nep-for New Economics Papers
on Forecasting
Issue of 2017‒02‒05
ten papers chosen by
Rob J Hyndman
Monash University

  1. Macroeconomic forecasting for Australia using a large number of predictors By Bin Jiang; George Athanasopoulos; Rob J Hyndman; Anastasios Panagiotelis; Farshid Vahid
  2. Nowcasting Building Permits with Google Trends By Coble, David; Pincheira, Pablo
  3. Mixed-frequency models for tracking short-term economic developments in Switzerland By Christian Hepenstrick and Rolf Scheufele Alain Galli
  4. Non-parameteric news impact curve: a variational approach By Matthieu Garcin; Clément Goulet
  5. Can U.S. EIA Retail Gasoline Price Forecasts Be Improved Upon? By Arunanondchai, Panit; Senia, Mark C.; Capps, Oral Jr
  6. The roles of inflation expectations, core inflation, and slack in real-time inflation forecasting By Kishor, N. Kundan; Koenig, Evan F.
  7. Commodity Futures and Forecasting Commodity Currencies By Francesco Ravazzolo; Tommy Sveen; Sepideh K. Zahiri
  8. New Estimates of the Redistributive Effects of Social Security By Li Tan; Cory Koedel;
  9. Forecasting Cash Rent Values By Benavidez, Justin; Hardin, Erin
  10. Indirect Inference Estimation of Mixed Frequency Stochastic Volatility State Space Models Using MIDAS Regressions and ARCH Models By Patrick Gagliardini; Eric Ghysels; Mirco Rubin

  1. By: Bin Jiang; George Athanasopoulos; Rob J Hyndman; Anastasios Panagiotelis; Farshid Vahid
    Abstract: A popular approach to forecasting macroeconomic variables is to utilize a large number of predictors. Several regularization and shrinkage methods can be used to exploit such high-dimensional datasets, and have been shown to improve forecast accuracy for the US economy. To assess whether similar results hold for economies with different characteristics, an Australian dataset containing observations on 151 aggregate and disaggregate economic series is introduced. An extensive empirical study is carried out investigating forecasts at different horizons, using a variety of methods and with information sets containing different numbers of predictors. The results share both differences and similarities with the conclusions from the literature on forecasting US macroeconomic variables. The major difference is that forecasts based on dynamic factor models perform relatively poorly compared to forecasts based on other methods which is the opposite of the conclusion made by Stock and Watson (2012) for the US. On the other hand, a conclusion that can be made for both the Australian and US data is that there is little to no improvement in forecast accuracy when the number of predictors is expanded beyond 20-40 variables.
    Keywords: Australian economy, Bayesian VAR, bagging, dynamic factor model, ridge regression, least angular regression, shrinkage, regularization.
    JEL: C52 C53 C55
    Date: 2017
  2. By: Coble, David; Pincheira, Pablo
    Abstract: We propose a useful way to predict building permits in the US, exploiting rich real-time data from web search queries. The time series on building permits is usually considered as a leading indicator of economic activity in the construction sector. Nevertheless, new data on building permits are released with a lag close to two months. Therefore, an accurate now-cast of this leading indicator is desirable. We show that models including Google search queries nowcast and forecast better than our good, not naïve, univariate benchmarks both in-sample and out-of-sample. We also show that our results are robust to different specifications, the use of rolling or recursive windows and, in some cases, to the forecasting horizon. Since Google queries information is free, our approach is a simple and inexpensive way to predict building permits in the United States.
    Keywords: Online Search; Prediction; Forecasting; Time Series; Building Permits; Real Estate; Google Trends.
    JEL: C10 C5 C53 F3 F37
    Date: 2017–02–01
  3. By: Christian Hepenstrick and Rolf Scheufele Alain Galli
    Abstract: We compare several methods for monitoring short-term economic developments in Switzerland. Based on a large mixed-frequency data set, the following approaches are presented and discussed: factor-based information combination approaches (including factor model versions based on the Kalman filter/smoother, a principal component based version and the three-pass regression filter), a model combination approach resting on MIDAS regression models and a model selection approach using a specific-to-general algorithm. In an out-of-sample GDP forecasting exercise, we show that the considered approaches clearly beat relevant benchmarks such as univariate time-series models and models that work with one or a small number of indicators. This suggests that a large data set is an important ingredient for successful real-time monitoring of the Swiss economy. The models using a large data set particularly outperform others during and after the Great Recession. Forecast pooling of the most-promising methods turns out to be the best option for obtaining a reliable nowcast for the Swiss economy.
    Keywords: Mixed frequency, GDP, nowcasting, forecasting, Switzerland
    JEL: C32 C53 E37
    Date: 2017
  4. By: Matthieu Garcin (Natixis Asset Management et LabEX ReFi); Clément Goulet (Centre d'Economie de la Sorbonne et LabEX ReFi)
    Abstract: In this paper, we propose an innovative methodology for modelling the news impact curve. The news impact curve provides a non-linear relation between past returns and current volatility and thus enables to forecast volatility. Our news impact curve is the solution of a dynamic optimization problem based on variational calculus. Consequently, it is a non-parametric and smooth curve. To our knowledge, this is the first time that such a method is used for volatility modelling. Applications on simulated heteroskedastic processes as well as on financial data show a better accuracy in estimation and forecast for this approach than for standard parametric (symmetric or asymmetric ARCH) or non-parametric (Kernel-ARCH) econometric techniques
    Keywords: Volatility modeling; news impact curve; calculus of variations; wavelet theory; ARCH
    JEL: C02 C14 C22 C51 C53 C58 C61
    Date: 2015–09
  5. By: Arunanondchai, Panit; Senia, Mark C.; Capps, Oral Jr
    Abstract: Perhaps the most widely followed price in the market is the price of crude oil. The volatility of this commodity is evident to consumers through the gasoline prices that consumers see on the retail side. The U.S. Energy Information Agency provides widely followed forecasts for the retail gasoline price (along with other energy products) produced with their short-term energy outlook (STEO) model. The purpose of this research is to compare a number of forecasts using different techniques to the STEO model. This is accomplished through the use of Holt Winters, structural, ARIMA, and vector error-correction models. We also construct a composite forecast by averaging the respective forecasts from the four models. From the empirical analysis, we find evidence from the structural model and the vector error-correction model that the movement in the gasoline prices can be explained by the West Texas Intermediate (WTI) benchmark and the spread between BRENT and WTI benchmarks. In terms of forecasting performance, the additive Holt Winters model outperforms the other models within sample. Out sample, the composite forecast is the best performing model. The composite forecast has a MAPE of 6.3% versus a MAPE of 8.1% from the STEO model.
    Keywords: Retail Gasoline Prices, Forecasting, short-term energy outlook model, Holt-Winters model, ARIMA model, structural model, vector error-correction model, Resource /Energy Economics and Policy, Q47,
    Date: 2017
  6. By: Kishor, N. Kundan (University of Wisconsin); Koenig, Evan F. (Federal Reserve Bank of Dallas)
    Abstract: Using state-space modeling, we extract information from surveys of long-term inflation expectations and multiple quarterly inflation series to undertake a real-time decomposition of quarterly headline PCE and GDP-deflator inflation rates into a common long-term trend, common cyclical component, and high-frequency noise components. We then explore alternative approaches to real-time forecasting of headline PCE inflation. We find that performance is enhanced if forecasting equations are estimated using inflation data that have been stripped of high-frequency noise. Performance can be further improved by including an unemployment-based measure of slack in the equations. The improvement is statistically significant relative to benchmark autoregressive models and also relative to professional forecasters at all but the shortest horizons. In contrast, introducing slack into models estimated using headline PCE inflation data or conventional core inflation data causes forecast performance to deteriorate. Finally, we demonstrate that forecasting models estimated using the Kishor-Koenig (2012) methodology-which mandates that each forecasting VAR be augmented with a flexible state-space model of data revisions-consistently outperform the corresponding conventionally estimated forecasting models.
    Keywords: Inflation; real-time forecasting; unobserved component model; slack
    JEL: E31 E37
    Date: 2016–11–01
  7. By: Francesco Ravazzolo; Tommy Sveen; Sepideh K. Zahiri
    Abstract: This paper analyzes the extent to which information in commodity futures markets is useful for out-of-sample forecasting of commodity currencies. In the earlier literature, commodity price changes are documented to be weak out-of-sample predictors of commodity currency return. In contrast, we find that the basis of several commodities may contain useful information, but the usefulness of any particular commodity basis varies over time and depends on the nature of the commodity. In particular, it seems the basis of commodities with relatively high storage costs tend to be more useful. We argue that high storage costs will tend to make the basis more prone to fluctuations in commodity risk and therefore provide information about the risk premium for commodity currencies. We implement forecast combination strategies that take full advantage
    Keywords: Exchange rate predictability, commodity futures market, commodity currencies, forecast combinations
    Date: 2016–11
  8. By: Li Tan (Department of Economics at the University of Missouri); Cory Koedel (Department of Economics and Truman School of Public Affairs, at the University of Missouri);
    Abstract: We forecast lifetime earnings of young workers to study the redistributive effects of Social Security, prospectively. Using data from an older generation of workers, we first establish that our forecasting method can recover the actual distribution of Average Indexed Monthly Earnings taken from Social Security Administration records. We then extend the method to forecast Social Security returns for recent cohorts and examine redistributive trends. Our methods and data are accessible, facilitating straightforward replications and extensions. Focusing on redistributions across race and education groups, and on men’s own benefits, we show that Social Security exhibits little progressivity, and little progressivity improvement, for recent cohorts.
    Keywords: Earnings forecast, Bayesian forecasting, Social Security, Social Security progressivity, Social Security projections
    JEL: H55 J18 J32
    Date: 2017–01
  9. By: Benavidez, Justin; Hardin, Erin
    Abstract: Stability in the farming sector provides stability in rural economies, with a varying but large portion of employment in rural communities across the nation directly related to agriculture or to the agricultural services and processing industries. Instability in the agricultural sector can send ripple effects throughout the economy through increased food and fiber prices. Additionally, there has been a movement towards land investment by equity firms. As rent i s the primary source of revenue, understanding movements in rent is useful for mitigating risk and understanding the market. The purpose of the following research is to address the deficit in recent forecast literature pertaining to land and cash rent prices and to identify the best methodology for forecasting. Tested methods include a Holt-Winters naive forecast, a structural model with lagged rent, farmland prices and crop prices as explanatory variables, an error correction model (ECM), an autoregressive integrated moving average (ARIMA) forecast model, and a composite forecast. Each model is evaluated using mean absolute percent error (MAPE) and root mean squared error (RMSE).
    Keywords: Rent, Land Value, Forecasting, Agribusiness, Agricultural Finance, Farm Management, Land Economics/Use,
    Date: 2017
  10. By: Patrick Gagliardini (University of Lugano and Swiss Finance Institute); Eric Ghysels (University of North Carolina Kenan-Flagler Business School, University of North Carolina (UNC)); Mirco Rubin (University of Bristol)
    Abstract: We examine the relationship between MIDAS regressions and the estimation of state space models applied to mixed frequency data. While in some cases the binding function is known, in general it is not, and therefore indirect inference is called for. The approach is appealing when we consider state space models which feature stochastic volatility, or other non-Gaussian and nonlinear settings where maximum likelihood methods require computationally demanding approximate filters. The stochastic volatility feature is particularly relevant when considering high frequency financial series. In addition, we propose a filtering scheme which relies on a combination of re-projection methods and now-casting MIDAS regressions with ARCH models. We assess the efficiency of our indirect inference estimator for the stochastic volatility model by comparing it with the Maximum Likelihood (ML) estimator in Monte Carlo simulation experiments. The ML estimate is computed with a simulation-based Expectation-Maximization (EM) algorithm, in which the smoothing distribution required in the E step is obtained via a particle forward-filtering/backward-smoothing algorithm. Our Monte Carlo simulations show that the Indirect Inference procedure is very appealing, as its statistical accuracy is close to that of MLE but the former procedure has clear advantages in terms of computational efficiency. An application to forecasting quarterly GDP growth in the Euro area with monthly macroeconomic indicators illustrates the usefulness of our procedure in empirical analysis.
    Keywords: Indirect inference, MIDAS regressions, State space model, Stochastic volatility, GDP forecasting.
    Date: 2016–07

This nep-for issue is ©2017 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.