nep-for New Economics Papers
on Forecasting
Issue of 2021‒02‒22
nine papers chosen by
Rob J Hyndman
Monash University

  1. "Nowcasting and forecasting GDP growth with machine-learning sentiment indicators". By Oscar Claveria; Enric Monte; Salvador Torra
  2. Smooth Robust Multi-Horizon Forecasts By Andrew B. Martinez; Jennifer L. Castle; David F. Hendry
  3. REST: Relational Event-driven Stock Trend Forecasting By Wentao Xu; Weiqing Liu; Chang Xu; Jiang Bian; Jian Yin; Tie-Yan Liu
  4. Real-Time Fixed-Target Statistical Prediction of Arctic Sea Ice Extent By Francis X. Diebold; Maximilian Gobel
  5. Financial Vulnerability and Volatility in Emerging Stock Markets: Evidence from GARCH-MIDAS Models By Riza Demirer; Rangan Gupta; He Li; Yu You
  6. Uncertainty and Forecastability of Regional Output Growth in the United Kingdom: Evidence from Machine Learning By Mehmet Balcilar; David Gabauer; Rangan Gupta; Christian Pierdzioch
  7. Peramalan Harga Emas Saat Pandemi Covid-19 Menggunakan Model Hybrid Autoregressive Integrated Moving Average - Support Vector Regression By Math., Jambura J.; Purnama, Drajat Indra
  8. Deep Learning for Market by Order Data By Zihao Zhang; Bryan Lim; Stefan Zohren
  9. Multi-Horizon Equity Returns Predictability via Machine Learning By Lenka Nechvatalova

  1. By: Oscar Claveria (AQR–IREA, Department of Econometrics, Statistics and Applied Economics, University of Barcelona, Diagonal 690, 08034 Barcelona, Spain.); Enric Monte (Department of Signal Theory and Communications, Polytechnic University of Catalunya (UPC).); Salvador Torra (Riskcenter–IREA, Department of Econometrics, Statistics and Applied Economics, University of Barcelona (UB).)
    Abstract: We apply the two-step machine-learning method proposed by Claveria et al. (2021) to generate country-specific sentiment indicators that provide estimates of year-on-year GDP growth rates. In the first step, by means of genetic programming, business and consumer expectations are evolved to derive sentiment indicators for 19 European economies. In the second step, the sentiment indicators are iteratively re-computed and combined each period to forecast yearly growth rates. To assess the performance of the proposed approach, we have designed two out-of-sample experiments: a nowcasting exercise in which we recursively generate estimates of GDP at the end of each quarter using the latest survey data available, and an iterative forecasting exercise for different forecast horizons We found that forecasts generated with the sentiment indicators outperform those obtained with time series models. These results show the potential of the methodology as a predictive tool.
    Keywords: Forecasting, Economic growth, Business and consumer expectations, Symbolic regression, Evolutionary algorithms, Genetic programming. JEL classification: C51, C55, C63, C83, C93.
    Date: 2021–02
  2. By: Andrew B. Martinez (Dept of Economics and Institute for New Economic Thinking at the Oxford Martin School, University of Oxford); Jennifer L. Castle (Magdalen College, Climate Econometrics and Institute for New Economic Thinking at the Oxford Martin School, University of Oxford); David F. Hendry (Nuffield College, Climate Econometrics and Institute for New Economic Thinking at the Oxford Martin School, University of Oxford)
    Abstract: We investigate whether smooth robust methods for forecasting can help mitigate pronounced and persistent failure across multiple forecast horizons. We demonstrate that naive predictors are interpretable as local estimators of the long-run relationship with the advantage of adapting quickly after a break, but at a cost of additional forecast error variance. Smoothing over naive estimates helps retain these advantages while reducing the costs, especially for longer forecast horizons. We derive the performance of these predictors after a location shift, and confirm the results using simulations. We apply smooth methods to forecasts of U.K. productivity and U.S. 10-year Treasury yields and show that they can dramatically reduce persistent forecast failure exhibited by forecasts from macroeconomic models and professional forecasters.
    Keywords: Location Shifts; Long differencing; Productivity forecasts; Robust forecasts. JEL codes: C51, C53
    Date: 2021–01–14
  3. By: Wentao Xu; Weiqing Liu; Chang Xu; Jiang Bian; Jian Yin; Tie-Yan Liu
    Abstract: Stock trend forecasting, aiming at predicting the stock future trends, is crucial for investors to seek maximized profits from the stock market. Many event-driven methods utilized the events extracted from news, social media, and discussion board to forecast the stock trend in recent years. However, existing event-driven methods have two main shortcomings: 1) overlooking the influence of event information differentiated by the stock-dependent properties; 2) neglecting the effect of event information from other related stocks. In this paper, we propose a relational event-driven stock trend forecasting (REST) framework, which can address the shortcoming of existing methods. To remedy the first shortcoming, we propose to model the stock context and learn the effect of event information on the stocks under different contexts. To address the second shortcoming, we construct a stock graph and design a new propagation layer to propagate the effect of event information from related stocks. The experimental studies on the real-world data demonstrate the efficiency of our REST framework. The results of investment simulation show that our framework can achieve a higher return of investment than baselines.
    Date: 2021–02
  4. By: Francis X. Diebold; Maximilian Gobel
    Abstract: We propose a simple statistical approach for fixed-target forecasting of Arctic sea ice extent, and we provide a case study of its real-time performance for target date September 2020. The real-time forecasting begins in early June and proceeds through late September. We visually detail the evolution of the statistically-optimal point, interval, and density forecasts as time passes, new information arrives, and the end of September approaches. Among other things, our visualizations may provide useful windows for assessing the agreement between dynamical climate models and observational data.
    Date: 2021–01
  5. By: Riza Demirer (Department of Economics & Finance, Southern Illinois University Edwardsville, Alumni Hall 3145, Edwardsville IL, 62026-1102, USA); Rangan Gupta (Department of Economics, University of Pretoria, Private Bag X20, Hatfield, 0028, South Africa); He Li (School of International Economics and Politics, Liaoning University, Shenyang, Liaoning, China); Yu You (Li Anmin Advanced Institute of Finance and Economics, Liaoning University, Shenyang, Liaoning, China)
    Abstract: This paper establishes a predictive relationship between financial vulnerability and volatility in emerging stock markets. Focusing on China and India and utilizing GARCH-MIDAS models, we show that incorporating financial vulnerability can substantially improve the forecasting power of standard macroeconomic fundamentals (output growth, inflation and monetary policy interest rate) for stock market volatility. The findings have significant implications for investors to improve the accuracy of volatility forecasts.
    Keywords: Stock Market Volatility, Financial Vulnerability, GARCH-MIDAS, Emerging Markets
    JEL: C32 C53 G15 G17
    Date: 2021–02
  6. By: Mehmet Balcilar (Eastern Mediterranean University, Famagusta, via Mersin 10, Northern Cyprus, Turkey); David Gabauer (Data Analysis Systems, Software Competence Center Hagenberg, Austria); Rangan Gupta (Department of Economics, University of Pretoria, Private Bag X20, Hatfield, 0028, South Africa); Christian Pierdzioch (Department of Economics, Helmut Schmidt University, Holstenhofweg 85, P.O.B. 700822, 22008 Hamburg, Germany)
    Abstract: Utilizing a machine-learning technique known as random forests, we study whether regional output growth uncertainty helps to improve the accuracy of forecasts of regional output growth for twelve regions of the United Kingdom using monthly data for the period from 1970 to 2020. We use a stochastic-volatility model to measure regional output growth uncertainty. We document the importance of interregional stochastic volatility spillovers and the direction of the transmission mechanism. Given this, our empirical results shed light on the contribution to forecast performance of own uncertainty associated with a particular region, output growth uncertainty of other regions, and output growth uncertainty as measured for London as well. We find that output growth uncertainty significantly improves forecast performance in several cases, where we also document cross-regional heterogeneity in this regard.
    Keywords: Regional Output Growth, Uncertainty, United Kingdom, Forecasting, Machine Learning
    JEL: C22 C53 D8 E32 E37 R11
    Date: 2021–02
  7. By: Math., Jambura J.; Purnama, Drajat Indra
    Abstract: Gold investment is one of the favorite investments during the Covid 19 pandemic as it is today. This is because the price of gold is relatively volatile but shows an increasing trend. Investors are required to be smart in investing in gold, able to predict future opportunities. One of the time series data forecasting models is the Autoregressive Integrated Moving Average (ARIMA) model. The ARIMA model is good for use on linear patterned data but if it is used on nonlinear data the accuracy decreases. To solve the problem of nonlinear data, you can use the Support Vector Regression (SVR) model. The linearity test on the gold price data shows that there are linear and nonlinear data patterns at the same time so that a combination of ARIMA and SVR is used, namely the ARIMA-SVR hybrid model. Forecasting results using the ARIMA-SVR hybrid model show better results than the ARIMA model. This is evidenced by the MAPE value of the ARIMA-SVR hybrid model which is smaller than the MAPE value of the ARIMA model. The MAPE value of the ARIMA-SVR hybrid model is 0.355 on the training data and 4.001 on the testing data, while the MAPE value of the ARIMA model is 0.903 in the training data and 4.076 in the testing data.
    Date: 2021–01–01
  8. By: Zihao Zhang; Bryan Lim; Stefan Zohren
    Abstract: Market by order (MBO) data - a detailed feed of individual trade instructions for a given stock on an exchange - is arguably one of the most granular sources of microstructure information. While limit order books (LOBs) are implicitly derived from it, MBO data is largely neglected by current academic literature which focuses primarily on LOB modelling. In this paper, we demonstrate the utility of MBO data for forecasting high-frequency price movements, providing an orthogonal source of information to LOB snapshots. We provide the first predictive analysis on MBO data by carefully introducing the data structure and presenting a specific normalisation scheme to consider level information in order books and to allow model training with multiple instruments. Through forecasting experiments using deep neural networks, we show that while MBO-driven and LOB-driven models individually provide similar performance, ensembles of the two can lead to improvements in forecasting accuracy -- indicating that MBO data is additive to LOB-based features.
    Date: 2021–02
  9. By: Lenka Nechvatalova (Institute of Economic Studies, Charles University and Institute of Information Theory and Automation, Czech Academy of Sciences Prague, Czech Republic)
    Abstract: We examine the predictability of expected stock returns across horizons using machine learning. We use neural networks, and gradient boosted regression trees on the U.S. and international equity datasets. We find that predictability of returns using neural networks models decreases with longer forecasting horizon. We also document the profitability of long-short portfolios, which were created using predictions of cumulative returns at various horizons, before and after accounting for transaction costs. There is a trade-off between higher transaction costs connected to frequent rebalancing and greater returns on shorter horizons. However, we show that increasing the forecasting horizon while matching the rebalancing period increases risk-adjusted returns after transaction cost for the U.S. We combine predictions of expected returns at multiple horizons using double-sorting and buy/hold spread, a turnover reducing strategy. Using double sorts significantly increases profitability on the U.S. sample. Buy/hold spread portfolios have better risk-adjusted profitability in the U.S.
    Keywords: Machine learning, asset pricing, horizon predictability, anomalies
    JEL: G11 G12 G15 C55
    Date: 2021–02

This nep-for issue is ©2021 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.