nep-for New Economics Papers
on Forecasting
Issue of 2020‒06‒22
ten papers chosen by
Rob J Hyndman
Monash University

  1. The impacts of asymmetry on modeling and forecasting realized volatility in Japanese stock markets By Daiki Maki; Yasushi Ota
  2. On the Performance of US Fiscal Forecasts: Government vs. Private Information By Zidong An; João Tovar Jalles
  3. Comparing Predictive Accuracy in the Presence of a Loss Function Shape Parameter By Sander Barendse; Andrew J. Patton
  4. The Role of Investor Sentiment in Forecasting Housing Returns in China: A Machine Learning Approach By Oguzhan Cepni; Rangan Gupta; Yigit Onay
  5. A Note on Uncertainty due to Infectious Diseases and Output Growth of the United States: A Mixed-Frequency Forecasting Experiment By Afees A. Salisu; Rangan Gupta; Riza Demirer
  6. Econometric Methods and Data Science Techniques: A Review of Two Strands of Literature and an Introduction to Hybrid Methods By Xie, Tian; Yu, Jun; Zeng, Tao
  7. Nowcasting Turkish GDP Growth with Targeted Predictors: Fill in the Blanks By Mahmut Gunay
  8. Forecasting Oil Volatility Using a GARCH-MIDAS Approach: The Role of Global Economic Conditions By Afees A. Salisu; Rangan Gupta; Elie Bouri
  9. Forecasting natural gas prices using highly flexible time-varying parameter models By Shen Gao; Chenghan Hou; Bao H. Nguyen
  10. Earnings Prediction with Deep Leaning By Lars Elend; Sebastian A. Tideman; Kerstin Lopatta; Oliver Kramer

  1. By: Daiki Maki; Yasushi Ota
    Abstract: This study investigates the impacts of asymmetry on the modeling and forecasting of realized volatility in the Japanese futures and spot stock markets. We employ heterogeneous autoregressive (HAR) models allowing for three types of asymmetry: positive and negative realized semivariance (RSV), asymmetric jumps, and leverage effects. The estimation results show that leverage effects clearly influence the modeling of realized volatility models. Leverage effects exist for both the spot and futures markets in the Nikkei 225. Although realized semivariance aids better modeling, the estimations of RSV models depend on whether these models have leverage effects. Asymmetric jump components do not have a clear influence on realized volatility models. While leverage effects and realized semivariance also improve the out-of-sample forecast performance of volatility models, asymmetric jumps are not useful for predictive ability. The empirical results of this study indicate that asymmetric information, in particular, leverage effects and realized semivariance, yield better modeling and more accurate forecast performance. Accordingly, asymmetric information should be included when we model and forecast the realized volatility of Japanese stock markets.
    Date: 2020–05
  2. By: Zidong An; João Tovar Jalles
    Abstract: This paper contributes to shed light on the quality and performance of US fiscal forecasts. The first part inspects the causes of official (CBO) fiscal forecasts revisions between 1984 and 2016 that are due to technical, economic or policy reasons. Both individual and cumulative means of forecast errors are relatively close to zero, particularly in the case of expenditures. CBO averages indicate net average downward revenue and expenditure revisions and net average upward deficit revisions. Focusing on the causes of the technical component, we uncover that its revisions are quite unpredictable which casts doubts on inferences about fiscal policy sustainability that rely on point estimates. Comparing official with private-sector (Consensus) forecasts, despite the informational advantages CBO might have, one cannot unequivocally say that one or the other is more accurate. Evidence also seems to suggest that CBO forecasts are consistently heavily biased towards optimism while this is less the case for Consensus forecasts. Not only is the extent of information rigidity is more prevalent in CBO forecasts, but evidence also seems to indicate that Consensus forecasts dominate CBO’s in terms of information content.
    Keywords: forecasting performance, encompassing tests, CBO, Consensus
    JEL: C53 E17 H62
    Date: 2020–05
  3. By: Sander Barendse; Andrew J. Patton
    Abstract: We develop tests for out-of-sample forecast comparisons based on loss functions that contain shape parameters. Examples include comparisons using average utility across a range of values for the level of risk aversion, comparisons of forecast accuracy using characteristics of a portfolio return across a range of values for the portfolio weight vector, and comparisons using a recently-proposed “Murphy diagrams†for classes of consistent scoring rules. An extensive Monte Carlo study verifies that our tests have good size and power properties in realistic sample sizes, particularly when compared with existing methods which break down when then number of values considered for the shape parameter grows. We present three empirical illustrations of the new test.
    Keywords: Forecasting, model selection, out-of-sample testing, nuisance parameters
    JEL: C53 C52 C12
    Date: 2020–05–27
  4. By: Oguzhan Cepni (Central Bank of the Republic of Turkey, Haci Bayram Mah. Istiklal Cad. No:10 06050, Ankara, Turkey); Rangan Gupta (Department of Economics, University of Pretoria, Pretoria, 0002, South Africa); Yigit Onay (Central Bank of the Republic of Turkey, Haci Bayram Mah. Istiklal Cad. No:10 06050, Ankara, Turkey)
    Abstract: This paper analyzes the predictive ability of aggregate and dis-aggregate proxies of investor sentiment, over and above standard macroeconomic predictors, in forecasting housing returns in China, using an array of machine learning models. Using a monthly out-of-sample period of 2011:01 to 2018:12, given an in-sample of 2006:01-2010:12, we find that indeed the new aligned investor sentiment index proposed in this paper has greater predictive power for housing returns than the a principal component analysis (PCA)-based sentiment index, used earlier in the literature. Moreover, shrinkage models utilizing the dis-aggregate sentiment proxies do not result in forecast improvement indicating that aligned sentiment index optimally exploits information in the dis-aggregate proxies of investor sentiment. Furthermore, when we let the machine learning models to choose from all key control variables and the aligned sentiment index, the forecasting accuracy is improved at all forecasting horizons, rather than just the short-run as witnessed under standard predictive regressions. This result suggests that machine learning methods are flexible enough to capture both structural change and time-varying information in a set of predictors simultaneously to forecast housing returns of China in a precise manner. Given the role of the real estate market in China’s economic growth, our result of accurate forecasting of housing returns, based on investor sentiment and macroeconomic variables using state-of-the-art machine learning methods, has important implications for both investors and policymakers.
    Keywords: Housing prices, Investor sentiment, Bayesian shrinkage, Time-varying parameter model
    JEL: C22 C32 C52 G12 R31
    Date: 2020–06
  5. 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); Riza Demirer (Department of Economics and Finance, Southern Illinois University Edwardsville, Edwardsville, IL 62026-1102, USA)
    Abstract: Utilizing a mixed data sampling (MIDAS) approach, we show that a daily newspaper based index of uncertainty associated with infectious diseases can be used to predict, both in- and out-of-sample, low-frequency movements of output growth for the United States (US). The predictability of monthly industrial production growth and quarterly real Gross Domestic Product (GDP) growth during the current period of heightened economic uncertainty due to the COVID-19 pandemic is likely to be of tremendous value to policymakers.
    Keywords: Infectious Diseases Related Uncertainty, Output Growth, Forecast, Mixed-Frequency
    JEL: C22 C53 D80 E23 E32
    Date: 2020–05
  6. By: Xie, Tian (Shanghai University of Finance and Economics); Yu, Jun (School of Economics, Singapore Management University); Zeng, Tao (Zhejiang University)
    Abstract: The data market has been growing at an exceptional pace. Consequently, more sophisticated strategies to conduct economic forecasts have been introduced with machine learning techniques. Does machine learning pose a threat to conventional econometric methods in terms of forecasting? Moreover, does machine learning present great opportunities to cross-fertilize the field of econometric forecasting? In this report, we develop a pedagogical framework that identifies complementarity and bridges between the two strands of literature. Existing econometric methods and machine learning techniques for economic forecasting are reviewed and compared. The advantages and disadvantages of these two classes of methods are discussed. A class of hybrid methods that combine conventional econometrics and machine learning are introduced. New directions for integrating the above two are suggested. The out-of-sample performance of alternatives is compared when they are employed to forecast the Chicago Board Options Exchange Volatility Index and the harmonized index of consumer prices for the euro area. In the first exercise, econometric methods seem to work better, whereas machine learning methods generally dominate in the second empirical application.
    Date: 2020–05–30
  7. By: Mahmut Gunay
    Abstract: This paper analyzes four dimensions of forecasting GDP growth using monthly data. Firstly, we use AR, VAR, BVAR, mean-growth and zero month-on-month change for forecasting the missing monthly data at the end of forecasting sample due to asynchronous nature of the release of the indicators. Second dimension is using a relatively large data set and testing some indicators that are not frequently used for forecasting GDP growth but due to timeliness have the potential to contribute to the forecasting performance. We analyze data from a career website, freight information from maritime transportation, capacity utilization of available plane seats, tax revenues of the central government and credit and debit card transaction volumes. Third dimension is comparing the performance of model averaging and factor models that are used to incorporate information content of large data sets to the forecasting process. Finally, we look at the forecasting performance of a core data set that is selected by a shrinkage method, namely LASSO. Our findings show that using VAR models with financial and survey indicators for forecasting missing monthly data improves short term GDP forecasting performance relative to other alternatives. We find that forecasting using targeted predictors rather than using an unscreened large data set helps to reduce forecasting errors considerably. Factor model approach performs better than forecast combination. So, using a targeted data set for factor extraction and forecasting missing monthly data with VAR performs relatively better than other specifications for producing timely and accurate nowcasts.
    Keywords: GDP forecasting, Bridge models, Factor models, LASSO, Targeted predictors
    JEL: C52 C53 E20
    Date: 2020
  8. 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 offer two main innovations. First, we subject six alternative indicators of global economic activity, including the one recently developed by Baumeister et al. (2020), to empirical tests of their relative predictive powers for crude oil market volatility. Second, we accommodate all the relevant series at their available data frequencies using the GARCH-MIDAS approach, thereby circumventing information loss and any associated bias. We find evidence in support of the ability of global economic activity to predict energy market volatility. Our forecast evaluation of the various indicators places a higher weight on the newly developed indicator of global economic activity by Baumeister et al. (2020), based on a set of 16 variables covering multiple dimensions of the global economy, than other indicators. The results leading to these conclusions are robust to multiple forecast horizons and consistent across alternative energy sources.
    Keywords: Energy Markets Volatility, Global Economic Conditions, Mixed-Frequency
    JEL: C32 C53 E32 Q41
    Date: 2020–05
  9. By: Shen Gao; Chenghan Hou; Bao H. Nguyen
    Abstract: The growing disintegration between the natural gas and oil prices, together with shale revolution and market financialization, lead to continued fundamental changes in the natural gas markets. To capture these structural changes, this paper considers a wide set of highly flexible time-varying parameter models to evaluate the out-of-sample forecasting performance of the natural gas spot prices across the US, European and Japanese markets. The results show that for both Japan and EU markets, the best forecasting performance is found when the model allows for drastic changes in the conditional mean and gradual changes in the conditional volatility. For the US market, however, no model performs systematically better than the simple autoregressive model. Full sample estimation results further confirm that allowing t-distributed error is important in modelling the natural gas prices, especially for EU markets.
    Keywords: Natural gas price, Structural breaks, Forecasting, Time-varying parameter, Markov switching, Stochastic volatility
    JEL: C32 E32 Q43
    Date: 2020–03
  10. By: Lars Elend; Sebastian A. Tideman; Kerstin Lopatta; Oliver Kramer
    Abstract: In the financial sector, a reliable forecast the future financial performance of a company is of great importance for investors' investment decisions. In this paper we compare long-term short-term memory (LSTM) networks to temporal convolution network (TCNs) in the prediction of future earnings per share (EPS). The experimental analysis is based on quarterly financial reporting data and daily stock market returns. For a broad sample of US firms, we find that both LSTMs outperform the naive persistent model with up to 30.0% more accurate predictions, while TCNs achieve and an improvement of 30.8%. Both types of networks are at least as accurate as analysts and exceed them by up to 12.2% (LSTM) and 13.2% (TCN).
    Date: 2020–06

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.