The State’s Politics of “Fake Data”

Paper Title

The State’s Politics of “Fake Data”

Publication Info

• Topic area: Examination of the production, negotiation, and functionality of "fake data" within state bureaucracies.
• Keywords: Fake data, state governance, data politics, ethnography, bureaucratic processes, data fakeness, representational accuracy, institutional functionality, critical data studies, sociotechnical systems.

Background and Problem

• Problem / challenge: Existing frameworks treat "fake data" as a binary failure of representation, ignoring its relational, processual, and performative dimensions. This perspective fails to account for how fake data function within state systems and why they are often normalized.
• Significance: Understanding the production and use of fake data is critical for improving state governance, ensuring accountability, and addressing public trust in state institutions.
• Motivation and related work: Prior research has focused on detecting and correcting fake data as a technical problem, often overlooking the organizational and political contexts that produce and sustain it. This paper builds on critical data studies and sociotechnical systems theory to explore the institutional logics and practices underlying fake data.

Solution

• Proposed approach: A relational, processual, and performative framework for understanding the production and functionality of fake data in state bureaucracies.
• Novelty:
  1. Reframes fake data as a relational and processual category rather than a binary failure of representation.
  2. Identifies four critical moments—creation, correction, collusion, and augmentation—through which fake data are produced and negotiated.
  3. Provides a cross-national ethnographic analysis of fake data in two distinct contexts: Chinese volunteering documentation and the 2020 US Census.
  4. Proposes policy and design recommendations for contextual data governance and uncertainty-forward system design.
• Procedure and key techniques:
  • Ethnographic fieldwork in two contexts: Chinese grassroots committees documenting volunteering activities and the US Census Bureau's data production processes.
  • Analysis of 211 semi-structured interviews and observational data to trace how fake data are created, processed, and mobilized.
  • Comparative analysis to identify commonalities and differences in the production of fake data across authoritarian and democratic bureaucracies.

Results

• Concrete findings:
  • Fake data are produced through four interconnected moments: creation (e.g., fabricating or adjusting data to meet quotas), correction (e.g., aligning data with bureaucratic expectations), collusion (e.g., coordinated efforts to maintain systemic coherence), and augmentation (e.g., cumulative distortions through processing).
  • Fake data often serve institutional needs, such as demonstrating compliance with political mandates or achieving statistical plausibility.
  • Bureaucrats prioritize functional validity over representational accuracy, viewing fake data as necessary for system functionality.
• Advantage over baselines:
  • Moves beyond binary frameworks of real vs. fake data by emphasizing the relational and performative aspects of data production.
  • Highlights the institutional logics and power dynamics that normalize fake data, offering a more nuanced understanding than prior problem-oriented approaches.
• Experiments / evaluation:
  • Ethnographic research in two political contexts: 10 months of fieldwork in Chinese grassroots committees and a four-year study of the US Census Bureau.
  • Analysis of data processing practices, interviews with bureaucrats, and public debates about fake data.
• Limitations and future work:
  • The study focuses on two cases, which may not capture all variations in state data practices globally.
  • Future research could explore how public perceptions of fake data influence state legitimacy and governance.
  • Additional work is needed to develop practical tools for managing uncertainty and contextualizing fake data in public-sector systems.

Summary

This paper examines the production and functionality of fake data in state bureaucracies through ethnographic studies of Chinese volunteering documentation and the 2020 US Census. It identifies four critical moments—creation, correction, collusion, and augmentation—through which fake data are produced and negotiated. The authors argue that fake data are relational, processual, and performative, serving institutional needs rather than merely failing at representation. By reframing fake data as a routine feature of state systems, the paper challenges binary frameworks and offers policy and design recommendations for contextual data governance and uncertainty-forward systems. These findings highlight the need to balance functional validity with democratic accountability in state data practices.