Mechanism Design for Investment Regulation under Herding

This paper proposes a regulator-leader-follower trilateral game framework based on optimal control theory to design quantitative investment regulation mechanisms that effectively mitigate herding behavior, minimize regulatory costs, and maximize social welfare.

The Gist

Imagine a bustling financial market as a giant, noisy dance floor. In this room, there are three main characters:

1. The Leader: A cool, rational dancer who knows exactly what moves to make based on the music (the market data). They don't care about what others are doing; they just follow the beat.
2. The Follower: A dancer who is a bit unsure. They could listen to the music, but they are easily distracted. They tend to copy the Leader's moves, even if the Leader is dancing in a way that might not be best for the Follower's own safety. This is Herding.
3. The Regulator: The DJ or the bouncer. Their job isn't to dance, but to make sure the whole party stays safe, fun, and doesn't turn into a chaotic mosh pit (a market crash).

The Problem: The "Copycat" Trap

Sometimes, copying is good. If everyone agrees on a move, the dance floor looks smooth. But often, the Follower copies the Leader blindly.

• The Issue: If the Leader is brave and dances near the edge of the stage (taking high risks), the Follower might copy them and fall off, even though the Follower is actually a very cautious person who should be dancing safely in the middle.
• The Result: The whole room becomes unstable. Prices go up and down wildly, and people lose money.

The Regulator wants to stop this blind copying, but they have a problem: They can't read minds. They don't know exactly how much the Follower likes to copy (their "herd coefficient"). If the Regulator tries to force everyone to stop copying, they might annoy the cautious dancers who want to copy, or waste money trying to fix people who don't need fixing.

The Solution: A Smart "Rule Book" (Mechanism Design)

Instead of just shouting "Stop copying!" (which is vague and often ignored), the Regulator designs a clever incentive system. Think of it like a video game with a specific set of rules to encourage good behavior.

The Regulator uses a three-step strategy:

1. The "Truth-Telling" Trick

Since the Regulator can't see the Follower's true fear level, they ask the Follower to raise their hand and say, "I am a big copier" or "I am a small copier."

• The Catch: The Regulator designs the rules so that it is always in the Follower's best interest to tell the truth. If they lie and say they are less of a copier than they really are, they get a worse deal. This is called Incentive Compatibility.

2. The "Switch" Strategy

The Regulator doesn't treat everyone the same. They use a smart switch based on two things:

• Risk Tolerance: Is the Follower more scared of falling than the Leader?
• Herding Strength: How badly do they want to copy?

The Logic of the Switch:

• Scenario A (The Cautious Follower): If the Follower is naturally very careful (risk-averse) and the Leader is bold, the Follower usually wants to copy the Leader to be bold. Wait, that's bad! The Regulator says, "Okay, you need help." They apply a strong policy (like a gentle nudge or a warning) to stop the Follower from copying the risky Leader.
• Scenario B (The Bold Follower): If the Follower is already brave and the Leader is cautious, the Follower might copy the Leader and become too safe, missing out on good opportunities. The Regulator says, "No need to interfere." They let them be.
• Scenario C (The Cost Check): Fixing the dance floor costs money (time, effort, resources). The Regulator calculates: "Is the Follower copying so much that fixing it will save more money than it costs to fix?" If the copying is weak, the cost to fix it is too high, so they do nothing. If the copying is strong, they step in.

3. The "Lollipop" (Compensation)

Changing someone's behavior is annoying. If the Regulator tells the Follower, "Stop copying that risky guy," the Follower might feel like they are losing out on potential fun (utility).

• To make the Follower agree to the rules, the Regulator offers a reward (like a subsidy or lower fees).
• The reward is calculated precisely: It's just enough to make the Follower feel happy about following the new rules, but not so much that the Regulator goes broke.

The Big Picture: What Happens?

When this system works:

1. The Follower tells the truth about how much they like to copy.
2. The Regulator only steps in when it's truly necessary (when the Follower is too risky and copying too hard) and only when it's worth the cost.
3. The Result: The Follower stops blindly following the Leader into danger. They make their own better decisions. The "dance floor" becomes more stable, and the total amount of money in the market (Social Welfare) goes up.

Summary in One Sentence

The paper proposes a smart, "pay-for-performance" system where a regulator only steps in to stop dangerous copying when it's actually worth the cost, using rewards to convince investors to tell the truth and make better choices for themselves.

Technical Summary

1. Problem Statement

The paper addresses the challenge of regulating herding behavior in financial markets, where investors (followers) imitate the decisions of influential investors (leaders) rather than relying on their own analysis. While moderate herding can aid information aggregation, excessive herding leads to market distortion, asset bubbles, and crashes.

Key Challenges Identified:

• Asymmetric Information: The regulator cannot directly observe an investor's "herd coefficient" (the intensity of their tendency to follow others), which is private information.
• Misaligned Objectives: Individual investors seek to maximize personal utility, while regulators aim to maximize social welfare (aggregate fund) and minimize regulatory costs.
• Limitations of Traditional Tools: Existing tools (e.g., information disclosure, transaction restrictions) often lack theoretical guarantees for effectiveness and are imprecise.

The core problem is to design a quantitative regulatory mechanism that induces followers to truthfully report their herding levels and adjust their investment decisions to improve social welfare, while satisfying Individual Rationality (IR) and Incentive Compatibility (IC) constraints.

2. Methodology

The authors propose a Regulator-Leader-Follower Trilateral Game Framework integrating Optimal Control Theory and Mechanism Design Theory.

A. Market Model

• Assets: One risk-free asset (rate $r$) and one risky asset (excess return $\nu$, volatility $\sigma$).
• Leader: A rational investor maximizing expected utility with an exponential utility function. The leader's optimal decision $\tilde{\pi}_t$ is derived analytically (Merton's solution).
• Follower: An investor who maximizes expected utility but also minimizes the deviation from the leader's decision. The follower's objective includes a "herd coefficient" $\eta$ and a penalty term $\delta(\pi)$ representing the deviation cost.
• Regulator: Designs a mechanism to influence the follower.

B. The Mechanism Design

The regulator implements a mechanism $\gamma = (q, c)$ consisting of:

1. Policy ($q$): A regulatory intervention (e.g., rational recommendations, mandatory disclosures) that reduces the follower's effective herd coefficient by an amount $u(q)$. This incurs a regulatory cost $w(q)$.
2. Compensation ($c$): A transfer (e.g., subsidies, fee reductions) $v(c)$ provided to the follower to ensure they accept the regulation.
3. Reporting: The follower reports a herd coefficient $\hat{\eta}$. The mechanism must ensure Truth-Telling ($\hat{\eta} = \eta$) via the IC constraint.

C. Solution Approach

The problem is modeled as a Stackelberg Game:

1. Follower's Problem (P2): Given a mechanism, the follower chooses an investment strategy $\pi$ and a reporting function $s$ to maximize utility.
2. Regulator's Problem (P1): The regulator chooses the mechanism $\gamma$ to maximize social welfare (expected terminal fund of the follower) minus regulatory costs, subject to IR and IC constraints.
3. Revelation Principle: The search is restricted to "truth-telling mechanisms" where $s^*(\eta) = \eta$.
4. Backward Induction:
   • Solve for the follower's optimal response $\omega(\eta)$ given a mechanism.
   • Derive the IR and IC constraints for the compensation $c(\eta)$.
   • Solve for the optimal policy $q^*(\eta)$ and compensation $c^*(\eta)$.

3. Key Contributions

1. Novel Framework: First integration of optimal control-based investment theory with mechanism design specifically for regulating herding behavior under asymmetric information.
2. Optimal Mechanism Derivation: Theoretically derives the closed-form expressions for the optimal regulatory policy and compensation, proving that the regulator should only intervene under specific conditions.
3. Threshold-Based Intervention: Identifies that regulation is not universally applied but depends on a switch-like structure determined by risk coefficients and a herd-coefficient threshold.
4. Quantitative Trade-off Analysis: Explicitly models the trade-off between regulatory costs and economic gains, showing how cost thresholds dictate the scope of intervention.

4. Key Results

A. Optimal Policy Structure (Switch-like)

The optimal policy $q^*(\eta)$ exhibits a binary, switch-like behavior:

• No Intervention: If the follower is more risk-averse than the leader ($\alpha \geq \tilde{\alpha}$) OR if the follower's herd coefficient is below a critical threshold ($\eta \leq \check{\eta}$). In these cases, herding either aligns with the regulator's goal (increasing risky asset allocation) or the cost of intervention outweighs the benefit.
• Maximum Intervention: If the follower is more risk-taking than the leader ($\alpha < \tilde{\alpha}$) AND the herd coefficient is high ($\eta > \check{\eta}$). Here, herding causes the follower to under-allocate to the risky asset (relative to the social optimum), so the regulator imposes the strongest policy to reduce $\eta$.

B. The Threshold $\check{\eta}$

The threshold $\check{\eta}$ is defined by the inverse of the economic gain function: $\check{\eta} = g^{-1}(\kappa)$.

• It represents the minimum herd intensity required for the economic gain (increase in social welfare) to exceed the constant regulatory cost $\kappa$.
• As regulatory costs increase, the threshold $\check{\eta}$ increases, meaning the regulator becomes more selective, intervening only in cases of severe herding.

C. Optimal Compensation

The optimal compensation $c^*(\eta)$ consists of two parts:

1. Constant Component ($\chi$): Ensures the Individual Rationality (IR) constraint (the follower is better off or indifferent compared to no regulation).
2. Variable Component: Increases with the herd coefficient $\eta$ to satisfy the Incentive Compatibility (IC) constraint. It offsets the utility loss the follower experiences due to the reduction in their herding tendency.

• Compensation is only required when the policy is active (i.e., when $\alpha < \tilde{\alpha}$ and $\eta > \check{\eta}$).

D. Impact on Social Welfare

• Economic Gain ($g(\eta)$): The improvement in social welfare is zero if no intervention occurs.
• Positive Gain: When intervention occurs, $g(\eta) \geq 0$. Crucially, the gain is an increasing function of the herd coefficient $\eta$. This implies that the more severe the herding problem, the greater the benefit of the regulatory mechanism.

5. Significance

• Theoretical Rigor: Provides a mathematically rigorous foundation for financial regulation, moving beyond heuristic rules to optimal control solutions with guaranteed constraints (IR/IC).
• Policy Efficiency: Demonstrates that "one-size-fits-all" regulation is suboptimal. Regulators should adopt a targeted approach, intervening only when investors are risk-taking and herding is strong, thereby saving regulatory resources.
• Practical Application: Offers a blueprint for designing financial policies (e.g., disclosure mandates, advisory systems) that account for investor psychology (herding) and private information, ensuring that regulations are both effective and accepted by market participants.
• Cost-Benefit Clarity: Explicitly quantifies the relationship between regulatory costs and the intensity of herding, helping policymakers determine when intervention is economically justified.