To What Extent Can We Use Google Trends to Predict Inflation Statistically?
摘要
Inflation is a direct expression of the increase in the Consumer Price Index, which can strongly impact the general living standard. Therefore, a method that can predict the future inflation rate is significant to improve the stability of the economy internationally. Unlike the previous attempts at predicting future inflation, in this paper, data from Google Trends is used, whose data source is reliable and fast. By comparing our results with the real historical inflation data, we found that OLS and AR models both show similar results. While using both a fixed method and a method with rolling coefficients, a better result is obtained compared to the AR(2) model. However, estimating through Google Trend data has its own weakness in long-term prediction, as it has more decreasing accuracy.