Turkey | Big Data presentation at the Bank of Canada, Federal Reserve and Bank of Italy

Summary

BBVA Research present at the Conference on Non-traditional Data, Machine Learning and Natural Language Processing in Macroeconomics. We show evidence that Big Data from financial transactions is very useful for Nowcasting models of consumption and investment and reinforces the power of our traditional models.

Key points

• Key points:
• Financial Transactions´ BigData improve accuracy of Nowcasting models in Turkey. It is useful more than 50% of the time (even with prevalence)
• The contribution is more relevant during the first 45 days (when Hard relevant Data is scarce) and uncertain crisis times.
• The Standard Nowcasting Models as Dynamic Factor Model (DFM) & Bayesian VARs (BVAR) appears to be a good alternative model even in a volatile environment.
• Nowcast combination outperform most of the single models in many cases but not in the short term.

Geographies

Topics

Tags