Agenda

From Governing by Numbers to Governing by Data: Public Service Regulation by Algorithm
h. 13.00 - 14.00
Management Lectures by Andrea Mennicken, London School of Economics and Political Science, London

Regulatory bodies across sectors and countries are paying increasing attention to the potential of ‘big data’ for the purposes of regulation. The prospect of moving to ‘real time’ measurement is alluring, as is algorithmic regulation’s potential for prediction, for example with regards to the prediction of failure (e.g., operational and financial failure in hospitals, schools, universities, prisons and other regulated entities). Thanks to the combination of algorithms, superior computing power, unprecedented volumes of digital data and 'the cloud' regulatory oversight can be complemented with simulations of the future that previously remained outside regulatory capacity (and imagination). Furthermore, algorithmic regulation promises an answer to frustrations with existing regulatory practice: regulatory regimes are often criticised for failings in information exchange; too many opportunities for gaming; inconsistencies in the regulatory approach (e.g. human-led inspections are seen as open to arbitrary judgements); and too much focus on process rather than outcomes.

This talk explores potentials and challenges posed by the utilisation of ‘big data’ and machine learning algorithms in the regulation of public services. First, it examines different, at times conflicting, political imaginaries that have come to be attached to attempts aimed at governing by ‘big data’. Here it is highlighted that the widespread enthusiasm for algorithmic regulation hides much deeper differences in worldviews about regulatory approaches, and that advancing the utilization of algorithmic regulation has the potential to transform existing mixes of regulatory approaches in non-anticipated ways. Second, it is shown that governing by ‘big data’ presents distinct administrative challenges in terms of knowledge creation, coordination and integration. Finally, it is argued that the use of machine learning algorithms in public service regulation requires renewed attention to questions of the ‘regulation of regulators’, and that questions of algorithmic regulation need to be explored in the context of understandings of legitimate use of state power.

The talk draws on different examples from regulatory practice, especially in the UK. It draws on findings from a cross-sectoral research project which investigated the role of quantification and big data in higher education, the correctional services and hospitals (Economic and Social Research Council Grant ES/N018869/1; project title: QUAD — Quantification, Administrative Capacity and Democracy).