Centrality-based Capital Allocations

IMF Working Papers, December 24, 2014

Source details

Canonical URL

Centrality-based Capital Allocations

Other formats

• Markdown version
• Structured JSON version
• Bundle manifest

Bibliographic details

• Authors: Adrian Alter, Ben Craig, Peter Raupach
• Published: December 24, 2014
• Series: IMF Working Papers
• DOI: https://doi.org/10.5089/9781498315548.001

Summary and research question

• Objective: Analyze the effect of capital rules on a banking system connected through correlated credit exposures and interbank lending.
• Approach: Design capital rules that combine individual bank characteristics and interconnectivity measures of interbank lending to minimize a measure of system‑wide losses.
• Data: Detailed German Credit Register used for estimation.
• Authors and date: Adrian Alter, Ben Craig, Peter Raupach, December 24, 2014.

Key findings

• Capital rules based on eigenvectors dominate any other centrality measure evaluated.
• Closeness centrality is the second‑best performing centrality measure.
• Compared to the baseline case, capital reallocation based on the Adjacency Eigenvector saves about 15% in system losses as measured by expected bankruptcy costs.

Methodology and model features

• The analysis models a banking system with:
  • Correlated credit exposures across banks.
  • Interbank lending links that generate interconnectivity.
• Capital allocation rules incorporate:
  • Individual bank characteristics (e.g., bank PDs, bank equity holder implications).
  • Network centrality measures derived from interbank exposure matrices (including eigenvector and closeness measures).
• Objective function: Minimize a system‑wide loss metric (expected bankruptcy costs).

Policy implications and recommendations

• Incorporating network centrality, particularly eigenvector centrality from the interbank adjacency matrix, into capital allocation rules can materially reduce system losses.
• Regulatory capital reallocation that weights banks by their position in the interbank exposure network (Adjacency Eigenvector) is a superior policy lever relative to allocations that ignore network centrality or rely on weaker centrality measures.
• Policymakers should consider network‑aware capital requirements as a tool to mitigate systemic risk arising from interconnected credit exposures and interbank markets.

Content in this bundle

• _wp14237 - Section VI. we present our main results, and we make some final remarks in Section VII..
  • _wp14237 - Section VI. we present our main results, and we make some final remarks in Section VII.. (Markdown version)
  • _wp14237 - Section VI. we present our main results, and we make some final remarks in Section VII.. (PDF)