nep-for New Economics Papers
on Forecasting
Issue of 2025–05–19
eleven papers chosen by
Malte Knüppel, Deutsche Bundesbank


  1. Modeling and Forecasting Realized Volatility with Multivariate Fractional Brownian Motion By Markus Bibinger; Jun Yu; Chen Zhang
  2. Integrated GARCH-GRU in Financial Volatility Forecasting By Jingyi Wei; Steve Yang; Zhenyu Cui
  3. Machine Learning and the Forecastability of Cross-Sectional Realized Variance: The Role of Realized Moments By Vasilios Plakandaras; Matteo Bonato; Rangan Gupta; Oguzhan Cepni
  4. A Network Approach to Volatility Diffusion and Forecasting in Global Financial Markets By Matteo Orlandini; Sebastiano Michele Zema; Mauro Napoletano; Giorgio Fagiolo
  5. Forecasting Spot and Futures Price Volatility of Agricultural Commodities: The Role of Climate-Related Migration Uncertainty By Afees A. Salisu; Ahamuefula E. Ogbonna; Rangan Gupta; Elie Bouri
  6. MLOps Monitoring at Scale for Digital Platforms By Yu Jeffrey Hu; Jeroen Rombouts; Ines Wilms
  7. Forecasting U.S. equity market volatility with attention and sentiment to the economy By Martina Halouskov\'a; \v{S}tefan Ly\'ocsa
  8. The Climate Adaptation Feedback By Alexander C. Abajian; Tamma Carleton; Kyle C. Meng; Olivier Deschenes
  9. Robust Tests for Factor-Augmented Regressions with an Application to the novel EA-MD Dataset By Alessandro Morico; Ovidijus Stauskas
  10. Informer in Algorithmic Investment Strategies on High Frequency Bitcoin Data By Filip Stefaniuk; Robert \'Slepaczuk
  11. A Causal Perspective of Stock Prediction Models By Songci Xu; Qiangqiang Cheng; Chi-Guhn Lee

  1. By: Markus Bibinger; Jun Yu; Chen Zhang
    Abstract: A multivariate fractional Brownian motion (mfBm) with component-wise Hurst exponents is used to model and forecast realized volatility. We investigate the interplay between correlation coefficients and Hurst exponents and propose a novel estimation method for all model parameters, establishing consistency and asymptotic normality of the estimators. Additionally, we develop a time-reversibility test, which is typically not rejected by real volatility data. When the data-generating process is a time-reversible mfBm, we derive optimal forecasting formulae and analyze their properties. A key insight is that an mfBm with different Hurst exponents and non-zero correlations can reduce forecasting errors compared to a one-dimensional model. Consistent with optimal forecasting theory, out-of-sample forecasts using the time-reversible mfBm show improvements over univariate fBm, particularly when the estimated Hurst exponents differ significantly. Empirical results demonstrate that mfBm-based forecasts outperform the (vector) HAR model.
    Date: 2025–04
    URL: https://d.repec.org/n?u=RePEc:arx:papers:2504.15985
  2. By: Jingyi Wei; Steve Yang; Zhenyu Cui
    Abstract: In this study, we propose a novel integrated Generalized Autoregressive Conditional Heteroskedasticity-Gated Recurrent Unit (GARCH-GRU) model for financial volatility modeling and forecasting. The model embeds the GARCH(1, 1) formulation directly into the GRU cell architecture, yielding a unified recurrent unit that jointly captures both traditional econometric properties and complex temporal dynamics. This hybrid structure leverages the strengths of GARCH in modeling key stylized facts of financial volatility, such as clustering and persistence, while utilizing the GRU's capacity to learn nonlinear dependencies from sequential data. Compared to the GARCH-LSTM counterpart, the GARCH-GRU model demonstrates superior computational efficiency, requiring significantly less training time, while maintaining and improving forecasting accuracy. Empirical evaluation across multiple financial datasets confirms the model's robust outperformance in terms of mean squared error (MSE) and mean absolute error (MAE) relative to a range of benchmarks, including standard neural networks, alternative hybrid architectures, and classical GARCH-type models. As an application, we compute Value-at-Risk (VaR) using the model's volatility forecasts and observe lower violation ratios, further validating the predictive reliability of the proposed framework in practical risk management settings.
    Date: 2025–04
    URL: https://d.repec.org/n?u=RePEc:arx:papers:2504.09380
  3. By: Vasilios Plakandaras (Department of Economics, Democritus University of Thrace, Komotini, Greece); Matteo Bonato (Department of Economics and Econometrics, University of Johannesburg, Auckland Park, South Africa; IPAG Business School, 184 Boulevard Saint-Germain, 75006 Paris, France); Rangan Gupta (Department of Economics, University of Pretoria, Private Bag X20, Hatfield 0028, South Africa); Oguzhan Cepni (Ostim Technical University, Ankara, Turkiye; University of Edinburgh Business School, Centre for Business, Climate Change, and Sustainability; Department of Economics, Copenhagen Business School, Denmark)
    Abstract: This paper forecasts monthly cross-sectional realized variance (RV) for U.S. equities across 49 industries and all 50 states. We exploit information in both own-market and cross-market (oil) realized moments (semi-variance, leverage, skewness, kurtosis, and upside and downside tail risk) as predictors. To accommodate cross-sectional dependence, we compare standard econometric panel models with machine-learning approaches and introduce a new machine-learning technique tailored specifically to panel data. Using observations from April 1994 through April 2023, the panel-dedicated machine-learning model consistently outperforms all other methods, while oil-related moments add little incremental predictive power beyond own-market moments. Short-horizon forecasts successfully capture immediate shocks, whereas longer-horizon forecasts reflect broader structural economic changes. These results carry important implications for portfolio allocation and risk management.
    Keywords: Cross-sectional realized variance, Realized moments, Machine learning, Forecasting
    JEL: C33 C53 G10 G17
    Date: 2025–04
    URL: https://d.repec.org/n?u=RePEc:pre:wpaper:202518
  4. By: Matteo Orlandini (Université Côte d'Azur, CNRS, GREDEG, France; Institute of Economics, Scuola Superiore Sant'Anna, Italy); Sebastiano Michele Zema (Scuola Normale Superiore, Italy); Mauro Napoletano (Université Côte d'Azur, CNRS, GREDEG, France; Sciences Po, OFCE, France; Institute of Economics, Scuola Superiore Sant'Anna, Italy); Giorgio Fagiolo (Institute of Economics, Scuola Superiore Sant'Anna, Italy)
    Abstract: The node degree distribution of an inferred financial network is often characterized by a small number of nodes with a large number of connections and many nodes with few connections. To date, there is no empirical evidence on how this stylized statistical fact can be useful in predicting fluctuations of financial assets. In this paper, we explore this possibility by modifying well-known time-series models and augmenting them with covariates from a reconstructed network, selecting nodes that are identified as the most connected to the index of interest. We then analyze the out-of-sample performance of these models across different volatility proxies. The results show that nodes belonging to the right tail of the degree distribution possess high predictive power over financial aggregates, independently of the volatility measure used. Our findings suggest that incorporating the topological information that arises from this statistical regularity in financial networks can enhance the accuracy of traditional predictive models.
    Keywords: Volatility forecasting, Network-augmented models, Cross-border volatility spillovers, Equity indexes
    JEL: G17 G11
    Date: 2025–05
    URL: https://d.repec.org/n?u=RePEc:gre:wpaper:2025-19
  5. By: Afees A. Salisu (Centre for Econometrics and Applied Research, Ibadan, Nigeria; Department of Economics, University of Pretoria, Pretoria, 0002, South Africa); Ahamuefula E. Ogbonna (Centre for Econometrics and Applied Research, Ibadan, Nigeria); Rangan Gupta (Department of Economics, University of Pretoria, Private Bag X20, Hatfield 0028, South Africa); Elie Bouri (School of Business, Lebanese American University, Lebanon)
    Abstract: We evaluate the predictive ability of the newly developed climate-related migration uncertainty index (CMUI) and its two components, the climate uncertainty index (CUI) and the migration uncertainty index (MUI), for the return volatility of agricultural commodity prices in both futures and spot markets. Employing a GARCH-MIDAS model, based on mixed data frequencies covering the period from 1977Q4 (with the earliest daily observation on October 3, 1977) to 2024Q1 (with the latest daily observation on March 29, 2024), we conduct both statistical and economic evaluations, including the Modified Diebold-Mariano test, Model Confidence Set procedure, and risk-adjusted performance metrics. The results demonstrate that integrating CUI, MUI, and CMUI into the predictive model of the return volatility of agricultural commodity prices significantly improves forecast accuracy relative to the conventional GARCH-MIDAS-RV benchmark. These findings suggest that the climate and migration related uncertainty indices are both statistically significant and economically relevant, offering enhanced predictive power and investment performance.
    Keywords: Climate-related Migration Uncertainty Index, Climate Uncertainty Index, Migration Uncertainty Index, Agricultural commodity prices, GARCH-MIDAS, Forecast evaluation, Economic Significance
    JEL: C53 D8 F22 Q02 Q13
    Date: 2025–04
    URL: https://d.repec.org/n?u=RePEc:pre:wpaper:202516
  6. By: Yu Jeffrey Hu; Jeroen Rombouts; Ines Wilms
    Abstract: Machine learning models are widely recognized for their strong performance in forecasting. To keep that performance in streaming data settings, they have to be monitored and frequently re-trained. This can be done with machine learning operations (MLOps) techniques under supervision of an MLOps engineer. However, in digital platform settings where the number of data streams is typically large and unstable, standard monitoring becomes either suboptimal or too labor intensive for the MLOps engineer. As a consequence, companies often fall back on very simple worse performing ML models without monitoring. We solve this problem by adopting a design science approach and introducing a new monitoring framework, the Machine Learning Monitoring Agent (MLMA), that is designed to work at scale for any ML model with reasonable labor cost. A key feature of our framework concerns test-based automated re-training based on a data-adaptive reference loss batch. The MLOps engineer is kept in the loop via key metrics and also acts, pro-actively or retrospectively, to maintain performance of the ML model in the production stage. We conduct a large-scale test at a last-mile delivery platform to empirically validate our monitoring framework.
    Date: 2025–04
    URL: https://d.repec.org/n?u=RePEc:arx:papers:2504.16789
  7. By: Martina Halouskov\'a; \v{S}tefan Ly\'ocsa
    Abstract: Macroeconomic variables are known to significantly impact equity markets, but their predictive power for price fluctuations has been underexplored due to challenges such as infrequency and variability in timing of announcements, changing market expectations, and the gradual pricing in of news. To address these concerns, we estimate the public's attention and sentiment towards ten scheduled macroeconomic variables using social media, news articles, information consumption data, and a search engine. We use standard and machine-learning methods and show that we are able to improve volatility forecasts for almost all 404 major U.S. stocks in our sample. Models that use sentiment to macroeconomic announcements consistently improve volatility forecasts across all economic sectors, with the greatest improvement of 14.99% on average against the benchmark method - on days of extreme price variation. The magnitude of improvements varies with the data source used to estimate attention and sentiment, and is found within machine-learning models.
    Date: 2025–03
    URL: https://d.repec.org/n?u=RePEc:arx:papers:2503.19767
  8. By: Alexander C. Abajian; Tamma Carleton; Kyle C. Meng; Olivier Deschenes
    Abstract: Many behavioral responses to climate change are carbon-intensive, raising concerns that adaptation may cause additional warming. The sign and magnitude of this feedback depend on how increased emissions from cooling balance against reduced emissions from heating across space and time. We present an empirical approach that forecasts the effect of future adaptive energy use on global average temperature over the 21st century. We find energy-based adaptation will lower global mean surface temperature in 2099 by 0.12 degrees Celsius (0.07 degrees Celsius) relative to baseline projections under RCP8.5 (RCP4.5) and avoid 1.8 (0.6) trillion USD ($2019) in damages. Energy-based adaptation lowers business-as-usual emissions for 85% of countries, reducing the mitigation required to meet their unilateral Nationally Determined Contributions under the UNFCCC by 20% on average. These findings indicate that while business-as-usual adaptive energy use is unlikely to accelerate warming, it raises important implications for countries’ existing mitigation commitments.
    JEL: Q4 Q5
    Date: 2025–02
    URL: https://d.repec.org/n?u=RePEc:nbr:nberwo:33531
  9. By: Alessandro Morico; Ovidijus Stauskas
    Abstract: We present four novel tests of equal predictive accuracy and encompassing for out-of-sample forecasts based on factor-augmented regression. We extend the work of Pitarakis (2023a, b) to develop the inferential theory of predictive regressions with generated regressors which are estimated by using Common Correlated Effects (henceforth CCE) - a technique that utilizes cross-sectional averages of grouped series. It is particularly useful since large datasets of such structure are becoming increasingly popular. Under our framework, CCE-based tests are asymptotically normal and robust to overspecification of the number of factors, which is in stark contrast to existing methodologies in the CCE context. Our tests are highly applicable in practice as they accommodate for different predictor types (e.g., stationary and highly persistent factors), and remain invariant to the location of structural breaks in loadings. Extensive Monte Carlo simulations indicate that our tests exhibit excellent local power properties. Finally, we apply our tests to a novel EA-MD-QD dataset by Barigozzi et al. (2024b), which covers Euro Area as a whole and primary member countries. We demonstrate that CCE factors offer a substantial predictive power even under varying data persistence and structural breaks.
    Date: 2025–04
    URL: https://d.repec.org/n?u=RePEc:arx:papers:2504.08455
  10. By: Filip Stefaniuk; Robert \'Slepaczuk
    Abstract: The article investigates the usage of Informer architecture for building automated trading strategies for high frequency Bitcoin data. Three strategies using Informer model with different loss functions: Root Mean Squared Error (RMSE), Generalized Mean Absolute Directional Loss (GMADL) and Quantile loss, are proposed and evaluated against the Buy and Hold benchmark and two benchmark strategies based on technical indicators. The evaluation is conducted using data of various frequencies: 5 minute, 15 minute, and 30 minute intervals, over the 6 different periods. Although the Informer-based model with Quantile loss did not outperform the benchmark, two other models achieved better results. The performance of the model using RMSE loss worsens when used with higher frequency data while the model that uses novel GMADL loss function is benefiting from higher frequency data and when trained on 5 minute interval it beat all the other strategies on most of the testing periods. The primary contribution of this study is the application and assessment of the RMSE, GMADL, and Quantile loss functions with the Informer model to forecast future returns, subsequently using these forecasts to develop automated trading strategies. The research provides evidence that employing an Informer model trained with the GMADL loss function can result in superior trading outcomes compared to the buy-and-hold approach.
    Date: 2025–03
    URL: https://d.repec.org/n?u=RePEc:arx:papers:2503.18096
  11. By: Songci Xu; Qiangqiang Cheng; Chi-Guhn Lee
    Abstract: In the realm of stock prediction, machine learning models encounter considerable obstacles due to the inherent low signal-to-noise ratio and the nonstationary nature of financial markets. These challenges often result in spurious correlations and unstable predictive relationships, leading to poor performance of models when applied to out-of-sample (OOS) domains. To address these issues, we investigate \textit{Domain Generalization} techniques, with a particular focus on causal representation learning to improve a prediction model's generalizability to OOS domains. By leveraging multi-factor models from econometrics, we introduce a novel error bound that explicitly incorporates causal relationships. In addition, we present the connection between the proposed error bound and market nonstationarity. We also develop a \textit{Causal Discovery} technique to discover invariant feature representations, which effectively mitigates the proposed error bound, and the influence of spurious correlations on causal discovery is rigorously examined. Our theoretical findings are substantiated by numerical results, showcasing the effectiveness of our approach in enhancing the generalizability of stock prediction models.
    Date: 2025–03
    URL: https://d.repec.org/n?u=RePEc:arx:papers:2503.20987

This nep-for issue is ©2025 by Malte Knüppel. It is provided as is without any express or implied warranty. It may be freely redistributed in whole or in part for any purpose. If distributed in part, please include this notice.
General information on the NEP project can be found at https://nep.repec.org. For comments please write to the director of NEP, Marco Novarese at <director@nep.repec.org>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.