nep-for New Economics Papers
on Forecasting
Issue of 2016‒03‒29
fourteen papers chosen by
Rob J Hyndman
Monash University

  1. Forecasting Unemployment with Google Searches By Tuhkuri, Joonas
  2. Real-time forecasting with a MIDAS VAR By Mikosch, Heiner; Neuwirth, Stefan
  3. Nowcasting Real GDP growth in South Africa By Alain Kabundi, Elmarie Nel and Franz Ruch
  4. Forecasting the Term Structure of Interest Rates with Potentially Misspecified Models By Eo, Yunjong; Kang, Kyu Ho
  5. VAR Models with Non-Gaussian Shocks By Ching-Wai (Jeremy) Chiu; Haroon Mumtaz; Gabor Pinter
  6. The impact of forecasting errors on warehouse labor efficiency By Kim, T.Y.; Dekker, R.; Heij, C.
  7. An adaptive approach to forecasting three key macroeconomic variables for transitional China By Niu, Linlin; Xu, Xiu; Chen, Ying
  8. Credit Funding and Banking Fragility: An Empirical Analysis for Emerging Economies By Alexander Guarín-López; Ignacio Lozano-Espitia
  9. State Correlation and Forecasting: A Bayesian Approach Using Unobserved Components Models By Luis Uzeda
  10. Order Invariant Evaluation of Multivariate Density Forecasts By Dovern, Jonas; Manner, Hans
  12. Spare part demand forecasting for consumer goods using installed base information By Kim, T.Y.; Dekker, R.; Heij, C.
  13. Can currency in circulation predict South African economic activity? By Cobus Vermeulen, Adél Bosch, Fanie Joubert and Jannie Rossouw
  14. Two historical changes in the narrative of energy forecasts By Minh Ha-Duong; Franck Nadaud; Martin Jegard

  1. By: Tuhkuri, Joonas
    Abstract: Data on Google searches help predict the unemployment rate in the U.S. But the predictive power of Google searches is limited to short-term predictions, the value of Google data for forecasting purposes is episodic, and the improvements in forecasting accuracy are only modest. The results, obtained by (pseudo) out-of-sample forecast comparison, are robust to a state-level fixed effects model and to different search terms. Joint analysis by cross-correlation function and Granger non-causality tests verifies that Google searches anticipate the unemployment rate. The results illustrate both the potentials and limitations of using big data to predict economic indicators.
    Keywords: Big Data, Google, Internet, Nowcasting, Forecasting, Unemployment
    JEL: C22 C53 C82 E27
    Date: 2016–03–02
  2. By: Mikosch, Heiner; Neuwirth, Stefan
    Abstract: This paper presents a MIDAS type mixed frequency VAR forecasting model. First, we propose a general and compact mixed frequency VAR framework using a stacked vector approach. Second, we integrate the mixed frequency VAR with a MIDAS type Almon lag polynomial scheme which is designed to reduce the parameter space while keeping models fexible. We show how to recast the resulting non-linear MIDAS type mixed frequency VAR into a linear equation system that can be easily estimated. A pseudo out-of-sample forecasting exercise with US real-time data yields that the mixed frequency VAR substantially improves predictive accuracy upon a standard VAR for dierent VAR specications. Forecast errors for, e.g., GDP growth decrease by 30 to 60 percent for forecast horizons up to six months and by around 20 percent for a forecast horizon of one year.
    Keywords: Forecasting, mixed frequency data, MIDAS, VAR, real time
    JEL: C53 E27
    Date: 2015–04–13
  3. By: Alain Kabundi, Elmarie Nel and Franz Ruch
    Abstract: This paper uses nowcasting to forecast real GDP growth in South Africa from 2010Q1 to 2014Q3 in real time. Such an approach exploits the ‡ow of high-frequency information underlying the state of the economy. It overcomes one of the major challenges faced by forecasters, policymakers, and economic agents - having a clear view of the state of the economy in real time. This is often not the case as many economic variables are only available at low frequency and with considerable lags, making it di¢ cult to have information on the state of the economyeven after the end of the quarter. The pseudo out-of-sample forecasts show that he nowcasting model’s performance is comparable to those of professional forecasters even though the latter enhance their forecasting accuracy with judgement. The nowcast model also outperforms all other benchmark models by a signi…cant margin.
    Keywords: Nowcasting, Factor Model, Bayesian VAR, forecasting
    JEL: E52 C53 C33
    Date: 2016
  4. By: Eo, Yunjong; Kang, Kyu Ho
    Abstract: This paper assesses the predictive gains of the pooling method in yield curve prediction. We consider three individual yield curve prediction models: the dynamic Nelson-Siegel model (DNS) and the arbitrage-free Nelson-Siegel model in addition to the random walk (RW) model as a benchmark. Despite the popularity of these three frameworks, none of them dominates the others across all maturities and forecast horizons. This fact indicates that those models are potentially misspecified. We investigate whether combining the possibly misspecified models in a linear form helps improve the predictive accuracy. To do this, we evaluate the out-of-sample forecasts of the pooled models in comparison with the individual models. In terms of density prediction, the pooled model of the DNS and RW models consistently outperforms those individual models regardless of maturities and forecast horizons. Our findings strongly suggest that one needs to try the pooling method rather than choosing one of the alternative models.
    Keywords: Model combination; Bayesian MCMC method; Markov switching process; Dynamic Nelson-Siegel; Affine term structure model
    Date: 2016–03
  5. By: Ching-Wai (Jeremy) Chiu (Bank of England); Haroon Mumtaz (Queen Mary University of London); Gabor Pinter (Bank of England)
    Abstract: We introduce a Bayesian VAR model with non-Gaussian disturbances that are modelled with a finite mixture of normal distributions. Importantly, we allow for regime switching among the different components of the mixture of normals. Our model is highly flexible and can capture distributions that are fat-tailed, skewed and even multimodal. We show that our model can generate large out-of-sample forecast gains relative to standard forecasting models, especially during tranquil periods. Our model forecasts are also competitive with those generated by the conventional VAR model with stochastic volatility.
    Keywords: Bayesian VAR, Non-Gaussian shocks, Density Forecasting
    JEL: C11 C32 C52
    Date: 2902
  6. By: Kim, T.Y.; Dekker, R.; Heij, C.
    Abstract: Efficiency of outbound warehouse operations depends on the management of demand forecasts and associated labor planning. A case study in consumer electronics shows that warehouse management systematically over-forecasts actual orders, by 3% on average and by 6-12% in busy periods (at the end of each month and also in the months September, October, and November). A time series model that corrects order forecasts for the biases in preceding weeks reduces the bias to less than 2%, both on average and also in busy periods. The arrangements with the labor provider imply potential benefits of intentional over- forecasting and the associated ample labor supply for the warehouse. As compared to under-forecasted days, labor productivity on over-forecasted days is higher by 12% for loading activities and by 4% for picking and total outbound activities. Similar productivity gains are found if unbiased forecasts are compared with the optimal bias obtained from non- linear models estimated from daily data on bias and labor efficiency. The positive effects of intentional over-forecasting on productivity are confirmed in a structural equations model. By following similar methodologies as described in this paper, warehouse managers can determine the amount of intentional forecast bias that works best for their situation. The information required for this kind of evidence-based labor management consists of historical data on order sizes, forecasts, and labor productivity, and the outcomes depend on the available hiring strategies and cost structures.
    Keywords: Decision support, warehouse planning, forecasting, labor efficiency, case study, time series
    Date: 2016–02–24
  7. By: Niu, Linlin; Xu, Xiu; Chen, Ying
    Abstract: We propose the use of a local autoregressive (LAR) model for adaptive estimation and forecasting of three of China’s key macroeconomic variables: GDP growth, inflation and the 7-day interbank lending rate. The approach takes into account possible structural changes in the data-generating process to select a local homogeneous interval for model estimation, and is particularly well-suited to a transition economy experiencing ongoing shifts in policy and structural adjustment. Our results indicate that the proposed method outperforms alternative models and forecast methods, especially for forecast horizons of 3 to 12 months. Our 1-quarter ahead adaptive forecasts even match the performance of the well-known CMRC Langrun survey forecast. The selected homogeneous intervals indicate gradual changes in growth of industrial production driven by constant evolution of the real economy in China, as well as abrupt changes in interestrate and inflation dynamics that capture monetary policy shifts.
    Keywords: Chinese economy, local parametric models, forecasting
    JEL: E43 E47
    Date: 2015–04–10
  8. By: Alexander Guarín-López (Banco de la República de Colombia); Ignacio Lozano-Espitia (Banco de la República de Colombia)
    Abstract: This paper proposes an empirical model to identify and forecast banking fragility episodes using information on the credit funding sources. We predict the probability of occurrence of such episodes 0, 3 and 6 months ahead employing a Bayesian Model Averaging of logistic regressions. The exercises use monthly balance sheet data since the middle of the nineties for the banking system of nine merging economies: Brazil, Colombia, Croatia, Czech Republic, Mexico, Peru, Poland, Taiwan and Turkey. Our findings suggest that the increasing use of wholesale funds to support credit expansion provides warning signals of banking frailness. The in-sample and out-of-sample predictions indicate that the proposed technique is a suitable tool for forecasting short-term financial fragility events. Therefore, monitoring these funds through our tool could become useful in prudential practice. Classification JEL:C11, C52, C53, G01, G21
    Keywords: credit cycle, financial stability, wholesale funds, balance sheet, logistic model regression, Bayesian model averaging.
    Date: 2016–03
  9. By: Luis Uzeda
    Abstract: Implications to signal extraction that arise from specifying unobserved components (UC) models with correlated or orthogonal innovations have been well-investigated. In contrast, an analogous statement for forecasting evaluation cannot be made. This paper attempts to fill this gap in light of the recent resurgence of studies adopting UC models for forecasting purposes. In particular, four correlation structures are entertained: orthogonal, correlated, perfectly correlated innovations as well as a novel approach which combines features from two contrasting cases, namely, orthogonal and perfectly correlated innovations. Parameter space restrictions associated with different correlation structures and their connection with forecasting are discussed within a Bayesian framework. Introducing perfectly correlated innovations, however, reduces the covariance matrix rank. To accommodate that, a Markov Chain Monte Carlo sampler which builds upon properties of Toeplitz matrices and recent advances in precision-based algorithms is developed. Our results for several measures of U.S. inflation indicate that the correlation structure between state variables has important implications for forecasting performance as well as estimates of trend inflation.
    Keywords: Bayesian, Markov Chain Monte Carlo, State Space, Unobserved ComponentsModels, ARIMA, Reduced Rank, Precision, Forecasting
    JEL: C11 C15 C51 C53
    Date: 2016–03
  10. By: Dovern, Jonas; Manner, Hans
    Abstract: We derive new tests for proper calibration of multivariate density forecasts based on Rosenblatt probability integral transforms. These tests have the advantage that they i) do not depend on the ordering of variables in the forecasting model, ii) are applicable to densities of arbitrary dimensions, and iii) have superior power relative to existing approaches. We furthermore develop adjusted tests that allow for estimated parameters and, consequently, can be used as in-sample specification tests. We demonstrate the problems of existing tests and how our new approaches can overcome those using Monte Carlo Simulation as well as two applications based on multivariate GARCH-based models for stock market returns and on a macroeconomic Bayesian vectorautoregressive model.
    Keywords: density calibration; goodness-of-fit test; predictive density; Rosenblatt transformation
    Date: 2016–03–08
  11. By: Murat Midilic (-)
    Abstract: This study applies the Iteratively Weighted Least Squares (IWLS) algorithm to a Smooth Transition Autoregressive (STAR) model with conditional variance. Monte Carlo simulations are performed to measure the performance of the algorithm, to compare its performance with the performances of established methods in the literature, and to see the effect of initial value selection method. Simulation results show that low bias and mean squared error are received for the slope parameter estimator from the IWLS algorithm when the real value of the slope parameter is low. In an empirical illustration, STAR-GARCH model is used to forecast daily US Dollar/Australian Dollar and FTSE Small Cap index returns. 1-day ahead out-of-sample forecast results show that forecast performance of the STAR-GARCH model improves with the IWLS algorithm and the model performs better that the benchmark model.
    Keywords: STAR, GARCH, iteratively weighted least squares, Australian Dollar,FTSE
    JEL: C15 C51 C53 C58 C87 F31
    Date: 2016–01
  12. By: Kim, T.Y.; Dekker, R.; Heij, C.
    Abstract: When stopping production, the manufacturer has to decide on the lot size in the final production run to cover spare part demand during the end-of-life phase. This decision can be supported by forecasting how much demand is expected in the future. Forecasts can be obtained from the installed base of the product, that is, the number of products still in use. Consumer decisions on whether or not to repair a malfunctioning product depend on the specific product and spare part. Further, consumers may differ in their decisions, for example, for products with fast innovations and changing social trends. Consumer behavior can be accounted for by using appropriate types of installed base, for example, full installed base for cheap but essential spare parts of expensive products, and warranty installed base for expensive spare parts of products with short lifecycle. The paper presents a general methodology for installed base forecasting of end-of-life spare part demand and formulates research hypotheses on which of four installed base types performs best under which conditions. The methodology is illustrated by case studies for eighteen spare parts of six products from a consumer electronics company. The research hypotheses are supported in the majority of cases, and forecasts obtained from installed base are substantially better than simple black box forecasts. Incorporating past sales via installed base supports final production decisions to satisfy future consumer demand for spare parts.
    Keywords: Installed base forecast, end-of-life service, decision support, consumer goods, spare parts
    Date: 2016–02–25
  13. By: Cobus Vermeulen, Adél Bosch, Fanie Joubert and Jannie Rossouw
    Abstract: The money supply can be broadly defined as consisting of currency and deposits. While currency forms but a small portion of the total money supply, it can be a crucial determinant of spending behaviour and subsequently economic activity. The ability of the money supply to predict an up- or downswing in economic activity, as measured by a positive or negative output gap, is evaluated over a sample period 1980 – 2012. Two models are estimated, one using only the currency component and a second using the total money supply (M3). It is found that the growth rate of real currency in circulation is reasonably accurate in predicting economic activity 6 months ahead, whereas the total money supply can predict economic activity up to 9 months ahead. It is concluded that currency in circulation can be a valuable additional source of information to policymakers and can complement other approaches of forecasting economic activity.
    Keywords: Business Cycle, Output gap, currency in circulation, probit
    JEL: C25 E32 E37 E51
    Date: 2016
  14. By: Minh Ha-Duong (CIRED - Centre International de Recherche sur l'Environnement et le Développement - EHESS - École des hautes études en sciences sociales - AgroParisTech - AgroParisTech - CIRAD - Centre de coopération internationale en recherche agronomique pour le développement - École des Ponts ParisTech (ENPC) - CNRS - Centre National de la Recherche Scientifique, CleanED - Clean Energy and Sustainable Development Lab - USTH - Université des Sciences et des Technologies de Hanoi); Franck Nadaud (CIRED - Centre International de Recherche sur l'Environnement et le Développement - EHESS - École des hautes études en sciences sociales - AgroParisTech - AgroParisTech - CIRAD - Centre de coopération internationale en recherche agronomique pour le développement - École des Ponts ParisTech (ENPC) - CNRS - Centre National de la Recherche Scientifique); Martin Jegard (CIRED - Centre International de Recherche sur l'Environnement et le Développement - EHESS - École des hautes études en sciences sociales - AgroParisTech - AgroParisTech - CIRAD - Centre de coopération internationale en recherche agronomique pour le développement - École des Ponts ParisTech (ENPC) - CNRS - Centre National de la Recherche Scientifique)
    Abstract: A collection of 417 energy scenarios was assembled and harmonized to compare what they said about nuclear, fossil and renewable energy thirty years from their publication. Based on data analysis, we divide the recent history of the energy forecasting in three periods. The first is defined by a decline in nuclear optimism, approximately until 1990. The second by a stability of forecasts, approximately until 2005. The third by a rise in the forecasted share of renewable energy sources. We also find that forecasts tend to cohere, that is they have a low dispersion within periods compared to the change across periods.
    Keywords: energy,scenario,periodization
    Date: 2016–02–17

This nep-for issue is ©2016 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.