nep-for New Economics Papers
on Forecasting
Issue of 2020‒06‒08
ten papers chosen by
Rob J Hyndman
Monash University

  1. Forecasting Inflation with the New Keynesian Phillips Curve: Frequency Matters By Manuel M. F. Martins; Fabio Verona
  2. Forecasting Inflation in a Data-Rich Environment: The Benefits of Machine Learning Methods By Marcelo Madeiros; Gabriel Vasconcelos; Álvaro Veiga; Eduardo Zilberman
  3. Scenario-driven forecasting: Modeling peaks and paths. Insights from the COVID-19 Pandemic in Belgium By Kristof Decock; Koenraad Debackere; Anne Mieke Vandamme; Bart Van Looy
  4. Forecasting Truck Parking Using Fourier Transformations By Sadek, Bassel A; Martin, Elliot W; Shaheen, Susan A
  5. Covid-19 outbreak and beyond: The information content of registered short-time workers for GDP now- and forecasting. By Sylvia Kaufmann
  6. Probabilistic Multi-Step-Ahead Short-Term Water Demand Forecasting with Lasso By Jens Kley-Holsteg; Florian Ziel
  7. A Dynamic Conditional Approach to Portfolio Weights Forecasting By Fabrizio Cipollini; Giampiero Gallo; Alessandro Palandri
  8. Predicting the Long-term Stock Market Volatility: A GARCH-MIDAS Model with Variable Selection By Tong Fang; Tae-Hwy Lee; Zhi Su
  9. Nonparametric Expected Shortfall Forecasting Incorporating Weighted Quantiles By Giuseppe Storti; Chao Wang
  10. Extracting Information of the Economic Activity from Business and Consumer Surveys in an Emerging Economy (Chile) By Camila Figueroa; Michael Pedersen

  1. By: Manuel M. F. Martins (Faculty of Economics, University of Porto and CEF.UP); Fabio Verona (Bank of Finland - Monetary Policy and Research Department and University of Porto - CEF.UP)
    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.
    Keywords: Inflation forecasting; New Keynesian Phillips curve; Frequency domain; Wavelets
    JEL: C53 E31 E37
    Date: 2020–04
  2. By: Marcelo Madeiros; Gabriel Vasconcelos; Álvaro Veiga; Eduardo Zilberman
    Abstract: Inflation forecasting is an important but difficult task. Here, we explore advances in machine learning (ML) methods and the availability of new datasets to forecast US inflation. Despite the skepticism in the previous literature, we show that ML models with a large number of covariates are systematically more accurate than the benchmarks. The ML method that deserves more attention is the random forest model, which dominates all other models. Its good performance is due not only to its specific method of variable selection but also the potential nonlinearities between past key macroeconomic variables and inflation.
    Date: 2019–05
  3. By: Kristof Decock; Koenraad Debackere; Anne Mieke Vandamme; Bart Van Looy
    Abstract: The recent ‘outburst’ of COVID-19 spurred efforts to model and forecast its diffusion patterns, either in terms of infections, people in need of medical assistance (ICU occupation) or casualties. Forecasting patterns and their implied end states remains cumbersome when few (stochastic) data points are available during the early stage of diffusion processes. Extrapolations based on compounded growth rates do not account for inflection points nor end-states. In order to remedy this situation, we advance a set of heuristics which combine forecasting and scenario thinking. Inspired by scenario thinking we allow for a broad range of end states (and their implied growth dynamics, parameters) which are consecutively being assessed in terms of how well they coincide with actual observations. When applying this approach to the diffusion of COVID-19, it becomes clear that combining potential end states with unfolding trajectories provides a better-informed decision space as short term predictions are accurate, while a portfolio of different end states informs the long view. The creation of such a decision space requires temporal distance. Only to the extent that one refrains from incorporating more recent data, more plausible end states become visible. Such dynamic approach also allows one to assess the potential effects of mitigating measures. As such, our contribution implies a plea for dynamically blending forecasting algorithms and scenario-oriented thinking, rather than conceiving them as substitutes or complements.
    Date: 2020–05–28
  4. By: Sadek, Bassel A; Martin, Elliot W; Shaheen, Susan A
    Abstract: Truck-based transportation is the predominant mode used to transport goods and raw materials within the United States. While trucks play a major role in local commerce, a significant portion of truck activity is also long haul in nature. Long-haul truck drivers are continuously faced with the problem of not being able to secure a safe parking spot since many rest areas become fully occupied, and information about parking and availability is limited. Truck drivers faced with full parking lots/facilities either continue driving until a safe parking spot is located or park illegally. Both scenarios pose a hazard to the truck driver, as well as the surrounding road users. Disseminating forecasts of parking availability to truck drivers may help mitigate this hazard, since many truck drivers plan their parking in advance of arrival. Building on 1 year of nearly continuous truck parking data collection, this paper proposes and demonstrates a method for developing a dynamic forecasting model that can predict truck parking occupancy for any specified time within the present day, using only truck parking occupancy data from a trucking logistics facility in the northern San Joaquin Valley during 2016. Different versions of the dynamic model were studied and verified against successive weekdays with performance measured using the root-mean-square error (RMSE). Results indicated that for a particular day, the maximum error can range between 13 and 40 trucks, about 5% of the absolute maximum capacity of the facility.
    Keywords: Engineering
    Date: 2020–08–01
  5. By: Sylvia Kaufmann (Study Center Gerzensee)
    Abstract: The number of employees historically filed and registered from January to April 2020 for short-time compensation is used to obtain a nowcast for GDP growth in the first quarter and an outlook until the third quarter 2021. We purge the monthly log level series from the systematic component to extract unexpected changes or shocks to log short-time workers. These monthly shocks are included in a univariate model for quarterly GDP growth to capture timely, current-quarter unexpected changes in growth dynamics. Included shocks explain additionally 24% in GDP growth variation. The model is able to forecast quite precisely the decrease in GDP during the financial crisis. It predicts a mean decline in GDP of 5.7% over the next two quarters. Without additional growth stimulus, the GDP level forecast remains persistently 4% lower in the long run. The uncertainty is large, as the 95% highest forecast density interval covers a decrease in GDP as large as 9%. A recovery to pre-crisis GDP level in 2021 lies only in the upper tail of the 95% highest forecast density interval.
    Date: 2020–05
  6. By: Jens Kley-Holsteg; Florian Ziel
    Abstract: Water demand is a highly important variable for operational control and decision making. Hence, the development of accurate forecasts is a valuable field of research to further improve the efficiency of water utilities. Focusing on probabilistic multi-step-ahead forecasting, a time series model is introduced, to capture typical autoregressive, calendar and seasonal effects, to account for time-varying variance, and to quantify the uncertainty and path-dependency of the water demand process. To deal with the high complexity of the water demand process a high-dimensional feature space is applied, which is efficiently tuned by an automatic shrinkage and selection operator (lasso). It allows to obtain an accurate, simple interpretable and fast computable forecasting model, which is well suited for real-time applications. The complete probabilistic forecasting framework allows not only for simulating the mean and the marginal properties, but also the correlation structure between hours within the forecasting horizon. For practitioners, complete probabilistic multi-step-ahead forecasts are of considerable relevance as they provide additional information about the expected aggregated or cumulative water demand, so that a statement can be made about the probability with which a water storage capacity can guarantee the supply over a certain period of time. This information allows to better control storage capacities and to better ensure the smooth operation of pumps. To appropriately evaluate the forecasting performance of the considered models, the energy score (ES) as a strictly proper multidimensional evaluation criterion, is introduced. The methodology is applied to the hourly water demand data of a German water supplier.
    Date: 2020–05
  7. By: Fabrizio Cipollini (Dipartimento di Statistica, Informatica, Applicazioni "G. Parenti", Università di Firenze); Giampiero Gallo (New York University in Florence); Alessandro Palandri (Dipartimento di Statistica, Informatica, Applicazioni "G. Parenti", Università di Firenze)
    Abstract: We build the time series of optimal realized portfolio weights from high-frequency data and we suggest a novel Dynamic Conditional Weights (DCW) model for their dynamics. DCW is benchmarked against popular model-based and model-free specifications in terms of weights forecasts and portfolio allocations. Next to portfolio variance, certainty equivalent and turnover, we introduce the break-even transaction costs as an additional measure that identifies the range of transaction costs for which one allocation is preferred to another. By comparing minimum-variance portfolios built on the components of the Dow Jones 30 Index, the proposed DCW overall attains the best allocations with respect to the measures considered, for any degree of risk-aversion, transaction costs and exposure.
    Keywords: Portfolio Allocation, Realized Volatility, Realized Correlations, Dynamic Conditional Modeling, Portfolio Weights Modeling
    JEL: C32 C53 G11 G17
    Date: 2020–05
  8. By: Tong Fang (Shandong University); Tae-Hwy Lee (Department of Economics, University of California Riverside); Zhi Su (Central University of Finance and Economics)
    Abstract: We consider a GARCH-MIDAS model with short-term and long-term volatility components, in which the long-term volatility component depends on many macroeconomic and financial variables. We select the variables that exhibit the strongest effects on the long-term stock market volatility via maximizing the penalized log-likelihood function with an Adaptive-Lasso penalty. The GARCH-MIDAS model with variable selection enables us to incorporate many variables in a single model without estimating a large number of parameters. In the empirical analysis, three variables (namely, housing starts, default spread and realized volatility) are selected from a large set of macroeconomic and financial variables. The recursive out-of-sample forecasting evaluation shows that variable selection significantly improves the predictive ability of the GARCH-MIDAS model for the long-term stock market volatility.
    Keywords: Stock market volatility, GARCH-MIDAS model, Variable selection, Penalized maximum likelihood, Adaptive-Lasso
    JEL: C32 C51 C53 G12
    Date: 2020–05
  9. By: Giuseppe Storti; Chao Wang
    Abstract: A new semi-parametric Expected Shortfall (ES) estimation and forecasting framework is proposed. The proposed approach is based on a two step estimation procedure. The first step involves the estimation of Value-at-Risk (VaR) at different levels through a set of quantile time series regressions. Then, the ES is computed as a weighted average of the estimated quantiles. The quantiles weighting structure is parsimoniously parameterized by means of a Beta function whose coefficients are optimized by minimizing a joint VaR and ES loss function of the Fissler-Ziegel class. The properties of the proposed approach are first evaluated with an extensive simulation study using various data generating processes. Two forecasting studies with different out-of-sample sizes are conducted, one of which focuses on the 2008 Global Financial Crisis (GFC) period. The proposed models are applied to 7 stock market indices and their forecasting performances are compared to those of a range of parametric, non-parametric and semi-parametric models, including GARCH, Conditional AutoRegressive Expectile (CARE, Taylor 2008), joint VaR and ES quantile regression models (Taylor, 2019) and simple average of quantiles. The results of the forecasting experiments provide clear evidence in support of the proposed models.
    Date: 2020–05
  10. By: Camila Figueroa; Michael Pedersen
    Abstract: The present paper discusses the extent to which business and consumer survey observations are useful for predicting the Chilean activity. The two surveys examined are called IMCE and IPEC, after their Spanish abbreviations, for the business and consumer survey, respectively. The baseline exercises consist in simple calculations of cross correlations between the surveys and activity variables, test for Granger causality and augmentation of autoregressive activity models with survey data to evaluate if the now- and forecast performances are improved. The evidence suggests that both surveys, in general, contain useful information for making predictions of the Chilean activity, particularly for the longer horizons. An additional exercise indicates that the data in the two surveys are complementary in the sense that the longer horizon forecasts improve further when both of them are included in the econometric model.
    Date: 2019–05

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.