nep-for New Economics Papers
on Forecasting
Issue of 2020‒05‒04
twelve papers chosen by
Rob J Hyndman
Monash University

  1. The Power of Narratives in Economic Forecasts By Christopher A. Hollrah; Steven A. Sharpe; Nitish R. Sinha
  2. Designing a NISQ reservoir with maximal memory capacity for volatility forecasting By Samudra Dasgupta; Kathleen E. Hamilton; Pavel Lougovski; Arnab Banerjee
  3. A New Metric for Lumpy and Intermittent Demand Forecasts: Stock-keeping-oriented Prediction Error Costs By Dominik Martin; Philipp Spitzer; Niklas K\"uhl
  4. Forecasting inflation with the New Keynesian Phillips curve : Frequency matters By Martins, Manuel M. F.; Verona, Fabio
  5. Forecasting tourism with targeted predictors in a data-rich environment By António Rua; Carlos Melo Gouveia; Nuno Lourenço
  6. Modern currency exchange rate behaviour and proposed trend-like forecasting model By Tweneboah Senzu, Emmanuel
  7. Time-frequency forecast of the equity premium By Faria, Gonçalo; Verona, Fabio
  8. Identifying Risk Factors and Their Premia: A Study on Electricity Prices By Wei Wei; Asger Lunde
  9. Volatility Forecasting in European Government Bond Markets By Özbekler, Ali Gencay; Kontonikas, Alexandros; Triantafyllou, Athanasios
  10. Forecasting directional movements of stock prices for intraday trading using LSTM and random forests By Pushpendu Ghosh; Ariel Neufeld; Jajati Keshari Sahoo
  11. Time-Varying Predictability of Financial Stress on Inequality in United Kingdom By Edmond Berisha; David Gabauer; Rangan Gupta; Jacobus Nel
  12. Estimate of underreporting of COVID-19 in Brazil by Acute Respiratory Syndrome hospitalization reports By Leonardo Costa Ribeiro; Américo Tristão Bernardes

  1. By: Christopher A. Hollrah; Steven A. Sharpe; Nitish R. Sinha
    Abstract: We apply textual analysis tools to the narratives that accompany Federal Reserve Board economic forecasts to measure the degree of optimism versus pessimism expressed in those narratives. Text sentiment is strongly correlated with the accompanying economic point forecasts, positively for GDP forecasts and negatively for unemployment and inflation forecasts. Moreover, our sentiment measure predicts errors in FRB and private forecasts for GDP growth and unemployment up to four quarters out. Furthermore, stronger sentiment predicts tighter than expected monetary policy and higher future stock returns. Quantile regressions indicate that most of sentiment’s forecasting power arises from signaling downside risks to the economy and stock prices.
    Keywords: Text analysis; Economic forecasts; Monetary policy; Stock returns; Narratives
    JEL: C53 E17 E27 E37 E52 G14
    Date: 2020–01–03
  2. By: Samudra Dasgupta; Kathleen E. Hamilton; Pavel Lougovski; Arnab Banerjee
    Abstract: Quantitative risk management, particularly volatility forecasting, is critically important to traders, portfolio managers as well as policy makers. In this paper, we applied quantum reservoir computing for forecasting VIX (the CBOE volatility index), a highly non-linear and memory intensive `real-life' signal that is driven by market dynamics and trader psychology and cannot be expressed by a deterministic equation. As a first step, we lay out the systematic design considerations for using a NISQ reservoir as a computing engine (which should be useful for practitioners). We then show how to experimentally evaluate the memory capacity of various reservoir topologies (using IBM-Q's Rochester device) to identify the configuration with maximum memory capacity. Once the optimal design is selected, the forecast is produced by a linear combination of the average spin of a 6-qubit quantum register trained using VIX and SPX data from year 1990 onwards. We test the forecast performance over the sub-prime mortgage crisis period (Dec 2007 - Jun 2009). Our results show a remarkable ability to predict the volatility during the Great Recession using today's NISQs.
    Date: 2020–04
  3. By: Dominik Martin; Philipp Spitzer; Niklas K\"uhl
    Abstract: Forecasts of product demand are essential for short- and long-term optimization of logistics and production. Thus, the most accurate prediction possible is desirable. In order to optimally train predictive models, the deviation of the forecast compared to the actual demand needs to be assessed by a proper metric. However, if a metric does not represent the actual prediction error, predictive models are insufficiently optimized and, consequently, will yield inaccurate predictions. The most common metrics such as MAPE or RMSE, however, are not suitable for the evaluation of forecasting errors, especially for lumpy and intermittent demand patterns, as they do not sufficiently account for, e.g., temporal shifts (prediction before or after actual demand) or cost-related aspects. Therefore, we propose a novel metric that, in addition to statistical considerations, also addresses business aspects. Additionally, we evaluate the metric based on simulated and real demand time series from the automotive aftermarket.
    Date: 2020–04
  4. By: Martins, Manuel M. F.; Verona, Fabio
    Abstract: We show that the New Keynesian Phillips Curve (NKPC) outperforms standard benchmarks in forecasting U.S. inflation once frequency-domain information is taken into account. We do so by decomposing the time series (of inflation and its predictors) into several frequency bands and forecasting separately each frequency component of inflation. The largest statistically significant forecasting gains are achieved with a model that forecasts the lowest frequency component of inflation (corresponding to cycles longer than 16 years) flexibly using information from all frequency components of the NKPC inflation predictors. Its performance is particularly good in the returning to recovery from the Great Recession.
    JEL: C53 E31 E37
    Date: 2020–04–21
  5. By: António Rua; Carlos Melo Gouveia; Nuno Lourenço
    Abstract: Along with the deepening of globalization and economic integration, economic agents face the challenge on how to extract useful information from large panels of data for forecasting purposes. Herein, we lay out a modelling strategy to explore the predictive content of large datasets for tourism forecasting. In particular, we assess the role of multi-country datasets to nowcast and forecast tourism by resorting to factor models with targeted predictors to cope with such a data-rich environment. Drawing on business and consumer surveys for Portugal and its main tourism source markets, we document the usefulness of factor models to forecast tourism exports up to several months ahead. Moreover, we find that forecast performance is enhanced if predictors are chosen before factors are estimated.
    JEL: C53 F47
    Date: 2020
  6. By: Tweneboah Senzu, Emmanuel
    Abstract: The study examined high volatile assets, specifically the currency exchange rate of the open financial market. Takes into consideration the five most traded paired currencies of the global financial market. And observed, generally, the dataset of the unit currency exchange rate exhibit homoscedastic qualities making it appropriate for the use of auto-regression integrated moving average as a reliable model forecast for future pricing of the volatile assets. However, the current model prediction addresses only the magnitude of asset price ignoring its direction, which is the paramount challenge of forecasters. Hence the paper resolves such weakness of the model by introducing a momentum model as a complementary tool to the ARIMA model to determine not only price magnitude but the vector direction of volatile asset pricing relative to the market, dependent on its lagged values.
    Keywords: Forecast, Momentum-model, Exchange rate, Homoscedacity, ARIMA, GARCH, Hard-currency
    JEL: G02 G11 G12 G15 G17
    Date: 2020–05–01
  7. By: Faria, Gonçalo; Verona, Fabio
    Abstract: Any time series can be decomposed into cyclical components fluctuating at different frequencies. Accordingly, in this paper we propose a method to forecast the stock market's equity premium which exploits the frequency relationship between the equity premium and several predictor variables. We evaluate a large set of models and find that, by selecting the relevant frequencies for equity premium forecasting, this method significantly improves in both statistical and economic sense upon standard time series forecasting methods. This improvement is robust regardless of the predictor used, the out-of-sample period considered, and the frequency of the data used.
    JEL: C58 G11 G17
    Date: 2020–04–27
  8. By: Wei Wei; Asger Lunde
    Abstract: We propose a multi-factor model and an estimation method based on particle MCMC to identify risk factors in electricity prices. Our model identifies long-run prices, shortrun deviations, and spikes as three main risk factors in electricity spot prices. Under our model, different risk factors have distinct impacts on futures prices and can carry different risk premia. We generalize the Fama-French regressions to analyze properties of true risk premia. We show that model specification plays an important role in detecting time varying risk premia. Using spot and futures prices in the Germany/Austria market, we demonstrate that our proposed model surpasses alternative models that have less risk factors in forecasting spot prices and in detecting time varying risk premia.
    Keywords: Risk factors, risk premia, futures, particle filter, MCMC.
    JEL: C51 G13 Q4
    Date: 2020
  9. By: Özbekler, Ali Gencay; Kontonikas, Alexandros; Triantafyllou, Athanasios
    Abstract: In this paper we examine the predictive power of the Heterogeneous Autoregressive (HAR) model on Treasury bond return volatility of major European government bond markets. The HAR-type volatility forecasting models show that short term and medium term volatility is a robust and statistically significant predictor of the term structure of intradayvolatility of bonds with maturities ranging from 1-year up to 30-years. When decomposing volatility into its continuous and discontinuous (jump) component, we find that the jump tail risk component is a significant predictor of bond market volatility. We lastly show that approximately half of the monetary policy announcement dates coincide with the presence of jumps in bond returns, and the pre-announcement drift is present in the bond market. Hence, the monetary policy announcements are important determinant of European bond market volatility.
    Keywords: Treasury Bonds, Jumps, Realized Volatility, Macroeconomic Announcements, Volatility Forecasting
    Date: 2020–04–24
  10. By: Pushpendu Ghosh; Ariel Neufeld; Jajati Keshari Sahoo
    Abstract: We employ both random forests and LSTM networks (more precisely CuDNNLSTM) as training methodologies to analyze their effectiveness in forecasting out-of-sample directional movements of constituent stocks of the S&P 500 from January 1993 till December 2018 for intraday trading. We introduce a multi-feature setting consisting not only of the returns with respect to the closing prices, but also with respect to the opening prices and intraday returns. As trading strategy, we use Krauss et al. (2017) and Fischer & Krauss (2018) as benchmark and, on each trading day, buy the 10 stocks with the highest probability and sell short the 10 stocks with the lowest probability to outperform the market in terms of intraday returns -- all with equal monetary weight. Our empirical results show that the multi-feature setting provides a daily return, prior to transaction costs, of 0.64% using LSTM networks, and 0.54% using random forests. Hence we outperform the single-feature setting in Fischer & Krauss (2018) and Krauss et al. (2017) consisting only of the daily returns with respect to the closing prices, having corresponding daily returns of 0.41% and of 0.39% with respect to LSTM and random forests, respectively.
    Date: 2020–04
  11. By: Edmond Berisha (Feliciano School of Business, Montclair State University, Montclair, NJ 07043, USA); David Gabauer (Institute of Applied Statistics, Johannes Kepler University, Altenbergerstraße 69, 4040 Linz, Austria); Rangan Gupta (Department of Economics, University of Pretoria, Pretoria, 0002, South Africa); Jacobus Nel (Department of Economics, University of Pretoria, Pretoria, 0002, South Africa)
    Abstract: In this paper, we analyze time-varying predictability of financial stress for growth in income (and consumption) inequality of the United Kingdom (UK) based on a high-frequency (quarterly) data set over 1975:2 to 2016:1. Results indicate that a well-established index of financial stress, derived from the European Central Bank, has a strong predictive power on growth rate of income (and to some extent consumption) inequality in the UK. Interestingly, the strength of the predictive power is found to be higher towards the beginning and end of the sample period corresponding to highly stressed financial markets in the UK. In addition, based on time-varying impulse response functions, we find that higher financial stress corresponds with subsequent increases in income inequality. Finally, the FSI is found to produce forecasting gains for the growth of income inequality over an out-of-sample period, especially at medium to long-runs.
    Keywords: Financial Stress, Inequality, Time-Varying Predictions
    JEL: C32 C53 D31 G01
    Date: 2020–04
  12. By: Leonardo Costa Ribeiro (Cedeplar-UFMG); Américo Tristão Bernardes (UFOP)
    Abstract: The number of COVID-19 infected people in each country is a crucial factor to determine public policies. It guides the governments to strengthen movement restrictions of people or to relieve it. The number of infected people is very important to forecast the needs of the health systems, which are collapsing in many countries. Thus, underreporting of infected people is a huge problem, since authorities do not know the real problem and act in darkness. In the present work, we discuss this subject for the Brazilian case. We take the time series of acute respiratory syndromes reported in the health public system in the last ten years and estimated the number for March/20 when the COVID-19 appeared in Brazil. Our results show a 7.7:1 rate of underreporting, meaning that the real cases in Brazil should be, at least, seven times the publicized number.
    Keywords: Corona virus, COVID-19; Underreporting; Brazil
    JEL: C15 I18
    Date: 2020–04

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.