💰 quantitative finance

From Gravity to Confinement: Wealth Redistribution as Optimal Drift Design in the Fokker-Planck Framework

This paper models wealth redistribution as an optimal control problem for the Fokker-Planck equation, demonstrating that while proportional taxes act as a non-distortionary but non-redistributive "gravitational" drift, progressive taxes function as a "confining potential" that effectively reduces inequality by reshaping the wealth distribution's tail, with optimal policy balancing redistribution speed against economic costs in a self-consistent general equilibrium framework.

Original authors: Anders G Frøseth

Published 2026-07-08

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Anders G Frøseth

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the economy's wealth distribution not as a static pile of money, but as a swirling cloud of particles (people) moving around. Some are drifting upward (getting richer), some are drifting downward (getting poorer), and they are constantly bumping into each other (market volatility).

This paper uses a physics tool called the Fokker–Planck equation to model how this cloud moves. Think of this equation as a weather forecast for wealth: it predicts how the "cloud" of rich and poor people will look in the future based on two main forces:

Drift: The average direction people are moving (are they generally getting richer or poorer?).
Diffusion: How much they are jiggling around (volatility and risk).

Here is the paper's story, broken down into simple concepts:

1. The Problem: The "Gravity" Tax Doesn't Work

The author starts by looking at a Proportional Wealth Tax (a flat tax where everyone pays the same percentage, say 2%).

The Analogy: Imagine a uniform gravitational field pulling every particle in the cloud downward at the exact same speed.
The Result: If you pull everyone down by the same amount, the shape of the cloud doesn't change. The rich are still just as far above the poor as they were before; they just all have slightly less money.
The Catch: Because the shape doesn't change, the Gini coefficient (a number measuring inequality) stays exactly the same. The paper proves that a flat tax is "neutral." It doesn't distort investment choices, but it also doesn't fix inequality through the market. It only works if the government takes that tax money and gives it back to the poor (the "fiscal channel"), but the tax itself doesn't change the wealth distribution.

2. The Solution: The "Trap" Tax

To actually change the shape of the cloud and reduce inequality, you need to break that symmetry. The paper suggests a Progressive Wealth Tax (where the tax rate gets higher as you get richer).

The Analogy: Instead of a uniform gravity pulling everyone down equally, imagine a harmonic trap or a giant spring.
- If you are in the middle, the spring is loose.
- If you drift too far to the right (become super rich), the spring pulls you back hard.
- If you are on the left (poor), the spring barely touches you.
The Result: This "confining potential" stops the rich from drifting too far away. Instead of a "Pareto" distribution (where a few people have massive, endless tails of wealth), the cloud becomes a Log-Normal distribution (a bell curve). The "tail" of the super-rich gets chopped off, and the cloud becomes more compact.
Speed: The paper argues this works much faster than waiting for people to die and be replaced by new generations (which is how flat taxes eventually change things). A progressive tax actively compresses the distribution within a political timeframe (like 10–15 years).

3. The "Traffic Jam" of Reality (Leakage)

The paper acknowledges that in the real world, people aren't perfect particles. They try to escape the trap.

Migration: If the tax is too high, the "particles" (rich people) might just leave the country. In physics terms, the wall of the trap becomes a leaky door.
Evasion: People might hide their wealth or restructure their finances to look poorer. This weakens the "spring" force exactly where it needs to be strongest.
The Fix: The paper suggests that policymakers need to design the tax stronger than the ideal math says, to account for these leaks.

4. The Feedback Loop (General Equilibrium)

The paper also looks at what happens when the whole system reacts.

The Logic: If you squeeze the wealth distribution too hard, the total amount of capital in the economy might drop. If capital drops, the return on capital (interest/profits) might go up because money is scarcer.
The Result: This creates a natural limit. You can't just squeeze the distribution infinitely; the economy fights back by making the "drift" (returns) stronger, which partially cancels out your tax efforts. The paper calls this "diminishing returns to progressivity."

5. The Ultimate Goal: Optimal Control

Finally, the paper treats tax design as a control problem (like steering a spaceship).

The Question: "What is the perfect shape of the tax 'spring' to get the wealth cloud to look exactly like we want (e.g., a specific level of equality) without causing too much economic damage?"
The Answer: It provides a mathematical formula to find that perfect balance. It allows a government to say, "We want inequality to be this low, and we are willing to pay this much in economic cost," and the math tells them exactly how progressive the tax should be.

Summary

Flat Tax: Like gravity. Pulls everyone down equally. Doesn't change the gap between rich and poor.
Progressive Tax: Like a spring trap. Pushes the rich back toward the middle. Changes the shape of the wealth distribution and reduces inequality quickly.
The Catch: Rich people might run away (migration) or hide money (evasion), and the economy might react by changing interest rates.
The Takeaway: To fix inequality effectively, you need a tax that acts like a "confining potential" (a progressive trap), not just a flat weight. The paper gives the math to design that trap perfectly.

Technical Summary: From Gravity to Confinement

Problem Statement
The paper addresses the redistributive efficacy of wealth tax design, specifically investigating why proportional wealth taxes often fail to reduce inequality through market channels despite generating fiscal revenue. It posits that the standard economic intuition—that taxing wealth reduces inequality—is mathematically constrained by the dynamics of random multiplicative growth. The core problem is to characterize the specific modifications required in the wealth accumulation process to achieve redistribution within policy-relevant timescales, distinguishing between "neutral" taxes that preserve inequality structures and "confining" taxes that actively reshape the distribution.

Methodology
The study employs the Fokker–Planck equation as the primary analytical framework to model the evolution of the wealth distribution density, $\pi(x, t)$ , where $x$ represents log-wealth. The methodology integrates concepts from statistical physics, stochastic calculus, and optimal control theory:

Baseline Dynamics: The paper establishes a baseline model of geometric Brownian motion for individual wealth, augmented by demographic turnover (birth/death of agents). This setup reproduces the empirically observed Pareto-tailed steady-state distribution.
Symmetry Analysis: It utilizes drift-shift symmetry to demonstrate that a proportional tax acts as a uniform reduction in the drift coefficient ( $v \to v - \tau_w$ ) without altering the diffusion coefficient ( $D$ ).
Spectral Theory: The convergence speed of the wealth distribution to a new steady state is analyzed via the spectral gap of the Fokker–Planck operator. This allows for the calculation of transition half-lives ( $t_{1/2}$ ) for different tax regimes.
Optimal Control Formulation: The redistribution problem is formulated as an optimal control problem. The objective is to find a drift modification $\delta v(x)$ that minimizes the distance between the resulting wealth distribution and a target distribution $\pi^*$ , subject to a regularization penalty $\lambda$ representing the economic and political costs of intervention (distortion, evasion, migration).
General Equilibrium (GE) Extension: The model is extended to a McKean–Vlasov framework where the drift coefficient becomes a functional of the aggregate wealth distribution. This captures the feedback loop where taxation alters aggregate capital, which in turn changes the marginal product of capital and the pre-tax return on wealth.
Empirical Strategy: The paper outlines a strategy for estimating the drift and diffusion functions non-parametrically from panel data using score-based diffusion models and neural stochastic differential equations (SDEs), leveraging Norwegian wealth register data.

Key Contributions and Results

The Redistribution Paradox (Market Channel): The paper proves that a proportional wealth tax preserves the Gini coefficient at all finite times. Because the tax induces a uniform drift shift, it scales all wealth levels proportionally, leaving the relative distribution (and thus the Gini) unchanged. While the steady-state Pareto exponent eventually changes, the transition is governed by the demographic turnover rate ( $\delta$ ), resulting in convergence timescales of decades to centuries—far exceeding typical policy horizons.
Confining Potential (Progressive Tax): The paper demonstrates that to achieve redistribution through the market channel, the drift-shift symmetry must be broken. A progressive tax acts as a confining potential (a state-dependent drift, $v(x) = v_0 - \kappa(x - \bar{x})$ $v (x) = v_{0} - κ (x - \overset{x}{ˉ})$ ), analogous to a harmonic trap in physics.
- Result: This transforms the steady-state distribution from a heavy-tailed Pareto distribution to a log-normal distribution (Gaussian in log-wealth).
- Result: The Gini coefficient becomes a closed-form, monotonically decreasing function of the progressivity parameter $\kappa$ .
- Result: Convergence to the new steady state is driven by the progressivity rate $\kappa$ rather than demographic turnover, allowing for meaningful redistribution within policy-relevant timescales (e.g., 14 years for $\kappa=0.05$ ).
Taxonomy of Interventions: The paper classifies tax designs into five distinct Fokker–Planck modifications:
1. Uniform Drift Shift: Proportional tax (Preserves Gini, slow convergence).
2. State-Dependent Drift: Progressive tax (Log-normal steady state, controllable Gini, fast convergence).
3. Source-Sink Terms: Transfers/UBI (Directly reshapes density).
4. Absorbing Boundary: Wealth cap (Truncates the tail).
5. State-Dependent Diffusion: Volatility policy (Indirectly thins tails).
General Equilibrium Feedback: In the GE extension, the paper finds that increasing progressivity compresses the wealth distribution, reducing aggregate capital $K$ . Due to diminishing returns, this raises the pre-tax return $r$ , which increases the baseline drift $v$ . This creates a diminishing return to progressivity: the GE feedback partially offsets the confining force, imposing a natural economic limit on redistribution independent of political constraints.
Leakage Channels: The analysis incorporates "leakage" via migration (permeable boundaries) and tax evasion (attenuated drift). These channels effectively reduce the realized progressivity parameter, requiring policymakers to over-design the nominal tax rate to achieve the desired effective confinement.

Significance and Claims
The paper claims to provide a unified mathematical language for comparing tax designs, moving beyond qualitative arguments to precise, quantitative predictions regarding distributional outcomes and convergence speeds. Its primary significance lies in:

Clarifying the Mechanism of Redistribution: It rigorously demonstrates that the "neutrality" of proportional taxes (which makes them non-distortionary for portfolio choice) is mathematically identical to their inability to redistribute inequality through the market channel. Redistribution requires breaking this symmetry via a confining potential.
Quantifying Timescales: It explicitly calculates the spectral gap to show why proportional taxes are ineffective on short horizons, whereas progressive taxes can alter inequality within a single electoral cycle.
Optimal Control Framework: It frames tax design as a well-posed optimal control problem for a partial differential equation, allowing for the explicit trade-off between redistributive goals (target Gini) and intervention costs (regularization parameter $\lambda$ ).
Integration of Channels: It unifies the market channel (drift modification) and the fiscal channel (source-sink terms) within a single equation, while also modeling their interaction through general equilibrium feedback loops.

The paper concludes that while the framework is mathematically agnostic regarding the normative choice of a target distribution (e.g., utilitarian vs. Rawlsian), it provides the necessary machinery to translate any such normative position into a concrete policy design with explicit costs and timescales.