pith. sign in

arxiv: 2605.20828 · v2 · pith:POH7JUW4new · submitted 2026-05-20 · 📊 stat.ME

Adaptive Test for Jump

Huifang Ma , Long Feng This is my paper

Pith reviewed 2026-05-22 09:35 UTC · model grok-4.3

classification 📊 stat.ME
keywords jump testhigh-frequency datasemimartingalesCauchy combinationasymptotic independencemicrostructure noisefinite-activity jumps
0
0 comments X

The pith

Combining the Aït-Sahalia--Jacod ratio statistic and Lee--Mykland extreme-return statistic via the Cauchy rule produces an adaptive jump test with closed-form calibration and explicit power for high-frequency semimartingales.

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

The paper constructs a jump test for discretely sampled high-frequency semimartingales that allows stochastic drift, volatility, and leverage. It merges an existing ratio-based statistic with an existing extreme-return statistic through the Cauchy combination rule. Asymptotic independence holds under the continuous-path null and under dense local alternatives, which permits exact analytic calibration of critical values and closed-form power functions. The resulting test is consistent against finite-activity jumps and extends directly to additive microstructure noise.

Core claim

By applying the Cauchy combination rule to the Aït-Sahalia--Jacod ratio statistic and the Lee--Mykland extreme-return statistic, the procedure yields asymptotic independence under the continuous-path null for processes with stochastic Itô drift, volatility, and leverage; the combined test therefore admits an analytically calibrated rejection rule together with closed-form power expressions and remains consistent when jumps occur only finitely often.

What carries the argument

The combined test statistic obtained by feeding the Aït-Sahalia--Jacod ratio statistic and the Lee--Mykland extreme-return statistic into the Cauchy combination rule.

If this is right

  • Critical values can be obtained from the Cauchy distribution without simulation or bootstrap.
  • Power against dense local alternatives can be expressed in closed form.
  • The test remains consistent against jumps of finite activity.
  • The same combination extends immediately to observations contaminated by additive microstructure noise.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The procedure may deliver better size-power balance than either component test alone when jumps are sparse.
  • The analytic calibration could be reused as a building block inside sequential monitoring schemes for intraday data.
  • Extension to multivariate semimartingales would require only verifying that the same two statistics remain asymptotically independent in higher dimensions.

Load-bearing premise

The two component statistics are asymptotically independent under the continuous-path null even when drift, volatility, and leverage are all stochastic.

What would settle it

A Monte Carlo experiment in which the finite-sample distribution of the combined p-value deviates materially from uniformity under a continuous semimartingale path with stochastic volatility would refute the claimed asymptotic independence.

Figures

Figures reproduced from arXiv: 2605.20828 by Huifang Ma, Long Feng.

Figure 1
Figure 1. Figure 1: Power curves at the 5% level under the dense-local alternative with Yim ∼ N(0, 400) and 5 secs. 23 [PITH_FULL_IMAGE:figures/full_fig_p023_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Power curves at the 5% level under the dense-local alternative with Yim ∼ N(0, 400) and 10 secs. 24 [PITH_FULL_IMAGE:figures/full_fig_p024_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Sparse-alternative power at the 5% level, 5-second sampling 25 [PITH_FULL_IMAGE:figures/full_fig_p025_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Sparse-alternative power at the 5% level, 10-second sampling 26 [PITH_FULL_IMAGE:figures/full_fig_p026_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Histograms of the daily noise estimates qˆ for SPY, QQQ, IWM, AAPL, MSFT, and NVDA. 30 [PITH_FULL_IMAGE:figures/full_fig_p030_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Intraday price paths on March 9, 2020, with the LM-based jump location marked for each asset. [PITH_FULL_IMAGE:figures/full_fig_p032_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Intraday price paths on November 10, 2022, with the LM-based jump location marked for each [PITH_FULL_IMAGE:figures/full_fig_p033_7.png] view at source ↗
read the original abstract

We develop an adaptive jump test for discretely observed high-frequency semimartingales by combining the A"it-Sahalia--Jacod ratio statistic (A"it-Sahalia and Jacod, 2009) and the Lee--Mykland extreme-return statistic (Lee and Mykland, 2008) with the Cauchy combination rule. Allowing stochastic It^o drift, volatility, and leverage, we show asymptotic independence under the continuous-path null and dense local alternatives, yielding an analytically calibrated test with closed-form power; under finite-activity jumps, the test is consistent. We also extend the method to additive microstructure noise. Simulations show that the combined procedure performs well under both dense and sparse alternatives and is typically best overall.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper develops an adaptive jump test for discretely observed high-frequency semimartingales by combining the Aït-Sahalia--Jacod ratio statistic and the Lee--Mykland extreme-return statistic via the Cauchy combination rule. It claims to establish asymptotic independence of the normalized statistics under the continuous-path null (allowing stochastic Itô drift, volatility, and leverage), yielding an analytically calibrated test with closed-form power; the procedure is consistent under finite-activity jumps and is extended to additive microstructure noise. Monte Carlo simulations indicate competitive performance under both dense and sparse alternatives.

Significance. If the joint convergence and independence result is rigorously established, the contribution is a practical, closed-form calibrated adaptive test that combines two standard high-frequency statistics without requiring bootstrap or tuning parameters. This would be useful for empirical work in financial econometrics where explicit power expressions aid design and interpretation, and the extension to noise is a practical addition.

major comments (2)
  1. [§3] §3 (main theorem on joint convergence): The claim that the normalized Aït-Sahalia--Jacod ratio and Lee--Mykland maximum-return statistics converge jointly to independent limits under stochastic volatility and non-zero leverage is load-bearing for the closed-form Cauchy calibration and power formula. The argument must explicitly address why shared randomness from the volatility path (via quadratic covariation) does not induce asymptotic dependence; a sketch relying on stable convergence or local constancy of volatility on the scale of the largest increment should be expanded with the precise limiting joint characteristic function or covariance calculation.
  2. [§4.2] §4.2 (power derivation): The closed-form power expression under dense local alternatives is derived from the independence assumption; if the joint limit is not exactly independent, this expression becomes approximate. The manuscript should state the precise conditions (e.g., on the rate of volatility variation) under which the independence holds exactly versus approximately.
minor comments (2)
  1. [Table 1] Table 1 (size results): Report Monte Carlo standard errors for the empirical rejection frequencies to allow assessment of simulation variability.
  2. [§2] Notation in §2: Define the local volatility estimator in the Lee--Mykland statistic more explicitly (including the bandwidth choice) to facilitate replication.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments on our manuscript. We address each major comment below and will revise the paper to provide the requested expansions and clarifications.

read point-by-point responses

  1. Referee: [§3] §3 (main theorem on joint convergence): The claim that the normalized Aït-Sahalia--Jacod ratio and Lee--Mykland maximum-return statistics converge jointly to independent limits under stochastic volatility and non-zero leverage is load-bearing for the closed-form Cauchy calibration and power formula. The argument must explicitly address why shared randomness from the volatility path (via quadratic covariation) does not induce asymptotic dependence; a sketch relying on stable convergence or local constancy of volatility on the scale of the largest increment should be expanded with the precise limiting joint characteristic function or covariance calculation.

    Authors: We appreciate the referee pointing out the need for greater explicitness here. The current sketch in §3 relies on stable convergence in law together with local constancy of volatility over intervals of length equal to the mesh size. In the revision we will expand this argument by deriving the joint limiting characteristic function explicitly. The calculation shows that the cross term vanishes because the normalized Aït-Sahalia–Jacod ratio converges to a continuous functional of the entire volatility path (an average), while the normalized Lee–Mykland statistic is asymptotically driven by the single largest increment at a random location; these two functionals are asymptotically orthogonal under the continuous semimartingale measure even when volatility and leverage are stochastic. The expanded derivation will appear in §3 together with a short appendix containing the characteristic-function algebra. revision: yes

  2. Referee: [§4.2] §4.2 (power derivation): The closed-form power expression under dense local alternatives is derived from the independence assumption; if the joint limit is not exactly independent, this expression becomes approximate. The manuscript should state the precise conditions (e.g., on the rate of volatility variation) under which the independence holds exactly versus approximately.

    Authors: We agree that the distinction between exact and approximate independence should be stated clearly. The independence is exact when the volatility process satisfies a local constancy condition at the observation frequency (specifically, its modulus of continuity is o_p(Δ_n^{1/2}) uniformly on compact sets). Under weaker Hölder conditions on volatility the joint convergence still holds but the limiting variables are only asymptotically independent, rendering the power formula approximate with an error of higher order. The revision will add a remark in §4.2 that spells out these conditions and notes the order of the approximation error under standard semimartingale assumptions on volatility. revision: yes

Circularity Check

0 steps flagged

No significant circularity; independence result presented as derived

full rationale

The paper states that it shows asymptotic independence of the Aït-Sahalia–Jacod ratio and Lee–Mykland extreme-return statistics under the continuous semimartingale null allowing stochastic drift, volatility, and leverage. This independence is then used to calibrate the Cauchy combination rule and obtain closed-form power. No quoted equation or step reduces the claimed independence or the resulting test to a fitted parameter, a self-definition, or a load-bearing self-citation whose content is itself unverified. The derivation is therefore treated as self-contained once the joint convergence is accepted; any doubt about the validity of that convergence belongs to correctness rather than circularity.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on standard asymptotic theory for semimartingales and the specific independence property of the two statistics; no free parameters or new entities are introduced in the abstract.

axioms (1)
  • domain assumption Asymptotic independence of Aït-Sahalia--Jacod and Lee--Mykland statistics under continuous-path null for Itô processes with stochastic drift, volatility, and leverage
    Invoked to enable Cauchy combination with closed-form critical values and power.

pith-pipeline@v0.9.0 · 5638 in / 1239 out tokens · 34104 ms · 2026-05-22T09:35:21.490890+00:00 · methodology

Review history (2 revisions) →

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

  • IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear
    ?
    unclear

    Relation between the paper passage and the cited Recognition theorem.

    We develop an adaptive jump test ... by combining the Aït-Sahalia–Jacod ratio statistic ... and the Lee–Mykland extreme-return statistic ... with the Cauchy combination rule. ... show asymptotic independence under the continuous-path null

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

50 extracted references · 50 canonical work pages

  1. [1]

    2020 , howpublished =

    brtnsmth , title =. 2020 , howpublished =

    work page 2020

  2. [2]

    and Xue, L

    Li, D. and Xue, L. , title =. 2015 , journal =

    work page 2015

  3. [3]

    High-frequency financial econometrics , publisher =

    A. High-frequency financial econometrics , publisher =

    work page

  4. [4]

    and Heyde, C

    Hall, P. and Heyde, C. C. , title =

    work page

  5. [5]

    , title =

    Vershynin, R. , title =

    work page

  6. [6]

    van der Vaart, A. W. , title =

    work page

  7. [7]

    Protter, P. E. , title =

    work page

  8. [8]

    Petrov, V. V. , title =

    work page

  9. [9]

    Tony Cai and Weidong Liu and Yin Xia , title=

    T. Tony Cai and Weidong Liu and Yin Xia , title=. Journal of the Royal Statistical Society. Series B (Statistical Methodology) , year=

    work page

  10. [10]

    Peter , title =

    Zu, Yang and Boswijk, H. Peter , title =. Journal of Econometrics , year =

    work page

  11. [11]

    Testing for jumps in a discretely observed process , journal =

    A. Testing for jumps in a discretely observed process , journal =. 2009 , volume =

    work page 2009

  12. [12]

    Ultra high frequency volatility estimation with dependent microstructure noise , journal =

    A. Ultra high frequency volatility estimation with dependent microstructure noise , journal =. 2011 , volume =

    work page 2011

  13. [13]

    Merrick and Laeven, Roger J

    Li, Z. Merrick and Laeven, Roger J. A. and Vellekoop, Michel H. , title =. Journal of Econometrics , year =

    work page

  14. [14]

    The Annals of Statistics , year =

    Jacod, Jean and Todorov, Viktor , title =. The Annals of Statistics , year =

    work page

  15. [15]

    Journal of Econometrics , year =

    Bibinger, Markus and Winkelmann, Lars , title =. Journal of Econometrics , year =

    work page

  16. [16]

    Testing for jumps in noisy high frequency data , journal =

    A. Testing for jumps in noisy high frequency data , journal =. 2012 , volume =

    work page 2012

  17. [17]

    and Shephard, Neil , title =

    Barndorff-Nielsen, Ole E. and Shephard, Neil , title =. Journal of Financial Econometrics , year =

    work page

  18. [18]

    Journal of Econometrics , year =

    Bibinger, Markus and Hautsch, Nikolaus and Ristig, Alexander , title =. Journal of Econometrics , year =

    work page

  19. [19]

    Journal of Financial Econometrics , year =

    Kwok, Simon , title =. Journal of Financial Econometrics , year =

    work page

  20. [20]

    and Hannig, Jan , title =

    Lee, Suzanne S. and Hannig, Jan , title =. Journal of Financial Economics , year =

    work page

  21. [21]

    and Mykland, Per A

    Lee, Suzanne S. and Mykland, Per A. , title =. The Review of Financial Studies , year =

    work page

  22. [22]

    and Mykland, Per A

    Lee, Suzanne S. and Mykland, Per A. , title =. Journal of Econometrics , year =

    work page

  23. [23]

    Journal of Econometrics , year =

    Li, Qiyuan and Li, Yifan and Nolte, Ingmar and Nolte, Sandra and Yu, Shifan , title =. Journal of Econometrics , year =

    work page

  24. [24]

    Journal of the American Statistical Association , year =

    Liu, Yaowu and Xie, Jun , title =. Journal of the American Statistical Association , year =

    work page

  25. [25]

    and Forbes, Catherine S

    Maneesoonthorn, Worapree and Martin, Gael M. and Forbes, Catherine S. , title =. Journal of Econometrics , year =

    work page

  26. [26]

    and Oomen, Roel C

    Jiang, George J. and Oomen, Roel C. A. , title =. Journal of Econometrics , year =

    work page

  27. [27]

    and Podolskij, Mark and Vetter, Mathias , title =

    Jacod, Jean and Li, Yingying and Mykland, Per A. and Podolskij, Mark and Vetter, Mathias , title =. Stochastic Processes and their Applications , year =

    work page

  28. [28]

    and Zhang, Lan , title =

    Mykland, Per A. and Zhang, Lan , title =. Journal of Econometrics , year =

    work page

  29. [29]

    The Annals of Probability , year =

    Hsing, Tailen , title =. The Annals of Probability , year =

    work page

  30. [30]

    Journal of Applied Probability , year =

    Ho, Hwai-Chung and Hsing, Tailen , title =. Journal of Applied Probability , year =

    work page

  31. [31]

    Biometrika , year =

    Xu, Gongjun and Lin, Lifeng and Wei, Peng and Pan, Wei , title =. Biometrika , year =

    work page

  32. [32]

    The Annals of Statistics , year =

    He, Yinqiu and Xu, Gongjun and Wu, Chong and Pan, Wei , title =. The Annals of Statistics , year =

    work page

  33. [33]

    The Annals of Statistics , year =

    Feng, Long and Jiang, Tiefeng and Liu, Binghui and Xiong, Wei , title =. The Annals of Statistics , year =

    work page

  34. [34]

    Statistica Sinica , year =

    Feng, Long and Jiang, Tiefeng and Li, Xiaoyun and Liu, Binghui , title =. Statistica Sinica , year =

    work page

  35. [35]

    How often to sample a continuous-time process in the presence of market microstructure noise , journal =

    A. How often to sample a continuous-time process in the presence of market microstructure noise , journal =. 2005 , volume =

    work page 2005

  36. [36]

    and Lunde, Asger , title =

    Hansen, Peter R. and Lunde, Asger , title =. Journal of Business & Economic Statistics , year =

    work page

  37. [37]

    Journal of Econometrics , year =

    Wang, Hongfei and Liu, Binghui and Feng, Long and Ma, Yanyuan , title =. Journal of Econometrics , year =

    work page

  38. [38]

    Telling from discrete data whether the underlying continuous-time model is a diffusion , journal =

    A. Telling from discrete data whether the underlying continuous-time model is a diffusion , journal =. 2002 , volume =

    work page 2002

  39. [39]

    Analyzing the spectrum of asset returns: Jump and volatility components in high frequency data , journal =

    A. Analyzing the spectrum of asset returns: Jump and volatility components in high frequency data , journal =. 2012 , volume =

    work page 2012

  40. [40]

    and Bollerslev, Tim and Diebold, Francis X

    Andersen, Torben G. and Bollerslev, Tim and Diebold, Francis X. , title =. The Review of Economics and Statistics , year =

    work page

  41. [41]

    Management Science , year =

    Bajgrowicz, Pierre and Scaillet, Olivier and Treccani, Adrien , title =. Management Science , year =

    work page

  42. [42]

    Christensen, Kim and Oomen, Roel C. A. and Podolskij, Mark , title =. Journal of Financial Economics , year =

    work page

  43. [43]

    Threshold bipower variation and the impact of jumps on volatility forecasting , journal =

    Corsi, Fulvio and Pirino, Davide and Ren. Threshold bipower variation and the impact of jumps on volatility forecasting , journal =. 2010 , volume =

    work page 2010

  44. [44]

    Journal of Business & Economic Statistics , year =

    Dumitru, Ana-Maria and Urga, Giovanni , title =. Journal of Business & Economic Statistics , year =

    work page

  45. [45]

    Journal of Econometrics , year =

    Dungey, Mardi and Erdemlioglu, Deniz and Matei, Marius and Yang, Xiye , title =. Journal of Econometrics , year =

    work page

  46. [46]

    Journal of the American Statistical Association , year =

    Fan, Jianqing and Wang, Yaozhong , title =. Journal of the American Statistical Association , year =

    work page

  47. [47]

    Journal of Financial Econometrics , year =

    Huang, Xin and Tauchen, George , title =. Journal of Financial Econometrics , year =

    work page

  48. [48]

    Scandinavian Journal of Statistics , year =

    Mancini, Cecilia , title =. Scandinavian Journal of Statistics , year =

    work page

  49. [49]

    Statistical Inference for Stochastic Processes , year =

    Podolskij, Mark and Ziggel, Daniel , title =. Statistical Inference for Stochastic Processes , year =

    work page

  50. [50]

    Statistica Neerlandica , year =

    Sun, Yucheng , title =. Statistica Neerlandica , year =

    work page