MOGU LI GOOGLE TREND PODACI POBOLJŠATI PROGNOZIRANJE PRINOSA NA ZAGREBAČKOJ BURZI?
CAN GOOGLE TRENDS DATA ENHANCE FORECASTING RETURNS ON ZAGREB STOCK EXCHANGE?

Summary/Abstract: Successful portfolio management presents a huge challenge nowadays. Today, investors have available a greater amount of investment possibilities, as well as information, mathematical and statistical tools, models and methods. This research seeks to give an overview how to include online search volume on the most popular search engine Google in modelling and forecasting return and risk on the Zagreb Stock Exchange. Based upon monthly data from January 2004 until September 2018, several specifications of ARMA-GARCH models have been estimated: without and with included search volume regarding stocks and Zagreb Stock Exchange. The results of the analysis have shown that models with included search volume are more successful in forecasting. Moreover, the results are confirmed by simulations of investment strategies out of sample, where portfolios based on forecasting with included Google search volume resulted with greater portfolio values. This was confirmed with the inclusion of transaction costs as well. However, there still remain many research possibilities in this area in order to achieve investment goals more quickly and successfully.

• Details
• Contents