Library Header Image
LSE Research Online LSE Library Services

Information regimes in government bureaucracies and 'digital decompression'

Dunleavy, Patrick ORCID: 0000-0002-2650-6398 (2022) Information regimes in government bureaucracies and 'digital decompression'. In: UK Political Studies Association Conference, 2022-04-11 - 2022-04-13, University of York, York, United Kingdom.

[img] Text (Dunleavy__information-regimes-in-government-bureaucracies--published) - Published Version
Download (784kB)


In all bureaucracies how information is acquired, stored, re-accessed and analysed creates an ‘information regime’ of crucial importance for the rational or efficient conduct of business. Government departments and agencies use a wide range of information practices that can seem simply heterogenous, highly specific or hard to characterize. Yet an essential move in all pre-digital forms of organization has been data compression, using administrative routines to reduce complex realities to data and information in formats and quantities that can be classified, indexed, filed and re-found when needed. Three conventional information regimes can be distinguished by their level and mode of compression. ‘Lossy’ data compression via drastic data selection and radical simplification, especially using open or gated-access forms, predominated in machine bureaucracies with hierarchic morphologies. By contrast, professional bureaucracies developed ‘lossless text/narrative/verbatim’compression for mission-critical tasks, relying on professional language, socialization, and knowledge development to summarize cases or events in more fully recoverable forms. With the advent of new public management and late twentieth century computerization/automation, hybrid forms of machine/professional bureaucracy developed, focusing on metrics-based compression (using pre-fixed statistics, key performance indicators and similar data) in a central governance role. In the current digital era governance wave technologies facilitating big data, artificial intelligence and data science approaches have made feasible a new information regime of ‘lossless’ uncompressed data and expanded data science, opening a potential for bureaucratic operations to alter in fundamental ways. Full digital data gathering or recording of interactions at the initial stage plus complete storing, organic indexing and new analytic capabilities can obviate much of the earlier need for data compression, and foster forms of post hoc knowledge development, e.g., via machine learning and algorithmic governance. This development will change most government bureaucracies somewhat, but how far still remains unclear.

Item Type: Conference or Workshop Item (Paper)
Additional Information: © 2022 The Author
Divisions: Government
Subjects: T Technology
Q Science > QA Mathematics > QA76 Computer software
J Political Science > JF Political institutions (General)
Date Deposited: 28 Mar 2022 09:03
Last Modified: 28 Mar 2022 09:03

Actions (login required)

View Item View Item


Downloads per month over past year

View more statistics