📈 economics

By Law, Every Zero-Mean Risk Is the Difference of Two Equally Distributed Risks

The paper proves that any probability distribution on the real line with a mean of zero can be represented as the distribution of the difference between two independent and identically distributed random variables.

Original authors: Mark Whitmeyer

Published 2026-07-08

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Mark Whitmeyer

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 you have a scale that is perfectly balanced. On one side, you have a pile of "good" outcomes (positive numbers), and on the other, a pile of "bad" outcomes (negative numbers). If the total weight of the good stuff exactly cancels out the total weight of the bad stuff, the scale is at zero mean. In the world of probability, this is called a "zero-mean risk."

Mark Whitmeyer's paper is essentially a mathematical proof that says: No matter how complicated or weird that balanced scale looks, you can always build it by taking two identical, fair dice rolls and subtracting one from the other.

Here is the breakdown of the paper's logic using simple analogies:

1. The Big Claim: The "Subtraction Trick"

The paper claims that any random situation where the average result is zero can be created by a simple formula: $X - Y$ .

$X$ is a random number (like a dice roll).
$Y$ is another random number.
The Catch: $X$ and $Y$ must be "identically distributed." This means they come from the exact same "bag of possibilities." If you roll $X$ a thousand times, the pattern of results looks exactly the same as if you rolled $Y$ a thousand times.

Whitmeyer proves that you don't need two different, complex rules to create a balanced risk. You just need one rule applied twice, and then you subtract the second result from the first.

2. The "Smoothie" Analogy (The Continuous Case)

The paper focuses on a specific, tricky scenario where the risks are "nonatomic." Think of this like a smoothie made of infinite tiny droplets, rather than a salad made of distinct chunks of fruit.

The Problem: You have a smoothie where the sweet flavors (positive numbers) and sour flavors (negative numbers) perfectly cancel each other out.
The Solution: Whitmeyer shows you how to construct two identical glasses of liquid ( $X$ and $Y$ ) such that when you pour one into the other and subtract, you get your original smoothie.
The Mechanism: The paper uses a clever "rotation" trick. Imagine a circle. The author takes a drop of sweet liquid and a drop of sour liquid and spins them around the circle in a specific way to pair them up. By doing this for every possible drop, they create a perfect pairing where the "good" and "bad" cancel out exactly, leaving you with the original zero-mean result.

3. The "Building Blocks" (The Finite Case)

The paper also mentions a simpler case where the risks aren't a smoothie but a stack of distinct blocks (finite or uniform).

Here, the solution is like a circular puzzle. If you have a set of blocks that balance out, you can arrange them in a circle and take steps around the circle. The difference between where you start and where you end up (after a full cycle) creates the balanced risk. This part of the proof was already known by other researchers (Maccheroni et al.), and Whitmeyer confirms it fits into his broader theory.

4. Why This Matters (In the Paper's Context)

The paper doesn't talk about stock markets, insurance, or medical treatments. It is a pure math proof.

The Goal: To prove a specific lemma (a small helper theorem) used in a 2025 economics paper by Maccheroni and colleagues.
The Result: It confirms that the mathematical tools those economists were using are solid. It proves that their assumption—that any balanced risk can be split into two identical parts—is always true, whether the risk is a smooth, continuous curve or a set of distinct steps.

Summary

Think of this paper as a master locksmith proving that every locked door (a complex zero-mean risk) can be opened with a single, specific type of key (the difference of two identical variables).

Whitmeyer didn't just say "it works"; he showed the exact blueprint for how to cut that key, even for the most complex, smooth, and continuous locks. He proved that nature (or mathematics) never creates a perfectly balanced risk that cannot be explained as the difference between two identical, fair chances.

Technical Summary: "By Law, Every Zero-Mean Risk is the Difference of Two Equally Distributed Risks"

Problem Statement
The paper addresses a fundamental question in probability theory and risk analysis: whether every probability law on $\mathbb{R}$ with a zero mean can be represented as the distribution of the difference between two independent (or dependent) random variables that share the same marginal distribution. Specifically, given a mean-zero law $\mu$ , the objective is to construct identically distributed random variables $X$ and $Y$ such that the law of their difference, $X - Y$ , equals $\mu$ . This problem arises within the context of the "adequacy assumption" maintained in Maccheroni et al. (2025) regarding risk aversion and insurance propensity.

Methodology
The author, Mark Whitmeyer, provides a constructive proof that resolves the problem for the specific cases required by the adequacy assumption in Maccheroni et al. (2025). The methodology is bifurcated based on the nature of the probability measure $P$ :

The Finite-Uniform Case:
For measures satisfying the finite-uniform condition, the paper relies on the "cyclic partial-sum construction" already provided by Maccheroni et al. (2025). No new construction is required for this subset.
The Nonatomic Case:
For a nonatomic probability measure $P$ with mean zero, the author constructs a joint distribution $\nu$ on the product space $S = \mathbb{R}^2$ from which $X$ and $Y$ are derived. The construction proceeds as follows:
- Decomposition: The measure $\mu$ is decomposed into an atom at zero ( $p_0 = \mu(\{0\})$ ), a positive part $\mu_+ = \mu|_{(0, \infty)}$ , and a transformed negative part $\mu_- = (-\text{id})_\#(\mu|_{(-\infty, 0)})$ .
- Normalization: A scaling factor $a$ is defined as the first moment of the positive part: $a = \int_0^\infty x \mu_+(dx) = \int_0^\infty y \mu_-(dy)$ .
- Joint Measure Construction:
  - If $a = 0$ , the joint measure is simply $\nu = \delta_{(0,0)}$ , implying $X=Y=0$ almost surely.
  - If $a > 0$ , the joint measure $\nu$ is defined as a mixture of the atom at the origin and a continuous component. The continuous component is constructed by integrating over $(0, \infty)^2$ with respect to $\mu_+(dx)\mu_-(dy)$ . For each pair $(x, y)$ , a rotation $R_{x,y}$ is applied on the interval $[-y, x)$ , and the measure is averaged over this interval.
- Marginal Verification: The author demonstrates that under this construction, the marginals of $\nu$ are identical ( $X \stackrel{d}{=} Y$ ). Furthermore, the distribution of the difference $X - Y$ is shown to recover the original measure $\mu$ through a calculation involving the pushforward of the constructed measure.

Key Results
The primary result is a rigorous proof that for any mean-zero law $\mu$ on $\mathbb{R}$ , there exist identically distributed random variables $X$ and $Y$ such that $X - Y \stackrel{d}{=} \mu$ . The paper specifically validates the "nonatomic case" of Lemma 1 in Maccheroni et al. (2025).

Additionally, the paper establishes the integrability properties of the constructed variables:

If the target measure $\mu$ is bounded (i.e., supported on a compact interval), the constructed $X$ and $Y$ are essentially bounded ( $L^\infty$ ).
If the target measure $\mu$ has finite moments of all orders ( $M^\infty$ ), the constructed $X$ and $Y$ also possess finite moments of all orders.

Significance and Claims
The paper claims to provide the missing technical link required to fully substantiate the "adequacy assumption" in the economic framework proposed by Maccheroni et al. (2025). By proving that the decomposition holds for both the finite-uniform and nonatomic cases, the paper asserts that these two cases "exhaust" the assumptions necessary for the economic model in the cited work.

The author modestly frames the contribution as a proof of a specific lemma (Lemma 1) within the referenced economic literature, rather than a broad new theory in probability. The work is dedicated to "KS" and acknowledges the use of ChatGPT for proof-checking. The paper concludes that the construction successfully demonstrates that every zero-mean risk can indeed be represented as the difference of two equally distributed risks, thereby validating the theoretical underpinnings of the referenced economic analysis.