|
on Econometrics |
| By: | Alexander Mayer; Tatsushi Oka; Dominik Wied |
| Abstract: | We propose a score test for dependence predictability in conditional copulas that is robust to temporal instabilities. Our semiparametric procedure accommodates flexible dynamics in the marginal processes and remains agnostic about the copula family by leveraging distributional regression techniques together with a local Gaussian representation of the copula link function. We derive the limiting distribution of our test statistic and propose a resampling scheme based on recent results for the moving block bootstrap of multi-stage estimators. Monte Carlo simulations and an empirical application illustrate the finite-sample performance of our methods. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.01721 |
| By: | Bruno Ferman |
| Abstract: | Design-based simulations - procedures that hold realized outcomes fixed and generate variation by resampling treatment assignment or shocks - are widely used in both methodological and applied work to assess inference procedures. This paper studies the extent to which such simulations are informative about inference validity. Focusing on shift-share designs, we show that standard simulations that fix outcomes and resample shocks may rely on a data-generating process that is not aligned with the true one. In particular, these simulations confound true treatment effects with error dependence, potentially overstating inference distortions due to spatial correlation. We propose alternative simulation designs that circumvent this problem and illustrate their use in prominent empirical applications. Our results highlight that the usefulness of design-based simulations depends critically on how closely the simulated data-generating process aligns with the true one. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.11381 |
| By: | Stephen Salerno; Zhenke Wu; Tyler McCormick |
| Abstract: | When outcome data are expensive or onerous to collect, scientists increasingly substitute predictions from machine learning and AI models for unlabeled cases, a process which has consequences for downstream statistical inference. While recent methods provide valid uncertainty quantification under independent sampling, real-world applications involve missing at random (MAR) labeling and spatial dependence. For inference in this setting, we propose a doubly robust estimator with cross-fit nuisances. We show that cross-fitting induces fold-level correlation that distorts spatial variance estimators, producing unstable or overly conservative confidence intervals. To address this, we propose a jackknife spatial heteroscedasticity and autocorrelation consistent (HAC) variance correction that separates spatial dependence from fold-induced noise. Under standard identification and dependence conditions, the resulting intervals are asymptotically valid. Simulations and benchmark datasets show substantial improvement in finite-sample calibration, particularly under MAR labeling and clustered sampling. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.11368 |
| By: | Abid Hussain; Michail Tsagris |
| Abstract: | Detecting simultaneous shifts in location and scale between two populations is a common challenge in statistical inference, particularly in fields like biomedicine where right-skewed data distributions are prevalent. The classical Lepage test, which combines the Wilcoxon-Mann-Whitney and Ansari-Bradley tests, can be suboptimal under these conditions due to its restrictive assumptions of equal variances and medians. This study systematically evaluates enhanced Lepage-type test statistics that incorporate modern robust components for improved performance with right-skewed data. We combine the Fligner-Policello test and Fong-Huang variance estimator for the location component with a novel empirical variance estimator for the Ansari-Bradley scale component, relaxing assumptions of equal variances and medians. Extensive Monte Carlo simulations across exponential, gamma, chi-square, lognormal, and Weibull distributions demonstrate that tests incorporating both robust components achieve power |
| Keywords: | Fligner-Policello test; Fong-Huang estimator; Lepage-type statistics; location-scale model; robust inference; right-skewed data; variance estimation |
| JEL: | C12 C14 C15 |
| Date: | 2026–03–07 |
| URL: | https://d.repec.org/n?u=RePEc:crt:wpaper:2601 |
| By: | Alexander Aue; Sebastian K\"uhnert; Gregory Rice; Jeremy VanderDoes |
| Abstract: | AutoRegressive Conditional Heteroscedasticity (ARCH) models are standard for modeling time series exhibiting volatility, with a rich literature in univariate and multivariate settings. In recent years, these models have been extended to function spaces. However, functional ARCH and generalized ARCH (GARCH) processes established in the literature have thus far been restricted to model ``pointwise'' variances. In this paper, we propose a new ARCH framework for data residing in general separable Hilbert spaces that accounts for the full evolution of the conditional covariance operator. We define a general operator-level ARCH model. For a simplified Constant Conditional Correlation version of the model, we establish conditions under which such models admit strictly and weakly stationary solutions, finite moments, and weak serial dependence. Additionally, we derive consistent Yule--Walker-type estimators of the infinite-dimensional model parameters. The practical relevance of the model is illustrated through simulations and a data application to high-frequency cumulative intraday returns. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.10272 |
| By: | Michail Tsagris; Yannis Pantazis |
| Abstract: | Compositional data–vectors of non-negative components summing to unity–frequently arise in scientific applications where covariates influence the relative proportions of components, yet traditional regression approaches face challenges regarding the unit-sum constraint and zero values. This paper revisits the α–regression framework, which uses a flexible power transformation parameterized by α to interpolate between raw data analysis and log-ratio methods, naturally handling zeros without imputation while allowing data-driven transformation selection. We formulate α–regression as a non-linear least squares problem, study its asymptotic properties, provide efficient estimation via the Levenberg-Marquardt algorithm, and derive marginal effects for interpretation. |
| Keywords: | compositional data, α–transformation, spatial regression |
| JEL: | C21 C31 C51 R15 |
| Date: | 2026–03–07 |
| URL: | https://d.repec.org/n?u=RePEc:crt:wpaper:2603 |
| By: | Dante Amengual (CEMFI, Centro de Estudios Monetarios y Financieros); Gabriele Fiorentini (Università di Firenze and RCEA); Enrique Sentana (CEMFI, Centro de Estudios Monetarios y Financieros) |
| Abstract: | The EM principle implies the moments underlying the information matrix test for switching regressions are the expectation given the data of the moments one would test if one knew the subpopulation each observation originated from. Thus, we identify components related to conditional heteroskedasticity, conditional and unconditional skewness, and unconditional kurtosis of regression residuals within each regime. Simulations indicate analytical expressions for the asymptotic covariance matrix of those moments adjusted for sampling variability in parameter estimators provide reliable finite sample sizes and good power against various alternatives, especially combined with the parametric bootstrap. We apply the test to cross-country convergence regressions. |
| Keywords: | Asymmetry, convergence regressions, expectation - maximization principle, heteroskedasticity, incomplete data, kurtosis. |
| JEL: | C24 C34 C52 O47 |
| Date: | 2026–02 |
| URL: | https://d.repec.org/n?u=RePEc:cmf:wpaper:wp2026_2601 |
| By: | Han Chen (College of Finance and Statistics, Hunan University); Yijie Fei (College of Finance and Statistics, Hunan University); Jun Yu (Faculty of Business Administration, University of Macau) |
| Abstract: | Modeling the dynamics of correlations of multiple time series is an important yet difficult task, especially when the dimension is not confined to be low. In this paper, we propose a new multivariate stochastic volatility model featuring a block correlation structure. Our specification is built upon the new parametrization of the correlation matrix of Archakov & Hansen (2021) and extends the MSV-GFT model introduced in Chen et al. (2025). A Particle Gibbs Ancestor Sampling (PGAS) method is proposed to conduct the Bayesian analysis. It is shown to perform well for our model in finite samples. An empirical application based on a dozen U.S. stocks shows that our new model outperforms alternative specifications in terms of both the in-sample performance and the out-of-sample performance. |
| Keywords: | Block correlation matrix; Generalized Fisher transformation; Markov chain Monte Carlo; Multivariate stochastic volatility; Particle filter |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:boa:wpaper:202638 |
| By: | Fotso, Chris Toumping; Özer, Yeliz; Palumbo, Dario; Sibbertsen, Philipp |
| Abstract: | A dynamic modelling for heavy-tailed cylindrical time series is developed by combining score-driven models with a generalised Pareto-type cylindrical distribution. The proposed specification extends existing cylindrical models by allowing location, scale, concentration, andcrucially, the tail index of the linear component through the conditional distribution of speed to vary according to its score. Whereas the Weibull-von Mises model, whose linear componentexhibits exponentially decaying tails, the GPar specification admits polynomial tail decay. An explicit expression for the time-varying circular-linear dependence measure is also derived. The methodology is applied to high-frequency data from two onshore wind turbines in Germany. The empirical results indicate that allowing time-varying tail thickness leads to overall improvements compared to the Weibull-von Mises model. The proposedmodelprovidesaflexibleandcomputationallytractableframeworkforanalysing heavy-tailed cylindrical time series in environmental and energy applications. |
| Keywords: | cylindrical distributions, dynamic correlation, generalised Pareto, score-driven models, Weibull-von Mises, wind energy. |
| JEL: | C13 C18 C22 C46 Q42 |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:han:dpaper:dp-745 |
| By: | Milos Ciganovic; Federico D'Amario; Massimiliano Tancioni |
| Abstract: | We modify the Double Machine Learning estimator to broaden its applicability to macroeconomic time-series settings. A deterministic cross-fitting step, termed Reverse Cross-Fitting, leverages the time-reversibility of stationary series to improve sample utilization and efficiency. We detail and prove the conditions under which the estimator is asymptotically valid. We then demonstrate, through simulations, that its performance remains valid in realistic finite samples and is robust to model misspecification and violations of assumptions, such as heteroskedasticity. In high dimensions, predictive metrics for tuning nuisance learners do not generally minimize bias in the causal score. We propose a calibration rule targeting a "Goldilocks zone", a region of tuning parameters that delivers stable, partialled-out signals and reduced small-sample bias. Finally, we apply our procedure to residualized Local Projections to estimate the dynamic effects of a rise in Tier 1 regulatory capital. The results underscore the usefulness of the methodology for inference in macroeconomic applications. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.10999 |
| By: | Alexander Chudik; Lutz Kilian |
| Abstract: | This paper proposes mean group and pooled estimators of impulse responses based on mixed-frequency auxiliary distributed lag (DL), autoregressive distributed lag (ARDL) or vector autoregressive distributed lag (VARDL) estimating equations. Our setup assumes that the data are generated by a high-frequency VAR process. While the shock of interest is directly observed at high frequency, the outcome variable is only observed as a temporally aggregated variable at a lower frequency. We derive the asymptotic distributions of the six proposed estimators. Monte Carlo experiments show that pooled estimators generally perform better than the corresponding mean group estimators for relevant sample sizes. An empirical illustration to the pass-through from daily wholesale gasoline price shocks to monthly consumer price inflation illustrates the usefulness of the proposed methods. |
| Keywords: | Mixed frequencies; temporal aggregation; impulse responses; shock sequences; distributed lag (DL); autoregression distributed lag (ARDL); vector autoregression distributed lag (VARDL) |
| Date: | 2026–02–17 |
| URL: | https://d.repec.org/n?u=RePEc:fip:feddwp:102857 |
| By: | Lucas Kock; David T. Frazier; Michael Stanley Smith; David J. Nott |
| Abstract: | Copula models of multivariate data are popular because they allow separate specification of marginal distributions and the copula function. These components can be treated as inter-related modules in a modified Bayesian inference approach called ''cutting feedback'' that is robust to their misspecification. Recent work uses a two module approach, where all $d$ marginals form a single module, to robustify inference for the marginals against copula function misspecification, or vice versa. However, marginals can exhibit differing levels of misspecification, making it attractive to assign each its own module with an individual influence parameter controlling its contribution to a joint semi-modular inference (SMI) posterior. This generalizes existing two module SMI methods, which interpolate between cut and conventional posteriors using a single influence parameter. We develop a novel copula SMI method and select the influence parameters using Bayesian optimization. It provides an efficient continuous relaxation of the discrete optimization problem over $2^d$ cut/uncut configurations. We establish theoretical properties of the resulting semi-modular posterior and demonstrate the approach on simulated and real data. The real data application uses a skew-normal copula model of asymmetric dependence between equity volatility and bond yields, where robustifying copula estimation against marginal misspecification is strongly motivated. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.11457 |
| By: | Francesco Giancaterini; Alain Hecq; Joann Jasiak; Aryan Manafi Neyazi |
| Abstract: | This paper introduces a regularized test of the null hypothesis of the absence of linear and nonlinear serial dependence for high-dimensional non-Gaussian time series. Our approach extends the portmanteau test introduced in Jasiak and Neyazi (2023) to the high-dimensional setting. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.10152 |
| By: | Renan Serenini (University of Rome); Frantisek Masek (National Bank of Slovakia) |
| Abstract: | We propose a spatial extension of the Synthetic Difference-in-Differences (Sy- DiD) estimator developed by Arkhangelsky et al. (2021). Our estimator addresses violations of the StableUnit TreatmentValue Assumption (SUTVA) that arisewhen treatment effects spill over to other units. Spillovers to units in the donor pool can lead to biased and inconsistent estimates of the Average Treatment Effect on the Treated (ATT), while spillovers outside the donor pool leave the ATT identi- fiable but prevent identification of the Average Treatment Effect (ATE). Building on the framework of the Spatial Difference-in-Differences estimator introduced by Delgado and Florax (2015), we develop a method that decomposes the ATE into direct (ATT) and indirect treatment effects. We demonstrate that our esti- mator improves the identification of the indirect effect relative to the standard Spatial Difference-in-Differences approach, while retaining the robustness and favorable properties of the Synthetic Difference-in-Differences method for esti- mating the direct effect. |
| JEL: | C21 C23 D62 I18 |
| Date: | 2025–08 |
| URL: | https://d.repec.org/n?u=RePEc:svk:wpaper:1126 |
| By: | Antoine Burg (CEREMADE - CEntre de REcherches en MAthématiques de la DEcision - Université Paris Dauphine-PSL - PSL - Université Paris Sciences et Lettres - CNRS - Centre National de la Recherche Scientifique, SCOR SE [Paris]); Christophe Dutang (ASAR - Applied Statistics And Reliability - ASAR - LJK - Laboratoire Jean Kuntzmann - Inria - Institut National de Recherche en Informatique et en Automatique - CNRS - Centre National de la Recherche Scientifique - UGA - Université Grenoble Alpes - Grenoble INP - Institut polytechnique de Grenoble - Grenoble Institute of Technology - UGA - Université Grenoble Alpes) |
| Abstract: | The maximum likelihood estimator (MLE) remains the most frequently used method to estimate the parameters of generalized linear models. But even for distributions within the exponential family, MLEs are not always tractable and need to be computed with time consuming numerical methods like the Iterative Weighted Least Square algorithm. In order to improve the computation time, closed-form estimators have been found in case of categorical explanatory variables for univariate random variables of one-parameter exponential type. In the context of multivariate generalized linear models (MGLM), we propose a new way to look at the score in case of single categorical variables for any distribution in the exponential family. Firstly, we derive closed-form MLE for MGLM assuming multinomial and negative multinomial distributions. Secondly, we deduce similar results for the multivariate normal distributions. For the Dirichlet distribution, we propose a closed-form estimator, yet not MLE, for which we prove the consistency. We illustrate the computation time gains on simulated datasets: closed-form estimators are about 1000 times faster, especially for high dimension. Closed-form estimators are computed in constant times.. Finally, we show the relevancy of the proposed estimator on real-world datasets by modeling cause-of-death mortality in US. We are able to catch the first-order effects of covid between 2019 and 2021. |
| Keywords: | maximum likelihood estimation, multi-parameter exponential family, closed-form estimators, computation time, actuarial mortality model |
| Date: | 2026–02–28 |
| URL: | https://d.repec.org/n?u=RePEc:hal:journl:hal-05539060 |
| By: | Campbell R. Harvey; Alessio Sancetta; Yuqian Zhao |
| Abstract: | Researchers generally acknowledge that statistical tests must be adjusted when hundreds of factors and trading strategies have been examined. But how should these adjustments be made? Existing methods are often misunderstood or misapplied. We show that proper inference requires accounting for dependence across tests, correctly specifying the null distribution, and mitigating sample-selection bias. We develop a simple framework that avoids assumptions about the total number of tests run and yields a lower bound on valid significance thresholds - implying that researchers should employ a t-statistic cutoff of at least 3.0. In addition, we advocate using the local False Discovery Rate, which provides the probability that the null hypothesis is true for a given test-statistic realization - information that a conventional p-value cannot supply. |
| JEL: | C12 C58 G11 G12 |
| Date: | 2026–02 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:34898 |
| By: | Niko Hauzenberger Massimiliano Marcellino Michael Pfarrhofer Anna Stelzer |
| Abstract: | We develop Bayesian machine learning methods for mixed frequency data. This involves handling frequency mismatches and specifying functional relationships between (possibly many) predictors and low frequency dependent variables. We use Gaussian Processes (GPs) in direct nonlinear predictive regressions, and compress higher frequency variables in a structured way. This yields a set of kernels for GPs with distinct properties and implications. We evaluate the proposed framework in an out-of-sample exercise focusing on quarterly US GDP growth and inflation. Our approach leverages high-dimensional mixed frequency data in a computationally efficient way, and offers robustness and gains in predictive accuracy along several dimensions. |
| Keywords: | Bayesian nonparametrics, direct forecasting, nowcasting, dimension reduction, MIDAS |
| JEL: | C11 C22 C53 E31 E37 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:baf:cbafwp:cbafwp26265 |
| By: | Yuichiro Otani |
| Abstract: | Local regression is widely used to explore spatial heterogeneity, but anisotropic or effectively low-dimensional neighborhoods can produce ill-conditioned local solves, causing coefficient variation driven by numerical artifacts rather than substantive structure. Such instability is often hidden when estimation relies on implicit tuning or optimization without exposing local diagnostics. This paper proposes Gimbal Regression (GR), a deterministic, geometry-aware local regression framework for stable and auditable estimation. GR constructs directional weights from neighborhood geometry using explicit orientation objects and deterministic safeguards, and computes local coefficients by a closed-form solve. Theoretical results are stated conditional on the realized neighborhood configuration, under which the estimator is a deterministic linear operator with finite-perturbation stability bounds. Simulations and empirical examples demonstrate predictable computation, transparent diagnostics, and improved numerical stability relative to common local regression baselines. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.10382 |
| By: | Michail Tsagris; Omar Alzeley |
| Abstract: | Simplicia-simplicial regression concerns statistical modeling scenarios in which both the predictors and the responses are vectors constrained to lie on the simplex. Fiksel et al. (2022) introduced a transformationfree linear regression framework for this setting, wherein the regression coefficients are estimated by minimizing the Kullback-Leibler divergence between the observed and fitted compositions, using an expectation-maximization (EM) algorithm for optimization. In this work, we reformulate the problem as a constrained logistic regression model, in line with the methodological perspective of Tsagris (2025), and we obtain parameter estimates via constrained iteratively reweighted least squares. Simulation results indicate that the proposed procedure substantially improves computational efficiency-yielding speed gains ranging from 6×−−326×-while providing estimates that closely approximate those obtained from the EM-based approach. |
| Keywords: | compositional data, iteratively reweighted least squares, quadratic programming |
| JEL: | C14 C15 C51 C63 |
| Date: | 2026–03–07 |
| URL: | https://d.repec.org/n?u=RePEc:crt:wpaper:2602 |
| By: | Chun Pang Chow; Hiroyuki Kasahara; Yoichi Sugita |
| Abstract: | We establish nonparametric identification of production functions, total factor productivity (TFP), price markups, and firms' output prices and quantities, as well as consumer demand, using firm-level revenue data, without observing output quantity, in a monopolistically competitive environment with a fully nonparametric demand system. This result overturns the widely held view -- formalized by Bond, Hashemi, Kaplan, and Zoch (2021) -- that output elasticities and markups are not nonparametrically identifiable from revenue data without quantity information. Under the additional restriction that demand satisfies the homothetic single-aggregator (HSA) structure of Matsuyama and Ushchev (2017), we further nonparametrically identify the representative consumer's utility function from firm-level revenue data. This new identification result enables counterfactual welfare analysis without parametric assumptions on preferences. We propose a semiparametric estimator that is feasible for standard firm-level datasets under a Cobb--Douglas production specification. Monte Carlo simulations show that the estimator performs well, while treating revenue as output induces substantial bias. Applying the estimator to Chilean manufacturing data, we reject the CES specification in favor of HSA, and find that market power reduces welfare by approximately 3%--6% of industry revenue in the three largest manufacturing industries in 1996. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.01492 |
| By: | Abdulrahman Alswaidan; Jeffrey D. Varner |
| Abstract: | Generating synthetic financial time series that preserve statistical properties of real market data is essential for stress testing, risk model validation, and scenario design. Existing approaches, from parametric models to deep generative networks, struggle to simultaneously reproduce heavy-tailed distributions, negligible linear autocorrelation, and persistent volatility clustering. We propose a hybrid hidden Markov framework that discretizes continuous excess growth rates into Laplace quantile-defined market states and augments regime switching with a Poisson-driven jump-duration mechanism to enforce realistic tail-state dwell times. Parameters are estimated by direct transition counting, bypassing the Baum-Welch EM algorithm. Synthetic data quality is evaluated using Kolmogorov-Smirnov and Anderson-Darling pass rates for distributional fidelity, and ACF mean absolute error for temporal structure. Applied to ten years of SPY data across 1, 000 simulated paths, the framework achieves KS and AD pass rates exceeding 97% and 91% in-sample and 94% out-of-sample (calendar year 2025), partially reproducing the ARCH effect that standard regime-switching models miss. No single model dominates all quality dimensions: GARCH(1, 1) reproduces volatility clustering more accurately but fails distributional tests (5.5% KS pass rate), while the standard HMM without jumps achieves higher distributional fidelity but cannot generate persistent high-volatility regimes. The proposed framework offers the best joint quality profile across distributional, temporal, and tail-coverage metrics. A Single-Index Model extension propagates the SPY factor path to a 424-asset universe, enabling scalable correlated synthetic path generation while preserving cross-sectional correlation structure. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.10202 |
| By: | Yinhuan Li; Chenxin Lyu; Ruodu Wang |
| Abstract: | Risk forecasts in financial regulation and internal management are calculated through historical data. The unknown structural changes of financial data poses a substantial challenge in selecting an appropriate look-back window for risk modeling and forecasting. We develop a data-driven online learning method, called the bootstrap-based adaptive window selection (BAWS), that adaptively determines the window size in a sequential manner. A central component of BAWS is to compare the realized scores against a data-dependent threshold, which is evaluate based on an idea of bootstrap. The proposed method is applicable to the forecast of risk measures that are elicitable individually or jointly, such as the Value-at-Risk (VaR) and the pair of the VaR and the corresponding Expected Shortfall. Through simulation studies and empirical analyses, we demonstrate that BAWS generally outperforms the standard rolling window approach and the recently developed method of stability-based adaptive window selection, especially when there are structural changes in the data-generating process. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.01157 |
| By: | Bruno Petrungaro; Anthony C. Constantinou |
| Abstract: | Causal machine learning (ML) recovers graphical structures that inform us about potential cause-and-effect relationships. Most progress has focused on cross-sectional data with no explicit time order, whereas recovering causal structures from time series data remains the subject of ongoing research in causal ML. In addition to traditional causal ML, this study assesses econometric methods that some argue can recover causal structures from time series data. The use of these methods can be explained by the significant attention the field of econometrics has given to causality, and specifically to time series, over the years. This presents the possibility of comparing the causal discovery performance between econometric and traditional causal ML algorithms. We seek to understand if there are lessons to be incorporated into causal ML from econometrics, and provide code to translate the results of these econometric methods to the most widely used Bayesian Network R library, bnlearn. We investigate the benefits and challenges that these algorithms present in supporting policy decision-making, using the real-world case of COVID-19 in the UK as an example. Four econometric methods are evaluated in terms of graphical structure, model dimensionality, and their ability to recover causal effects, and these results are compared with those of eleven causal ML algorithms. Amongst our main results, we see that econometric methods provide clear rules for temporal structures, whereas causal-ML algorithms offer broader discovery by exploring a larger space of graph structures that tends to lead to denser graphs that capture more identifiable causal relationships. |
| Date: | 2026–02 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.00041 |
| By: | Montes-Galdón, Carlos; Paredes, Joan; Wolf, Elias |
| Abstract: | We introduce a novel methodology, ”parametric tilting, ” for incorporating external information into econometric model-based density forecasts. Unlike traditional entropic tilting, which can generate unrealistic or unstable distributions under certain conditions, parametric tilting ensures more reliable and numerically stable results. Our approach leverages the flexibility of the skew-T distribution, which captures key moments of macroeconomic time series, and minimizes the Kullback-Leibler divergence between the target and model-based distributions. This method overcomes limitations of entropic tilting, such as multimodal or degenerate distributions, providing a robust alternative for policymakers and researchers aiming to integrate external views into probabilistic forecasting frameworks. JEL Classification: C14, C53, E52 |
| Keywords: | entropic tilting, forecasting, Kullback-Leibler information criterion |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:ecb:ecbwps:20263200 |
| By: | Adir Saly-Kaufmann; Kieran Wood; Jan Peter-Calliess; Stefan Zohren |
| Abstract: | We present a large scale benchmark of modern deep learning architectures for a financial time series prediction and position sizing task, with a primary focus on Sharpe ratio optimization. Evaluating linear models, recurrent networks, transformer based architectures, state space models, and recent sequence representation approaches, we assess out of sample performance on a daily futures dataset spanning commodities, equity indices, bonds, and FX spanning 2010 to 2025. Our evaluation goes beyond average returns and includes statistical significance, downside and tail risk measures, breakeven transaction cost analysis, robustness to random seed selection, and computational efficiency. We find that models explicitly designed to learn rich temporal representations consistently outperform linear benchmarks and generic deep learning models, which often lead the ranking in standard time series benchmarks. Hybrid models such as VSN with LSTM, a combination of Variable Selection Networks (VSN) and LSTMs, achieves the highest overall Sharpe ratio, while VSN with xLSTM and LSTM with PatchTST exhibit superior downside adjusted characteristics. xLSTM demonstrates the largest breakeven transaction cost buffer, indicating improved robustness to trading frictions. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.01820 |