Paper: arXiv 2504.20488
Authors: Hernán Larralde, Roberto Mota Navarro
Abstract
We show that assuming that the returns are independent when conditioned on the value of their variance (volatility), which itself varies in time randomly, then the distribution of returns is well described by the statistics of the sum of conditionally independent random variables. In particular, we show that the distribution of returns can be cast in a simple scaling form, and that its functional form is directly related to the distribution of the volatilities. This approach explains the presence of power-law tails in the returns as a direct consequence of the presence of a power law tail in the distribution of volatilities. It also provides the form of the distribution of Bitcoin returns, which behaves as a stretched exponential, as a consequence of the fact that the Bitcoin volatilities distribution is also closely described by a stretched exponential. We test our predictions with data from the S&P 500 index, Apple and Paramount stocks; and Bitcoin.
Complexity vs Empirical Score
- Math Complexity: 8.5/10
- Empirical Rigor: 7.0/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: The paper employs advanced statistical physics and probability theory to model returns as conditionally independent sums, involving heavy integral transforms and scaling laws. It tests these predictions on real-world data (S&P 500, Apple, Paramount, Bitcoin) with systematic fitting and data collapse analysis, though it lacks executable backtests or code.
Research Flowchart
flowchart TD
A["Research Goal: Model financial return distributions<br>using conditionally independent random variables"] --> B["Methodology: Assumption of Conditional Independence<br>returns are independent given time-varying volatility"]
B --> C["Computational Process: Model Sum of Conditionally<br>Independent Variables & Derive Scaling Form"]
C --> D{"Data Input: Empirical Asset Returns"}
D --> E["Asset 1: S&P 500 Index"]
D --> F["Asset 2: Apple & Paramount Stocks"]
D --> G["Asset 3: Bitcoin"]
E --> H["Key Findings & Outcomes"]
F --> H
G --> H
subgraph H ["Outcomes"]
I["Power-law tails in equities<br>explained by power-law volatility distribution"]
J["Stretched exponential in Bitcoin<br>explained by stretched exponential volatility distribution"]
K["General scaling form derived<br>for return distributions"]
end
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes a scaling form for return distributions based on conditional independence given volatility, linking the return shape to the volatility distribution.
Source sentence
We show that assuming that the returns are independent when conditioned on the value of their variance (volatility), which itself varies in time randomly, then the distribution of returns
Assumes 4
- Returns are independent when conditioned on the volatility value. model
Source sentence
These individual returns are considered to be independent random variables when conditioned on the value of the volatility that describes the fluctuations during the period of time over which they are sampled.
- Volatility is essentially constant over short time scales. model
Source sentence
that describes the fluctuations of the individual returns rτ(t) over the n time steps to yield the compound return rnτ(t) is essentially constant.
- Mean return conditional on volatility is zero. statistical
Source sentence
if the mean ⟨rτ(t)|σ⟩= 0, the linear auto correlation function is zero for all lags greater than zero
- Linear autocorrelation of returns is zero for lags > 0. statistical
Source sentence
the linear auto correlation function is zero for all lags greater than zero, in agreement with the stylized facts.
Data
- S&P 500 index market-prices public
Source sentence
We test our predictions with data from the S&P 500 index, Apple and Paramount stocks; and Bitcoin.
- Apple stock market-prices public
Source sentence
We test our predictions with data from the S&P 500 index, Apple and Paramount stocks; and Bitcoin.
- Paramount stock market-prices public
Source sentence
We test our predictions with data from the S&P 500 index, Apple and Paramount stocks; and Bitcoin.
- Bitcoin crypto-onchain public
Source sentence
We test our predictions with data from the S&P 500 index, Apple and Paramount stocks; and Bitcoin.
Costs
Not applicable
Finds 5
- Return distributions over different time intervals obey a simple scaling law derived from conditional independence. theoretical
Source sentence
we show that the distributions of returns over different time intervals obey a simple scaling law.
- Power‑law tails in returns arise from power‑law tails in the volatility distribution. theoretical
Source sentence
the heavy (power law) tails of the return distribution, when they occur, are essentially a consequence of the volatility distribution.
- Bitcoin returns follow a stretched exponential shape because Bitcoin volatility follows a stretched exponential distribution. theoretical
Source sentence
It also provides the form of the distribution of Bitcoin returns, which behaves as a stretched exponential, as a consequence of the fact that the Bitcoin volatilities distribution is also closely described by a stretched exponential.
- The model loses accuracy below 5 minutes for stocks and below 90 minutes for Bitcoin due to intraday autocorrelations. negative
Source sentence
For the case of stocks (S&P 500, AAPL and PARA), this means going below 5 minutes. For the case of BTC, the loss of accuracy starts below 90 minutes.
- Outliers in Bitcoin volatility are rare and have negligible probability mass, so the stretched exponential fit remains a good model for the bulk of the tail. statistical
Source sentence
the outliers represent a very small portion of the events, accumulating a minuscule probability mass in relation to the probability mass accrued in the exponential part.
Checked by others
Supported by
- Matrix H-theory approach to stock market fluctuations
Compare the findings
This paper: Power‑law tails in returns arise from power‑law tails in the volatility distribution.
That paper: Empirical aggregated returns exhibit heavy tails deviating from a Gaussian, while short‑term (10‑day) returns appear Gaussian.
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.
Cited References (21)
- Exact extreme, order, and sum statistics in a class of strongly correlated systems (2024)
- Scaling of the distribution of price fluctuations of individual companies (1999)
- Empirical properties of asset returns: stylized facts and statistical issues (2001)
- Econophysics review: I. Empirical facts (2011)
- Power-law distributions in empirical data (2009)
- powerlaw: a Python package for analysis of heavy-tailed distributions (2014)
- Statistical properties of the volatility of price fluctuations (1999)
- Scaling behaviour in the dynamics of an economic index (1995)
and 13 more
- Empirical properties of volume dynamics in the limit order book (2025)
- Power laws in economics and finance (2009)
- Inverse cubic law for the distribution of stock price variations (1998)
- Scaling of the distribution of fluctuations of financial market indices (1999)
- What really causes large price changes? (2004)
- Stylized Facts of High-Frequency Bitcoin Time Series — → 2402.11930
- Modeling and simulation of financial returns under non-Gaussian distributions — → 2302.02769
- Financial return distributions: Past, present, and COVID-19 (2021)
- Multiscale characteristics of the emerging global cryptocurrency market (2021)
- A theory of power-law distributions in financial market fluctuations (2003)
- Self-organization for the stylized facts and finite-size effects in a financial-market model (2009)
- Minimal agent based model for financial markets II: statistical properties of the linear and multiplicative dynamics (2009)
- Time-varying volatility and the power law distribution of stock returns (2018)
2 of 21 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.