|
on Econometric Time Series |
|
Issue of 2026–06–22
thirteen papers chosen by Simon Sosvilla-Rivero, Instituto Complutense de Análisis Económico |
| By: | Marie Corillon; Stephan Smeekes; Ines Wilms |
| Abstract: | High-dimensional time series regressions are often regularized to produce sparse coefficients. We show that temporal aggregation provides a powerful alternative to reduce dimensionality in high-order autoregressions and mixed-frequency regressions. To this end, we propose StarTime (Sparse Tree-based Aggregation for Time Series), a convex penalization method that uses a temporal tree to arrange lags hierarchically from high to low frequency. StarTime then flexibly selects coefficients to be aggregated at possibly varying frequencies, sparse or a combination thereof. We provide new error bounds for StarTime, demonstrate improved estimation accuracy and recovery of aggregation and sparsity in simulations relative to benchmarks, and illustrate StarTime's relevance for financial and macroeconomic applications. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.03665 |
| By: | Lei Jia; Shouri Hu; Zhaoxing Gao |
| Abstract: | This paper develops a canonical-correlation-based method for detecting structural changes in high-dimensional transformed factor models. The proposed approach exploits the low-rank canonical-correlation structure induced by dynamically dependent common factors, while serially uncorrelated idiosyncratic components correspond to a noise subspace with zero canonical correlations. We construct an eigenvalue-ratio criterion that measures residual dynamic dependence in the estimated noise subspace and identifies the true change point under sufficient separation of the regime-specific loading spaces or dynamic canonical correlation structures. Since the change-point location and the regime-specific factor numbers are both unknown, we further propose an alternating iterative estimation procedure that updates them sequentially until convergence. Under suitable mixing and moment conditions, we establish asymptotic properties of the proposed estimators, with convergence rates depending explicitly on factor strength, cross-sectional dimension, and sample size. Monte Carlo experiments and empirical applications to intraday stock returns and U.S. temperature series demonstrate the finite-sample |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.01553 |
| By: | Hyeon-seung Huh (Yonsei University); David Kim (University of Sydney) |
| Abstract: | We propose an alternative sampling scheme for sign-restricted SVARs that addresses two fundamental challenges in the standard approach: prior dependence and shock masquerading. The key idea is to utilize a direct sampling of structural coefficients in the SVAR combined with the robust inference of Giacomini and Kitagawa (2021). The scheme delivers large sets of exactly identified models, and the rotation matrix is uniquely solved via a direct, non-iterative linear algorithm. Sign restrictions are then used as a post-identification filter to ensure economic plausibility. This design is capable of eliminating prior dependence concerns, tightening the credible bounds, mitigating shock masquerading, and improving computational efficiency. We demonstrate the practical utility of the alternative sampling scheme through an application to the U.S. SVAR of Peersman (2005). |
| Keywords: | Structural vector autoregressions, Exact identification, Sign restrictions, Haar prior, Robust inference, Shock masquerading, Givens matrix |
| JEL: | C32 C36 C51 E32 E52 |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:yon:wpaper:2026rwp-293 |
| By: | del Barrio Castro, Tomás; Escribano, Álvaro; Özer, Yeliz; Sibbertsen Philipp |
| Abstract: | Long paleoclimate time series combine strong persistence, multiple orbital-scale periodicities, and structural changes across climate states. These features complicate the statistical analysis of deep-time proxy records, since common spectral peaks do not necessarily imply stable relationships between variables. We analyze Cenozoic climate variability using the Cenozoic Global Reference benthic isotope record over the last 67.1 million years. The cleaned and interpolated d18O and d13C series are studied in a regime-based setting, with segments defined by major Cenozoic climate-state transitions to investigate if a stable coupling occurs between the isotope proxies and Earth's astronomical variables. The analysis combines long-memory estimation, sequential frequency identification, frequency-adapted unit-root and stationarity testing at zero and harmonic frequencies, and cyclical fractional cointegration. This allows us to distinguish zero-frequency persistence from persistent cyclical behaviour and to test whether shared cyclical frequencies correspond to stable frequency-specific relationships between the proxies and orbital reference variables. The results show that the proxy series are not well described by a simple stationary versus unit-root dichotomy. Instead, they exhibit persistent long-memory dynamics with pronounced cyclical structure. Shared spectral peaks occur across several climate-state segments, yet only selected frequencies support stable fractional cointegration. Thus, isotope proxies and orbital reference variables may overlap spectrally without necessarily forming a stable long memory relationship. A key finding is that the additional split at the Eocene-Oligocene transition reveals a change in the frequency-specific relationship between d18O and d13C. Before the transition, stable fractional cointegration is associated with an orbital-scale band, whereas after the transition it shifts toward multi-million-year variability. This points to a substantial reorganisation of the frequency-specific coupling between the oxygen- and carbon-isotope records after the transition. |
| Keywords: | CENOGRID, Cyclical Fractional Cointegration, Deep-Time Paleoclimate |
| JEL: | C22 C32 Q54 |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:han:dpaper:dp-749 |
| By: | Saakstra, Sake |
| Abstract: | We propose a time-varying-parameter (TVP) state-space approach to survival modelling under sparse-event data conditions, in which the conditional hazard depends on a generative-process parameter whose evolution is driven by the score of the predictive likelihood. The score-driven Generalized Autoregressive Score (GAS) specification provides a parsimonious and asymptotically optimal mechanism for parameter time-variation, requiring only the score and a one-parameter persistence specification. The methodology is motivated by, and applied to, the empirical setting of irreversible clean-technology investments under transition uncertainty, where the policy-conditional hazard of project cancellation is plausibly regime-dependent rather than constant - a setting in which constant-parameter survival models systematically under-estimate the time-variation of the operating economic mechanism. Three rival TVP specifications are compared: M1 with constant parameter, M2 with parameter-driven block-step transitions, and M3 with observation-driven GAS persistence. The three are tested for out-of-sample forecast accuracy via three independent designs - a 75-25 time split, 5-fold within-project block cross-validation, and rolling one-step-ahead full-sample prediction. Significance is assessed by the Diebold-Mariano-Harvey-Leybourne-Newbold test on per-observation Bernoulli log-loss. The score-driven specification wins uniformly in point estimates across all three designs and is the unique element of the Hansen-Lunde-Nason Model Confidence Set at alpha = 0.10 under the rolling-one-step design. The DM-HLN test statistics for M3 versus M1 and M3 versus M2 are +5.59 and +4.80 respectively (p |
| Keywords: | time-varying parameter, score-driven model, state-space, survival modelling, Diebold-Mariano test, hydrogen |
| JEL: | C32 C41 Q42 |
| Date: | 2026–05–28 |
| URL: | https://d.repec.org/n?u=RePEc:pra:mprapa:129308 |
| By: | Jan Rovirosa; Jesse Schmolze |
| Abstract: | Modeling non-stationary stochastic systems requires balancing the representational capacity of deep learning with the structural transparency of classical probabilistic models. Markov transition matrices provide such a framework, but traditional frequency-based estimation collapses at high resolutions due to data sparsity. We propose a hybrid approach that parameterizes the manifold of stochastic matrices through a neural network, enabling estimation of time-inhomogeneous Markov chains in sparse-data regimes, and use financial markets as a testbed to investigate the Markov state variable as a critical inductive bias. We show that conditioning on realized volatility produces a more internally consistent Markovian structure than return-based states, achieving a $5.6\%$ reduction in Chapman-Kolmogorov discrepancy and superior held-out likelihood in 9 of 10 assets. Unlike black-box sequence models, our approach generates explicit matrices amenable to direct geometric analysis, surfacing structural findings such as the universal homogenization of transition probabilities under high-volatility regimes. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2605.30943 |
| By: | Xinyue Fang; Robert \'Slepaczuk |
| Abstract: | This study investigates whether regime-dependent volatility forecasting and machine-learning-based return prediction can be jointly integrated to improve both statistical forecasting performance and economic strategy outcomes in equity markets. Using high-frequency CSI 300 Index data from 2005 to 2023, a sequential twostage framework is developed. In the first stage, realized volatility is modeled using regime-augmented HARQ specifications combined with Markov-switching GJR-GARCH filtering to capture long-memory dynamics, asymmetry, and structural market regimes. In the second stage, volatility forecasts, regime indicators, and return-related predictors are incorporated into an XGBoost return-prediction model estimated through a strictly walk-forward out-of-sample procedure. The empirical results demonstrate that regime-aware volatility forecasting consistently outperforms baseline HARQ models across forecast evaluation metrics and is generally supported by formal forecast comparison tests. In contrast, return predictability remains weak, state-dependent, and concentrated primarily in low-volatility regimes. Although naive predictive trading strategies generally fail after accounting for realistic transaction costs, carefully designed implementations incorporating volatility scaling, low-volatility gating, threshold calibration, and turnover controls can improve defensive economic performance. The findings suggest that the practical value of predictive systems in financial markets may depend less on generating strong unconditional return forecasts and more on transforming weak state-dependent signals into economically robust portfolio allocation rules. Overall, the study contributes by integrating econometric volatility modeling, regime classification, machine-learning return prediction, and implementation realism within a unified framework. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.09478 |
| By: | Vasnev, Andrey; Liu, Chu-An |
| Abstract: | This paper proposes corrected forecast combinations when the original combined forecast errors are serially dependent. Motivated by the classic Bates and Granger (1969) example, we show that combined forecast errors can be strongly autocorrelated and that a simple correction – adding a fraction of the previous combined error to the next-period combined forecast – can deliver sizable improvements in forecast accuracy, often exceeding the original gains from combining. We formalize the approach within the conditional-risk framework of Gibbs and Vasnev (2024), in which the combined error decomposes into a predictable component (measurable at the forecast origin) and an innovation. We then link this correction to efficient estimation of combination weights under time-series dependence via GLS, allowing joint estimation of weights and an error-covariance structure. Using the U.S. Survey of Professional Forecasters for major macroeconomic indices across various subsamples (including pre/post-2000, GFC, and COVID), we find that a parsimonious correction of the mean forecast with a coefficient around 0.5 is a robust starting point and often yields material improvements in forecast accuracy. For optimal-weight forecasts, the correction substantially mitigates the forecast combination puzzle by turning poorly performing out-of-sample optimal-weight combinations into competitive forecasts. |
| Keywords: | Monetary policy indicators, China, forecast combination, optimal weights |
| Date: | 2026–01–21 |
| URL: | https://d.repec.org/n?u=RePEc:syb:wpbsba:2123/34743 |
| By: | Konrad J. Mueller; Nikita Zozoulenko; Ben Wood; Thomas Cass; Lukas Gonon |
| Abstract: | Generating realistic financial time series is challenging as training data is often limited to a single historical path. With such scarce data, overfitting is hard to avoid, especially under adversarial training where a trained discriminator can memorize the training samples. To mitigate this, recent approaches train generators to minimize the discrepancy between untrained feature representations of real and generated time series. In these works, the feature maps are based on path signatures, which can fail to capture relevant time series properties at tractable truncation depths. In this work, we instead train generators by matching random convolutional features of real and generated time series. Existing random convolutional feature maps, such as Rocket and Hydra, have been shown to provide informative representations of real-world time series, but cannot supervise generative models because they are non-differentiable. We introduce SOCK (SOft Competing Kernels), a fully differentiable random convolutional feature map, suited to train generative time series models. We show that generators trained by matching random SOCK features consistently outperform signature and diffusion baselines across a wide range of small-sample financial datasets. We further demonstrate SOCK's expressiveness on two-sample hypothesis testing and time series classification tasks, where SOCK matches or outperforms existing unsupervised feature maps. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.05138 |
| By: | Mingxuan Yi; Vidal Mehra; Jing Chen; John Cartlidge |
| Abstract: | Regime shifts in financial markets reorganise the joint dynamics of asset prices and macro variables, breaking any single-regime calibration. They are nonetheless difficult to detect reliably because the data signal is noisy and heavily multicollinear, while the contemporaneous text that announces them is unstructured. Standard regime shift detection methods rely solely on structured time-series data and ignore policy communications, even though these texts often signal shifts before they materialise in observed prices. We propose a text-enhanced regime shift detection pipeline that combines large language model (LLM) reasoning over central-bank communications with statistical validation on multivariate financial time series. The framework is detector-agnostic: text-proposed candidates are validated using a bootstrap likelihood-ratio test on a vector autoregression (VAR), while data-driven candidates from arbitrary regime detectors are ratified through a lenient LLM text check. We evaluate the framework on 2010-2024 FOMC minutes paired with a 14-variable U.S. Treasury and macroeconomic panel, using four interchangeable data-driven detectors. The proposed pipeline achieves F1 = 0.82 against a verified anchor list of monetary-policy regime shifts, with same-day modal detection latency and consistently stronger performance than pure data-driven baselines. The results demonstrate that combining unstructured policy text with statistical structural-break detection improves the robustness and interpretability of regime shift identification in financial markets. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2605.30363 |
| By: | Akash Deep; Nicholas Appiah; Svetlozar T. Rachev |
| Abstract: | This paper studies the joint role of long-memory dynamics, rough-volatility behavior, and persistence-based forecasting features in equity volatility modeling. We combine semiparametric long-memory estimation, rough-volatility diagnostics, and structured forecasting regressions to examine whether persistence measures contain economically meaningful forecasting information beyond conventional volatility predictors. Using a panel of 115 S&P500 constituents from November 2001 through April 2026, we document that volatility proxies exhibit substantial long-memory behavior and locally rough dynamics. The cross-sectional mean Geweke-Porter-Hudak estimate of the memory parameter is $\hat{d} = 0.226$, while the corresponding local-Whittle estimate is $\hat{d} = 0.440$, with statistical significance observed across nearly the entire panel. Rolling estimates of persistence rise substantially during the global financial crisis and the COVID period and display a positive contemporaneous association with the VIX. We then examine whether persistence-related features improve out-of-sample volatility forecasts beyond standard HAR and HAR-X benchmarks. Incorporating cross-sectional persistence aggregates, sectoral persistence measures, and persistence-by-stress interaction terms produces moderate but statistically significant forecasting improvements, particularly at longer horizons and during stress regimes. Forecast gains are strongest during periods of elevated market volatility and in volatility-managed portfolio applications. The results suggest that persistence measures may serve as useful reduced-form indicators of the duration and propagation of uncertainty in financial markets, although the paper does not claim structural identification of the economic mechanisms generating persistence. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2605.24285 |
| By: | Aoxin Zhang; Yuhan Cheng; Kwanting Leung |
| Abstract: | Benchmarking forecasting architectures for daily equity portfolios is not just a prediction exercise. It also asks which model remains usable after preferences, costs, and portfolio constraints are imposed. We build a CRSP daily-stock benchmark for 15 deep and statistical time-series architectures over 2018--2024. The protocol combines common-window decile portfolios, stochastic multi-criteria acceptability analysis, a deployment-adjusted acceptability index, and a constrained quadratic portfolio layer with capacity, beta, industry, risk, leverage, and turnover controls. The index starts from the SMAA rank-acceptability distribution and downweights models whose criteria-level wins produce high portfolio regret; its Gibbs form is characterized as an entropic update from the SMAA prior. Empirically, no architecture dominates the raw benchmark: TransEnc-8 has the largest rank-1 acceptability, 0.352, and no model exceeds about 0.36. Rankings vary with preferences, market state, feature universe, and transaction costs. In the promoted five-model constrained-portfolio comparison, TransEnc-8 is selected throughout, while return-oriented raw rankings can favor TS-RIDGE. Broad-universe decile signals can survive costs, but the baseline constrained-QP net Sharpe at 20 bps is negative for every promoted model. The benchmark supports model selection and diagnosis rather than a standalone trading-strategy claim. |
| Date: | 2026–06 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.09420 |
| By: | Daniel Cunha Oliveira; Kieran Wood; Stefan Zohren; Mihai Cucuringu; Andr\'e Fujita |
| Abstract: | Deep learning models show promise in financial forecasting, yet their generalization is often undermined by small datasets, noisy signals, and non-stationarity. While meta-learning and related techniques mitigate some of these issues, they typically do not account for a core limitation in macro-financial prediction: the scarcity of distinct macroeconomic regimes that drive asset returns. We introduce HANET (Hierarchical Attention Network), a hybrid LSTM-based architecture that integrates macroeconomic domain knowledge through attention over long-run macro contexts while preserving high-frequency market dynamics. HANET organizes information in a hierarchical mixed-frequency structure, with daily asset-return signals nested within monthly macroeconomic windows, and introduces a Hierarchical Cross-Attention mechanism that reconciles low-frequency macro signals with high-frequency returns without discarding granular daily information. By framing regime selection as attention over macroeconomic contexts, the model adapts to scarce and shifting regimes. Empirically, across 55 liquid futures spanning multiple asset classes, HANET consistently outperforms neural forecasters that ignore macroeconomic information, particularly during turbulent periods, improving risk-adjusted returns and mitigating losses. Ablation studies show that these gains rely on structured macro conditioning rather than naive feature augmentation: an LSTM with the same macro representation performs poorly, and shuffling macro contexts substantially degrades performance. Finally, HANET provides interpretability through attention weights, highlighting which historical regimes are most influential for each forecast and linking macro conditions to portfolio outcomes. These results establish HANET as a systematic approach to integrating macroeconomic information into attention-based deep learning for financial forecasting. |
| Date: | 2026–05 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2606.00624 |