Assessing the predictive capability of Google search volumes on Philippine peso exchange rates
Date
2016-06
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The onset of the computer age and developments in information technology have
changed the way people gather information. Web search engines such as Google and
Yahoo! provide efficient platforms for acquiring relevant data. Web search query
volumes are now being made available through websites such as Google Trends. This
study aims to utilize time series data on weekly Google searches in order to determine
whether the number of searches for selected keywords related to the foreign exchange
market is a significant indicator of movements in the said market. The study applies a
two-step approach consisting of GARCH (1,1) and OLS regression. Results show that
the GARCH (1,1) conditional variance is not the sole and unbiased predictor of
foreign exchange rate volatility. Additionally, tests show that geopolitical search
terms give better predictions of foreign exchange rate volatility beyond the GARCH
(1,1) model as compared to economic-related keywords based on worldwide Google
search volumes, while a combination of geopolitical- and economic-related keywords
give the best predictions for Philippine search volumes.
Description
Keywords
foreign exchange, market volatility, web search volumes, Google trends