Can search data help forecast inflation? Evidence from a 13-country panel

With the wide use of Internet, search data are expected to reflect people's expectations and be used as a predictor for selective macro-economic indicators. In order to improve the accuracy of forecasting, as high frequency data, Google search volume index (GSVI) has been used to forecast macro-economic variables. This paper applies the Panel Vector Autoregressive Model (PVAR) to examine the dynamic relationship between GSVI and Consumer Price Index (CPI). A Mixed Data Sampling (MIDAS) model with GSVI is proposed to forecast CPI of 13 countries, including 9 developed and 4 developing economics. Empirical results indicate that GSVI has high correlation with CPI; the response of GSVI to a CPI shock is positive and strongly significant, vice versa. However, the feedback of GSVI to a CPI shock is much stronger for developing economies compared with developed economies. And the paper proves the usefulness of search data for inflation forecasting with 13 countries' monthly inflation data from June 2012 to June 2017, the MIDAS models including GSVI outperform the multivariate regression models on average and verify the effectiveness of treating search data as an efficient indicator for inflation forecasting.

[1]  Eric Ghysels,et al.  Forecasting Professional Forecasters , 2006 .

[2]  N. Askitas,et al.  Google Econometrics and Unemployment Forecasting , 2009, SSRN Electronic Journal.

[3]  Inessa Love,et al.  The dynamics of exchange rate volatility: A panel VAR approach , 2014 .

[4]  David M. Pennock,et al.  Predicting consumer behavior with Web search , 2010, Proceedings of the National Academy of Sciences.

[5]  Saeed Moshiri,et al.  Neural Network versus Econometric Models in Forecasting Inflation , 1999 .

[6]  H. Varian,et al.  Predicting the Present with Google Trends , 2009 .

[7]  Nikos Askitas,et al.  Google Econometrics and Unemployment Forecasting , 2009 .

[8]  Luís Catela Nunes,et al.  Nowcasting quarterly GDP growth in a monthly coincident indicator model , 2005 .

[9]  Giselle C. Guzman,et al.  Internet Search Behavior as an Economic Forecasting Tool: The Case of Inflation Expectations , 2011 .

[10]  W. Newey,et al.  Estimating vector autoregressions with panel data , 1988 .

[11]  Adnan Haider,et al.  Neural network models for inflation forecasting: an appraisal , 2012 .

[12]  E. Ghysels,et al.  Série Scientifique Scientific Series Predicting Volatility: Getting the Most out of Return Data Sampled at Different Frequencies , 2022 .

[13]  M. Aiken.  Using a neural network to forecast inflation , 1999 .

[14]  Inessa Love,et al.  Financial development and dynamic investment behavior: Evidence from panel VAR , 2006 .

[15]  Chien Chin Chen,et al.  Business Cycle Indication Using Query Logs of Search Engines , 2010, 2010 International Conference on P2P, Parallel, Grid, Cloud and Internet Computing.

[16]  Jeffrey S. Racine,et al.  Semiparametric ARX neural-network models with an application to forecasting inflation , 2001, IEEE Trans. Neural Networks.

[17]  Michael P. Clements,et al.  Macroeconomic Forecasting With Mixed-Frequency Data , 2008 .

[18]  Rob Law,et al.  Forecasting tourism demand with composite search index , 2017 .

[19]  E. Ghysels,et al.  MIDAS Regressions: Further Results and New Directions , 2006 .

[20]  Matteo Manera,et al.  Oil Prices, Inflation and Interest Rates in a Structural Cointegrated VAR Model for the G-7 Countries , 2005 .

[21]  Roselyne Joyeux,et al.  Macro fundamentals as a source of stock market volatility in China: A GARCH-MIDAS approach , 2013 .

[22]  Massimiliano Marcellino,et al.  Midas Vs. Mixed-Frequency VAR: Nowcasting GDP in the Euro Area , 2009 .

[23]  J. Stock,et al.  Macroeconomic Forecasting Using Diffusion Indexes , 2002 .

[24]  J. Junttila.  Structural breaks, ARIMA model and Finnish inflation forecasts , 2001 .

[25]  Jian Ma,et al.  A MIDAS modelling framework for Chinese inflation index forecast incorporating Google search data , 2015, Electron. Commer. Res. Appl..