nep-for New Economics Papers
on Forecasting
Issue of 2022‒03‒07
eleven papers chosen by
Rob J Hyndman
Monash University

  1. Testing the Forecasting Power of Global Economic Conditions for the Volatility of International REITs using a GARCH-MIDAS Approach By Afees A. Salisu; Rangan Gupta; Elie Bouri
  2. Forecasting Environmental Data: An example to ground-level ozone concentration surfaces By Alexander Gleim; Nazarii Salish
  3. Long Short-Term Memory Neural Network for Financial Time Series By Carmina Fjellstr\"om
  4. Assessing Point Forecast Bias Across Multiple Time Series: Measures and Visual Tools By Davydenko, Andrey; Goodwin, Paul
  5. 2T-POT Hawkes model for dynamic left- and right-tail quantile forecasts of financial returns: out-of-sample validation of self-exciting extremes versus conditional volatility By Matthew F. Tomlinson; David Greenwood; Marcin Mucha-Kruczynski
  6. Simulating Using Deep Learning The World Trade Forecasting of Export-Import Exchange Rate Convergence Factor During COVID-19 By Effat Ara Easmin Lucky; Md. Mahadi Hasan Sany; Mumenunnesa Keya; Md. Moshiur Rahaman; Umme Habiba Happy; Sharun Akter Khushbu; Md. Arid Hasan
  7. Mis-specified Forecasts and Myopia in an Estimated New Keynesian Model By Ina Hajdini
  8. Predicting The Stock Trend Using News Sentiment Analysis and Technical Indicators in Spark By Taylan Kabbani; Fatih Enes Usta
  9. Strengths and weaknesses of the logistic function used in forecasting By MODIS, THEODORE
  10. Forecasting the distribution of long-horizon returns with time-varying volatility By Hwai-Chung Ho
  11. Building a predictive machine learning model of gentrification in Sydney By Thackway, William; Ng, Matthew Kok Ming; Lee, Chyi Lin; Pettit, Christopher

  1. By: Afees A. Salisu (Centre for Econometric and Allied Research, University of Ibadan, Ibadan, Nigeria; Department of Economics, University of Pretoria, Private Bag X20, Hatfield, 0028, South Africa); 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 examine the power of global economic conditions (GECON) in forecasting the daily return volatility of various international Real Estate Investment Trusts (REITs) indices. To this end, we use the GARCH-MIDAS framework due to the mixed frequencies of the variables under study and given its merit of circumventing the problems of information loss due to data aggregation and biases through data disaggregation. The results show evidence of forecast gains in the model that accommodates GECON, and significant in-sample forecastability where improvements in global economic conditions lower the risk associated with the international REITs particularly in the US and emerging markets. Further analysis shows the possibility of gaining higher returns on REITs by exploiting the information contents of GECON. A robustness analysis indicates that other measures of global economic conditions such as Global Weakness Index (GWI) and Global Intensity Index (GII) contain lower forecasting power than GECON but with significant improvements in their forecast outcomes when combined with the latter using the principal components analysis. Consequently, monitoring the global economic dynamics via GECON as well as other indices (GWI and GII) is crucial for optimal investment decisions.
    Keywords: REITs volatility, global economic conditions, mixed data analysis, GARCH-MIDAS model, forecasting
    JEL: C32 C53 E32 R30
    Date: 2022–02
  2. By: Alexander Gleim; Nazarii Salish
    Abstract: Environmental problems are receiving increasing attention in socio-economic and health studies. This in turn fosters advances in recording and data collection of many related real-life processes. Available tools for data processing are often found too restrictive as they do not account for the rich nature of such data sets. In this paper, we propose a new statistical perspective on forecasting spatial environmental data collected sequentially over time. We treat this data set as a surface (functional) time series with a possibly complicated geographical domain. By employing novel techniques from functional data analysis we develop a new forecasting methodology. Our approach consists of two steps. In the first step, time series of surfaces are reconstructed from measurements sampled over some spatial domain using a finite element spline smoother. In the second step, we adapt the dynamic functional factor model to forecast a surface time series. The advantage of this approach is that we can account for and explore simultaneously spatial as well as temporal dependencies in the data. A forecasting study of ground-level ozone concentration over the geographical domain of Germany demonstrates the practical value of this new perspective, where we compare our approach with standard functional benchmark models.
    Date: 2022–02
  3. By: Carmina Fjellstr\"om
    Abstract: Performance forecasting is an age-old problem in economics and finance. Recently, developments in machine learning and neural networks have given rise to non-linear time series models that provide modern and promising alternatives to traditional methods of analysis. In this paper, we present an ensemble of independent and parallel long short-term memory (LSTM) neural networks for the prediction of stock price movement. LSTMs have been shown to be especially suited for time series data due to their ability to incorporate past information, while neural network ensembles have been found to reduce variability in results and improve generalization. A binary classification problem based on the median of returns is used, and the ensemble's forecast depends on a threshold value, which is the minimum number of LSTMs required to agree upon the result. The model is applied to the constituents of the smaller, less efficient Stockholm OMX30 instead of other major market indices such as the DJIA and S&P500 commonly found in literature. With a straightforward trading strategy, comparisons with a randomly chosen portfolio and a portfolio containing all the stocks in the index show that the portfolio resulting from the LSTM ensemble provides better average daily returns and higher cumulative returns over time. Moreover, the LSTM portfolio also exhibits less volatility, leading to higher risk-return ratios.
    Date: 2022–01
  4. By: Davydenko, Andrey; Goodwin, Paul
    Abstract: Measuring bias is important as it helps identify flaws in quantitative forecasting methods or judgmental forecasts. It can, therefore, potentially help improve forecasts. Despite this, bias tends to be under represented in the literature: many studies focus solely on measuring accuracy. Methods for assessing bias in single series are relatively well known and well researched, but for datasets containing thousands of observations for multiple series, the methodology for measuring and reporting bias is less obvious. We compare alternative approaches against a number of criteria when rolling origin point forecasts are available for different forecasting methods and for multiple horizons over multiple series. We focus on relatively simple, yet interpretable and easy to implement metrics and visualization tools that are likely to be applicable in practice. To study the statistical properties of alternative measures we use theoretical concepts and simulation experiments based on artificial data with predetermined features. We describe the difference between mean and median bias, describe the connection between metrics for accuracy and bias, provide suitable bias measures depending on the loss function used to optimise forecasts, and suggest which measures for accuracy should be used to accompany bias indicators. We propose several new measures and provide our recommendations on how to evaluate forecast bias across multiple series. /// Note: This is the final version of the paper, which appeared in the International Journal of Statistics and Probability. The first draft of this paper was uploaded to on 11 May, 2021: /// Copyrights: This is an open access article distributed under the terms and conditions of the Creative Commons Attribution license ( 4.0/).
    Date: 2021–08–19
  5. By: Matthew F. Tomlinson; David Greenwood; Marcin Mucha-Kruczynski
    Abstract: Dynamic extreme value analysis offers a promising approach to the forecasting of the extreme tail events that often dominate systemic risk. We extend the two-tailed peaks-over-threshold (2T-POT) Hawkes model as a tool for dynamic quantile forecasting for both the left and right tails of a univariate time series; this is applied to the daily log-returns of six large cap indices using a wide range of exceedance thresholds (from the 1.25% to 25.00% mirrored quantiles). Out-of-sample convergence tests find that the 2T-POT Hawkes model offers more accurate and reliable forecasting of next step ahead extreme left- and right-tail quantiles -- as measured by (mirrored) value-at-risk and expected shortfall at the 2.5% coverage level and below -- compared against GARCH-type models. Quantitatively similar asymmetries in the parameters of the Hawkes arrival process are found across all six indices, adding further empirical support to a temporal leverage effect in which the impact of losses is not only greater but also more immediate. Our results suggest that asymmetric Hawkes-type arrival dynamics are a better approximation of the true data generating process for extreme daily log-returns than GARCH-type variance dynamics and, therefore, that the 2T-POT Hawkes model presents a better performing alternative to GARCH-type models.
    Date: 2022–02
  6. By: Effat Ara Easmin Lucky; Md. Mahadi Hasan Sany; Mumenunnesa Keya; Md. Moshiur Rahaman; Umme Habiba Happy; Sharun Akter Khushbu; Md. Arid Hasan
    Abstract: By trade we usually mean the exchange of goods between states and countries. International trade acts as a barometer of the economic prosperity index and every country is overly dependent on resources, so international trade is essential. Trade is significant to the global health crisis, saving lives and livelihoods. By collecting the dataset called "Effects of COVID19 on trade" from the state website NZ Tatauranga Aotearoa, we have developed a sustainable prediction process on the effects of COVID-19 in world trade using a deep learning model. In the research, we have given a 180-day trade forecast where the ups and downs of daily imports and exports have been accurately predicted in the Covid-19 period. In order to fulfill this prediction, we have taken data from 1st January 2015 to 30th May 2021 for all countries, all commodities, and all transport systems and have recovered what the world trade situation will be in the next 180 days during the Covid-19 period. The deep learning method has received equal attention from both investors and researchers in the field of in-depth observation. This study predicts global trade using the Long-Short Term Memory. Time series analysis can be useful to see how a given asset, security, or economy changes over time. Time series analysis plays an important role in past analysis to get different predictions of the future and it can be observed that some factors affect a particular variable from period to period. Through the time series it is possible to observe how various economic changes or trade effects change over time. By reviewing these changes, one can be aware of the steps to be taken in the future and a country can be more careful in terms of imports and exports accordingly. From our time series analysis, it can be said that the LSTM model has given a very gracious thought of the future world import and export situation in terms of trade.
    Date: 2022–01
  7. By: Ina Hajdini
    Abstract: The paper considers a New Keynesian framework in which agents form expectations based on a combination of mis-specified forecasts and myopia. The proposed expectations formation process is found to be consistent with all three empirical facts on consensus inflation forecasts, namely, that forecasters under-react to ex-ante forecast revisions, that forecasters over-react to recent events, and that the response of forecast errors to a shock initially under-shoots but then over-shoots. The paper then derives the general equilibrium solution consistent with the proposed expectations formation process and estimates the model with likelihood-based Bayesian methods, yielding three novel results: (i) The data strongly prefer the combination of autoregressive mis-specified forecasting rules and myopia over other alternatives; (ii) The best fitting expectations formation process for both households and firms is characterized by high degrees of myopia and simple AR(1) forecasting rules; (iii) Frictions such as habit in consumption, which are typically necessary for models with Full-information Rational Expectations, are significantly less important, because the proposed expectations generate substantial internal persistence and amplification to exogenous shocks. Simulated inflation expectations data from the estimated general equilibrium model reflect the three empirical facts on forecasting data.
    Keywords: Myopia; Survey of Professional Forecasters; Bayesian Estimation; Internal Propagation
    JEL: C11 C53 D84 E13 E30 E50 E70 E52
    Date: 2022–02–16
  8. By: Taylan Kabbani (Ozyegin University; Huawei Turkey R&D Center); Fatih Enes Usta (Marmara University)
    Abstract: Predicting the stock market trend has always been challenging since its movement is affected by many factors. Here, we approach the future trend prediction problem as a machine learning classification problem by creating tomorrow_trend feature as our label to be predicted. Different features are given to help the machine learning model predict the label of a given day; whether it is an uptrend or downtrend, those features are technical indicators generated from the stock's price history. In addition, as financial news plays a vital role in changing the investor's behavior, the overall sentiment score on a given day is created from all news released on that day and added to the model as another feature. Three different machine learning models are tested in Spark (big-data computing platform), Logistic Regression, Random Forest, and Gradient Boosting Machine. Random Forest was the best performing model with a 63.58% test accuracy.
    Date: 2022–01
    Abstract: This work describes strengths and weaknesses of the logistic function used in forecasting from a theoretical and a practical point of view. Theoretical topics treated are: generalizing the concept of competition, dividing the growth cycle in four "seasons", and using logistics simply qualitatively to obtain rare insights and intuitive understanding. Practical topics addresses are: determination of the uncertainties, how to decide whether to fit cumulative or per unit of time data, and how to deal with a bias toward a low ceiling. This article is my contribution to a massive review article with title "Forecasting: theory and practice" published in the International Journal of Forecasting.
    Date: 2022–01–19
  10. By: Hwai-Chung Ho
    Abstract: The study of long-horizon returns has received a great deal of attention in recent years (see, for example, Boudoukh, Richardson, and Whitelaw (2008), Neuberger (2012) and Lee (2013), Fama and French (2018)). While most of the discussions are concerned with some practical issues in investment, few have touched the important aspect on risk management. The approach adopted in this article is to predict the future distribution of the returns of a fixed long-horizon by which the risk measures of interest that come in the form of a distributional functional such as the value at risk (VaR) and the conditional tail expectation (CTE) can be easily derived. The characteristic feature of our approach which requires no specification of the volatility dynamics nor parametric assumptions of the shock distribution extends the work by Ho et al. (2016) and Ho ( 2017) to a more general volatility dynamics that includes both the widely-used SV model and the GARCH model (Bollerslev, 1986) as special cases.
    Date: 2022–01
  11. By: Thackway, William; Ng, Matthew Kok Ming (University of New South Wales); Lee, Chyi Lin; Pettit, Christopher
    Abstract: In an era of rapid urbanisation and increasing wealth, gentrification is an urban phenomenon impacting many cities around the world. The ability of policymakers and planners to better understand and address gentrification-induced displacement hinges upon proactive intervention strategies. It is in this context that we build a tree-based machine learning (ML) model to predict neighbourhood change in Sydney. Change, in this context, is proxied by the Socioeconomic Index for Advantage and Disadvantage, in addition to census and other ancillary predictors. Our models predict gentrification from 2011-2016 with a balanced accuracy of 74.7%. Additionally, the use of an additive explanation tool enables individual prediction explanations and advanced feature contribution analysis. Using the ML model, we predict future gentrification in Sydney up to 2021. The predictions confirm that gentrification is expanding outwards from the city centre. A spill-over effect is predicted to the south, west and north-west of former gentrifying hotspots. The findings are expected to provide policymakers with a tool to better forecast where likely areas of gentrification will occur. This future insight can then inform suitable policy interventions and responses in planning for more equitable cities outcomes, specifically for vulnerable communities impacted by gentrification and neighbourhood change.
    Date: 2021–12–16

This nep-for issue is ©2022 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.