|
on Econometrics |
| By: | Li-Hsien Sun; Zong-Yuan Huang; Yi-Ling Huang; Chi-Yang Chiu; Ning Ning |
| Abstract: | We study offline change-point estimation for time series data exhibiting nonlinear serial dependence. To address this problem, we propose a copula-based Markov chain model with Weibull marginal distributions, which is suitable for modeling nonnegative data such as event times and volatility measures. Nonlinear dependence is incorporated through the Clayton and Joe copulas, allowing the model to capture asymmetric lower-tail and upper-tail dependence structures, respectively. We derive the corresponding likelihood function and estimate the change point and model parameters using maximum likelihood estimation implemented through the Newton--Raphson algorithm. Confidence intervals are constructed via a parametric bootstrap Monte Carlo procedure. Extensive numerical studies are conducted to evaluate the finite-sample performance and robustness of the proposed method under different dependence structures and copula misspecification scenarios. The results demonstrate that the proposed estimators perform well in terms of RMSE and relative error, particularly for the estimation of the change point. An empirical application to the VIX index during the COVID-19 pandemic further illustrates the practical usefulness of the proposed approach in detecting structural changes in both the marginal distributions and serial dependence structure. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2605.29541 |
| By: | David J. Price |
| Abstract: | Power laws are common in economics, as in city and firm sizes, and can cause extreme heteroskedasticity. I show that estimators based on observations exhibiting this extreme heteroskedasticity may not be consistent or asymptotically normal and may have unreliable confidence intervals. These problems can occur even without heteroskedasticity if weighted estimators are used. I construct a quasi-maximum likelihood estimator to form more accurate estimates and more reliable inference. This estimator is broadly useful when weighting is considered to improve estimators' precision. Simulations confirm it improves estimation precision and inference, while a replication shows it can lead to substantially different results. |
| Keywords: | heteroskedasticity, power laws, asymptotics, quasi-maximum likelihood |
| JEL: | C12 C13 C18 |
| Date: | 2026–05–26 |
| URL: | https://d.repec.org/n?u=RePEc:tor:tecipa:tecipa-822 |
| By: | Giorgi Nikolaishvili (Wake Forest University) |
| Abstract: | Nonparametric local projections estimate impulse responses without imposing parametric assumptions on the response function. Existing plug-in implementations identify the response through a nonparametric regression of future outcomes on the structural shock. This paper shows that the same response function can also be identified by reweighting outcomes according to how a structural shock shifts the shock density. Combining the two representations yields a doubly robust estimator: a nonparametric regression estimate augmented with a residual correction based on shock density reweighting. Consistency requires only that either the outcome regression or the density ratio be consistently estimated, making the method less vulnerable to smoothing, approximation, and specification errors. The correction also improves the calibration of confidence intervals, both by reducing centering bias and by producing a score whose variation the standard error fully reflects. In simulations, the residual correction removes persistent regression bias and substantially improves empirical coverage. |
| Keywords: | local projections; double robustness; orthogonal estimation; impulse responses |
| JEL: | C14 C22 C32 |
| Date: | 2026–05–01 |
| URL: | https://d.repec.org/n?u=RePEc:ris:wfuewp:022595 |
| By: | Omer Faruk Akbal; Max-Sebastian Dovi |
| Abstract: | We study inference via heteroskedasticity in linear models commonly used for macroeconomic policy analysis, where covariate endogeneity must often be addressed with limited time and data. Our framework nests standard heteroskedasticity-based approaches, allows for new non-nested restrictions, and does not require ex-ante regime labelling. We propose an easily implementable weak-identification-robust test and derive sufficient conditions for its validity. Simulation results show good size and power properties in a wide range of settings. Empirical applications to the fuel-price passthrough in Sierra Leone, the effect of remittances on consumption in the Philippines, and exchange-rate passthroughs in many countries illustrate the versatility and scalability of our approach. |
| Keywords: | Heteroskedasticity; weak identification; Anderson-Rubin; two-step inference |
| Date: | 2026–05–22 |
| URL: | https://d.repec.org/n?u=RePEc:imf:imfwpa:2026/100 |
| By: | Chebbi, Ali |
| Abstract: | Standard trend–cycle extraction methods in macroeconomics rely on assumptions of global smoothness and time-invariant dynamics that become restrictive in the presence of structural breaks and regime-dependent behavior. These limitations affect both cyclical measurement and the stability of inferred long-run relationships under cointegration. This paper develops a fuzzy clustering–based filtering framework that provides a regime-sensitive decomposition of macroeconomic time series. While fuzzy methods have been used in applied filtering, their asymptotic properties—particularly under structural breaks and cointegration—have not been formally characterized. The proposed approach assigns observations probabilistically across latent regimes, allowing for smooth transitions and mixed states. We establish √T-consistency of the cyclical component, vanishing endpoint bias, and preservation of cointegrating relationships. The filter is embedded in a regime-dependent cointegrated system (MS-F-VECM), allowing both short-run dynamics and long-run equilibria to vary across regimes. Monte Carlo simulations confirm strong finite-sample performance in terms of break detection and cointegration preservation. An application to Eurozone data (1999–2023) shows that standard measures of comovement are not invariant but reflect regime-dependent aggregation. The contribution is methodological: a unified framework for regime-dependent filtering and cointegration analysis. |
| Keywords: | Fuzzy filtering; asymptotic theory; structural breaks; cointegration; regime dependence; nonlinear time series. |
| JEL: | C22 C32 C51 E32 |
| Date: | 2026–04–01 |
| URL: | https://d.repec.org/n?u=RePEc:pra:mprapa:128903 |
| By: | Gerdie Everaert (-) |
| Abstract: | Local projections (LPs) are a widely used method for estimating impulse responses. While LP estimators are consistent as the number of time periods T tends to infinity under standard conditions, they exhibit non-negligible bias in finite samples. This bias is particularly relevant in panel data settings, where the number of individuals N may be large but T relatively small. In this paper, we derive an analytical expression for the incidental-parameters bias in the LP estimator with individual fixed effects. We show that this bias exhibits a non-monotone pattern across the projection horizon: it increases at intermediate horizons, where it can substantially exceed the standard dynamic panel bias even for moderate T, before declining at longer horizons. We propose an iterative bias-correction procedure that, when suitably initialized, effectively eliminates the incidental-parameters bias across the entire projection horizon. |
| Keywords: | Local projections, panel data, fixed effects, incidental parameters bias, bias correction |
| JEL: | C32 C33 C13 |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:rug:rugwps:26/1145 |
| By: | Runyu Dai; Yasumasa Matsuda |
| Abstract: | Estimating large volatility matrices is essential to finance research in the era of big data. We propose a unified estimate that embeds a dynamic conditional correlation GARCH (DCC-GARCH) structure into a mixed factor model comprising both observable and weak latent factors, with the residual covariance estimated through an adaptively thresholded sparse matrix based on the extended Principal Orthogonal Complement Thresholding (ePOET) framework. The resulting method, termed DCC-ePOET, jointly captures pervasive signals from both types of factors and dynamic idiosyncratic co-volatilities. It resolves the singularity issue that arises in high-dimensional settings where the cross-sectional dimension N exceeds the serial dimension T, while remaining computationally feasible. Monte Carlo simulations confirm the good finite sample performance of DCC-ePOET across various dimensions. An out-of-sample minimum variance portfolio analysis using S&P 500 data demonstrates the usefulness of DCC-ePOET in practice. |
| Date: | 2026–05–17 |
| URL: | https://d.repec.org/n?u=RePEc:toh:dssraa:152 |
| By: | Giorgi Nikolaishvili (Wake Forest University); Noah D. Gade (Wake Forest University) |
| Abstract: | Impulse response analysis builds economic narratives by scanning a large set of coefficients for significant effects. Pointwise inference ignores this multiplicity, so the false rejection rate grows unbounded with the response family. Simultaneous inference bounds the probability of even a single false rejection, which yields increasingly uninformative results as the family expands. Researchers are left to choose between overstating their evidence and understating it. We propose false discovery and false coverage control as a more appropriate target: bounding the expected share of false rejections among responses declared significant, with calibrated post-selection confidence intervals. Neither guarantee deteriorates as the response family grows, so researchers are not penalized for investigating thoroughly. The procedure integrates into standard VAR and local projection bootstrap workflows. Applications show that this inference strategy recovers effects lost under simultaneous bands while discarding fragile pointwise findings, in some cases materially altering the economic narrative. |
| Keywords: | impulse response; multiple testing; false discovery; VAR; local projection |
| JEL: | C12 C22 C32 E00 |
| Date: | 2026–04–17 |
| URL: | https://d.repec.org/n?u=RePEc:ris:wfuewp:022594 |
| By: | Dongwoo Kim |
| Abstract: | Many selection problems are multilayered: agents first decide whether to participate and then sort among ordered or unordered categories. This paper shows that the sorting layer changes the geometry of identification. Unlike binary selection, in which selection bias can be summarized by a scalar control function, ordered and multinomial sorting generally produce multi-index control functions whose dimension determines the continuous covariate variation needed for identification. I establish matched non-identification and point-identification results for both architectures, showing how nonlinearity in the selection structure can substitute for excluded variables. I also show how additional structural restrictions reduce the control-function dimension and make estimation practical. I propose √n-consistent two-step sieve plug-in estimators and apply the framework to gender wage gaps among Korean college graduates. Accounting for sorting reshapes the entry-level gap along the firm-size margin, where the corrected female coefficient turns positive for large-firm employment. |
| Date: | 2026–05–27 |
| URL: | https://d.repec.org/n?u=RePEc:azt:cemmap:10/26 |
| By: | Vidal Llauradó, Joan |
| Abstract: | This paper develops a feasible estimator for the source-screened latent contagion object isolated in the first two papers and applies it to a balanced Oxford-Man realized-volatility panel of eight global equity indices. Starting from the reduced local Gaussian block experiment, it represents local alternatives by covariance derivatives, removes the target-only tangent space, and estimates the remaining source-screened component with a low-dimensional projected covariance-score GMM statistic. The paper derives the projected-score geometry, proves the associated local Gaussian efficiency, rough-regime projected-rank, pilot-adaptive transfer, and uniform minimax results, and validates the implementation in synthetic experiments using closed-form information and noncentrality constants. In the Oxford-Man application, estimated physical-measure roughness lies between about 0.04 and 0.09 across the panel, with H_P approximately 0.071 for SPX, while the full-sample directed contagion map is dense and economically informative through intensity ranking and rolling stability rather than sparse edge selection. The paper closes the trilogy with a feasible estimator, a validation protocol, and a real-data physical-measure application, while leaving matched option-panel P/Q classification for later work. |
| Keywords: | latent volatility contagion; projected score estimator; covariance score GMM; source-screened inference; rough volatility; realized volatility; Oxford-Man realized library; local Gaussian experiments; nuisance-orthogonal estimation; financial econometrics |
| JEL: | C13 C14 C58 G12 G17 |
| Date: | 2026–04–12 |
| URL: | https://d.repec.org/n?u=RePEc:pra:mprapa:128738 |
| By: | Rob J. Hyndman; David T. Frazier |
| Abstract: | Anomaly detection methods are widely used but often rely on ad hoc rules or strong assumptions, and they often focus on tail events, missing inlier anomalies that occur in low-density gaps between modes. We propose a unified framework that defines an anomaly as an observation with unusually low probability under a possibly misspecified model. For each observation we compute its surprisal, defined as the negative log generalized density, and define an anomaly score as the probability of a surprisal at least as large as that observed. This reduces anomaly detection for complex univariate or multivariate data to estimating the upper tail of a univariate surprisal distribution. We develop two model-robust estimators of these tail probabilities: an empirical estimator based on the observed surprisal distribution and an extreme-value estimator that fits a Generalized Pareto Distribution above a high threshold. Simulations and applications to French mortality and Test-cricket data show the approach remains effective under substantial model misspecification. |
| Keywords: | anomaly detection, surprisal, outlier detection, generalized Pareto distribution, extreme value theory, tail bounds |
| JEL: | C10 C14 C46 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:msh:ebswps:2026-3 |
| By: | Beniamino Hadj-Amar; Jack Jewson |
| Abstract: | The Bayesian approach provides powerful methods for variable selection. The ability to incorporate sparsity through prior beliefs and account for parameter uncertainty allows Bayesian variable selection to consistently identify which variables are active and exhibit strong finite-sample performance. However, Bayesian methods require the correct specification of full likelihoods for the data, and there is increasing awareness of the problems that model misspecification causes for variable selection. Current approaches to mitigate misspecification either require complex models, detracting from the interpretability of the variable selection task, or move outside rigorous Bayesian uncertainty quantification and provide no recognised method for variable selection. This paper establishes the model quasi-posterior as a principled tool for variable selection. |
| Keywords: | Bayesian variable selection, quasi-posterior, model misspecification, sparse regression, posterior consistency, Bayesian inference |
| JEL: | C11 C15 C52 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:msh:ebswps:2026-2 |
| By: | Jin Seo Cho (Yonsei University); Peter C. B. Phillips (Yale University, University of Auckland & Singapore Management University) |
| Abstract: | In GMM estimation it is well known that if the number of moment conditions grows with the sample size, GMM asymptotics differ from the standard case with moment size fixed as the sample size tends to infinity. The present work explores infinite dimensional GMM estimation under various conditions on the moment conditions and the weight matrix. Our approach employs a partial sum process formed by the moment conditions to represent high dimensional moments and an invariance principle to capture the infinite dimensional asymptotics as the moment size grows. Next, the GMM weight matrix is assumed to converge to one of two kernels at the limit: a continuous kernel or the Dirac delta function. Combining these different conditions enables development of a large sample theory for most efficient GMM estimation. The effects of permuting the moment conditions on GMM efficiency are also explored. The resulting theory is applied to weak instrumental variable estimation and the Angrist and Krueger(1991) data are re-analyzed in an empirical application of the new methods. |
| Keywords: | Infinite dimensional GMM; Invariance principle; Neumann’s series expansion; Stochastic integral; Weak IV, 2SLS. |
| JEL: | C13 C18 C36 C55 E24 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:yon:wpaper:2026rwp-289 |
| By: | De Jonghe, Olivier; Lewis, Daniel |
| Abstract: | We propose a new model in which relationship-specific effects or shocks are identified in a bipartite network under mild covariance restrictions, generalizing the influential Abowd et al. (1999) framework. For example, separate demand shocks are identified for each bank from which a firm borrows. We show how previous approaches break down when confronted with such heterogeneity, while our novel identification strategy yields a simple estimator that is consistent and asymptotically normal, under weaker network density assumptions than previous approaches. The methodology performs well in empirically-calibrated simulations. We apply our approach to identify relationship-level credit demand and supply shocks for thousands of firms and banks across nine Euro-area countries and three distinct economic episodes. We formally reject the Abowd et al. (1999) assumptions in nearly every country-period and show that within-firm/bank shock variation is of comparable scale to between firm/bank variation. We document considerable bias in Abowd et al. (1999) style estimates and associated regressions, while finding significant deleterious effects of the post-2022 monetary contraction on exposed firms. We highlight novel heterogeneity in the transmission of monetary policy. JEL Classification: C33, C58, E44, G21, G30 |
| Keywords: | corporate credit, demand shock, higher moments, identification, networks, networks, supply shock, two-way fixed effects |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:ecb:ecbwps:20263238 |
| By: | Jaap H. Abbring; Yifan Yu |
| Abstract: | This paper studies interdependent durations as equilibrium outcomes of a synchronization game, a continuous-time stopping game in which the incentive to stop increases when other players stop. We allow the payoffs to vary with both common shocks and observed and unobserved agent characteristics. The common shocks follow a spectrally negative L\'evy process, a semiparametric process that includes Brownian motion as a special case but may also have jumps. We show that equilibrium outcomes can be represented as interdependent hitting times and use this to establish the game's nonparametric identification from data on stopping times and covariates. We develop maximum simulated likelihood and method of simulated moments estimators and evaluate their finite-sample and computational performance in Monte Carlo experiments. The results provide a tractable framework for identifying and estimating synchronization games from interdependent duration data. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.06251 |
| By: | Zijun Jia; Sharon X. Lee |
| Abstract: | Risk management is an important part of financial practice, essential for protecting assets and investments in modern-day volatile markets. This paper proposes a mixture of mirrored Weibull (MMW) distribution for modelling stock returns and estimating risk measures. Unlike common practices which are typically based on the normal distribution, the MMW model can flexibly accommodate non-normal features frequently exhibited in financial data. It also enjoys appealing properties such as having a simple density expression and fast parameter estimation. We demonstrate the effectiveness of our model by assessing its performance in Value-at-Risk (VaR) estimation of three S&P500 stocks. The MMW model compares favourably to Gaussian mixture model and t-mixture model, with significant improvements in VaR estimation and prediction. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2605.20142 |
| By: | Costa, Manon; Gadat, Sébastien; Gendre, Xavier; Klein, Thierry |
| Abstract: | The main objective of this paper is to propose a new approach for estimating the entire collection of Sobol’ indices simultaneously. Our approach exploits the fact that Sobol’ indices can be rewritten as solutions to an optimization problem over a simplex, to construct an online sequence of estimators using a stochastic mirror descent algorithm. We prove that our estimation procedure is consistent and provide a non-asymptotic upper bound for its rate of convergence. Furthermore, we demonstrate the numerical accuracy of our method and compare it with other classical estimation procedures. |
| Keywords: | global sensitivity analysis; Sobol’ indices; stochastic Mirror descent algorithm |
| Date: | 2026–06–01 |
| URL: | https://d.repec.org/n?u=RePEc:tse:wpaper:131794 |
| By: | Dang, Hai-Anh H.; Do, Minh; Lahiri, Partha; Gualavisi, Melany; Newhouse, David; Kilic, Talip; Lanjouw, Peter; Van der Weide, Roy |
| Abstract: | This paper uses five rounds of Mexican and Brazilian census extracts to evaluate the accuracy of different model specifications and estimation methods that use survey and census data to generate small area estimates of poverty. Models that utilize more granular data for prediction (household- and/or village-level predictors) tend to produce more accurate estimates of poverty than models estimated only using area-level predictors. Differences in accuracy across models and methods that utilize household or village level predictors are minor. Models that omit household-level predictors tend to be more robust than unit-level models to the use of old census data and classical measurement error in survey predictors. The performance of the Fay-Herriot area-level model falls in the presence of sample selection bias and small sample sizes. Rescaling sample weights is important in Mexico, where the sample is informative within areas. Applying raw sample weights without rescaling in this case greatly reduces the accuracy of estimates from linear models and distorts methodological comparisons. Overall, no one approach dominates across all contexts, but when sample weights are rescaled there is no downside to using more granular data for prediction. |
| Date: | 2026–05–27 |
| URL: | https://d.repec.org/n?u=RePEc:wbk:wbrwps:11396 |
| By: | Li, Z. M.; Linton, O. B.; Zhai, Y.; Zhang, H. |
| Abstract: | We propose a new family of liquidity measures—including order imbalance metrics—based on the dispersion and persistence of transitory gaps between transaction prices and the underlying efficient price. We devise an estimation method that renders these latent gaps observable, allowing plug-in estimates of the new measures from intraday trades alone, along with an inference method that allows us to quantify the sampling uncertainty in our estimates. We apply the approach to the S&P 500 equity portfolio, as well as to individual stocks. We use event study methodology to capture heterogeneous liquidity responses to FOMC announcements, which reveals distinct order-persistence patterns on surprise versus non-surprise days, highlighting how markets anticipate and react to monetary policy via the liquidity channel. |
| Keywords: | Market liquidity, FOMC Announcements, Spot Estimation, Monetary Policy Surprises, Order Imbalance, High-Frequency Identification |
| Date: | 2026–01–18 |
| URL: | https://d.repec.org/n?u=RePEc:cam:camdae:2639 |
| By: | Jaqueson K. Galimberti |
| Abstract: | This paper provides an analysis of the relevance of initial states for the estimation of natural rates of interest. It focuses on U.S. data using an established semi-structural macroeconomic model with time-varying trends. Alternative methods for the specification of initial states are reviewed and evaluated. The results indicate that initial states can significantly impact end-of-sample estimates of the natural rate of interest, with alternative initials leading to estimates about 40 to 130 basis points lower than original estimates. Re-estimating the model with alternative initials also leads to more volatile natural rate estimates. Key dimensions of the initialization issue are discussed, including the uncertainty around initial estimates, the use of diffuse prior initials, and jointly estimated initials. An extension of the original method using an unobserved component model that makes all initial estimates data-dependent is found to provide the most robust model and state estimates relative to varying sample definitions. |
| Keywords: | monetary policy stance, state space models, filtering, uncertainty |
| JEL: | C32 E43 E52 |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:een:camaaa:2026-38 |
| By: | De Castro, Yohann; Gadat, Sébastien; Marteau, Clément |
| Abstract: | We investigate the global optimization of the objective function arising in continuous sparse regression, specifically the Beurling LASSO (BLASSO), over the space of measures. While Conic Particle Gradient Descent (CPGD) methods are computationally efficient, they may become trapped in local minima due to the non-convexity of the parameterization. To overcome this limitation, we introduce Fast Spawn & Prune (FS&P), a stochastic algorithm that extends FastPart introduced in De Castro et al. (2025a) and combines CPGD with a birth–death process. The birth mechanism ensures asymptotic global exploration by introducing particles in regions where first-order optimality conditions are violated, while the death process preserves computational efficiency by pruning non-informative particles. We provide the first theoretical guarantee of global convergence for this class of discrete-time stochastic algorithms, without requiring exponentially large initializations. Furthermore, we derive convergence rates for the excess risk, thereby quantifying the trade-off between global exploration and local refinement. Moreover, we also propose a horizon-free variant that does not require prior knowledge of the iteration budget. |
| Keywords: | continuous sparse regression; conic particle gradient descent; birth and death; process; global convergence; stochastic optimization |
| Date: | 2026–06–01 |
| URL: | https://d.repec.org/n?u=RePEc:tse:wpaper:131793 |
| By: | Otneim, Håkon (Dept. of Business and Management Science, Norwegian School of Economics) |
| Abstract: | This paper addresses grade comparability across exam cohorts when assessors and item sets change from year to year. Ordinal item scores reflect a mixture of student ability, item difficulty, and assessor severity; separating these components requires linking assumptions rarely verified empirically. We fit a sequence of Bayesian cumulative logit models to item-level scores from nine cohorts of an undergraduate statistics course. The setting is fully observational with no cross-grading and only partial assessor overlap, so cross-cohort alignment relies on repeated content used as anchors and on shared assessors. Sequential model expansion guided by posterior predictive checks reveals that treating anchors as having fixed difficulty across cohorts is inconsistent with the data. A soft-linking formulation, where linked items share a baseline difficulty but admit cohort-specific deviations regularised toward zero, removes the systematic misfit without discarding anchor information. Approximate cross-validation confirms that each modelling step improves out-of-sample predictive accuracy. Student ability estimates are robust to anchor specification (pairwise correlations exceeding 0.996), whereas cohort location estimates shift materially, which is a finding with direct consequences for grading policy. Using the recovered ability scale, we construct counterfactual grades and show that assessor severity is the dominant predictor of individual grade movement. |
| Keywords: | comparable grading; many-facet measurement; soft anchors; cumulative logit model; Bayesian hierarchical models; PSIS-LOO cross-validation |
| JEL: | C11 C25 C52 I21 I23 |
| Date: | 2026–05–22 |
| URL: | https://d.repec.org/n?u=RePEc:hhs:nhhfms:2026_006 |
| By: | Jayesh Chaudhary |
| Abstract: | Financial volatility exhibits substantial non-stationarity, making single-regime models inadequate for characterising changing market conditions. This paper proposes a triple-timeframe Markov-Switching GARCH (MS-GARCH) framework for volatility regime detection in EUR/USD across daily, four-hour, and hourly horizons. Three independent AR(1)-MS-GARCH models are estimated to capture macro, meso, and micro regime dynamics, while Filardo-style time-varying transition probabilities (TVTP) are incorporated at the shorter horizons through composite stress indicators. The resulting regime probabilities are combined through an outer-product construction into a 27-state cross-scale probability tensor. Using EUR/USD data from 2015-2025, the framework produces statistically distinct Calm, Turbulent, and Crisis regimes and achieves superior out-of-sample volatility forecasting performance relative to a conventional GARCH benchmark. The results suggest that volatility dynamics contain meaningful structure across multiple timescales and that modelling these scales separately provides a more informative representation of market conditions than a single-timescale approach. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.06190 |