Pith. sign in

REVIEW 4 major objections 5 minor 36 references

Price Discovery in Cryptocurrency Markets

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Centralized exchanges lead price discovery in cryptocurrency markets, with futures ahead of spot for Bitcoin except in stress events.

desk verdict Standard price-discovery methods applied to 2024 crypto data, but the VECM estimates contain internal inconsistencies that sink the central claim. read the letter →

arxiv 2506.08718 v1 pith:OGL5RFMU submitted 2025-06-10 q-fin.TR

Juan Plazuelo Pascual , Carlos Tardon Rubio , Juan Toro Cebada , Angel Hernando Veciana This is my paper
classification q-fin.TR MSC 62M1091B84
keywords pricediscoveryHasbrouckinformationsharecointegrationlead-laganalysisUniswapEthereumBitcoinfuturesvectorerrorcorrectionmodel
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper asks where the price of a cryptocurrency is actually set: on a centralized exchange like Binance, in a decentralized trading pool like Uniswap, in the spot market, or in futures. The authors fit cointegration-based models to ETH prices on Binance versus Uniswap v2 and to Bitcoin prices on CME futures versus Binance spot, across one long sample and five high-volatility windows in 2024. They find that for ETH the centralized market consistently leads, with information shares above 0.9 on four of the five event dates, and that for BTC futures generally lead spot with a smaller margin, except on August 5, 2024 when the spot market led instead. If correct, the results tell traders and DeFi oracle designers which market to treat as the reference price, and they show that short-window stress events can temporarily reverse the usual ordering.

What carries the argument

The quantitative engine is a bivariate vector error correction model (VECM) fitted to each pair of price series, plus three statistics derived from it. Hasbrouck's information share attributes the variance of the common efficient price to each market and yields a 0-to-1 dominance score. The Gonzalo-Granger permanent-transitory decomposition reads the leader off the error-correction vector alpha: the market that adjusts least to price gaps carries the common stochastic trend. The Hayashi-Yoshida lagged estimator handles asynchronous tick data and reports which market's movements precede the other's, summarized as a lead-lag ratio and a lead time in seconds. The three measures are deliberately redundant, and when they agree the paper treats the directional conclusion as robust.

What would settle it

Recompute the May 20 ETH event metrics without the cointegration assumption, for example with pure lead-lag correlations on returns or with a model that lets the series drift apart, and see whether Binance still leads. A second check is to apply the identical metric suite to 60-minute windows around each event, where cointegration is far more likely to hold, and test whether the August 5, 2024 spot-leads-futures flip for BTC persists.

Watch

Extended reading notes

Core claim

The paper's central claim is that the centralized market is the price leader for ETH, and that futures lead spot for BTC except under stress. In the long-run ETH analysis, the centralized market's Hasbrouck information share is 0.999 against essentially zero for Uniswap v2, and the Gonzalo-Granger adjustment vector points the same way: Uniswap does the correcting while Binance does the discovering. Across the five 2024 event windows, the centralized share stays above 0.9 in four cases and is 0.628 in the remaining one (May 20), with every Hayashi-Yoshida lead-lag ratio above 1, indicating Binance leads. For BTC, the CME futures market shows Hasbrouck shares between 0.52 and 0.56 and positive lead-lag times of 0.055 to 0.15 seconds on most dates, but on August 5, 2024 the futures share collapses to 0.158 and all three metrics place the lead in the Binance spot market. The paper reads this as futures leading more often, although not as decisively as centralized spot leads decentralized exchange.

Load-bearing premise

The results stand or fall on the assumption that the two price series are cointegrated inside each short 14-to-20-minute event window, which is what makes the VECM information shares interpretable; the paper's own Johansen test for the May 20, 2024 ETH window finds no cointegration at the 10% level, yet the analysis of that window proceeds anyway.

Editorial extensions

If this is right

  • Traders tracking ETH should treat Binance as the reference price and read Uniswap v2 deviations as lagged adjustments rather than independent information, because that is the direction the paper's three metrics point.
  • For BTC under normal conditions, CME futures order flow is the earlier signal, with measured lead times between 0.055 and 0.15 seconds over Binance spot.
  • The ordering is not fixed: on August 5, 2024, all three metrics place the lead in the Binance spot market, so stress events can temporarily invert the futures-spot hierarchy.
  • DeFi arbitrage between Uniswap v2 and Binance is rarely profitable once gas fees are included, so the price gaps the paper documents mostly reflect transaction costs rather than exploitable profit.
  • Applying the same metric suite to Uniswap v3 pools, which the paper notes hold more liquidity than v2, is the natural follow-up it explicitly proposes.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the centralized lead is caused by liquidity and speed, then the ETH ordering should weaken as Uniswap v3 and layer-2 venues gain depth; the paper only tests Uniswap v2, so this is a testable prediction of its own explanation.
  • The August 5 reversal suggests a rule the paper leaves implicit: when the shock arrives outside CME trading hours, the 24/7 spot market becomes the information source until futures reopen, which can be checked from the timestamps of the lead-lag flip.
  • The May 20 ETH window, where the Johansen test finds no cointegration yet the analysis proceeds, is the case to re-estimate without the cointegration assumption before treating the centralized-lead result as universal.
  • Combining all three venues in one model could reveal whether the spot-futures and centralized-decentralized leads are two faces of the same fact, namely Binance spot as the anchor price, or two separate mechanisms.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. This paper studies price discovery in cryptocurrency markets using VECM-based Hasbrouck information shares, Gonzalo-Granger permanent-transitory decompositions, and Hayashi-Yoshida lead-lag statistics. For ETH it compares Binance spot with Uniswap v2 prices over a one-year span and five 14–20 minute event windows in 2024; for BTC it compares CME Micro Bitcoin futures with Binance spot over the same event windows. The headline claims are that the centralized market consistently leads the decentralized market for ETH, and that futures generally lead spot for BTC, with August 5, 2024 as a notable exception in the BTC case. The paper also provides a detailed reconstruction of Uniswap v2/v3 order books and an arbitrage-cost analysis based on gas fees.

Significance. If the results were reliable, the paper would offer a useful benchmark for assessing AMM oracle quality and for understanding information flow between centralized and decentralized venues, and between futures and spot markets. The institutional and data-processing material on reconstructing Uniswap order books is a genuine contribution, and the use of multiple price-discovery measures is appropriate in principle. However, the empirical support as currently reported is not yet solid enough to carry the headline conclusions. Load-bearing problems in the event-window VECM estimation, in the internal consistency of the Hasbrouck shares, and in the unspecified synchronization of the two price series prevent the paper from establishing its central claims in its present form.

major comments (4)
  1. [Section 3.1.4, Table 3.11] The Johansen trace test for the May 20 ETH window rejects both r=0 (16.80 > 13.43) and r=1 (6.755 > 2.705) at the 10% level. The paper states that the two series are not found to be cointegrated by this test and that the analysis is performed anyway. With no single cointegrating vector, the VECM is not identified for the assumed one common trend, and the Hasbrouck and Gonzalo-Granger shares in Table 3.12 are not valid estimates. This event therefore cannot be counted as evidence for the universal 'centralized leads' claim unless the model is re-specified and the test issue is resolved.
  2. [Tables 3.9 and 3.12] The reported Hasbrouck information shares do not sum to one within each Cholesky ordering. For April 30, Table 3.9 gives 0.965 + 0.076 = 1.041 and 0.923 + 0.003 = 0.926. For May 20, Table 3.12 gives 0.628 + 0.416 = 1.044 and 0.583 + 0.371 = 0.954. Hasbrouck information shares partition the innovation variance of the common factor, so each ordering should sum to one. This indicates a computational or reporting error in a central metric and requires correction and re-estimation.
  3. [Section 3.1 and Section 2.8] The paper never specifies how the one-second Binance series is synchronized with the sparse, block-level Uniswap v2 observations before estimating the VECM. Hasbrouck and Gonzalo-Granger methods require a synchronous bivariate series, and the choice of interpolation, aggregation, or timestamp alignment can mechanically create or erase lead-lag signals. The authors should describe the exact synchronization procedure and show that the price-discovery conclusions are robust to alternative reasonable alignments.
  4. [Section 3.3 and Table 3.35] No confidence intervals or standard errors are reported for any information share, Gonzalo-Granger alpha test, or Hayashi-Yoshida metric. This is especially damaging for the BTC futures-leads-spot conclusion, where the information shares are 0.52–0.56 and the LLR values are 1.11–1.36, i.e., close to the null values of 0.5 and 1.0. Without precision measures, these estimates do not distinguish 'futures lead' from 'both markets contribute equally.' At a minimum, block-bootstrap or delta-method confidence intervals should be provided for each event.
minor comments (5)
  1. [Section 3.1.1, Table 3.2] The text reports a centralized information share of 0.9996 for both orderings and a decentralized share of 4.2071e-4, but Table 3.2 shows centralized shares of 0.999 and 0.982 and decentralized shares of 0.001 and 0.018. These numbers are inconsistent and should be reconciled.
  2. [Section 3.1.2, Table 3.4] The text states that the rank-1 trace statistic of 2.113 'exceeds the critical value (2.705)' and thus rejects r=1, but 2.113 is less than 2.705. The correct reading is that exactly one cointegrating relationship is supported; this reversed inequality should be fixed.
  3. [Section 3.3.4 and Table 3.35] The text reports an August 5 BTC lead-lag time of -0.057 seconds, while Table 3.35 lists -0.571 seconds. One of these is a typographical error and should be corrected.
  4. [Section 3.1 (event windows)] The paper asserts that the series are 'clearly non-stationary' but reports no unit-root test statistics for the event windows. The VECM setup depends on this assumption, so the supporting tests should be reported or referenced.
  5. [Section 3.1.2, Gas Fee Calculation] The arbitrage cost calculation uses a gas limit of 50,000, which is low for a Uniswap swap; if actual gas usage is larger, the conclusion that arbitrage is unprofitable would only be strengthened, but the cost figure should be justified or corrected.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the price-discovery conclusions are data-driven estimates interpreted with standard metrics, not derivations equivalent to the paper's own inputs.

full rationale

The paper's chain is standard empirical inference. Long-run and event-window VECMs are estimated from Binance/Uniswap and CME/Binance price series; Hasbrouck information shares, Gonzalo-Granger permanent-transitory decompositions, and Hayashi-Yoshida lead-lag ratios are then computed from the estimated adjustment coefficients, residual covariance, and tick returns. The central statements—'the centralized market incorporated the most information in most cases' and 'the futures market appears to lead more often'—are interpretations of those estimated metrics, not quantities defined to equal the inputs. There is no fitted parameter renamed as a prediction and no out-of-sample claim. The nearest concern is Section 3.1.4, where the reported Johansen test does not support a single cointegrating vector and the paper states 'we decide to perform the analysis anyway'; this is a maintained-assumption/validity problem, not circularity, because the information shares are not constructed from the cointegration test result. Similarly, the non-summing Hasbrouck shares in Table 3.12 and the incorrect verbal reading of Table 3.4's rank-1 statistic are statistical or reporting errors, not evidence that the conclusion reduces to the data by construction. No load-bearing self-citations, imported uniqueness theorems, or ansatz-smuggling citations appear. The comparison with prior Bitcoin futures literature is used only as external context. The derivation is therefore self-contained and non-circular.

Assumptions & free parameters 5 free parameters · 6 assumptions · 0 invented entities

The paper introduces no new theoretical entities; it applies existing econometric methods. The hidden assumptions are mostly about data synchronization, short-window cointegration, and treating different market microstructures as comparable price series. The fitted alpha coefficients, beta vector, lag order, and the arbitrary HY lag grid are the main quantities the conclusions depend on, and none of them is validated out of sample.

free parameters (5)
  • VECM alpha adjustment coefficients = long-run (-0.026, 0.958); event values e.g.
    Gonzalo-Granger conclusions are read directly off these fitted coefficients; they are estimated from the same data used to assess the conclusion, not independently predicted.
  • VECM cointegrating vector beta = [1, -0.999] for long-run ETH; other windows not reported
    The cointegrating vector is estimated and normalized from the data; the paper does not report its standard errors or estimation details.
  • VECM lag order = p = 1 for all windows
    A single lag is used throughout with no reported lag selection procedure; the choice affects the estimated alpha and information shares.
  • Hayashi-Yoshida lag grid = not disclosed (the text says 'arbitrarily choosing')
    Section 2.8.1 defines LLR over a grid of lags the authors choose arbitrarily; the exact grid is not given in the empirical sections, so LLR values are not reproducible.
  • Gas limit for arbitrage fee = 50,000 gas units
    Used to convert Gwei prices to USD fees; the paper applies it twice, reporting $30.58 in the text and $34.92 in Table 3.6, an unresolved inconsistency.
assumptions (6)
  • domain assumption Binance and Uniswap v2 ETH prices are I(1) and cointegrated, so a VECM(1) is the correct model.
    Section 3.1.1 and each event section assert non-stationarity and cointegration; the May 20 window's Johansen test contradicts this, yet the analysis proceeds.
  • domain assumption Sparse Uniswap v2 trade or block observations can be aligned with 1-second Binance data to form a bivariate time series.
    The paper never describes the synchronization procedure, but all VECM, Hasbrouck, and Gonzalo-Granger estimates require synchronous series.
  • domain assumption The WETH/USDT Uniswap v2 pool price and the Binance ETH price represent the same asset with a stationary difference.
    The analysis treats the two pools as perfect substitutes, ignoring pool fee, block-time staleness, and DEX-specific microstructure.
  • domain assumption The CME Micro Bitcoin futures price and Binance spot BTC price are cointegrated with a stationary basis, and no maturity or convenience-yield adjustment is needed.
    Section 3.3 uses raw futures and spot prices in a VECM without modeling the basis or contract expiration.
  • ad hoc to paper Long-run cointegration justifies VECM estimation on 14-20 minute event windows.
    Section 3.1.4 states that even though the window-level Johansen test fails, the series are 'clearly cointegrated in the long-run and we decide to perform the analysis anyway.' This assumption is load-bearing because one of the five ETH events is included in every summary table.
  • standard math Standard asymptotic theory for Johansen, Hasbrouck, and Gonzalo-Granger statistics applies to these short, heteroskedastic, high-frequency samples.
    The paper uses 10% critical values from standard tables with 43-178 DEX observations per event window, where asymptotic approximations are questionable.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Price Discovery in Cryptocurrency Markets." pith.science (2026). https://pith.science/paper/OGL5RFMU

@misc{pith2026250608718,
  author       = {Pith},
  title        = {Pith review of: Price Discovery in Cryptocurrency Markets},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OGL5RFMU}},
  note         = {Machine review of arXiv:2506.08718}
}
read the original abstract

This document analyzes price discovery in cryptocurrency markets by comparing centralized and decentralized exchanges, as well as spot and futures markets. The study focuses first on Ethereum (ETH) and then applies a similar approach to Bitcoin (BTC). Chapter 1 outlines the theoretical framework, emphasizing the structural differences between centralized exchanges and decentralized finance mechanisms, especially Automated Market Makers (AMMs). It also explains how to construct an order book from a liquidity pool in a decentralized setting for comparison with centralized exchanges. Chapter 2 describes the methodological tools used: Hasbrouck's Information Share, Gonzalo and Granger's Permanent-Transitory decomposition, and the Hayashi-Yoshida estimator. These are applied to explore lead-lag dynamics, cointegration, and price discovery across market types. Chapter 3 presents the empirical analysis. For ETH, it compares price dynamics on Binance and Uniswap v2 over a one-year period, focusing on five key events in 2024. For BTC, it analyzes the relationship between spot and futures prices on the CME. The study estimates lead-lag effects and cointegration in both cases. Results show that centralized markets typically lead in ETH price discovery. In futures markets, while they tend to lead overall, high-volatility periods produce mixed outcomes. The findings have key implications for traders and institutions regarding liquidity, arbitrage, and market efficiency. Various metrics are used to benchmark the performance of modified AMMs and to understand the interaction between decentralized and centralized structures.

Figures

Figures reproduced from arXiv: 2506.08718 by the authors.

Figure 1
Figure 1. [PITH_FULL_IMAGE:figures/full_fig_p017_1.png] view at source ↗
Figure 1
Figure 1. [PITH_FULL_IMAGE:figures/full_fig_p019_1.png] view at source ↗
Figure 1
Figure 1. [PITH_FULL_IMAGE:figures/full_fig_p021_1.png] view at source ↗
Figures from the paper (24 more)
Figure 2
Figure 2. Figure 2: illustrates the dynamics of the cointegrated VECM system. The common trend [PITH_FULL_IMAGE:figures/full_fig_p036_2.png]
Figure 2
Figure 2. Figure 2 [PITH_FULL_IMAGE:figures/full_fig_p040_2.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p047_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p048_3.png]
Figure 3.1
Figure 3.1. Figure 3.1 [PITH_FULL_IMAGE:figures/full_fig_p052_3_1.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p052_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p053_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p057_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p059_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p060_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p062_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p063_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p065_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p066_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p070_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p071_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p073_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p074_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p076_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p077_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p079_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p080_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p082_3.png]
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p083_3.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

36 extracted references · 31 canonical work pages

  1. [1]

    Uniswap v3 Core

    Hayden Adams et al. Uniswap v3 Core. 2021. URL: https://api.semanticscholar. org/CorpusID:232416764

    work page 2021

  2. [2]

    Lecture Notes

    Luigi Ambrosio. Lecture Notes. Google Sites. https://sites.google.com/site/ ambropo/LectureNotes. 2024

    work page 2024

  3. [3]

    Coexisting Exchange Platforms: Limit Order Books and Auto- mated Market Makers

    Jun Aoyagi and Yuki Ito. Coexisting Exchange Platforms: Limit Order Books and Auto- mated Market Makers . 2021. SSRN: 3808755. URL: https : / / ssrn . com / abstract = 3808755

    work page 2021

  4. [4]

    Price discovery and common factor models

    Richard T. Baillie et al. “Price discovery and common factor models”. In: Journal of Financial Markets 5.3 (2002). Price Discovery, pp. 309–321. ISSN : 1386-4181. DOI: https : / / doi . org / 10 . 1016 / S1386 - 4181(02 ) 00027 - 7. URL: https : / / www . sciencedirect.com/science/article/pii/S1386418102000277

    work page 2002

  5. [5]

    On The Quality Of Cryptocurrency Markets: Cen- tralized Versus Decentralized Exchanges

    Andrea Barbon and Angelo Ranaldo. On The Quality Of Cryptocurrency Markets: Cen- tralized Versus Decentralized Exchanges . 2021. arXiv: 2112 . 07386 [q-fin.TR]. URL: https://arxiv.org/abs/2112.07386

    arXiv 2021

  6. [6]

    Binance: Cryptocurrency Exchange

    Binance. Binance: Cryptocurrency Exchange. https://www.binance.com. 2024

    work page 2024

  7. [7]

    Campbell, Andrew W

    John Y. Campbell, Andrew W. Lo, and A.Craig MacKinlay. The Econometrics of Fi- nancial Markets. Princeton University Press, 1997. ISBN : 9780691043012. URL: http: //www.jstor.org/stable/j.ctt7skm5 (visited on 07/24/2024)

    work page 1997

  8. [8]

    Price Discovery on Decentralized Ex- changes

    Agostino Capponi, Ruizhe Jia, and Shihao Yu. Price Discovery on Decentralized Ex- changes. 2024. URL: https://ink.library.smu.edu.sg/lkcsb_research/7508

    work page 2024

Show all 36 references
  1. [9]

    Cartea, S

    Á. Cartea, S. Jaimungal, and J. Penalva. Algorithmic and High-Frequency Trading. Cam- bridge University Press, 2015. ISBN : 9781107091146. URL: https://books.google. es/books?id=5dMmCgAAQBAJ

    2015

  2. [10]

    Flash Boys 2.0: Frontrunning, Transaction Reordering, and Consensus Instability in Decentralized Exchanges

    Philip Daian et al. Flash Boys 2.0: Frontrunning, Transaction Reordering, and Consensus Instability in Decentralized Exchanges. 2019. arXiv: 1904.05234 [cs.CR]. URL: https: //arxiv.org/abs/1904.05234

    2019 arXiv

  3. [11]

    OxMetrics: An Interface to Empirical Modelling

    Jurgen A. Doornik and David F. Hendry. “OxMetrics: An Interface to Empirical Modelling”. In: 9th OxMetrics User Conference. 2018. 90 Crypto

    2018

  4. [12]

    Dynamic Conditional Correlation: A Simple Class of Multivariate Generalized Autoregressive Conditional Heteroskedasticity Models

    Robert F. Engle. “Dynamic Conditional Correlation: A Simple Class of Multivariate Generalized Autoregressive Conditional Heteroskedasticity Models”. In: Journal of Business & Economic Statistics 20.3 (2002), pp. 339–350

    2002

  5. [13]

    Etherscan: Ethereum Blockchain Explorer

    Etherscan. Etherscan: Ethereum Blockchain Explorer. https://etherscan.io. 2024

    2024

  6. [14]

    Modelling and measuring price dis- covery in commodity markets

    Isabel Figuerola-Ferretti and Jesus Gonzalo. “Modelling and measuring price dis- covery in commodity markets”. In: Journal of Econometrics 158.1 (2010), pp. 95–107

    2010

  7. [15]

    Ethereum poised for high volatility as Grayscale CEO departs several days be- fore SEC’s ETF decision

    FXStreet. Ethereum poised for high volatility as Grayscale CEO departs several days be- fore SEC’s ETF decision . Accessed: 2024-05-20. 2024. URL: https://www.fxstreet. com / cryptocurrencies / news / ethereum - poised - for - high - volatility - as - grayscale-ceo-departs-sev...

    2024

  8. [16]

    Estimation of Common Long-Memory Com- ponents in Cointegrated Systems

    Jesus Gonzalo and Clive Granger. “Estimation of Common Long-Memory Com- ponents in Cointegrated Systems”. In: Journal of Business & Economic Statistics 13.1 (1995), pp. 27–35. ISSN : 07350015. URL: http://www.jstor.org/stable/1392518 (visited on 07/18/2024)

    1995

  9. [17]

    Estimation of Common Long-Memory Com- ponents in Cointegrated Systems

    Jesus Gonzalo and Clive Granger. “Estimation of Common Long-Memory Com- ponents in Cointegrated Systems”. In: Journal of Business & Economic Statistics 13.1 (1995), pp. 27–35

    1995

  10. [18]

    Investigating Causal Relations by Econometric Models and Cross- spectral Methods

    C. W. J. Granger. “Investigating Causal Relations by Econometric Models and Cross- spectral Methods”. In:Econometrica 37.3 (1969), pp. 424–438.ISSN : 00129682, 14680262. URL: http://www.jstor.org/stable/1912791 (visited on 07/25/2024)

    1969

  11. [19]

    Trust in DeFi: An Empirical Study of the Decentralized Exchange

    Jianlei Han, Shiyang Huang, and Zhuo Zhong. Trust in DeFi: An Empirical Study of the Decentralized Exchange. 2021. SSRN: 3896461. URL: https://papers.ssrn.com/ sol3/papers.cfm?abstract_id=3896461

    2021

  12. [20]

    One Security, Many Markets: Determining the Contributions to Price Discovery

    Joel Hasbrouck. “One Security, Many Markets: Determining the Contributions to Price Discovery”. In: Journal of Finance 50.4 (1995), pp. 1175–1199

    1995

  13. [21]

    On covariance estimation of non-synchronously observed diffusion processes

    Takaki Hayashi and Nakahiro Yoshida. “On covariance estimation of non-synchronously observed diffusion processes”. In: Bernoulli 11.2 (2005), pp. 359–379. DOI: 10.3150/ bj/1116340299. URL: https://doi.org/10.3150/bj/1116340299

    2005

  14. [22]

    Behavior of Liquidity Providers in Decentralized Exchanges

    Lioba Heimbach, Ye Wang, and Roger Wattenhofer. Behavior of Liquidity Providers in Decentralized Exchanges. 2021. arXiv: 2105.13822 [q-fin.CP]. URL: https://arxiv. org/abs/2105.13822

    2021 arXiv

  15. [23]

    What role do futures markets play in Bitcoin pricing? Causality, cointegration and price discovery from a time-varying perspective?

    Yang Hu, Yang Greg Hou, and Les Oxley. “What role do futures markets play in Bitcoin pricing? Causality, cointegration and price discovery from a time-varying perspective?” In: International Review of Financial Analysis 72 (2020), p. 101569. ISSN : 1057-5219. DOI: https://doi....

    2020

  16. [24]

    High Frequency Lead/Lag Relationships—Empirical Facts

    Nicolas Huth and Frederic Abergel. “High Frequency Lead/Lag Relationships—Empirical Facts”. In: Quantitative Finance 14.4 (2014), pp. 497–507

    2014

  17. [25]

    Estimation and Hypothesis Testing of Cointegration Vectors in Gaussian Vector Autoregressive Models

    Søren Johansen. “Estimation and Hypothesis Testing of Cointegration Vectors in Gaussian Vector Autoregressive Models”. In:Econometrica 59.6 (1991), pp. 1551–1580. ISSN : 00129682, 14680262. URL: http://www.jstor.org/stable/2938278 (visited on 09/03/2024)

    1991

  18. [26]

    Likelihood-Based Inference in Cointegrated Vector Autoregressive Models

    Søren Johansen. Likelihood-Based Inference in Cointegrated Vector Autoregressive Models. Oxford University Press, 1996

    1996

  19. [27]

    The role of the constant and linear terms in cointegration analysis of nonstationary variables

    Søren Johansen. “The role of the constant and linear terms in cointegration analysis of nonstationary variables”. In: Econometric Reviews 13.2 (1994), pp. 205–229. DOI: 10.1080/07474939408800284

    1994 doi

  20. [28]

    Smart Contracts and Decentralized Fi- nance

    Kose John, Leonid Kogan, and Fahad Saleh. “Smart Contracts and Decentralized Fi- nance”. In: SSRN Electronic Journal(2022). DOI: 10.2139/ssrn.4222528. URL: https: //papers.ssrn.com/sol3/papers.cfm?abstract_id=4222528

    2022 doi

  21. [29]

    Fragmentation and optimal liq- uidity supply on decentralized exchanges

    Alfred Lehar, Christine Parlour, and Marius Zoican. Fragmentation and optimal liq- uidity supply on decentralized exchanges . 2024. arXiv: 2307 . 13772 [q-fin.TR]. URL: https://arxiv.org/abs/2307.13772

    2024 arXiv

  22. [30]

    Spoofing and Layering

    Gideon Mark. “Spoofing and Layering”. In: Journal of Corporation Law 45.2 (2020), pp. 399–430

    2020

  23. [31]

    Putnin, š

    T ¯alis J. Putnin, š. An Overview of Market Manipulation. SSRN, 2019, p. 46

    2019

  24. [32]

    Refinitiv: Financial Market Data and Infrastructure

    Refinitiv. Refinitiv: Financial Market Data and Infrastructure. https://www.refinitiv. com. 2024

    2024

  25. [33]

    Decentralized Finance: On Blockchain- and Smart Contract-Based Fi- nancial Markets

    Fabian Schär. “Decentralized Finance: On Blockchain- and Smart Contract-Based Fi- nancial Markets”. In: Review 103.2 (Apr. 2021), pp. 153–174. DOI: 10.20955/r.103. 153-74. URL: https://ideas.repec.org/a/fip/fedlrv/91428.html

    2021 doi

  26. [34]

    Statsmodels: Econometric and statistical mod- eling with Python

    Skipper Seabold and Josef Perktold. “Statsmodels: Econometric and statistical mod- eling with Python”. In: 9th Python in Science Conference. 2010

    2010

  27. [35]

    Investigating the Efficiency of Bitcoin Futures in Price Dis- covery

    Prashant Sharma et al. “Investigating the Efficiency of Bitcoin Futures in Price Dis- covery”. In: International Journal of Economics and Financial Issues 12.3 (May 2022), pp. 104–109. URL: https://ideas.repec.org/a/eco/journ1/2022-03-12.html

    2022

  28. [36]

    Ruey S. Tsay. Analysis of Financial Time Series. 3rd. Wiley, 2010, p. 712. 92 de 92

    2010

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.