Nowcasting Turkish unemployment using real time data from Google

Turkish unemployment data is published with a significant delay (nearly three months), making it difficult to analyze the labor market in advance. In order to assess the evolution of the unemployment in high frequency, we have developed a dynamic model combining both the Business Cycle and Google searches for jobs.

Key points

• Key points:
• Google Trends data in real time (daily) allows us to cover the gap in the timeline for the release of official data regarding the Turkish unemployment rate (almost three months), thereby improving our nowcasting accuracy
• Bayesian moving average techniques show that the own dynamics of unemployment, economic activity indicators such as Industrial Production and Capacity Utilization and Google Searches related to unemployment, are important variables for nowcasting the Turkish unemployment rate
• Nowcasting results for August, September and October show that unemployment is stabilizing slightly and improving in line with the renewed strength of economic activity and Google searches

Authors

Geographies

Topics

Tags