pith. machine review for the scientific record. sign in

arxiv: 2604.25668 · v1 · submitted 2026-04-28 · 🧮 math.ST · stat.TH

Recognition: unknown

Sharp adaptive nonparametric testing for constant volatility

Johannes Brutsche, Lukas Riepl

Pith reviewed 2026-05-07 14:33 UTC · model grok-4.3

classification 🧮 math.ST stat.TH
keywords nonparametric testingvolatility functionGaussian white noiseminimax optimalityadaptive testinginfill asymptoticsimplicit function theorem
0
0 comments X

The pith

A test for constant volatility achieves minimax optimality and adaptivity by measuring deviations as a ratio to the L2 average.

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

The paper develops a statistical test to determine whether the volatility function remains constant in the Gaussian white noise model observed at discrete times. It proves the test attains the minimax-optimal detection boundary while adapting automatically to the unknown smoothness level of the volatility. The analysis shows that the natural way to quantify a departure from constancy is through the ratio of the local volatility value to the square root of its integrated square. The sharp constants in these rates come from specially constructed alternatives whose height parameter is defined implicitly by solving an auxiliary equation for each sample size.

Core claim

Based on discrete observations, we develop a test to infer if the volatility function σ(·) within the nonparametric Gaussian white noise model dY_t = σ(t)dW_t is constant. The testing procedure is shown to be minimax-optimal and adaptive for infill asymptotics and these results entail that a deviation from the null hypothesis of constancy is best measured in terms of the ratio of σ(t) and its L²-average. The derivation of optimal constants requires the construction of hypotheses with height h(b), where the parameter b solves F_n(b)=0 for given functions F_n. Proving this equation to be solvable for each n∈N and establishing quantitative bounds of the solutions is built upon the implicit

What carries the argument

The family of alternative hypotheses whose height h(b) is chosen so that the auxiliary equation F_n(b)=0 holds, with solvability and bounds for b established by the implicit function theorem for every n.

Load-bearing premise

The observations are generated exactly from the Gaussian white noise model dY_t = σ(t) dW_t and the equation F_n(b)=0 admits a solution for every finite sample size n.

What would settle it

A Monte Carlo experiment in which the empirical power of the test lies strictly below the minimax boundary predicted by the solved value of b for alternatives scaled by the ratio σ(t) over its L2 average would refute the optimality claim.

Figures

Figures reproduced from arXiv: 2604.25668 by Johannes Brutsche, Lukas Riepl.

Figure 1
Figure 1. Figure 1: The sample path of a given Brownian motion (blue) together with the process view at source ↗
read the original abstract

Based on discrete observations, we develop a test to infer if the volatility function $\sigma(\cdot)$ within the nonparametric Gaussian white noise model $dY_t = \sigma(t)dW_t$ is constant. The testing procedure is shown to be minimax-optimal and adaptive for infill asymptotics and these results entail that a deviation from the null hypothesis of constancy is best measured in terms of the ratio of $\sigma(t)$ and its $L^2$-average. The derivation of optimal constants requires the construction of hypotheses with height $h(b)$, where the parameter $b$ solves $F_n(b)=0$ for given functions $F_n$. Proving this equation to be solvable for each $n\in\mathbb{N}$ and establishing quantitative bounds of the solutions is built upon the implicit function theorem.

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 / 1 minor

Summary. The manuscript develops a test for the null hypothesis that the volatility function σ(·) is constant in the nonparametric Gaussian white noise model dY_t = σ(t)dW_t observed at discrete times. It establishes minimax optimality and adaptivity of the procedure under infill asymptotics, showing that deviations from constancy are optimally detected via the ratio of σ(t) to its L²-average. The optimality constants are derived by constructing least-favorable alternatives using solutions b to F_n(b)=0, with existence and bounds justified by the implicit function theorem.

Significance. If the optimality and adaptation claims hold with the stated sharp constants, the work would advance nonparametric testing for stochastic volatility by providing adaptive procedures and a novel ratio-based characterization of the separation rate. The technical construction via the implicit function theorem for hypothesis generation is a potential strength if the uniform bounds are rigorously verified.

major comments (2)
  1. [Abstract] Abstract: The claim that existence for every n and quantitative bounds on solutions to F_n(b)=0 are obtained from the implicit function theorem requires explicit verification that |∂F_n/∂b| ≥ c > 0 uniformly in n near the root. The standard IFT guarantees only local existence and does not automatically supply the uniform-in-n control needed to justify the sharp minimax constants for the ratio-based separation in the infill regime.
  2. [Derivation of optimal constants] Derivation of optimal constants (via hypothesis construction with height h(b)): The minimax optimality and adaptation results rest on solvability of F_n(b)=0 for each n. Without a separate monotonicity argument or uniform derivative bound, the least-favorable alternatives may not yield the claimed sharp constants, undermining the central assertion that deviation is best measured by the ratio of σ(t) and its L²-average.
minor comments (1)
  1. [Model and setup] Clarify the precise statement of the infill asymptotic regime and the discrete observation scheme in the model definition to ensure consistency with the testing procedure.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments. The points raised about the application of the implicit function theorem and the justification of sharp constants are substantive and we address them directly below, indicating the revisions we will undertake.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The claim that existence for every n and quantitative bounds on solutions to F_n(b)=0 are obtained from the implicit function theorem requires explicit verification that |∂F_n/∂b| ≥ c > 0 uniformly in n near the root. The standard IFT guarantees only local existence and does not automatically supply the uniform-in-n control needed to justify the sharp minimax constants for the ratio-based separation in the infill regime.

    Authors: We agree that the standard implicit function theorem yields only local existence for each fixed n and that uniform-in-n control on the derivative is required to justify the sharp constants under infill asymptotics. In the manuscript the IFT is applied pointwise in n, with uniformity expected to follow from uniform convergence of F_n to a limiting F with F'(b*) bounded away from zero together with the explicit form of F_n. To make this rigorous we will insert a new lemma establishing the uniform lower bound |∂F_n/∂b| ≥ c > 0 for all sufficiently large n and b in a neighborhood of the root, using the explicit expression for F_n and standard domination arguments. This addition will supply the missing quantitative control without changing the main results. revision: yes

  2. Referee: [Derivation of optimal constants] Derivation of optimal constants (via hypothesis construction with height h(b)): The minimax optimality and adaptation results rest on solvability of F_n(b)=0 for each n. Without a separate monotonicity argument or uniform derivative bound, the least-favorable alternatives may not yield the claimed sharp constants, undermining the central assertion that deviation is best measured by the ratio of σ(t) and its L²-average.

    Authors: The referee is correct that both the uniform derivative bound and, where necessary, a monotonicity argument are needed to guarantee that the solutions b_n produce well-defined least-favorable alternatives with the claimed sharp constants. We will add the uniform derivative bound as described in the response to the first comment and, if a monotonicity statement is not already explicit, include a short argument showing that F_n is strictly monotone in a neighborhood of the root (again using the explicit form of F_n). These clarifications will confirm that the ratio-based separation rate is indeed optimal and that the adaptation results hold with the stated constants. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation applies external IFT to constructed hypotheses

full rationale

The paper defines hypotheses via height h(b) where b solves the newly introduced equation F_n(b)=0, then invokes the implicit function theorem (an external result) to prove solvability for each n and obtain quantitative bounds. This construction feeds into the minimax optimality and adaptation arguments for infill asymptotics, but does not reduce the final claims to a self-definition, a fitted parameter renamed as prediction, or a self-citation chain. The central result that deviations are best measured by the ratio of σ(t) to its L²-average follows from the asymptotic analysis of these alternatives and remains independent of the result itself. No load-bearing self-citations or ansatzes are present in the provided text.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on the Gaussian white noise model and the solvability of F_n(b)=0; no free parameters or invented entities are introduced in the abstract.

axioms (2)
  • domain assumption Observations follow the nonparametric Gaussian white noise model dY_t = σ(t) dW_t
    This is the explicit model stated as the basis for the testing procedure.
  • domain assumption Infill asymptotics govern the discrete observation scheme
    Required for the minimax and adaptive optimality statements.

pith-pipeline@v0.9.0 · 5423 in / 1283 out tokens · 123463 ms · 2026-05-07T14:33:33.517204+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

20 extracted references · 20 canonical work pages

  1. [1]

    Change-point inference on volatility in noisy I t\^o semimartingales

    Markus Bibinger and Mehmet Madensoy. Change-point inference on volatility in noisy I t\^o semimartingales. Stochastic Process. Appl., 129 0 (12): 0 4878--4925, 2019. ISSN 0304-4149,1879-209X. doi:10.1016/j.spa.2018.12.013. URL https://doi.org/10.1016/j.spa.2018.12.013

    work page doi:10.1016/j.spa.2018.12.013 2019

  2. [2]

    Nonparametric change-point analysis of volatility

    Markus Bibinger, Moritz Jirak, and Mathias Vetter. Nonparametric change-point analysis of volatility. Ann. Statist., 45 0 (4): 0 1542--1578, 2017. ISSN 0090-5364,2168-8966. doi:10.1214/16-AOS1499. URL https://doi.org/10.1214/16-AOS1499

    work page doi:10.1214/16-aos1499 2017

  3. [3]

    Sharp adaptive and pathwise stable similarity testing for scalar ergodic diffusions

    Johannes Brutsche and Angelika Rohde. Sharp adaptive and pathwise stable similarity testing for scalar ergodic diffusions. Ann. Statist., 52 0 (3): 0 1127--1151, 2024. ISSN 0090-5364,2168-8966. doi:10.1214/24-aos2386. URL https://doi.org/10.1214/24-aos2386

    work page doi:10.1214/24-aos2386 2024

  4. [4]

    A. C. Davison. Statistical models, volume 11 of Cambridge Series in Statistical and Probabilistic Mathematics. Cambridge University Press, Cambridge, 2003. ISBN 0-521-77339-3. doi:10.1017/CBO9780511815850. URL https://doi.org/10.1017/CBO9780511815850

    work page doi:10.1017/cbo9780511815850 2003

  5. [5]

    Testing the parametric form of the volatility in continuous time diffusion models---a stochastic process approach

    Holger Dette and Mark Podolskij. Testing the parametric form of the volatility in continuous time diffusion models---a stochastic process approach. J. Econometrics, 143 0 (1): 0 56--73, 2008. ISSN 0304-4076,1872-6895. doi:10.1016/j.jeconom.2007.08.002. URL https://doi.org/10.1016/j.jeconom.2007.08.002

    work page doi:10.1016/j.jeconom.2007.08.002 2008

  6. [6]

    On a test for a parametric form of volatility in continuous time financial models

    Holger Dette, von Lieres, and Carsten Wilkau. On a test for a parametric form of volatility in continuous time financial models. Finance Stoch., 7 0 (3): 0 363--384, 2003. ISSN 0949-2984,1432-1122. doi:10.1007/s007800200087. URL https://doi.org/10.1007/s007800200087

    work page doi:10.1007/s007800200087 2003

  7. [7]

    Estimation of integrated volatility in continuous-time financial models with applications to goodness-of-fit testing

    Holger Dette, Mark Podolskij, and Mathias Vetter. Estimation of integrated volatility in continuous-time financial models with applications to goodness-of-fit testing. Scand. J. Statist., 33 0 (2): 0 259--278, 2006. ISSN 0303-6898,1467-9469. doi:10.1111/j.1467-9469.2006.00479.x. URL https://doi.org/10.1111/j.1467-9469.2006.00479.x

    work page doi:10.1111/j.1467-9469.2006.00479.x 2006

  8. [8]

    David L. Donoho. Asymptotic minimax risk for sup-norm loss: solution via optimal recovery. Probab. Theory Related Fields, 99 0 (2): 0 145--170, 1994. ISSN 0178-8051,1432-2064. doi:10.1007/BF01199020. URL https://doi.org/10.1007/BF01199020

    work page doi:10.1007/bf01199020 1994

  9. [9]

    Spokoiny

    Lutz D\"umbgen and Vladimir G. Spokoiny. Multiscale testing of qualitative hypotheses. Ann. Statist., 29 0 (1): 0 124--152, 2001. ISSN 0090-5364,2168-8966. doi:10.1214/aos/996986504. URL https://doi.org/10.1214/aos/996986504

    work page doi:10.1214/aos/996986504 2001

  10. [10]

    umbgen and G\

    Lutz D\"umbgen and G\"unther Walther. Multiscale inference about a density, technical report. 2007. Available at: https://arxiv.org/abs/0706.3968v2

    work page arXiv 2007

  11. [11]

    umbgen and G\

    Lutz D\"umbgen and G\"unther Walther. Multiscale inference about a density. Ann. Statist., 36 0 (4): 0 1758--1785, 2008. ISSN 0090-5364,2168-8966. doi:10.1214/07-AOS521. URL https://doi.org/10.1214/07-AOS521

    work page doi:10.1214/07-aos521 2008

  12. [12]

    H\"ardle and E

    W. H\"ardle and E. Mammen. Comparing nonparametric versus parametric regression fits. Ann. Statist., 21 0 (4): 0 1926--1947, 1993. ISSN 0090-5364,2168-8966. doi:10.1214/aos/1176349403. URL https://doi.org/10.1214/aos/1176349403

    work page doi:10.1214/aos/1176349403 1926

  13. [13]

    URLhttps://link.springer.com/10.1007/978-3-030-61871-1

    Olav Kallenberg. Foundations of modern probability, volume 99 of Probability Theory and Stochastic Modelling. Springer, Cham, third edition, 2021. ISBN 978-3-030-61871-1; 978-3-030-61870-4. doi:10.1007/978-3-030-61871-1. URL https://doi.org/10.1007/978-3-030-61871-1

    work page doi:10.1007/978-3-030-61871-1 2021

  14. [14]

    Detection of a structural break in intraday volatility pattern

    Piotr Kokoszka, Tim Kutta, Neda Mohammadi, Haonan Wang, and Shixuan Wang. Detection of a structural break in intraday volatility pattern. Stochastic Process. Appl., 176: 0 Paper No. 104426, 15, 2024. ISSN 0304-4149,1879-209X. doi:10.1016/j.spa.2024.104426. URL https://doi.org/10.1016/j.spa.2024.104426

    work page doi:10.1016/j.spa.2024.104426 2024

  15. [15]

    Laurent and P

    B. Laurent and P. Massart. Adaptive estimation of a quadratic functional by model selection. Ann. Statist., 28 0 (5): 0 1302--1338, 2000. ISSN 0090-5364,2168-8966. doi:10.1214/aos/1015957395. URL https://doi.org/10.1214/aos/1015957395

    work page doi:10.1214/aos/1015957395 2000

  16. [16]

    Sergei L. Leonov. Remarks on extremal problems in nonparametric curve estimation. Statist. Probab. Lett., 43 0 (2): 0 169--178, 1999. ISSN 0167-7152,1879-2103. doi:10.1016/S0167-7152(98)00256-9. URL https://doi.org/10.1016/S0167-7152(98)00256-9

    work page doi:10.1016/s0167-7152(98)00256-9 1999

  17. [17]

    Adaptive goodness-of-fit tests based on signed ranks

    Angelika Rohde. Adaptive goodness-of-fit tests based on signed ranks. Ann. Statist., 36 0 (3): 0 1346--1374, 2008. ISSN 0090-5364,2168-8966. doi:10.1214/009053607000000992. URL https://doi.org/10.1214/009053607000000992

    work page doi:10.1214/009053607000000992 2008

  18. [18]

    Optimal calibration for multiple testing against local inhomogeneity in higher dimension

    Angelika Rohde. Optimal calibration for multiple testing against local inhomogeneity in higher dimension. Probab. Theory Related Fields, 149 0 (3-4): 0 515--559, 2011. ISSN 0178-8051,1432-2064. doi:10.1007/s00440-010-0263-1. URL https://doi.org/10.1007/s00440-010-0263-1

    work page doi:10.1007/s00440-010-0263-1 2011

  19. [19]

    2009 , isbn =

    Alexandre B. Tsybakov. Introduction to nonparametric estimation. Springer Series in Statistics. Springer, New York, 2009. ISBN 978-0-387-79051-0. doi:10.1007/b13794. URL https://doi.org/10.1007/b13794. Revised and extended from the 2004 French original, Translated by Vladimir Zaiats

    work page doi:10.1007/b13794 2009

  20. [20]

    A. W. van der Vaart and Jon A. Wellner. Weak convergence and empirical processes---with applications to statistics. Springer Series in Statistics. Springer, Cham, second edition, 2023. ISBN 978-3-031-29038-1; 978-3-031-29040-4. doi:10.1007/978-3-031-29040-4. URL https://doi.org/10.1007/978-3-031-29040-4

    work page doi:10.1007/978-3-031-29040-4 2023