Stochastic Maps, Wealth Distribution in Random Asset Exchange Models and the Marginal Utility of Relative Wealth

Sitabhra Sinha

DOI: https://doi.org/10.1238/Physica.Topical.106a00059

2003-04-15

Abstract:We look at how asset exchange models can be mapped to random iterated function systems (IFS) giving new insights into the dynamics of wealth accumulation in such models. In particular, we focus on the "yard-sale" (winner gets a random fraction of the poorer players wealth) and the "theft-and-fraud" (winner gets a random fraction of the loser's wealth) asset exchange models. Several special cases including 2-player and 3-player versions of these `games' allow us to connect the results with observed features in real economies, e.g., lock-in (positive feedback), etc. We then implement the realistic notion that a richer agent is less likely to be aggressive when bargaining over a small amount with a poorer player. When this simple feature is added to the yard-sale model, in addition to the accumulation of the total wealth by a single agent ("condensation"), we can see exponential and power-law distributions of wealth. Simulation results suggest that the power-law distribution occurs at the cross-over of the system from exponential phase to the condensate phase.

Statistical Mechanics,Adaptation and Self-Organizing Systems,General Finance

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to explore the wealth distribution mechanism in asset - exchange models, especially how to understand the wealth - accumulation dynamics in these models by introducing stochastic dynamical systems (such as stochastic iterated function systems, IFS). Specifically, the author focuses on the "yard - sale" model (the winner obtains a random part of the wealth of the poorer player) and the "theft - and - fraud" model (the winner obtains a random part of the wealth of the loser), and tries to explain why these models exhibit different wealth - distribution characteristics under different conditions.

### Summary of main problems:
1. **Understanding of wealth - distribution mechanism**:
- Researchers hope to gain new insights into wealth - accumulation dynamics by mapping asset - exchange models to stochastic iterated function systems (IFS).
- In particular, study the behavior of the "yard - sale" model and the "theft - and - fraud" model under different conditions.

2. **Simulation of real - economic phenomena**:
- The author attempts to connect the observed real - economic phenomena through these models, such as positive feedback ("lock - in" effect).
- They also explore how introducing more realistic assumptions, such as wealthy agents may be less aggressive when bargaining with poorer players, affects wealth distribution.

3. **Diversity of wealth distribution**:
- Researchers hope to understand whether these models can reproduce the wealth - distribution characteristics observed in actual society, such as exponential distribution and power - law distribution.
- By introducing the principle of diminishing marginal utility of relative wealth, researchers hope to explain the phenomenon of power - law distribution in the transition area from exponential distribution to the condensed phase (all wealth is concentrated in the hands of a few people).

### Formula representation:
- **Wealth - change formula**:
\[
x_i(t + 1)=x_i(t)+\Delta x; \quad x_j(t + 1)=x_j(t)-\Delta x
\]
where \(\Delta x\) is the net wealth exchanged between two agents.

- **Yard - Sale model**:
\[
\Delta x=\alpha \min(x_i(t), x_j(t))
\]

- **Theft - and - Fraud model**:
\[
\Delta x=\alpha x_j(t) \quad (\text{when agent } j \text{ loses})
\]

- **Fermi function for probability calculation**:
\[
p(i|i, j)=\frac{1}{1+\exp(\beta \left( \frac{x_i}{x_j}-1 \right))}
\]
where \(\beta\) parameterizes the importance of the relative wealth of both trading parties.

Through the exploration of these problems, the author aims to provide a new perspective for understanding the complex dynamics of wealth distribution and a theoretical basis for future research.