ETLAnow is an experiment at ETLA, The Research Institute of the Finnish Economy, to use big data in economic forecasting.
At the moment, ETLAnow predicts the unemployment rate in every EU country using real-time data from Google Trends database and the latest official data from Eurostat. The forecasts are automatically updated every morning. The forecasts are also available on Twitter. ETLAnow is a collaborative work of 29 European research institutions.
We have crafted several documents on the ETLAnow project:
- Nowcasting the Unemployment Rate in the EU with Seasonal BVAR and Google Search Data provides an overlook on how the latest version of the model works.
- ETLAnow: A Model for Forecasting with big data describes the background of the project and the methodology of the earlier model.
- Forecasting Unemployment with Google Searches provides an extensive analysis on unemployment forecasting using Google data.
- Earlier work also includes this brief, which is based on our first trial for real-time forecasting and an earlier report in Finnish.
The ETLAnow project has been originally developed by Joonas Tuhkuri (MIT), the latest version of the ETLAnow model has been developed by Jetro Anttonen and the project is maintained by Markku Lehmus at Etla.
The map above and tables below show the ETLAnow forecasts. The first row of a table gives the most recent official unemployment statistics in the EU from the European Union Labor Force Survey and the second row gives the ETLAnow forecasts.
The last row of each table compares the ETLAnow forecasts to the official unemployment rate one year ago. This comparison tells whether the ETLAnow forecasts rising or falling unemployment rate. For example, +0.1 indicates that the unemployment rate is expected to be 0.1 percentage points higher than a year ago. Similarly, -0.1 indicates that the unemployment rate is expected to fall as compared to the rate from a year ago. The forecasts and Google search terms used in the predictions are also available in the links below each table.
Last update for the automated estimates: 4.8.2020. Next forecast available: 5.8.2020.
ETLAnow is fully automatic and Etla does not make any modifications to its forecasts prior to publication. Simulated historical data for the forecasts are available for download from links below the tables. This data estimates what ETLAnow would have predicted since 2014, if it were in use. Ultimately, we can evaluate the accuracy of the ETLAnow forecasts every month when new official data becomes available. The reported previous forecasts starting from August 2018 have been released prior to the official data releases. ETLAnow model is a seasonal Bayesian vector autoregressive model extended with Google data and it uses harmonized and non-seasonally adjusted unemployment rates. More details are available in the publications listed at the top of the page.
Potential errors or other observations can be reported to firstname.lastname@example.org.
Copyright: All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.