The Reality of Our Wealth Distribution

I am not too sure if this is much of a surprise to many these days, but one of the “weirdest” lies that is often taught or said throughout either economic or political debates is the concept of the bell curve representation of wealth distribution. I am not going to sit here and bash that representation, though I’d like to go ahead and make sense of the reality of how wealth is truly distributed and why it concentrates the way it does.

To get this out of the way quickly, we can start with the representation as a bell curve. Most people often think of things like IQ, height, or in this case income, as a bell curve. This is characterized by the center being our mean, with some level of standard deviation determining the width of the curve and the tails representing the least likely outliers. This can be represented by the expression of the Gaussian curve as follows:

$$f(x)=\frac{1}{\sigma\sqrt{2\pi}}e^{-\frac{(x-\mu)^{2}}{2\sigma^{2}}}$$

where $\sigma$ is our standard deviation and $\mu$ is our mean. In the case of income, we know that the mean is simply the average income level. To find the probability of encountering someone within a given wealth bracket, all one has to do is integrate the density function:

$$P(a\le X\le b)=\int_{a}^{b}\frac{1}{\sigma\sqrt{2\pi}}e^{-\frac{(x-\mu)^{2}}{2\sigma^{2}}}dx$$

which resolves to:

$$P(a\le X\le b)=\frac{1}{2}\left[\text{erf}\left(\frac{b-\mu}{\sigma\sqrt{2}}\right)-\text{erf}\left(\frac{a-\mu}{\sigma\sqrt{2}}\right)\right]$$

where the error function is defined as:

$$\text{erf}(z)=\frac{2}{\sqrt{\pi}}\int_{0}^{z}e^{-t^{2}}dt$$

Now, before even evaluating this as a viable way of calculating our distribution, we need to first take into account the structural limitations of using a bell curve to begin with. One of the biggest issues with the Gaussian distribution is that it assumes perfect symmetry around our mean. The Gaussian function requires an identical mirror image going both to the left and the right of the average. What this means is that, mathematically speaking, the model assigns as much weight to negative wealth values as it does to fortunes climbing into the billions.

Furthermore, from a pure probability theory stance, under a Gaussian framework, billionaires shouldn’t exist. This is not a moral statement; it is quite the opposite. What I am saying here is that the mere existence of billionaires proves that the Gaussian distribution is entirely the wrong model to represent economic reality. To see this mathematically, look at how a Gaussian begins to behave as $x\rightarrow\infty$. As we move down the axis, the distribution decays exponentially:

$$e^{-\frac{(x-\mu)^{2}}{2\sigma^{2}}}$$

where the probability density converges to zero:

$$\lim_{x\rightarrow+\infty}\frac{1}{\sigma\sqrt{2\pi}}e^{-\frac{(x-\mu)^{2}}{2\sigma^{2}}}=0$$

This means that if wealth were Gaussian, finding someone with ultra-high net worth would be virtually impossible. By definition, under this curve, it would take an infinite amount of time or an infinite population to produce even one single billionaire. We know that is not the case, and we need to understand why.

Before I get into the power law, I will use a quick analogy to draw out the explicit difference between the Gaussian model and the power law. Imagine an “economy” contained entirely within a hypothetical bar. This bar has an owner and a room full of patrons. The main mechanism of this economy is simple: you earn money by playing a game of coin flipping and betting a fixed amount of cash. If we have a decent number of people in the bar, after a large number of games, the wealth will organize into a standard bell curve. There will be a massive cluster of average players holding the mean amount of money, a few people who hit bad streaks and have literally nothing, and a few lucky standouts with a lot.

Now imagine that the bar owner charges a flat entry fee for every game played. While his income is driven by the volume of economic activity occurring in the room, he is insulated from the random coin-flip outcomes determining the players’ fates. His wealth accumulation behaves differently; he starts to function as a centralized, connective hub for the network.

To make this toy model fit reality a bit better, let’s introduce two critical modifications. Imagine now that the players make bets proportional to their current wealth, and the owner charges a percentage-based fee on those bets. Instead of everyone experiencing linear, arithmetic growth, a handful of early winners will see their earnings compound in a geometric, multiplicative manner proportional to their current asset state. Early winners can instantly bet and win more as the games go on. Concurrently, the owner of the bar makes more money from every single transaction, extracting exponential yields from the wealthiest players.

The macroeconomic shift from a Gaussian distribution to a Pareto distribution is fundamentally a phase transition from linear, arithmetic accumulation to geometric, multiplicative feedback loops. By tying bet sizes and network fees directly to the current state-space value of the assets, the system eliminates the characteristic scale of the mean ($\mu$), giving rise to scale-invariant power-law dynamics where variance is no longer bounded.

This structural divergence highlights a fundamental phase transition in the network topology. In a basic economy, resource accumulation is strictly arithmetic and localized, giving rise to a standard, bounded Gaussian curve. However, when asset-backed leverage and percentage-based transaction fees are introduced, the system eliminates the characteristic scale of the mean ($\mu$).

Instead, the macro-economy splits into two distinct thermodynamic regimes divided by a critical threshold: $x_{\text{min}}$. Below $x_{\text{min}}$, capital dynamics remain bounded by linear constraints. But once an agent crosses this minimum wealth boundary, they enter a scale-invariant realm governed by geometric, multiplicative feedback loops. Past this phase boundary, the Gaussian model collapses entirely, and the distribution settles into a scale-free power law defined by the Pareto distribution:

$$f(x)=Cx^{-\alpha}$$

Integrating this probability density function from our lower boundary condition $x_{\text{min}}$ yields the complementary cumulative distribution function (CCDF), which maps the probability of encountering an individual whose wealth exceeds a specific level $x$:

$$P(X>x)=\left(\frac{x_{\text{min}}}{x}\right)^{\alpha-1}$$

Here, $x_{\text{min}}$ acts as the physical origin point where scale-invariant behavior initiates, while the Pareto index $\alpha$ dictates the systemic elasticity of the distribution. In physics terms, $\alpha$ represents the structural decay rate of the network; a lower alpha value signifies weaker resistance to resource concentration, allowing capital to cascade exponentially toward the absolute upper boundaries of the state-space.

While the unconstrained Pareto model allows for infinite variance, any real-world economic ecosystem possesses a hard saturation threshold. Because the central hub (the bar owner) continually siphons liquidity via percentage-based transaction friction, the total active wealth pool of the agents exponentially decays. This systemic depletion creates an upper boundary condition where transaction volume collapses, forcing the owner’s revenue growth curve to break away from exponential scaling and flatten into a resource-constrained asymptote. The following serves as a simple visualization of the analogy used for our bar “economy”.

One thing you will notice visually is the random nature of the players’ wealth and how some actually are able to obtain a rather large amount of wealth. You can also see the owner being able to be detached from the random behavior of the game itself.

For Those Who Care

If you’re wondering how the simulation engine above actually generates these geometry shifts under the hood, I decided to use a two-phase Markov chain scaling structure written in Python:

1. Phase 1 (Rounds 1 to 500): The system runs an arithmetic random walk. Every round, agents are randomly paired up to flip a coin with a flat bet of 1.50 and a flat entry fee of 0.10. Because the stakes are fixed and linear, the system preserves a characteristic scale. This is why Panel 3 retains its tight, bounded Gaussian distribution around the initial starter wealth pool ($100).
2. Phase 2 (Rounds 501 to 6000): We trigger a phase transition by shifting the rules from linear to multiplicative. Bets are now calculated as a percentage of the poorest paired agent’s net worth (5%), and the bar owner takes a percentage-based cut (1%).

The Math Mechanics of Panel 3

The far-right graph utilizes a logarithmic binning function:

$$\text{Bin Centers} = \gamma_i, \quad \text{Counts} = \log_{10}(N_i)$$

When you scrub the timeline slider to Round 6000, you are watching the probability density profile flatten out into a straight line on the log-log scale. In complexity statistics, a straight line on a log-log plot is the unmistakable, mathematical fingerprint of a scale-invariant power law ($f(x) = C x^{-\alpha}$). The slope of that declining line is your Pareto index ($\alpha$), which is the structural rate at which wealth violently concentrates into the central hub while the rest of the room’s liquidity faces exponential depletion.

Why the Simulation is Split into Two Phases

To genuinely isolate and demonstrate the exact catalyst behind wealth concentration, the script intentionally establishes an initial experimental control phase.

By running an arithmetic random walk for the first 500 rounds, we prove that linear transactions natively preserve a stable, bounded Gaussian scale. The sudden rule-shift at Round 501 doesn’t just display a power law; it captures a live thermodynamic phase transition. It forces the reader to watch the Gaussian boundary conditions violently fracture and realign into a scale-free Pareto distribution the exact millisecond the network symmetry is broken by geometric feedback loops.

Before finishing this off, I would like to quickly talk about our next topic: the utility of the bar owner in this ecosystem. While we are often told that this level of wealth concentration is driven purely by “greed,” it is actually extremely useful to look at the physics of the system that drives this outcome, and why, for a general market, it often leads to a more efficient use of both wealth and more importantly, information.