Principal Investigator


Project Goals

The 2008 financial crisis demonstrated that complex and opaque networks of credit relationships among firms can set the stage for sudden and unexpected propagation of financial uncertainty throughout the economy. The motivating goal of this project is to develop new models of dynamic credit networks to provide a basis for evaluating systemic risks of this kind, and for designing institutions and policies that improve robustness to asset price fluctuations and economic shocks of various kinds. The models are based on credit networks, a class of graph-based trust accounting mechanisms, developed in recent years by computer scientists and economists. Prior work in this research group studied strategic issues in credit network formation. The focus in this project is extending the formalism to represent financial credit relationships, and to analyze dynamic network structure and performance over time, as firms modify their credit relationships in response to circumstances and economic conditions.

The investigation takes an agent-based approach, which means that the key actors (banks, non-bank financial firms, and non-financial enterprises) are instantiated by computational objects executing strategies aimed to optimize objectives (profit) given available information. Modeling credit networks as multiagent systems affords heterogeneity: actors differing in objectives, information (access to data and observability of the environment), and  capabilities (financial and technological resources). Evidence about systemic properties is derived by simulation, as models incorporating such heterogeneity and moreover accommodating complex information and fine-grained dynamics are analytically intractable. Since there are infinite combinations of agent behaviors one might consider—leading to different conclusions—it is necessary to adopt principled criteria for selecting salient agent strategies. This project employs empirical game-theoretic analysis methods for this purpose.


The first paper from this project, Strategic Payment Routing in Financial Credit Networks, was presented by Frank Cheng at the ACM Conference on Economics and Computation, Maastricht, in July 2016. The key contribution of that paper is the extension of credit networks to incorporate interest rates, producing a representation we call financial credit networks. Supporting interest rates requires several conceptual extensions, including maintaining a distinction between credit lines and debt holdings, and taking account of relative rates in routing payments. The paper presents an algorithm for determining the maximum payment flow between a pair of nodes that respects interest rate monotonicity along paths. We investigate the liquidity properties of financial credit networks, and explore tradeoffs between alternative payment mechanisms (i.e., rules for choosing payment paths) and liquidity. Finally, we consider the strategic implications of allowing agents to choose payment mechanisms on social welfare.

In ongoing work, we are applying financial credit networks for modeling systemic risk. Our focus is on replicating and extending prior models of the implications of alternative mechanisms for regulation of capital.

Broader Impact

This research extends our ability to reason about complex credit relationships, which are arguably at the center of the calamitous financial crisis of 2008. Such a reasoning capacity can lead to new tools for risk management, supporting applications within individual financial firms as well as for central banks and other economic regulators.

The project has supported two graduate student researchers to date (listed above).

Related Publications

Related Presentations


This project is supported by the National Science Foundation and US Treasury Office of Financial Research under grant number IIS-1440360 from the CIFRAM program. (September 2014 through August 2017)

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.