nep-for New Economics Papers
on Forecasting
Issue of 2021‒10‒18
three papers chosen by
Rob J Hyndman
Monash University

  1. Do inflation expectations improve model-based inflation forecasts? By Bańbura, Marta; Leiva-Leon, Danilo; Menz, Jan-Oliver
  2. The time-varying evolution of inflation risks By Korobilis, Dimitris; Landau, Bettina; Musso, Alberto; Phella, Anthoulla
  3. Machine Learning, Deep Learning, and Hedonic Methods for Real Estate Price Prediction By Mahdieh Yazdani

  1. By: Bańbura, Marta; Leiva-Leon, Danilo; Menz, Jan-Oliver
    Abstract: Those of professional forecasters do. For a wide range of time series models for the euro area and its member states we find a higher average forecast accuracy of models that incorporate information on inflation expectations from the ECB’s SPF and Consensus Economics compared to their counterparts that do not. The gains in forecast accuracy from incorporating inflation expectations are typically not large but significant in some periods. Both short- and long-term expectations provide useful information. By contrast, incorporating expectations derived from financial market prices or those of firms and households does not lead to systematic improvements in forecast performance. Individual models we consider are typically better than univariate benchmarks but for the euro area the professional forecasters are more accurate, especially in recent years (not always for the countries). The analysis is undertaken for headline inflation and inflation excluding energy and food and both point and density forecast are evaluated using real-time data vintages over 2001-2019. JEL Classification: C53, E31, E37
    Keywords: Bayesian VAR, forecasting, inflation, inflation expectations, Phillips curve
    Date: 2021–10
  2. By: Korobilis, Dimitris; Landau, Bettina; Musso, Alberto; Phella, Anthoulla
    Abstract: This paper develops a Bayesian quantile regression model with time-varying parameters (TVPs) for forecasting inflation risks. The proposed parametric methodology bridges the empirically established benefits of TVP regressions for forecasting inflation with the ability of quantile regression to model flexibly the whole distribution of inflation. In order to make our approach accessible and empirically relevant for forecasting, we derive an efficient Gibbs sampler by transforming the state-space form of the TVP quantile regression into an equivalent high-dimensional regression form. An application of this methodology points to a good forecasting performance of quantile regressions with TVPs augmented with specific credit and money-based indicators for the prediction of the conditional distribution of inflation in the euro area, both in the short and longer run, and specifically for tail risks. JEL Classification: C11, C22, C52, C53, C55, E31, E37, E51
    Keywords: Bayesian shrinkage, euro area, Horseshoe, inflation tail risks, MCMC, quantile regression, time-varying parameters
    Date: 2021–10
  3. By: Mahdieh Yazdani
    Abstract: In recent years several complaints about racial discrimination in appraising home values have been accumulating. For several decades, to estimate the sale price of the residential properties, appraisers have been walking through the properties, observing the property, collecting data, and making use of the hedonic pricing models. However, this method bears some costs and by nature is subjective and biased. To minimize human involvement and the biases in the real estate appraisals and boost the accuracy of the real estate market price prediction models, in this research we design data-efficient learning machines capable of learning and extracting the relation or patterns between the inputs (features for the house) and output (value of the houses). We compare the performance of some machine learning and deep learning algorithms, specifically artificial neural networks, random forest, and k nearest neighbor approaches to that of hedonic method on house price prediction in the city of Boulder, Colorado. Even though this study has been done over the houses in the city of Boulder it can be generalized to the housing market in any cities. The results indicate non-linear association between the dwelling features and dwelling prices. In light of these findings, this study demonstrates that random forest and artificial neural networks algorithms can be better alternatives over the hedonic regression analysis for prediction of the house prices in the city of Boulder, Colorado.
    Date: 2021–10

This nep-for issue is ©2021 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.