social Sciences, Ramaiah University of Applied Sciences)