Authors : Geung-Hee Lee(Professor, Department of Information Statistics, Korean National Open University), SangPil Hwang(the Bank of Korea)

Abstract

     With the advent of big data, there are a lot of economic analyses going on using amorphous data. Such are internet search queries, dialogues on social networks and blog posts. In this paper, we introduce a search business index based on the internet search data provided by Naver Trends considering market shares in Korea. The index is compiled based on the difference between search query data related to the business boom and recession. To check the usefulness of the newly compiled index, various analyses have been carried out. The analyses show that the index is highly correlated with the economic sentiment index and leads the business cycle by 2 months. Moreover, the forecasting performance of models with the index can be compared with benchmark models such as the random walk model. The results show that models with the index outperform the random walk model and AR(1) model during the global financial crisis period. The Naver search business index would be helpful in evaluating business cycles and complementing the economic sentiment index.