Problems of Using Big Data in Econometric Research

Abstract

We live in the age of Big Data and it is no coincidence that big data has been dubbed the ‘new oil’. The term ‘big data’ (or metadata) was first used by Clifford Lynch in 2008. Emphasising that the volume of information in the world is growing rapidly, he defined big data as any heterogeneous data matrix exceeding 150 GB (75*231 Gb) per day. Until 2011, big data analysis was carried out only within the framework of scientific and statistical research. However, in early 2012, the volume of data started to be expressed in very large numbers and serious needs arose for their systematisation, classification and practical use. These needs raised the question of developing new methods of analysis to make more accurate and justified economic decisions. Thus, the need to build clear econometric models and detect patterns as a result of collecting, processing and analysing regular and large amounts of data arose. Since econometrics is a science that quantitatively and qualitatively studies economic relationships using mathematical methods, statistical models and computer science, in this paper we examine the problem of applying big data, which is part of Data Science, in econometric scientific research and in building econometric models. Because only by using econometric methods can the digital economy make more reliable forecasts, evaluate cause-effect relationships and calculate risks with high accuracy. This paper aims to highlight the importance and potential of using big data in econometric research and to provide recommendations for researchers and practitioners who are trying to use these data effectively in their work.