Information content of Google searches on job separation

Theoretical thesis.Spine title: The information content of Google searches on job separation.Bibliography: pages 61-68.1. Introduction -- 2. Literature review -- 3. Methodology -- 4. Results and discussion -- 5. Conclusion.There is a growing interest in alternative sources of data within the economic literature. These sources are referred to as 'Big Data'. Internet search queries are one such data source, providing an avenue to discern the real-time behaviour of users searching for information online. This thesis explores whether weekly Google search query data contain informationally-relevant signals on the composition of the US labour market; that is, search queries seemingly employed by Internet users in fear of losing their jobs, or planning to quit the labour force altogether, for example 'unemployment insurance'.A weekly composite search index is constructed from the search query data, in order to utilise all the data available in the given period. The relationship between Google search and the unemployment rate is modelled using a recently developed technique in the mixed-frequency time series literature to model the weekly Google search data and the corresponding job separations data, specifically, Ghysels et al's (2015) model. To assess the informational content of Internet search query data, a mixed-frequency Granger causality test is conducted.It is established that there is insuffcient evidence to suggest that Internet search query data is useful in predicting future job separation statistics.Mode of access: World wide web1 online resource (68 pages