Evolutionary game analysis of FinTech transformation: A social co-governance pattern of peer-to-peer lending market in China

Qi Wang, Xin Liu, Chenghu ZhangFrontiers in Psychology 2022 · DOI

This paper, published in Frontiers in Psychology, applies mathematical modeling to a pressing financial crisis: the winding down of the Peer-to-Peer (P2P) lending industry in China. As thousands of platforms face "benign exit" (a regulated shutdown), the authors use evolutionary game theory to determine how platforms, investors, and regulators can cooperate to prevent massive financial losses and social instability.

TL;DR

The researchers developed a tripartite evolutionary game model to find the optimal strategy for the "benign exit" of P2P lending platforms in China. They conclude that a "social co-governance" pattern—where regulators provide incentives, platforms dispose of assets transparently, and investors actively participate—is the only sustainable way to protect investor interests and ensure FinTech stability.

What Problem Does This Paper Address?

Between 2011 and 2022, China’s P2P lending market exploded, reaching over $1.27 trillion in loans. However, a lack of oversight led to rampant fraud, "vicious exits" (where platform owners vanish with the money), and massive investor losses. By 2021, the default rate hit a staggering 44.37%.

The Chinese government mandated a "benign exit" process to shut down these platforms while returning as much capital as possible to investors. The challenge is that traditional top-down regulation is often too slow or rigid. This paper addresses the gap in how to coordinate three conflicting groups:

  1. P2P Platforms: Who want to minimize costs and may be tempted to hide assets.
  2. Investors: Who want their money back but face high costs to organize and supervise platforms.
  3. Financial Regulators: Who want market stability but face high costs for intensive, one-on-one supervision.

Methods

The authors employ Tripartite Evolutionary Game Theory (EGT). Unlike classical game theory, which assumes players are perfectly rational and reach an equilibrium instantly, EGT assumes "bounded rationality." Players learn, adapt, and change their strategies over time based on payoffs.

The model tracks three populations and their strategy choices:

  • Platforms (x): Positive disposal (transparently returning funds) vs. Negative disposal.
  • Investors (y): Participating in co-governance vs. Not participating.
  • Regulators (z): Co-governance regulation (incentive-based) vs. Traditional regulation.

The researchers derived Replicated Dynamic Equations to model the evolution of these strategies and used a Jacobian Matrix to identify Evolutionary Stable Strategies (ESS)—states where no player has an incentive to change their behavior. Finally, they used MATLAB for numerical simulations to test how sensitive the system is to changes in penalties, rewards, and costs.

Datasets

The paper does not use a specific empirical training dataset. Instead, it utilizes:

  • Market Statistics: Historical data on the number of defaulted platforms (6,607) and geographical distribution in China to provide context.
  • Simulated Parameters: For the game analysis, the authors assigned theoretical values to variables (e.g., cost of disposal, reputation benefits, regulatory penalties) to observe how different scenarios (Scenarios 1–4) play out in a controlled mathematical environment.

Key Insights and Ideas

  • The "Win-Win-Win" is Possible: The model identifies four stable states, but only one is optimal: E8 (1, 1, 1). In this state, platforms choose positive disposal, investors participate, and regulators use co-governance.
  • The Threshold Effect: Penalties for "negative disposal" must exceed a specific threshold. If the fine is too low, platforms will mathematically always choose to default or hide assets, regardless of regulatory pressure.
  • Reputation as Capital: For platforms, the "reputation benefit" of a benign exit is a powerful motivator. If a platform exits "positively," the owners can potentially transition into other legitimate FinTech sectors. If they exit "viciously," the loss of reputation (and potential legal blacklisting) acts as a fatal blow to future business.
  • Initial Proportions Matter for Speed: The starting number of people choosing a "good" strategy doesn't change the final stable outcome, but it significantly changes how fast the market reaches stability. This suggests that early government intervention to set a "positive tone" is crucial.

Key Results

  • Optimal Strategy: The co-governance pattern effectively safeguards investor interests if incentive, penalty, and reputation mechanisms are well-designed.
  • Convergence: High initial participation from investors and regulators significantly accelerates the speed at which platforms move toward "positive disposal."
  • Cost Sensitivity: If the cost for investors to participate in supervision (e.g., legal fees, time) is too high, the system collapses into a "vicious exit" cycle where platforms cheat because they know no one is watching.

Limitations and Future Work

  • Simplified Players: The model treats all investors as a monolithic group. In reality, small-scale and large-scale investors have very different risk tolerances and bargaining power.
  • Static Parameters: The simulation assumes costs and rewards remain constant over the exit process, whereas in reality, asset values often depreciate rapidly during a shutdown.
  • Future Directions: The authors suggest incorporating more stakeholders, such as third-party auditors or "RegTech" (Regulatory Technology) tools, into the model to further reduce the costs of supervision. They also recommend calibrating the model with real-world case study data from specific platform closures.

Summary updated

Abstract

Benign exit has become the main theme of the transformation in China's peer-to-peer (P2P) lending industry. To protect the interests of investors in the benign exit process, this paper proposes a social co-governance pattern using a tripartite evolutionary game model to capture the behavior strategies of P2P lending platforms, investors, and financial regulators. The results demonstrate that there are four evolutionary stable strategies for the game model, among which the positive disposal of P2P lending platforms, the participation of the investors, and the co-governance policy of financial regulators is the optimal strategy in the benign exit process. The results also show that the initial proportion of P2P lending platforms, investors, and financial regulators would significantly affect the convergence speed of the evolutionary stable strategy. The proposed social co-governance pattern would effectively safeguard the interests of investors if incentive, penalty, and reputation mechanisms are well-designed. This paper provides in-depth implications for protecting investors' interests in the transformation of the P2P lending industry and enhancing the sustainable development of the FinTech industry.
BibTeX
@inproceedings{wang2022evolutionary,
  author = {Qi Wang and Xin Liu and Chenghu Zhang},
  title = {{Evolutionary game analysis of FinTech transformation: A social co-governance pattern of peer-to-peer lending market in China}},
  booktitle = {Frontiers in Psychology},
  year = {2022},
  doi = {10.3389/fpsyg.2022.954132},
}
Figure