nep-for New Economics Papers
on Forecasting
Issue of 2020‒06‒15
twenty papers chosen by
Rob J Hyndman
Monash University

  1. Does Judgment Improve Macroeconomic Density Forecasts? By Galvao, Ana Beatriz; Garratt, Anthony; Mitchell, James
  2. Forecast combinations in machine learning By Qiu, Yue; Xie, Tian; Yu, Jun
  3. Daily Middle-Term Probabilistic Forecasting of Power Consumption in North-East England By Roberto Baviera; Giuseppe Messuti
  4. Probabilistic multivariate electricity price forecasting using implicit generative ensemble post-processing By Tim Janke; Florian Steinke
  5. Higher moment constraints for predictive density combination By Laurent Pauwels; Peter Radchenko; Andrey L. Vasnev
  6. Equal predictability test for multi-step-ahead system forecasts invariant to linear transformations By Håvard Hungnes
  7. Common factors and the dynamics of cereal prices. A forecasting perspective By Marek Kwas; Alessia Paccagnini; Michal Rubaszek
  8. Machine Learning, the Treasury Yield Curve and Recession Forecasting By Michael Puglia; Adam Tucker
  9. The Role of Global Economic Conditions in Forecasting Gold Market Volatility: Evidence from a GARCH-MIDAS Approach By Afees A. Salisu; Rangan Gupta; Elie Bouri
  10. Temporal mixture ensemble models for intraday volume forecasting in cryptocurrency exchange markets By Nino Antulov-Fantulin; Tian Guo; Fabrizio Lillo
  11. Making text count: economic forecasting using newspaper text By Kalamara, Eleni; Turrell, Arthur; Redl, Chris; Kapetanios, George; Kapadia, Sujit
  12. Air Pollution, Affect, and Forecasting Bias: Evidence from Chinese Financial Analysts By Rui Dong; Raymond Fisman; Yongxiang Wang; Nianhang Xu
  13. Forecasting gasoline prices with mixed random forest error correction models By Wang, Dandan; Escribano Saez, Alvaro
  14. Using Machine Learning to Forecast Future Earnings By Xinyue Cui; Zhaoyu Xu; Yue Zhou
  15. On the external validity of experimental inflation forecasts: A comparison with five categories of field expectations By Camille Cornand; Paul Hubert
  16. Structural Scenario Analysis with SVARs By Antolin-Diaz, Juan; Petrella, Ivan; Rubio-Ramirez, Juan F.
  17. Macroeconomic Forecasting with Fractional Factor Models By Tobias Hartl
  18. High-dimensional macroeconomic forecasting using message passing algorithms By Dimitris Korobilis
  19. Application of GARCH Models For Volatility Modelling of Stock Market Returns: Evidences From BSE India By Neeti Mathur; Himanshu Mathur
  20. Predicting when peaks will occur, ex ante. Insights from the COVID-19 Pandemic in Italy and Belgium By Kristof Decock; Michela Bergamini; Koenraad Debackere; Enrico Lupi; Anne Mieke Vandamme; Bart Van Looy

  1. By: Galvao, Ana Beatriz (University of Warwick); Garratt, Anthony (University of Warwick); Mitchell, James (University of Warwick)
    Abstract: This paper presents empirical evidence on how judgmental adjustments affect the accuracy of macroeconomic density forecasts. We aim to separate the effects of judgments made about the first three moments of a set of professional forecasters’ density forecasts for UK output growth and inflation. Using entropic tilting methods, we evaluate whether judgmental adjustments about the mean, variance and skewness improve the accuracy of density forecasts from statistical models. We find that not all judgmental adjustments improve density forecasts: overall, density forecasts from statistical models prove hard to beat. Judgments about point forecasts tend to improve density forecast accuracy at short horizons and at times of heightened macroeconomic uncertainty. Judgments about the variance clearly hinder at short horizons, but can help deliver better tail risk forecasts at longer horizons. Finally, judgments about skew in general take value away, with gains seen only for longer horizon output growth forecasts when statistical models took longer to learn that downside risks had reduced with the end of the Great Recession.
    Keywords: density forecasting ; judgment forecasting ; skewness ; exponential tilting ; forecasting uncertainty ;
    JEL: C32 C53 E37
    Date: 2020
  2. By: Qiu, Yue (Shanghai University of International Business and Economics); Xie, Tian (Shanghai University of Finance and Economics); Yu, Jun (School of Economics, Singapore Management University)
    Abstract: This paper introduces novel methods to combine forecasts made by machine learning techniques. Machine learning methods have found many successful applications in predicting the response variable. However, they ignore model uncertainty when the relationship between the response variable and the predictors is nonlinear. To further improve the forecasting performance, we propose a general framework to combine multiple forecasts from machine learning techniques. Simulation studies show that the proposed machine-learning-based forecast combinations work well. In empirical applications to forecast key macroeconomic and financial variables, we find that the proposed methods can produce more accurate forecasts than individual machine learning techniques and the simple average method, later of which is known as hard to beat in the literature.
    Keywords: Model uncertainty; Machine learning; Nonlinearity; Forecast combinations
    JEL: C52 C53
    Date: 2020–05–11
  3. By: Roberto Baviera; Giuseppe Messuti
    Abstract: Probabilistic forecasting of power consumption in a middle-term horizon (months to a year) is a main challenge in the energy sector. It plays a key role in planning future generation plants and transmission grid. We propose a new model that incorporates trend and seasonality features as in traditional time-series analysis and weather conditions as explicative variables in a parsimonious machine learning approach, known as Gaussian Process. Applying to a daily power consumption dataset in North East England provided by one of the largest energy suppliers, we obtain promising results in Out-of-Sample density forecasts up to one year, even using a small dataset, with only a two-year In-Sample data. In order to verify the quality of the achieved power consumption probabilistic forecast we consider measures that are common in the energy sector as pinball loss and Winkler score and backtesting conditional and unconditional tests, standard in the banking sector after the introduction of Basel II Accords.
    Date: 2020–05
  4. By: Tim Janke; Florian Steinke
    Abstract: The reliable estimation of forecast uncertainties is crucial for risk-sensitive optimal decision making. In this paper, we propose implicit generative ensemble post-processing, a novel framework for multivariate probabilistic electricity price forecasting. We use a likelihood-free implicit generative model based on an ensemble of point forecasting models to generate multivariate electricity price scenarios with a coherent dependency structure as a representation of the joint predictive distribution. Our ensemble post-processing method outperforms well-established model combination benchmarks. This is demonstrated on a data set from the German day-ahead market. As our method works on top of an ensemble of domain-specific expert models, it can readily be deployed to other forecasting tasks.
    Date: 2020–05
  5. By: Laurent Pauwels; Peter Radchenko; Andrey L. Vasnev
    Abstract: The majority of financial data exhibit asymmetry and heavy tails, which makes forecasting the entire density critically important. Recently, a forecast combination methodology has been developed to combine predictive densities. We show that combining individual predictive densities that are skewed and/or heavy-tailed results in significantly reduced skewness and kurtosis. We propose a solution to overcome this problem by deriving optimal log score weights under Higher-order Moment Constraints (HMC). The statistical properties of these weights, such as consistency and asymptotic distribution, are investigated theoretically and through a simulation study. An empirical application that uses the S&P 500 daily index returns illustrates that the proposed HMC weight density combinations perform very well relative to other combination methods.
    Keywords: Forecast combinations, Predictive densities, Moment constraints, Financial data.
    JEL: C53 C58
    Date: 2020–05
  6. By: Håvard Hungnes (Statistics Norway)
    Abstract: The paper derives a test for equal predictability of multi-step-ahead system forecasts that is invariant to linear transformations. The test is a multivariate version of the Diebold-Mariano test. An invariant metric for multi-step-ahead system forecasts is necessary as the conclusions otherwise can depend on how the forecasts are reported (e.g., as in levels or differences; or log-levels or growth rates). The test is used in comparing quarterly multi-step-ahead system forecasts made by Statistics Norway with similar forecasts made by Norges Bank.
    Keywords: Macroeconomic forecasts; Econometric models; Forecast performance; Forecast evaluation; Forecast comparison
    JEL: C32 C53
    Date: 2020–05
  7. By: Marek Kwas; Alessia Paccagnini; Michal Rubaszek
    Abstract: This article investigates what determines the price dynamics of the main cereals: barley, maize, rice and wheat. Using an extensive dataset of monthly time series covering the years 1980 - 2019, we extract four different common factors explaining the dynamics of commodity prices, exchange rates, financial and macroeconomic indicators. Next, we examine whether these factors are useful in explaining the movements of cereal prices. We show that models incorporating all four factors outperform significantly the naive random walk model in out-of-sample forecasting competition, especially for longer horizons. However, they have only marginally better performance than a simpler model based on the commodity factor alone.
    Keywords: Cereal prices, Forecasting, Factor models, Autoregressive models.
    JEL: C32 C53 Q11
    Date: 2020–05
  8. By: Michael Puglia; Adam Tucker
    Abstract: We use machine learning methods to examine the power of Treasury term spreads and other financial market and macroeconomic variables to forecast US recessions, vis-à-vis probit regression. In particular we propose a novel strategy for conducting cross-validation on classifiers trained with macro/financial panel data of low frequency and compare the results to those obtained from standard k-folds cross-validation. Consistent with the existing literature we find that, in the time series setting, forecast accuracy estimates derived from k-folds are biased optimistically, and cross-validation strategies which eliminate data "peeking" produce lower, and perhaps more realistic, estimates of forecast accuracy. More strikingly, we also document rank reversal of probit, Random Forest, XGBoost, LightGBM, neural network and support-vector machine classifier forecast performance over the two cross-validation methodologies. That is, while a k-folds cross-validation indicates tha t the forecast accuracy of tree methods dominates that of neural networks, which in turn dominates that of probit regression, the more conservative cross-validation strategy we propose indicates the exact opposite, and that probit regression should be preferred over machine learning methods, at least in the context of the present problem. This latter result stands in contrast to a growing body of literature demonstrating that machine learning methods outperform many alternative classification algorithms and we discuss some possible reasons for our result. We also discuss techniques for conducting statistical inference on machine learning classifiers using Cochrane's Q and McNemar's tests; and use the SHapley Additive exPlanations (SHAP) framework to decompose US recession forecasts and analyze feature importance across business cycles.
    Keywords: Shapley; Probit; XGBoost; Treasury yield curve; Neural network; LightGBM; Recession; Tree ensemble; Support-vector machine; Random forest
    JEL: C45 C53 E37
    Date: 2020–05–20
  9. By: Afees A. Salisu (Department for Management of Science and Technology Development, Ton Duc Thang University, Ho Chi Minh City, Vietnam; Faculty of Business Administration, Ton Duc Thang University, Ho Chi Minh City, Vietnam); Rangan Gupta (Department of Economics, University of Pretoria, Pretoria, 0002, South Africa); Elie Bouri (USEK Business School, Holy Spirit University of Kaslik, Jounieh, Lebanon)
    Abstract: In this study, we examine the role of global economic conditions in forecasting gold market volatility using alternative measures. Based on the available data frequency for the relevant series, we adopt the GARCH-MIDAS approach which allows for mixed data frequencies. We find that global economic conditions contribute significantly to gold market volatility albeit with mixed outcomes. While the results lend support to the safe-haven properties of the gold market, the outcome is influenced by the choice of measure of global economic conditions. For completeness, we extend the analyses to other precious metals such as silver, platinum, palladium, and rhodium and find that global economic conditions forecast the volatility of gold returns better than other precious metals. Our results are robust to multiple forecast horizons and offer useful insights into plausible investment choices in the precious metals market.
    Keywords: Precious Metals Volatility, Global Economic Conditions, Mixed-Frequency
    JEL: C32 C53 E32 Q02
    Date: 2020–05
  10. By: Nino Antulov-Fantulin; Tian Guo; Fabrizio Lillo
    Abstract: We study the problem of the intraday short-term volume forecasting in cryptocurrency exchange markets. The predictions are built by using transaction and order book data from different markets where the exchange takes place. Methodologically, we propose a temporal mixture ensemble model, capable of adaptively exploiting, for the forecasting, different sources of data and providing a volume point estimate, as well as its uncertainty. We provide evidence of the outperformance of our model by comparing its outcomes with those obtained with different time series and machine learning methods. Finally, we discuss the difficulty of volume forecasting when large quantities are abruptly traded.
    Date: 2020–05
  11. By: Kalamara, Eleni (King’s College London); Turrell, Arthur (Bank of England); Redl, Chris (International Monetary Fund); Kapetanios, George (King’s College London); Kapadia, Sujit (European Central Bank)
    Abstract: We consider the best way to extract timely signals from newspaper text and use them to forecast macroeconomic variables using three popular UK newspapers that collectively represent UK newspaper readership in terms of political perspective and editorial style. We find that newspaper text can improve economic forecasts both in absolute and marginal terms. We introduce a powerful new method of incorporating text information in forecasts that combines counts of terms with supervised machine learning techniques. This method improves forecasts of macroeconomic variables including GDP, inflation, and unemployment, including relative to existing text-based methods. Forecast improvements occur when it matters most, during stressed periods.
    Keywords: Text; forecasting; machine learning
    JEL: C55 J42
    Date: 2020–05–22
  12. By: Rui Dong (University of China); Raymond Fisman (Boston University); Yongxiang Wang (University of Southern California); Nianhang Xu (Renmin University of China)
    Abstract: We document a negative relation between air pollution during corporate site visits by investment analysts and subsequent earnings forecasts. After accounting for analyst, weather , and firm characteristics, an extreme worsening of air quality from “good/excellent†to “severely polluted†is associated with a more than 1 percentage point lower profit forecast, relative to realized profits. We explore heterogeneity in the pollution-forecast relation to understand better the underlying mechanism. Pollution only affects forecasts that are announced in the weeks immediately following a visit, indicating that mood likely plays a role, and the effect of pollution is less pronounced when analysts from different brokerages visit on the same date, suggesting a debiasing effect of multiple perspectives. Finally, there is suggestive evidence of adaptability to environmental circumstances – forecasts from analysts based in high pollution cities are relatively unaffected by site visit pollution.
    Keywords: Pollution; Forecasting bias; Investment analysts; Adaptation
    JEL: D91 Q5
    Date: 2019–12
  13. By: Wang, Dandan; Escribano Saez, Alvaro
    Abstract: The use of machine learning (ML) models has been shown to have advantages over alternative and more traditional time series models in the presence of big data. One of the most successful ML forecasting procedures is the Random Forest (RF) machine learning algorithm. In this paper we propose a mixed RF approach for modeling departures from linearity, instead of starting with a completely nonlinear or nonparametric model. The methodology is applied to the weekly forecasts of gasoline prices that are cointegrated with international oil prices and exchange rates. The question of interest is whether gasoline prices react asymmetrically to increases in oil prices rather than to decreases in oil prices, the "rockets and feathers" hypothesis. In this literature most authors estimate parametric nonlinear error correction models using nonlinear least squares. Recent specifications for nonlinear error correction models include threshold autoregressive models (TAR), double threshold error correction models (ECM) or double threshold smooth transition autoregressive (STAR) models. In this paper, we describe the econometric methodology that combines linear dynamic autoregressive distributed lag (ARDL) models with cointegrated variables with added nonlinear components, or price asymmetries, estimated by the powerful tool of RF. We apply our mixed RF specification strategy to weekly prices of the Spanish gasoline market from 2010 to 2019. We show that the new mixed RF error correction model has important advantages over competing parametric and nonparametric models, in terms of the generality of model specification, estimation and forecasting.
    Keywords: Mixed Random Forest; Random Forest; Machine Learning; Nonlinear Error Correction; Cointegration; Rockets And Feathers Hypothesis; Forecasting Gasoline Prices
    JEL: L71 L13 D43 C53 C52 C24 B23
    Date: 2020–06–04
  14. By: Xinyue Cui; Zhaoyu Xu; Yue Zhou
    Abstract: In this essay, we have comprehensively evaluated the feasibility and suitability of adopting the Machine Learning Models on the forecast of corporation fundamentals (i.e. the earnings), where the prediction results of our method have been thoroughly compared with both analysts' consensus estimation and traditional statistical models. As a result, our model has already been proved to be capable of serving as a favorable auxiliary tool for analysts to conduct better predictions on company fundamentals. Compared with previous traditional statistical models being widely adopted in the industry like Logistic Regression, our method has already achieved satisfactory advancement on both the prediction accuracy and speed. Meanwhile, we are also confident enough that there are still vast potentialities for this model to evolve, where we do hope that in the near future, the machine learning model could generate even better performances compared with professional analysts.
    Date: 2020–05
  15. By: Camille Cornand (Centre National de la Recherche Scientifique (CNRS)); Paul Hubert (Observatoire français des conjonctures économiques)
    Abstract: Establishing the external validity of experimental inflation forecasts is essential if laboratory experiments are to be used as decision-making tools for monetary policy. Our contribution is to document whether different measures of inflation expectations, based on various categories of agents (participants in experiments, households, industry forecasters, professional forecasters, financial market participants and central bankers), share common patterns. We do so by analyzing the forecasting performance of these different categories of data, their deviations from full information rational expectations, and the variables that enter the determination of these expectations. Overall, the different categories of forecasts exhibit common features: forecast errors are comparably large and autocorrelated, and forecast errors and forecast revisions are predictable from past information, suggesting the presence of some form of bounded rationality or information imperfections. Finally, lagged inflation positively affects the determination of inflation expectations. While experimental forecasts are relatively comparable to survey and financial market data, more heterogeneity is observed compared to central bank forecasts.
    Keywords: Inflation expectations; Experimental forecasts; Survey forecasts; Market-based forecasts; Central bank forecasts
    JEL: E3 E5
    Date: 2020–01
  16. By: Antolin-Diaz, Juan (London Business School); Petrella, Ivan (University of Warwick); Rubio-Ramirez, Juan F. (Emory University)
    Abstract: Macroeconomists seeking to construct conditional forecasts often face a choice between taking a stand on the details of a fully-specified structural model or relying on empirical correlations from vector autoregressions and remain silent about the underlying causal mechanisms. This paper develops tools for constructing “structural scenarios” that can be given an economic interpretation using identified structural VARs. We provide a unified and transparent treatment of conditional forecasting and structural scenario analysis and relate our approach to entropic forecast tilting. We advocate for a careful treatment of uncertainty, making the methods suitable for density forecasting and risk assessment. We also propose a metric to assess and compare the plausibility of alternative scenarios. We illustrate our methods with two applications: assessing the power of forward guidance about future interest rate policies and stress testing the reaction of bank profitability to an economic recession.
    Keywords: conditional forecasts ; SVARs ; Bayesian Methods ; Forward Guidance ; stress testing ;
    JEL: E37
    Date: 2020
  17. By: Tobias Hartl
    Abstract: We combine high-dimensional factor models with fractional integration methods and derive models where nonstationary, potentially cointegrated data of different persistence is modelled as a function of common fractionally integrated factors. A two-stage estimator, that combines principal components and the Kalman filter, is proposed. The forecast performance is studied for a high-dimensional US macroeconomic data set, where we find that benefits from the fractional factor models can be substantial, as they outperform univariate autoregressions, principal components, and the factor-augmented error-correction model.
    Date: 2020–05
  18. By: Dimitris Korobilis
    Abstract: This paper proposes two distinct contributions to econometric analysis of large information sets and structural instabilities. First, it treats a regression model with time-varying coefficients, stochastic volatility and exogenous predictors, as an equivalent high-dimensional static regression problem with thousands of covariates. Inference in this specification proceeds using Bayesian hierarchical priors that shrink the high-dimensional vector of coefficients either towards zero or time-invariance. Second, it introduces the frameworks of factor graphs and message passing as a means of designing efficient Bayesian estimation algorithms. In particular, a Generalized Approximate Message Passing (GAMP) algorithm is derived that has low algorithmic complexity and is trivially parallelizable. The result is a comprehensive methodology that can be used to estimate time-varying parameter regressions with arbitrarily large number of exogenous predictors. In a forecasting exercise for U.S. price inflation this methodology is shown to work very well.
    Keywords: high-dimensional inference; factor graph; Belief Propagation; Bayesian shrinkage; time-varying parameter model
    JEL: C11 C22 C52 C55 C61
    Date: 2019–09
  19. By: Neeti Mathur (NIIT University, Neemrana); Himanshu Mathur (Bright Minds Education Society)
    Abstract: The National Stock Exchange and Bombay Stock Exchange are the two major stock exchanges in India. The Bombay Stock Exchange is the first stock exchange of Asia and 10th largest stock exchange in the world in the terms of market capitalisation. Stock markets significantly contributes in the economic development of India. The stock markets have volatile character which results into the uncertainty of the returns, volatility is caused by the variability in speculative market prices and the instability of business performance. Volatility plays a significant role in financial decisions of the investors, managers, policy makers and the researchers as it can assess the risk exposures in their investments and the uncertainty in stocks returns. The risk averse investor avoid investment in highly volatile market. The stock return forecasting leads to volatility forecasting. This paper has made an attempt to analyse the volatility with reference to Bombay Stock Exchange. The daily data of S&P Sensex 30 has been collected and used to calculate the volatility of stock market in India for last 3 years (April 2016 to March 2019). The preliminary analysis is done on the basis of descriptive statistics Stationery test, Normality test and serial correlation test. Volatility modelling is done by the ARCH and GARCH family models.The findings of the study will help investors in taking good investment decisions in Indian stock market in the presence of its volatile character.
    Keywords: ARCH Model, Custer Analysis Diversification, Expansion, Generalized ARCH Model (GARCH Model), Growth, Return, Risk
    JEL: C55 C19
    Date: 2020–02
  20. By: Kristof Decock; Michela Bergamini; Koenraad Debackere; Enrico Lupi; Anne Mieke Vandamme; Bart Van Looy
    Abstract: In this paper we advance a set of heuristics which allow to predict ex ante the peak of diffusion curves and apply these heuristics on the COVID-19 pandemic (casualties). The heuristics build on innovation diffusion models and combine an extensive grid search with a loss function. The grid search is designed such that multi-finality (different end states) can unfold; the loss function takes into account the fit with a limited set of available observations. No assumptions are made ex ante in terms of the timing of inflection points. As such, these heuristics combine scenario thinking with forecasting algorithms (scenario driven forecasting). The heuristics have been applied for both Italy and Belgium as a whole as well as for decomposed time series based on policy relevant cohorts (regions (IT); hospital / residential care centers (BE)). While actually observed peaks (including Black Friday in Italy) are consistently falling into the predicted time range, we also observe that the predictive validity increases when decomposed time series – coinciding with cohorts displaying different, but policy relevant, diffusion dynamics– are being introduced. As the heuristics implied are agnostic in terms of epidemiological parameters – which are unknown in the case of a novel pathogen – they provide a decision-making space which is highly informative in situations characterized by levels of Knightian uncertainty. As such, scenario driven forecasting might become a valuable alternative and complement both for more qualitative approaches as well as for hope (and guesses) in decision making contexts characterized by such profound uncertainty. Acknowledgment: This contribution benefited from useful input and reflections by Jorge Ricardo Blanco Nova (KU Leuven), Sien Luyten (Flanders Business School), Xiaoyan Song (KU Leuven) and Stijn Kelchtermans (KU Leuven) and feedback from our students at Flanders Business School (MBA; Panther Program) and KU Leuven (Strategy and Innovation; Technology Trends and Opportunities). We want to express our gratitude to the Rega Institute and the Institute for the Future for providing a context to validate our models, and to EURO POOL GROUP for funding part of the research reported here.
    Date: 2020–05–19

This nep-for issue is ©2020 by Rob J Hyndman. 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 For comments please write to the director of NEP, Marco Novarese at <>. 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.