Results of Nowcasting competitions - international trade

Back to introduction

The European Statistics Awards Programme on Nowcasting aims at improving timelines of European Statistics which is a recurring demand by policymakers and other users of European statistics.

We were looking for new approaches based on advanced modelling (possibly using alternative, almost real-time, information) that could give us accurate estimates of key indicators much faster than before.

The third round of the Nowcasting awards was divided into 3 separate competitions, each focusing on the nowcasting of one international trade indicator.

European International Trade Intra-EU exports

• [EU27_2020] European Union - 27 countries (from 2020) (Geopolitical entity (partner))
• [ET-T] Total (Indicator)
• [EXP] Export (Stock of flow)
• [MIO-EUR-NSA] Trade value - million of euro - not seasonally adjusted (Unit of measure)

European International Trade Extra-EU exports

• [EXT_EU27_2020] Extra-EU27 (from 2020) (Geopolitical entity (partner))
• [ET-T] Total (Indicator)
• [EXP] Export (Stock of flow)
• [MIO-EUR-NSA] Trade value - million of euro - not seasonally adjusted (Unit of measure)

European International Trade Extra-EU imports

• [EXT_EU27_2020] Extra-EU27 (from 2020) (Geopolitical entity (partner))
• [ET-T] Total (Indicator)
• [IMP] Import (Stock of flow)
• [MIO-EUR-NSA] Trade value - million of euro - not seasonally adjusted (Unit of measure)

Trade flows cover all goods entering (imports) or leaving (exports) the statistical territories of the EU Member States. Extra-EU refers to transactions with all countries outside of the EU: the rest of the world except for the European Union (EU) Member States.

As part of the competition requirements, teams were expected to submit 6 point estimates of monthly time series across all 27 EU countries. Teams were also required to provide comprehensive methodological descriptions, including source code.

The International Trade Nowcasting Challenge was launched in October 2024, with a final deadline for submissions in March 2025 and documentation in April 2025.

Participation was quite wide as a total of 35 teams, comprised of 66 individuals from 17 countries, signed up for this competition. The results of the evaluation are announced below.

The participants were competing for three types of awards:

• The Accuracy Award prizes were given to the top three teams whose entries yielded point estimates closest to the published indicator.
• The Reusability Award prize was given to the team whose entry shows great potential to be scaled up to European statistics production. The Reusability Award was intended to support the best described and documented solutions, with the most innovative, open approach using open data.
• The Innovativity Award prize was given to the team whose entry shows the most originality in its approach. The Innovativity award is intended to support the most thoroughly described and documented solutions, with the most innovative approach.

Accuracy Award winners - Intra-EU exports

Accuracy Award winners - Extra-EU exports

Accuracy Award winners - Extra-EU imports

Reusability award winners

Innovativity award winners

Solutions of the reusability and innovativity award winners

Since the Reusability and Innovativity awards were intended to support the most thoroughly described and documented solutions, with the most innovative, open approach, we are happy to share with you the solutions which won this prize in three competitions (INTRA-EU Exports; EXTRA-EU Exports; Extra-EU Imports).

This round of nowcasting competitions has concluded with one team claiming exceptional achievements!

Team TURKSTAT-DATG has won both the Reusability and Innovativity prizes across all three competitions. Their solution clearly stood out from other submissions in both innovation and clarity. In presenting their solution, team TURKSTAT-DATG demonstrated a textbook example of reproducible architecture, developing an innovative nowcasting solution that combines theoretical rigor with practical simplicity. Their methodology extends the well-established TRAMO/SEATS framework—widely used in official statistics for seasonal adjustment—by incorporating trade-weighted foreign variables such as relative prices, partner country industrial production, and exchange rates. The approach features both multivariate and univariate specifications, with their ensemble entry averaging forecasts from both models to balance complexity and robustness. Key strengths include exceptional replicability through open-source code and versioned data management, scalability via a modular, control-table-driven architecture that can easily accommodate new countries or indicators, interpretability with transparent, theory-driven predictors grounded in international trade economics, and maintainability using widely-adopted R packages and clear documentation.

The team demonstrated that simpler, transparent econometric models can outperform more complex alternatives like GVAR and LSTM while remaining fast - approximately 3 minutes runtime - cost-effective, and suitable for operational implementation in statistical offices. Their methodology offers a sustainable, production-ready solution that successfully balances statistical sophistication with practical usability.

Their solutions can be found in the following links:

Accuracy Award winners - Intra-EU exports

The GitHub repository containing the code and full project documentation is publicly accessible here:

• GitHub repository

Accuracy Award winners - Extra-EU exports

The GitHub repository containing the code and full project documentation is publicly accessible here:

• GitHub repository

Accuracy Award winners - Extra-EU imports

The GitHub repository containing the code and full project documentation is publicly accessible here:

• GitHub repository

Final Accuracy Rankings

Intra-EU exports

*The team did not submit additional documentation describing their approach, which was an eligibility condition in this round of nowcasting competitions, making them not eligible to receive a prize.

Extra-EU exports

*The team did not submit additional documentation describing their approach, which was an eligibility condition in this round of nowcasting competitions, making them not eligible to receive a prize.

Extra-EU imports

*The team did not submit additional documentation describing their approach, which was an eligibility condition in this round of nowcasting competitions, making them not eligible to receive a prize.