This thesis is a part of the Master of Science program at BI Norwegian School of Management. The school takes no responsibility for the methods used, results found and conclusions drawn. Acknowledgements This project has been a rewarding, fun, educational and exciting journey for both of us. We were both a bit nervous last spring when we chose to follow the tracks of Hal Varian and Hyunyoung Choi and challenged ourselves to write our thesis on forecasting economic variables based on Internet

more »

... queries. Their work was the first on the field, and we had apparently only a vague idea about how to tackle the challenge. However, we have never regretted our choice. It has been a true joy to write about this topic since little has been done so far and results are apparently promising. It has made us to think out of the box and to learn new methods we had never touched upon before. Though, we would have not been where we are today without certain persons who we would like to thank for their valuable support and advice. We would especially thank Professor Espen R. Moen who has, through his brilliant logic and knowledge, guided us throughout this journey. His engagement in our project, with challenging comments and advice, has been very important for our work. Furthermore we would also like to thank Professor Hilde C. Bjørnland for her support on the methodology and general comments on the thesis. We also greatly appreciate the advice and feedback from Bjørn Roger Wilhelmsen at First Securities. Lastly, I, Christian, would like to thank Linda for her support by taking more than her fair share of common responsibilities at times when work has been intense. I also truly appreciate your patience when "economics talk" has been overwhelming. That being said, all the work in this thesis is to be considered our own and we alone answer for any conclusions drawn. This goes also for any errors that could be found in our work. OSLO. Abstract This thesis explores whether online search queries, represented by Google search queries, contain information useful in forecasting short term unemployment figures in Norway or not. Based on earlier work utilizing online web queries this should be possible, even in small countries. Looking at job search theory supplied with intuition, words from the Norwegian Welfare Administration (NAV) and counseling from the Language Council of Norway we create four Google Indicators that we add to baseline models to check if this reduces the forecasting error (RMSE) of the models. Our findings supports our hypothesis, that Google search contain information useful when predicting short term changes in unemployment. Our top performing model improves the forecasting accuracy compared to its baseline model by 18.3% on average over twelve months. Our best models also outperform the leading indicator "published job advertisements". These are remarkable results given the noise in our data.