Measuring Brexit uncertainty and understanding the causes

Researchers have constructed a novel news-based indicator of Brexit uncertainty at both aggregate and topic-specific level for the UK economy (such as Northern Ireland, supply chain issues, and energy & climate), based on textual analysis and unsupervised machine learning methods. 

The project constructs novel measures of Brexit-related uncertainty based on textual analysis and unsupervised machine learning methods (see Chung et al., 2022). The Brexit Uncertainty index (BUI) commenced in 2013 and is consistent and comparable before, during, and after the EU referendum. Measures were constructed using 11 main UK newspapers as well as Twitter data.

These Brexit uncertainty indices (BUI) capture well UK firms’ concerns reported by the Bank of England’s Decision Maker Panel surveys. The monthly indices are cost-effective and can be updated in close to real time, which are advantageous to policy evaluation compared to traditional large firm surveys. They also allow for disentangling the Brexit effects from those of the COVID-19 pandemic. The pandemic accounted for around half of the overall uncertainty index post 2020, and this magnification effect varies significantly across policy areas.

Executive summary

  • The researchers present innovative real-time indicators of Brexit uncertainty based on UK newspaper coverage, which allow for disentangling the COVID effects.
  • The indices provide a constantly evolving picture of how uncertainty is changing and where the uncer-tainty originates, by decomposing uncertainty into different policy areas (such as Northern Ireland, supply chain issues, and energy & climate).
  • The pandemic accounted for around 50% of the overall uncertainty index post 2020, and this magnification effect varies across policy areas, with the most pronounced effect related to employment (COVID effect accounting for 77.9%), government spending (74.3%), and supply chains (69%).
  • These indices closely match UK firms’ concerns as reported by the Bank of England’s Decision Maker Panel (DMP) survey data.

 

Academics:

Wanyu Chung, Associate Professor of Economics, Univer-sity of Birmingham and Research Affiliate, CEPR
Duiyi Dai, Doctoral Student, University of Birmingham
Robert J R Elliott, Professor of Economics, University of Birmingham

Download the Executive Summary

Contact Public Affairs