Paper: arXiv 2412.13172

Authors: Victor Olkhov

Abstract

This paper derives the expressions of correlations between prices of two assets, returns of two assets, and price-return correlations of two assets that depend on statistical moments and correlations of the current values, past values, and volumes of their market trades. The usual frequency-based expressions of correlations of time series of prices and returns describe a partial case of our model when all trade volumes and past trade values are constant. Such an assumptions are rather far from market reality, and its use results in excess losses and wrong forecasts. Traders, banks, and funds that perform multi-million market transactions or manage billion-valued portfolios should consider the impact of large trade volumes on market prices and returns. The use of the market-based correlations of prices and returns of two assets is mandatory for them. The development of macroeconomic models and market forecasts like those being created by BlackRock’s Aladdin, JP Morgan, and the U.S. Fed., is impossible without the use of market-based correlations of prices and returns of two assets.

Complexity vs Empirical Score

  • Math Complexity: 7.5/10
  • Empirical Rigor: 2.0/10
  • Quadrant: Lab Rats — high math complexity, limited empirical testing

Why this score: The paper is heavily mathematical, deriving new correlation expressions using statistical moments, market trade equations, and detailed derivations with notation-heavy formulas. However, it lacks any empirical backtesting, data sources, or implementation details, focusing purely on theoretical model development.

Research Flowchart

  flowchart TD
  A["Research Goal<br>Determine correlation expressions<br>incorporating market impact"] --> B["Key Methodology<br>Derive statistical relationships<br>using joint distribution of trade values"]
  
  B --> C["Data Inputs<br>Current & Past Trade Prices<br>Trade Volumes for Assets A & B"]
  
  C --> D["Computational Process<br>Calculate moments & correlations<br>of price/return vectors"]
  
  D --> E["Key Findings<br>1. New correlation formulas<br>2. Classic models = special case<br>3. Essential for large portfolios"]
  
  E --> F["Outcomes<br>Improved forecasts & risk models<br>for institutional traders/banks"]

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 derives market-based expressions of correlations between prices, returns, and price‑return pairs of two assets, showing their dependence on statistical moments and trade volumes and past values.

Source sentence
This paper derives the expressions of correlations between prices of two assets, returns of two assets, and price-return correlations of two assets that depend on statistical moments and correlations of the current values, past values, and volumes of their market trades. Our paper derives market-based expressions of correlations between prices of two assets, returns of two assets, and price-return correlations of

Assumes 5

  • The time interval between trades ε is constant and there are N trades during the averaging interval Δ. model
    Source sentence
    Let us assume that the time interval ε between trades is constant and there were N trades during the averaging interval Δ (1.1): ∆= [𝑡− ∆ 2 ; 𝑡+ ∆ 2] ; 𝑡𝑖∈∆ ; 𝑡𝑖−𝑡𝑖−1 = 𝜀 ; 𝑖= 1,2, . . 𝑁 (1.1) For such a case, the usual expressions of frequency-based average price π(t|1) of asset 1 and average price π(t-β|2) of asset 2 take simple and familiar form (1.2): 𝜋(𝑡|1) = 1 𝑁∑ 𝑝(𝑡𝑖|1) 𝑁 𝑖=1 ;
  • All five prices are adjusted to the current time t. data
    Source sentence
    In this paper, we assume that all 5 prices are adjusted to the current time t.
  • If all trade volumes U(ti|1) are constant, the average price reduces to the familiar frequency‑based form. market
    Source sentence
    One can easily notice that if all trade volumes U(ti|1)=U, i=1,..N, are constant, the average price (1.9) takes its familiar 4 frequency-based form (1.2).
  • If all past trade values C0(ti,α|1) and C0(ti,β|2) are constant during Δ, the market‑based average returns become the frequency‑based averages. market
    Source sentence
    In the case that all past values C0(ti,α|1) (3.7) and C0(ti,β|2) (3.8) of N trades are constant during Δ (1.1), then VaWAR h(t,α|1) (3.9) and h(t,β|2) (3.10) become equal to frequency-based average returns ρ(t,α|1) and ρ(t,β|2) (3.2).
Show 1 more
  • When trade volumes and past values are constant, the derived market‑based correlations collapse to the usual frequency‑based correlations. model
    Source sentence
    If one assumes that all trade volumes and past values of both assets are constant during the averaging interval, all expressions of correlations take the form of usual frequency-based 11 correlations between time series of prices and returns of two assets.

Data

No dataset named in the text.

Costs

Not applicable The paper reports no trading or execution result.

Finds 2

  • The derived market‑based correlation expressions depend on statistical moments and correlations of current and past trade values and volumes, unlike traditional frequency‑based correlations. theoretical
    Source sentence
    This paper derives the expressions of correlations between prices of two assets, returns of two assets, and price-return correlations of two assets that depend on statistical moments and correlations of the current values, past values, and volumes of their market trades. 5. Conclusion We derived the expressions of market-based of correlations between prices and returns of two assets that reveal their complex
  • Assuming constant trade volumes and past values leads to frequency‑based correlation formulas, and neglecting volume randomness can cause large losses for large traders. negative
    Source sentence
    Such an assumptions are rather far from market reality, and its use results in excess losses and wrong forecasts. These non-market approximations could lead to huge losses.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 8 related papers). Coverage is partial, so this is neither replication nor refutation.

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.