People Search for Their Money: Search Volume and the Stock Market Prediction  

Oleh:

Oh, Jeongwoo

Jenis: Article from Proceeding
Dalam koleksi: SIBR-Thammasat 2014 Conference on Interdisciplinary Business & Economics Research June 5th- 7th, 2014 di Emerald Hotel Bangkok, page 1-2.
Topik: Search data; Stock market prediction; Co-movement; Causality
Fulltext: b14-141.pdf (16.58KB)

Isi artikel

Search data has attracted considerable amount of interest in applied studies in terms of reflecting the real time interest of people in large amount. Finance study also has increased its own attention toward the search data to capture many unknown investor’s interest. By using new type of data with the existing financial data, market prediction and finance research is expected to be enriched, providing more sophisticated information of the real time market to the scholars and investors. In this context, we try to verify the usefulness of internet based data in the finance study. Such objective is projected on the stock market prediction with Google trends, normalized time series index of the query volume based on the query share. We explore the effective way to use both search data and stock market data, testing the co-movement of both time series data stream and verifying the causality. IT big companies are selected to examine as the first step, since they are highly liquidated as Nasdaq-100 components and expected to be searched highly frequently online. In order to test the co-movement of both time series data stream, VAR (vector autoregressive) model is set up. As a result, strong co-movement between search volume and the trading volume is discovered. While search volume doesn’t show consistent results with the stock price, strong correlation with stock price is still detected for some individual companies. As the co-movement between search volume and the share trading volume is verified, Granger causality on this co-movement is tested accordingly. The result of the Granger causality test demonstrates the predictability of search data on the trading volume. We expect that what people search delivers new source of information about the investment intention, in the circumstance that the financial data is archived on the web in a huge amount and many people already use the internet-based source for their investment. Therefore, our work can provide information especially on the mass collective interest in the stock market, by focusing on the collective search activity of unknown users, instead of professional or institutional investors. Finally, our effort of such application is expected to brighten the potential of web-based information in explaining the real market. We hope that we raise active discussion on this topic so to bring more resources for the research as well.

Opini Anda

Klik untuk menuliskan opini Anda tentang koleksi ini!

Kembali