Paper: arXiv 2404.07222

Authors: Qi Deng, Zhong-guo Zhou

Abstract

We propose that the liquidity of an asset includes two components: liquidity jump and liquidity diffusion. We show that liquidity diffusion has a higher correlation with crypto wash trading than liquidity jump and demonstrate that treatment on wash trading significantly reduces the level of liquidity diffusion, but only marginally reduces that of liquidity jump. We confirm that the autoregressive models are highly effective in modeling the liquidity-adjusted return with and without the treatment on wash trading. We argue that treatment on wash trading is unnecessary in modeling established crypto assets that trade in unregulated but mainstream exchanges.

Complexity vs Empirical Score

  • Math Complexity: 4.5/10
  • Empirical Rigor: 6.0/10 (capped from 8.0: transaction costs not modelled; see penalty flags)
  • Quadrant: Street Traders — empirical and applied, lower math complexity (effective empirical rigor capped from 8.0)

Why this score: The paper uses established econometric models (ARMA-GARCH) and standard portfolio optimization (Mean-Variance) with relatively accessible mathematics, but it demonstrates strong empirical rigor through detailed treatment of high-frequency crypto data, backtesting portfolio strategies (LAMV vs. TMV), and quantitative analysis of wash trading’s impact on liquidity components.

Research Flowchart

  flowchart TD
  A["Research Goal: Decompose Liquidity & Model<br>Impact of Wash Trading Treatment"] --> B["Key Methodology: Proposed Liquidity Decomposition<br>into Jump & Diffusion Components"]
  B --> C["Data: Crypto Asset Trades & Liquidity Measures"]
  C --> D{"Computational Processes"}
  D --> E["Analyze Correlation with Wash Trading"]
  D --> F["Apply Treatment on Wash Trading"]
  D --> G["Fit Autoregressive Models on Returns"]
  E --> H["Key Findings: Diffusion correlates more<br>than Jump with Wash Trading"]
  F --> I["Treatment significantly reduces Diffusion,<br>only marginally reduces Jump"]
  G --> J["Autoregressive models remain highly effective<br>post-treatment on mainstream exchanges"]
  H --> K["Conclusion: Treatment unnecessary for<br>established crypto assets on mainstream exchanges"]
  I --> K
  J --> K

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 that asset liquidity can be split into two components, liquidity jump and liquidity diffusion.

Source sentence
We propose that the liquidity of an asset includes two components: liquidity jump and liquidity diffusion.

Assumes

No explicit assumptions found in the text.

Data

  • Binance tick-level trading data · Binance API order-book public
    April 27, 2019 to February 8, 2024 · 10 largest crypto assets by market value (ADA, BNB, BTC, ETC, ETH, LINK, LTC, MATIC, XMR,
    Source sentence
    We first collect tick-level trading data of these assets from the Binance API and aggregate the tick-level data to construct minute-level (intraday) amount.

Recognised sources: Binance

Costs

Not modeled A trading or execution result is reported without costs.

Finds 5

  • Liquidity diffusion correlates more strongly with crypto wash trading than liquidity jump. statistical
    Source sentence
    We show that liquidity diffusion has a higher correlation with crypto wash trading than liquidity jump
  • Treatment on wash trading markedly reduces liquidity diffusion but only marginally reduces liquidity jump. statistical
    Source sentence
    treatment on wash trading significantly reduces the level of liquidity diffusion, but only marginally reduces the level of liquidity jump
  • Autoregressive models (ARMA‑GARCH/EGARCH) are highly effective for modeling liquidity‑adjusted returns, with or without wash‑trading treatment. predictive-accuracy
    Source sentence
    We confirm that the autoregressive models are highly effective in modeling the liquidity-adjusted return with and without the treatment on wash trading
  • Liquidity‑adjusted Mean‑Variance (LAMV) portfolios outperform traditional MV portfolios; SR improves from 0.96 to 1.45 with treatment and from 1.41 to 1.81 without treatment after forecast enhancement. performance
    Source sentence
    With treatment on wash trading, the SR of the standard LAMV Portfolio 8 from 0.96 is greatly improved to 1.45 for the forecast enhanced LAMV Portfolio 11. Even without treatment, the SR has also greatly improved from 1.41 of the standard LAMV Portfolio 9 to 1.81 for the forecast enhanced LAMV Portfolio 12
  • Treatment on wash trading is unnecessary for modeling established crypto assets on mainstream unregulated exchanges. theoretical
    Source sentence
    We argue that treatment on wash trading is unnecessary in modeling established crypto assets that trade in unregulated but mainstream exchanges

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 28 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.