Google trends indices as factors in forecasting unemployment rates in the Philippines

• Main Authors: Bellosillo, Dianne Althea A., Bernardo, Learrah Mari A., Ng, Mark Stevens C., Villanueva, Celina Camille B.
• Format: text
• Published: Animo Repository 2022
• Subjects: Unemployment—Philippines—Forecasting; Labor Economics; Work, Economy and Organizations
• Online Access: https://animorepository.dlsu.edu.ph/etdb_econ/41; https://animorepository.dlsu.edu.ph/cgi/viewcontent.cgi?article=1041&context=etdb_econ
• Institution: De La Salle University

With search engines gaining traction for job seekers, Internet searches have become a viable data source in forecasting unemployment in developed countries. The project seeks to answer if search data from Google Trends are useful as a factor in forecasting the unemployment rates in the context of developing countries such as the Philippines. Search and matching theory is the basis for the use of Google Trends as the theory uses search intensity to model unemployment outcomes. The models used in forecasting are VAR and the ARIMA regression models. The data on the chosen variables are taken from the Google Trends website and the quarterly LFS. The RMSE, MAE, and MAPE error measures were applied to test the accuracy between the models’ forecasts. The tests show Google Trends models as more appropriate for short-term forecasting and may be less appropriate for long-term forecasting.