nep-for New Economics Papers
on Forecasting
Issue of 2018‒11‒12
nine papers chosen by
Rob J Hyndman
Monash University

  1. Forecasting Changes of Economic Inequality: A Boosting Approach By Christian Pierdzioch; Rangan Gupta; Hossein Hassani; Emmanuel Silva
  2. Predicting relative forecasting performance : An empirical investigation By Granziera, Eleonora; Sekhposyan, Tatevik
  3. Nowcasting Japanese GDPs By Kyosuke Chikamatsu, Naohisa Hirakata, Yosuke Kido, Kazuki Otaka
  4. A Co-Evolutionary, Long-Term, MacroEconomic Forecast for the UK Using Demographic Projections By Nick Jagger
  5. Nowcasting the Unemployment Rate in the EU with Seasonal BVAR and Google Search Data By Anttonen, Jetro
  6. Evaluating the Bank of Canada Staff Economic Projections Using a New Database of Real-Time Data and Forecasts By Julien Champagne; Guillaume Poulin-Bellisle; Rodrigo Sekkel
  7. Forecasting Realized Volatility Measures with Multivariate and Univariate Models: The Case of The US Banking Sector By Gianluca Cubadda; Alain Hecq; Antonio Riccardo
  8. “A new metric of consensus for Likert scales” By Oscar Claveria
  9. Bond Risk Premia and the ”Return Forecasting Factor” By Agustin Gutierrez; Constantino Hevia; Martin Sola

  1. By: Christian Pierdzioch (Department of Economics, Helmut Schmidt University, Hamburg, Germany); Rangan Gupta (Department of Economics, University of Pretoria, Pretoria, South Africa); Hossein Hassani (Research Institute for Energy Management and Planning, University of Tehran, Tehran, Iran); Emmanuel Silva (Fashion Business School, London College of Fashion, University of the Arts London, 272 High Holborn, London, WC1V 7EY)
    Abstract: We use a boosting algorithm to forecast changes in three income- and three consumption-based inequality measures. We study quarterly UK data covering the period from 1975Q1 to 2016Q1. We find that the boosted forecasting models, at forecasting horizons of up to one year, have predictive value for changes in the six different inequality measures. Evidence of predictability is strong when we use information criteria that result in relatively parsimonious forecasting models. In addition to lagged inequality measures, stock-market developments and fiscal deficits and, for the consumption-based inequality measures at a forecast horizon of four quarters, economic policy uncertainty and output growth turn out to be relatively important predictors.
    Keywords: Inequality, Predictability, Boosting, UK data
    JEL: C53 D63
    Date: 2018–10
  2. By: Granziera, Eleonora; Sekhposyan, Tatevik
    Abstract: The relative performance of forecasting models changes over time. This empirical observation raises two questions: is the relative performance itself predictable? If so, can it be exploited to improve forecast accuracy? We address these questions by evaluating the predictive ability of a wide range of economic variables for two key US macroeconomic aggregates, industrial production and inflation, relative to simple benchmarks. We find that business indicators, financial conditions, uncertainty as well as measures of past relative performance are generally useful for explaining the relative forecasting performance of the models. We further conduct a pseudo-real-time forecasting exercise, where we use the information about the conditional performance for model selection and model averaging. The newly proposed strategies deliver sizable improvements over competitive benchmark models and commonly used combination schemes. Gains are larger when model selection and averaging are based on financial conditions as well as past performance measured at the forecast origin date.
    JEL: C22 C52 C53
    Date: 2018–11–08
  3. By: Kyosuke Chikamatsu, Naohisa Hirakata, Yosuke Kido, Kazuki Otaka (Bank of Japan)
    Abstract: In this paper, we discuss the approaches to nowcasting Japanese GDPs, namely preliminary quarterly GDP estimates and revised annual GDP estimates. First, we look at nowcasting preliminary estimates of quarterly GDP using monthly indicators, ranging from hard data to soft data. In doing so, we compare a variety of mixed frequency approaches, a bridge equation approach, Mixed-Data Sampling (MIDAS) and factor-augmented version of these approaches, and also discuss the usefulness of forecast combination. Second, we work on nowcasting revised annual GDP, which is compiled with comprehensive statistics but only available after a considerable delay. In nowcasting the revised annual GDP, we employ several benchmarking methods, including Chow and Lin (1971), and examine the usefulness of monthly supply-side indicators to predict revised annual GDP. Our findings are summarized as follows. First, regarding nowcasting preliminary quarterly GDP, some of the mixed frequency models discussed in this paper record out-of-sample performance superior to an in-sample mean benchmark. Furthermore, there is a gain from combining model forecasts and professional forecasts. Second, regarding nowcasting revised annual GDP, some benchmarking models that exploit supply-side data serve as useful tools for predicting revised annual growth rates.
    Keywords: Nowcasting; Forecast Combination; Mixed-Data Sampling (MIDAS); Benchmarking
    JEL: C53
    Date: 2018–11–07
  4. By: Nick Jagger (SPRU, University of Sussex; University of Brighton)
    Abstract: This paper is based around outlining and illustrating the use of a co-evolutionary method for long-term macro-economic forecasting. The paper includes economic forecasts for the UK to 2060 using a novel approach based on Multichannel Singular Spectral Analysis (MSSA). The forecasts are based on projections of the working-age population and their educational attainment, as well as building on the historic trends of these variables. The variables forecasted are Gross Domestic Product (GDP), investment and productivity, based on historic time-series dating back to 1856, and their interactions with the projected variables. Other longterm forecasts for the UK are examined and the important impact of demographic change and plateauing educational attainment is assessed. Additionally, the power of the new MSSA forecasting technique proposed here is illustrated.
    Keywords: Co-evolutionary forecasting; Multichannel Singular Spectral Analysis; Demographics; Educational Attainment; Long-term macro-economic forecasting
    JEL: B15 B22 C14 C53 J11
    Date: 2018–10
  5. By: Anttonen, Jetro
    Abstract: Abstract In this paper a Bayesian vector autoregressive model for nowcasting the seasonally non-adjusted unemployment rate in EU-countries is developed. On top of the official statistical releases, the model utilizes Google search data and the effect of Google data on the forecasting performance of the model is assessed. The Google data is found to yield modest improvements in forecasting accuracy of the model. To the author’s knowledge, this is the first time the forecasting performance of the Google search data has been studied in the context of Bayesian vector autoregressive model. This paper also adds to the empirical literature on the hyperparameter choice with Bayesian vector autoregressive models. The hyperparameters are set according to the mode of the posterior distribution of the hyperparameters, and this is found to improve the out-of-sample forecasting accuracy of the model significantly, compared to the rule-of-thumb values often used in the literature.
    Keywords: Nowcasting, Forecasting, BVAR, Big Data, Unemployment
    JEL: C32 C53 C82 E27
    Date: 2018–11–05
  6. By: Julien Champagne; Guillaume Poulin-Bellisle; Rodrigo Sekkel
    Abstract: We present a novel database of real-time data and forecasts from the Bank of Canada’s staff economic projections. We then provide a forecast evaluation for GDP growth and CPI inflation since 1982: we compare the staff forecasts with those from commonly used time-series models estimated with real-time data and with forecasts from other professional forecasters and provide standard bias tests. Finally, we study changes in the predictability of the Canadian economy following the announcement of the inflation-targeting regime in 1991. Our database is unprecedented outside the United States, and our evidence is particularly interesting, as it includes over 30 years of staff forecasts, two severe recessions and different monetary policy regimes. The database will be made available publicly and updated annually.
    Keywords: Econometric and statistical methods, Economic models, Inflation targets, Monetary Policy
    JEL: C32 E17 E37
    Date: 2018
  7. By: Gianluca Cubadda (DEF & CEIS,University of Rome "Tor Vergata"); Alain Hecq (Maastricht University); Antonio Riccardo (ICE Data Services Italy)
    Abstract: This paper compares the forecasting performances of both univariate and multivariate models for realized volatilities series. We consider realized volatility measures of the returns of 13 major banks traded in the NYSE. Since our variables are characterized by the presence of long range dependence, we use several modelling approaches that are able to capture such feature. We look at the forecasting accuracy of the considered models to make inference on the underlying mechanism that has generated volatilities of the assets. Our main conclusion is that the contagion effect among the considered volatilities is small or, at least, not well captured by the considered multivariate models.
    Keywords: Consumption,asymmetry,expectations,noisy information
    JEL: C32
    Date: 2018–10–30
  8. By: Oscar Claveria (AQR-IREA, University of Barcelona)
    Abstract: In this study we present a metric of consensus for Likert-type scales. The measure gives the level of agreement as the percentage of consensus among respondents. The proposed framework allows to design a positional indicator that gives the degree of agreement for each item and for any given number of reply options. In order to assess the performance of the proposed metric of consensus, in an iterated one-period ahead forecasting experiment we test whether the inclusion of the degree of agreement in consumers’ expectations regarding the evolution of unemployment improves out-of-sample forecast accuracy in eight European countries. We find evidence that the degree of agreement among consumers contains useful information to predict unemployment rates in most countries. The obtained results show the usefulness of consensus-based metrics to track the evolution of economic variables.
    Keywords: Likert scales; consensus; geometry; economic tendency surveys; consumer expectations; unemployment JEL classification: C14; C51; C52; C53; D12; E24
    Date: 2018–10
  9. By: Agustin Gutierrez; Constantino Hevia; Martin Sola
    Abstract: The return forecasting factor is a linear combination of forward rates that seems to predict one-year excess bond returns of bond of all maturities better than traditional measures obtained from the yield curve. If this single factor actually captures all the relevant fluctuations in bond risk premia, then it should also summarize all the economically relevant variations in excess returns considering different holding periods. We find that it does not. We conclude that including the return forecasting factor as the main driver of risk premia in a term structure model, as has been suggested, is not supported by the data.
    Keywords: : Excess returns, bond risk premia, return forecasting factor, affine term structure models.
    Date: 2018–10

This nep-for issue is ©2018 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.