nep-for New Economics Papers
on Forecasting
Issue of 2015‒06‒27
twelve papers chosen by
Rob J Hyndman
Monash University

  1. Forecasting Core Inflation: The Case of South Africa By Franz Ruch; Mehmet Balcilar; Mampho P. Modise; Rangan Gupta
  2. Generalized Dynamic Factor Models and Volatilities: Estimation and Forecasting By Matteo Barigozzi; Marc Hallin
  3. Earnings Forecast Accuracy And Career Concerns By Roger, Tristan
  4. Forecasting Inflation in an Inflation Targeting Economy: Structural Versus Non-Structural Models By Rangan Gupta; Alessia Paccagnini; Charles Rahal
  5. Incorporating Economic Policy Uncertainty in US Equity Premium Models: A Nonlinear Predictability Analysis By Stelios Bekiros; Rangan Gupta; Anandamayee Majumdar
  6. Core Inflation and Trend Inflation By James H. Stock; Mark W. Watson
  7. Economic Relevance of Hidden Factors in International Bond Risk Premia By Tiozzo Pezzoli, Luca
  8. Predictive quantile regression with persistent covariates: IVX-QR approach By Lee, JiHyung
  9. Managing longevity risk By Li, Hong
  10. Commodity Currencies Revisited By Passari, Evgenia
  11. FRED-MD: A Monthly Database for Macroeconomic Research By McCracken, Michael W.; Ng, Serena
  12. Revisiting the long memory dynamics of implied-realized volatility relation: A new evidence from wavelet band spectrum regression By Barunik, Jozef; Barunikova, Michaela

  1. By: Franz Ruch (South African Reserve Bank); Mehmet Balcilar (Department of Economics, Eastern Mediterranean University, Famagusta, Northern Cyprus , via Mersin 10, Turkey; Department of Economics, University of Pretoria, Pretoria, 0002, South Africa.); Mampho P. Modise (National Treasury, 40 Church Square, Pretoria, 0002, South Africa); Rangan Gupta (Department of Economics, University of Pretoria)
    Abstract: Forecasting and estimating core inflation has recently gained attention, especially for inflation targeting countries, following research showing that targeting headline inflation may not be optimal; a Central Bank can miss the signal due to the noise. Despite its importance there is sparse literature on estimating and forecasting core inflation in South Africa, with the focus still on measuring core inflation. This paper emphasises predicting core inflation using large time-varying parameter vector autoregressive models (TVP-VARs), factor augmented VAR, and structural break models using quarterly data from 1981Q1 to 2013Q4. We use mean squared forecast errors (MSFE)and predictive likelihoods to evaluate the forecasts. In general, we find that (i) small TVP-VARs consistently outperform all other models; (ii) models where the errors are heteroscedastic do better than models with homoscedastic errors; (iii) models assuming that the forgetting factor remains 0.99 throughout the forecast period outperforms models that allow for the forgetting factors to change with time; and (iv) allowing for structural break does not improve the predictability of core inflation. Overall, our results imply that additional information on the growth rate of the economy and interest rate is sufficient to forecast core inflation accurately, but the relationship between these three variables needs to be modelled in a time-varying (nonlinear) fashion.
    Keywords: Core inflation, forecasting, small- and large-scale vector autoregressive models, constant and time-varying parameters
    JEL: C22 C32 E27 E31
    Date: 2015–06
  2. By: Matteo Barigozzi; Marc Hallin
    Keywords: volatility; dynamic factor models; GARCH models
    JEL: C32
    Date: 2015–06
  3. By: Roger, Tristan
    Abstract: Previous studies show that analysts' compensation is not linked to earnings forecast accuracy. We evidence however that analysts have incentives to issue accurate forecasts. We show that brokerage houses reward their best forecasters by assigning them to large, mature firms. Covering such firms increases the potential for future compensation as these firms generate a great deal of investment banking and trading activities. The coverage of such firms also increases analysts' exposure to large buy-side investors. We find that analysts covering large, mature firms are twice as likely to be recognized as star analysts by Institutional Investor. We explain our findings on forecast accuracy as the result of brokerage houses' concerns for reputation.
    Keywords: Forecast accuracy; Brokerage houses; Analysts' compensation; Reputation; Carrer;
    JEL: G23 G24 M12 M52
    Date: 2015–06
  4. By: Rangan Gupta (Department of Economics, University of Pretoria); Alessia Paccagnini (Department of Economics, Università degli Studi Milano - Bicocca); Charles Rahal (Department of Economics, University of Birmingham)
    Abstract: We propose a comparison between a group of nested and non-nested atheoretical and theoretical models in forecasting the inflation rate for South Africa, an inflation-targeting country. In a pseudo real-time environment, our results show that for shorter horizons, the atheoretical models, such as Vector Error Correction Models, with and without factors, perform better, while for longer horizons, theoretical (DSGE based) models outperform their competitors.
    Keywords: Inflation, South Africa, Structural, Atheoretical, Factors, DSGE
    JEL: C11 C32 C52
    Date: 2015–06
  5. By: Stelios Bekiros (European University Institute (EUI) and IPAG Business School); Rangan Gupta (Department of Economics, University of Pretoria and IPAG Business School); Anandamayee Majumdar (Center for Advanced Statistics and Econometrics, Soochow University, Suzhou, China)
    Abstract: Information on economic policy uncertainty does matter in predicting the US equity premium, especially when accounting for structural instabilities and omitted nonlinearities in their relationship, via a quantile predictive regression approach over the monthly period 1900:1-2014:2. Unlike as suggested by a linear mean-based predictive model, the extended quantile regression model with the incorporation of the EPU proxy, enhances significantly the out-of-sample stock return predictability. This is observed especially when the market is neutral, exhibits a side or mildly upward trending behavior, yet not when the market appears to turn highly bullish.
    Keywords: stock markets, economic uncertainty, predictability, quantile regression
    JEL: C22 C53 E60 G10
    Date: 2015–06
  6. By: James H. Stock; Mark W. Watson
    Abstract: An important input to monetary policymaking is estimating the current level of inflation. This paper examines empirically whether the measurement of trend inflation can be improved by using disaggregated data on sectoral inflation to construct indexes akin to core inflation, but with time-varying distributed lags of weights, where the sectoral weight depends on the time-varying volatility and persistence of the sectoral inflation series, and on the comovement among sectors. The model is estimated using U.S. data on 17 components of the personal consumption expenditure inflation index. The modeling framework is a dynamic factor model with time-varying coefficients and stochastic volatility as in del Negro and Otrok (2008); this is the multivariate extension of the univariate unobserved components-stochastic volatility model of trend inflation in Stock and Watson (2007). Our main empirical results are (i) the resulting multivariate estimate of trend inflation is similar to the univariate estimate of trend inflation computed using core PCE inflation (excluding food and energy) in the first half of the sample, but introduces food in the second half of the sample: early in the sample, food inflation was noisy and a poor indicator of trend inflation, but now food inflation is less volatile, more persistent, and a useful indicator; (ii) the model-based filtering uncertainty about trend inflation is substantially reduced by using the disaggregated series in a multivariate model, relative to computing the trend using only headline inflation; (iii) the multivariate trend and the univariate trend constructed using core measures of inflation forecast average inflation over the 1-3 year horizon more accurately than a variety of other benchmark inflation measures, although there is considerable sampling uncertainty in these forecast comparisons.
    JEL: E31
    Date: 2015–06
  7. By: Tiozzo Pezzoli, Luca
    Abstract: This paper investigates the relevance of hidden factors in international bond risk premia to forecast future excess bond returns and macroeconomic variables such as economic growth and ination rate. Using maximum likelihood estimation of a linear Gaussian state-space model, adopted to explain the dynamics of expected excess bond returns of a given country, associated selection criteria detect as relevant, factors otherwise judged negligible by the classical explained variance approach adopted by Cochrane and Piazzesi (2005) and Cochrane and Piazzesi (2008). We call these factors hidden, meaning that they are not visible through the lens of a principal component analysis of expected excess bond returns. We find that these hidden factors are useful predictors of both future economic growth and ination rate given that they add forecasting power over and above the information contained both in the Cochrane and Piazzesi (2008) and in yield curve factors. These empirical findings are robust across different sample periods and countries as well as with respect to the interpolation technique used in the construction of the international bond yield data sets.
    Keywords: Financial econometrics; Interest rates; International finance;
    JEL: C52 E43 G12 G15
    Date: 2014–12
  8. By: Lee, JiHyung
    Abstract: This paper develops econometric methods for inference and prediction in quantile regression (QR) allowing for persistent predictors. Conventional QR econometric techniques lose their validity when predictors are highly persistent. I adopt and extend a methodology called IVX filtering (Magdalinos and Phillips, 2009) that is designed to handle predictor variables with various degrees of persistence. The proposed IVX-QR methods correct the distortion arising from persistent multivariate predictors while preserving discriminatory power. Simulations confirm that IVX-QR methods inherit the robust properties of QR. These methods are employed to examine the predictability of US stock returns at various quantile levels.
    Keywords: IVX filtering, Local to unity, Multivariate predictors, Predictive regression, Quantile regression.
    JEL: C1 C22
    Date: 2015–04–28
  9. By: Li, Hong (Tilburg University, School of Economics and Management)
    Abstract: The thesis first examines the choice of sample size for mortality forecasting, and then deal with the hedging of longevity risk using longevity-linked instruments. Chapter 2 proposes a Bayesian learning approach to determine the (posterior distribution of) the sample sizes for mortality forecasting using mortality models based on linear extrapolation approaches. Chapter 3 studies the static robust management of longevity risk in the situation that the hedger does not have precise knowledge of the underlying probability distribution of the future mortality rates. Mean-variance and mean-conditional-value-at-risk objective functions are used. Chapter 4 focuses on the dynamic hedging of longevity risk in the case where the trading frequency of the longevity-linked derivatives is limited. A minimum-variance objective function is used, and time-consistent hedging strategies are derived in both the benchmark case, where all assets can be traded continuously, and a constrained case, where the longevity-linked derivatives can only be traded at a low and deterministic frequency.
    Date: 2015
  10. By: Passari, Evgenia
    Abstract: I build an exchange rate strategy that trades currencies conditional on changes in the global prices of commodity indices; hence, termed “commodity strategy”. First, I document that commodity prices have signicant out-of-sample forecasting ability for the future exchange rates of several commodity exporters and importers at the daily frequency. However, I report that the reverse forecasting relationship does not survive out-of-sample testing. Second, I find a signicant cross-sectional spread, in both spot and excess returns, of 6% p.a. between the currencies that are predicted to appreciate and those that are predicted to depreciate. The returns appear to be uncorrelated to those of popular exchange rate strategies such as the carry trade and currency momentum. Furthermore, the spread in returns is not explained by traditional risk factors; however, it is partly accounted for by the strategy’s high transaction costs. Net probability can be restored by either implementing a simple market timing rule or by investing in developed markets with low costs and high liquidity.
    Keywords: Foreign Exchange; Commodity Currencies; Asset Pricing;
    JEL: F31 F37 G10 G11
    Date: 2015–06
  11. By: McCracken, Michael W. (Federal Reserve Bank of St. Louis); Ng, Serena (Department of Economics, Columbia University)
    Abstract: This paper describes a large, monthly frequency, macroeconomic database with the goal of establishing a convenient starting point for empirical analysis that requires "big data." The dataset mimics the coverage of those already used in the literature but has three appealing features. First, it is designed to be updated monthly using the FRED database. Second, it will be publicly accessible, facilitating comparison of related research and replication of empirical work. Third, it will relieve researchers from having to manage data changes and revisions. We show that factors extracted from our dataset share the same predictive content as those based on various vintages of the so-called Stock-Watson dataset. In addition, we suggest that diffusion indexes constructed as the partial sum of the factor estimates can potentially be useful for the study of business cycle chronology.
    Keywords: diffusion index; forecasting; big data; factors.
    JEL: C30 C33 G11 G12
    Date: 2015–06–15
  12. By: Barunik, Jozef; Barunikova, Michaela
    Abstract: This paper revisits the fractional co-integrating relationship between ex-ante implied volatility and ex-post realized volatility. Previous studies on stock index options have found biases and inefficiencies in implied volatility as a forecast of future volatility. It is argued that the concept of corridor implied volatility (CIV) should be used instead of the popular model-free option-implied volatility (MFIV) when assessing the relation as the latter may introduce bias to the estimation. In addition, a new tool for the estimation of fractional co-integrating relation between implied and realized volatility based on wavelets, a wavelet band least squares (WBLS) uncovers that corridor implied volatility is an unbiased forecast of future volatility in the long run. The main advantage of WBLS in comparison to other methods is that it allows us to work conveniently with potentially non-stationary volatility due to the properties of wavelets and allows us to study the relation at different investment horizons. In the estimation, we use the S&P 500 and DAX monthly and biweekly option prices covering the recent financial crisis, and we conclude that the dependence comes solely from the lower frequencies of the spectra representing long horizons. The findings enable improvement of future volatility forecasts by discarding the bias coming from the short horizons.
    Keywords: wavelet band spectrum regression,corridor implied volatility,realized volatility,fractional cointegration
    JEL: C14 C22 C51 C52 G14
    Date: 2015

This nep-for issue is ©2015 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.