Leading indicators of regional resilience

Resilience is a central issue to policy debates on sub-national economic inequality and ‘levelling up’. In thinking about economic resilience, measuring and analysing a wider range of metrics is essential. However, a fuller understanding of regional economic resilience requires consideration of other dimensions of resilience, and associated measures that can capture these.

In seeking to better understand regional economic resilience in the UK, two main types of issues emerge.

First is the question of how we think about the concept of regional economic resilience, along with the coverage, availability and timeliness of the regional economic data that can be used to understand this concept.

Second, is the question of how this data can be used to understand the resilience of different parts of the country to a range of economic shocks. This requires new methods development.

This project covers two key components:

1. Exploring the dimensions of regional economic resilience that are important and scoping out how these can be analysed.
2. Developing a methodology that enables us to measure dimensions of regional economic resilience in a new way and to understand the effect of economic shocks on sub-national economic resilience.

Each component of the project will need very different methodologies:

Research question 1 – What are the key components of regional economic resilience, and how can we measure them? 

We will conduct desk-based research combined with a series of workshop-based consultations with stakeholders. These activities will give us an understanding of the key elements of regional economic resilience that are important, both in terms of the existing evidence base but also to users. Having defined these dimensions, we would assess the availability and suitability of data on these outcomes. We will work with ONS colleagues to identify relevant stakeholders – and to ensure that there is coordination with existing work in this area in ONS.

Research question 2 – Can we develop methods that enable us to rigorously explore how UK-wide economic shocks differentially affect sub-national resilience? 

The project also aims to address the challenge of analysing regional economic resilience using econometric methods. We will explore how different types of econometric model be applied. These will be time series models, using mixed-frequency data (typically annual and quarterly data). We have a particular focus on combining aggregate measures such as GVA with distributional data such as on the income distribution, thus combining macroeconomic and microeconomic data into a single model. Developing this class of model will require a series of method developments.

There are two models that we are developing, each makes use of these mixed frequency micro and macro variables in a different way. The first is known as a functional VAR – which are a recent development in the academic literature. The other is known as a pseudo VAR – which is a new type of model developed by the research team. Both models reflect that we are working at a sub-national level, and also consideration of how best to exploit the UK data in the model, whilst ensuring that our regional estimates are consistent with the aggregate UK estimates and data.

This project has examined how regional economic resilience has been previously conceptualised and measured, focusing on both the underlying concepts and the methods used to assess them. Through a series of deliverables, we have documented this landscape and identified ways to operationalise resilience using our existing and ongoing research. A key element of our approach is the use of causal analysis to examine the impact of economic shocks on regional economic aggregates—principally GDP—moving beyond the static measurement approaches that dominate much of the current literature and practice. This builds on our earlier work modelling regional GDP.

In parallel, we have developed a mixed-frequency functional VAR (fVAR) model. Applied to the UK, this model illustrates an alternative way to assess resilience to economic shocks. A distinguishing feature of the fVAR model is its ability to capture distributional economic outcomes. It enables analysis of the effects of different types of shocks at various points in the outcome distribution. We argue that this provides a more nuanced and informative understanding of economic resilience than existing approaches, which typically focus only on aggregate outcomes. We have also set out how we could take this approach forward at a sub-national level.

The focus of this project so far has been on understanding and extending the evidence base on measuring regional economic resilience and developing proof of concept tools to do so. In future work we would seek to develop applications in support of the work of key policy stakeholders. Consistent with our previous work, we are keen to support ONS and other colleagues in adopting and using the models and methods developed. As part of this project we held two workshops with policy stakeholders, and from this we gathered important insights that could inform future applications of these approaches.

De Polis, A., Koop, G., McIntyre, S. and Mitchell, J. ‘Mixed Frequency Functional VARs for Nowcasting the income Distribution in the UK’ Contributed session I: National Accounts/prices. ESCoE Conference on Economic Measurement, King’s College London, 21-23 May 2025

De Polis, A., Koop, G., McIntyre, S. and Mitchell, J. ‘Measuring Regional Resilience to Economic Shocks in the UK’ Contributed session C: Labour markets/households/inequality. ESCoE Conference on Economic Measurement, King’s College London, 21-23 May 2025

De Polis, A. ‘Mixed Frequency Functional VARs for Nowcasting the Income Distribution in the UK’ Workshop on Macroeconomic Analysis and Forecasting for Policy and Practice, Department of Economics, Strathclyde Business School, Glasgow, 28-29 November 2024

De Polis, A. Koop, G., McIntyre, S. and Mitchell, J. ‘Mixed Frequency Functional VARs for Nowcasting the Income Distribution in the UK’ ESCoE Early Career Researcher and PhD Workshop, King’s College London, 8 November 2024

Koop, G., McIntyre, S., Mitchell, J., Poon, A. and Wu, P. ‘Measuring Sub-Regional Economic Activity: Missing Frequencies and Missing Data’ ESCoE Discussion Paper, ESCoE DP 2024-11, 30 September 2024

Koop, G. ‘Incorporating Micro Data into Macro Models using Pseudo Vars’ Keynote. Seminar Presentation University of Salzburg. 11 June 2024

De Polis, A., Koop, G., McIntyre, S., Mitchell, J. and Wu, P. ‘Mixed Frequency Functional VARs for Nowcasting and Structural Analysis of the Income Distribution in the UK’, Seminar, University of Verona, 6 June 2024

Koop, G., McIntyre, S., Mitchell, J. and Wu, P. ‘Incorporating Micro Data into Macro Models using Pseudo VARs’ Parallel session 3: The 14th RCEA Bayesian Econometrics Workshop, Brunel University London, 20-22 May 2024

Koop, G., McIntyre, S., Mitchell, J. and Wu, P. ‘Incorporating Micro Data into Macro Models using Pseudo VARs’ Contributed session I: Measuring labour markets. ESCoE Conference on Economic Measurement, Alliance Business School, University of Manchester 15-17 May 2024

Koop, G., McIntyre, S., Mitchell, J. Poon, A. and Wu, P. ‘Measuring Sub-Regional Economic Activity: Missing Frequencies and Missing Data’ Contributed session N: Subnational statistics. ESCoE Conference on Economic Measurement, Alliance Business School, University of Manchester 15-17 May 2024

De Polis, A., Koop, G., McIntyre, S., Mitchell, J. and Wu, P. ‘Mixed Frequency Functional VARs for Nowcasting and Structural Analysis of the Income Distribution in the UK’ Contributed session C: Inequality and life satisfaction. ESCoE Conference on Economic Measurement, Alliance Business School, University of Manchester 15-17 May 2024

Koop, G. ‘Incorporating Micro Data into Macro Models using Pseudo Vars’ Keynote. EconDat Conference 2024, King’s College London, 9-10 May 2024

Koop, G., McIntyre, S., Mitchell, J., Poon, A. and Wu, P. (2023) ‘Incorporating short data into large mixed-frequency vector autoregressions for regional nowcasting’ Journal of the Royal Statistical Society Series A: Statistics in Society, https://doi.org/10.1093/jrsssa/qnad130