Not logged in.
Quick Search - Contribution
Contribution Details
| Type | Bachelor's Thesis |
| Scope | Discipline-based scholarship |
| Title | Can User Trends in the Internet explain and predict Sector-Based Stock Market Movements? |
| Organization Unit | |
| Authors |
|
| Supervisors |
|
| Language |
|
| Institution | University of Zurich |
| Faculty | Faculty of Business, Economics and Informatics |
| Number of Pages | 31 |
| Date | 2017 |
| Zusammenfassung | A large number of investors in financial markets gather information before trading. Most of the collecting of useful information is done over the internet, using a search engine like Google. Google Trends is a free service provided by Google, which gives insight into people’s searches on its platform. Therefore, it might be possible to gather information on investor’s decisions by using Google Trends. By analysing the data, it should be possible to forecast stock price movements. This thesis focuses on negative word data from Google Trends and researches the possibility to use it as an indicator for market movements in different U.S. industry indices. The hypothesis is that if investors search for more negative words, markets are expected to lose value but prices should rise if negative search volumes decline. This was tested on the 21 different S&P Select Industry Indices using a list composed of 18 negative words from the paper written by Tim Loughran and Bill McDonald in 2011. The word debt was also included, being the word yielding the highest return in a similar research done by Preis, Moat, & Stanley in 2013. The timeframe for the data analysis was chosen to be the period from 2007-2017, data was collected on a weekly frequency. The period was split into two five-year spans. Initially a linear regression model of the data was analysed. During the first period (2007-2012) five of the 19 negative words on the list were statistically significant in more than 50% of the 21 tested indices. For the second period (2012-2017), that number was seven words. In both periods, a significance level of 5% was chosen. Those new lists were then used for further analysis. First of all, new Google Trends datasets from the same period were downloaded on different days. This data was then reevaluated in the linear regression model and in most cases, nothing changed in the statistical significance of the words. New datasets were used to see if the models were able to accurately predict the indices. This did not yield any useful results concerning a good fit of the models, especially not for the second period 2012-2017. This may indicate that the information gained from the linear regression model should only be used for the similar investment strategy proposed by Preis, Moat & Stanley’s paper in 2013. The trading strategy analysed for this thesis utilizes the statistical significant words gained from the linear regression models in this research and uses those as indicators for investing in different U.S. stock market indices. Depending on the Google Data a weeklong short-, long- or hold-position is taken. Results from this indicate that there might be connections between Google search query volume changes and the different U.S. industries. Especially during times of deflation and crises the investment strategy based on Google Trends data could outperform the buy-and-hold strategy of the benchmarks. During periods where markets were growing the returns from the active investment strategy yielded worse returns than the passive strategy. A reason for this might be that the simple investment strategy analysed in this paper does not include a growth constant for the economy during thriving market periods. This may indicate a rather strong bias during times of crises of this investment strategy. The basis for this research comes from different independent researches, where similar interactions between internet data and the economy were analysed. Comparable papers concluded that at least some correlation exists. But this paper has no final answer to the question whether some of those strategies can and should be considered for real world investment. |
| Export |
BibTeX
|