Abstract
In interconnected financial systems, the failure of one firm can trigger a domino effect of defaults, exacerbated by liquidation costs. This paper explores "prepayments"—early debt settlement using external assets—as a mechanism to halt default contagion. We demonstrate that while prepayments can maximize social welfare, finding the optimal set of prepayments is NP-complete. We further analyze the strategic landscape using game theory, calculating the Price of Anarchy and Stability, and employ Empirical Game-Theoretic Analysis (EGTA) to evaluate heuristic strategies under local information. Our results show that strategic prepayments are most effective when default costs are high and firms employ solvency-aware heuristics.
Introduction
Financial credit networks are defined by mutual debt obligations. When a firm becomes insolvent, its inability to pay creditors can lead to a "default contagion." This systemic risk is intensified by default costs: the loss of asset value during bankruptcy liquidation. Prepayments offer a strategic intervention where firms pay off debts before the formal clearing process begins. By injecting undiscounted cash into the network, a firm might prevent the bankruptcy of its creditors, potentially receiving that liquidity back through cycles in the debt graph.
Complexity of Optimal Prepayment
A central question for regulators is whether a system-wide optimal set of prepayments can be efficiently determined. We define the problem of maximizing the total assets or total equity in a network with default costs. Through a reduction from the partition problem, we prove that these decision problems are NP-complete. This suggests that even with full information, a central authority faces significant computational hurdles in optimizing systemic stability via prepayments.
Game-Theoretic Analysis
We model prepayments as a strategic game where firms choose which creditors to prepay to maximize their own utility (Total Assets or Equity). The strategic landscape varies significantly based on the utility function. For equity-based utility, a Pure Strategy Nash Equilibrium (PSNE) always exists, and the best equilibrium aligns with the social optimum. However, for asset-based utility, common for insolvent firms, the gap between selfish behavior and social optimality can be substantial.
As shown in Figure 6, the efficiency of the network is highly sensitive to the default cost parameter $\alpha$. When $\alpha$ is low (high default costs), the incentives for prepayment are strongest, but the potential for inefficient selfish outcomes is also at its peak. As $\alpha$ approaches 1, the "penalty" for default vanishes, and the incentive to prepay—and thus the strategic complexity—diminishes.
Empirical Evaluation of Heuristics
In reality, firms lack global knowledge of the credit network. We utilize Empirical Game-Theoretic Analysis (EGTA) to evaluate nine heuristic strategies ($h_1$ to $h_9$) in synthetic networks. These heuristics range from simple "Do Nothing" approaches to sophisticated greedy algorithms with solvency checks.
Our simulations (Figure 5) highlight that $h_9$ (Heuristic Belief with Solvency Check) is a dominant strategy. This heuristic involves estimating the assets of lenders and ensuring the prepayment does not bankrupt the sender. The robustness of "No Operation" in many scenarios underscores the stringent conditions required for a prepayment to be mutually beneficial: the presence of a debt cycle and a significant reduction in contagion costs.
Conclusion
This research provides a foundational analysis of prepayments as a tool for financial stability. While the underlying optimization is NP-complete, strategic prepayments can significantly improve outcomes in high-friction environments. Future work should explore the impact of fractional prepayments and the potential for cooperative coordination among firms in shared debt cycles.
Summary updated