Theory and Applications in Macroeconometric Modelling

The motivation behind this thesis is rooted in the critical need to enhance the understanding and application of Global Vector Autoregressive (GVAR) models in macroeconometric analysis, particularly in the context of global economic interconnectedness and regional economic differentiation within the United Kingdom. Despite the widespread adoption of GVAR models for their robustness in capturing global economic dynamics, there remains a notable research gap in their comparative effectiveness, ability to identify structural shocks, and application in regional economic analysis. This thesis is motivated by the imperative to address these gaps, thereby advancing the theoretical and practical utility of GVAR models in economic forecasting and policy formulation.

Research Questions Addressed:
Comparative Effectiveness: A significant gap in the literature is the lack of a comprehensive comparison of GVAR models with other macroeconometric frameworks. While various models are employed for forecasting and scenario analysis, their comparative effectiveness, especially in the context of GVAR's adaptability and robustness across diverse datasets, remains insufficiently explored. This thesis aims to fill this gap by providing a detailed comparative analysis, thereby positioning GVAR models within the broader macroeconometric modelling landscape.

Structural Shock Identification: The literature on GVAR models primarily utilises generalised impulse response functions (GIRFs), which, despite their practicality, fall short of distinguishing between shock types, leading to potential ambiguities in policy implications. This thesis addresses the critical need for a methodological shift towards identifying and analysing structural shocks in a manner that aligns GVAR models closer to DSGE models. By extending GVAR models to estimate structural shocks, this research contributes to refining shock analysis and enhancing the interpretability of economic dynamics.

Regional Economic Analysis within the UK: Another profound gap is the absence of GVAR applications in dissecting regional economic dynamics within the UK. Existing macroeconometric models primarily focus on national or global scales, often overlooking the nuanced economic interplays at the regional level. This oversight is particularly significant in the context of the UK, where regional economies exhibit distinct characteristics and shock responses. By developing a GVAR-based regional model for the UK, this thesis pioneers a framework that provides deeper insights into regional economic interdependencies and the differential impacts of shocks, offering valuable guidance for region-specific policy interventions.

Theories Tested:
The research critically evaluates the theoretical underpinnings of GVAR models, comparing these with other macroeconometric frameworks such as unrestricted and structural VAR models, and Factor-Augmented VAR (FAVAR) models. It tests the hypothesis that GVAR models, when extended to identify structural shocks, can offer comparable, if not superior, insights into economic dynamics relative to DSGE models. Additionally, the thesis posits that a regional GVAR model can effectively capture the economic interdependencies and differential shock impacts across UK regions.

Methodologies Employed:
A mixed-method approach is adopted, comprising quantitative econometric analysis and qualitative theoretical exploration. The study employs comparative analysis, structural shock estimation techniques, and regional economic modelling using the GVAR framework. Methodological innovations include the extension of GVAR models for structural shock identification and the creation of a novel regional economic model for the UK.

Data Description:
The thesis utilises a comprehensive dataset encompassing global economic indicators, COVID-19 pandemic-related economic data, and regional economic data within the UK. This dataset facilitates a broad analysis of GVAR models' forecasting abilities and their effectiveness in structural shock identification and regional economic analysis.

Key Findings:
GVAR models demonstrate a comparative advantage in forecasting and scenario analysis across diverse economic datasets, showcasing superior adaptability and robustness. The extension of GVAR models to estimate structural shocks enhances their analytical capabilities, aligning them closer to the insights provided by DSGE models, particularly evident in the analysis of COVID-19 pandemic data. The development of a GVAR-based regional model for the UK offers novel insights into regional economic interdependencies and the varied impacts of shocks, underscoring the importance of tailored regional policy interventions.

This thesis undertakes a comprehensive examination of Global Vector Autoregressive (GVAR) models, focusing on their theoretical underpinnings, comparative efficacy, and applicability in macroeconometric modelling, particularly in forecasting, structural shock identification, and regional economic analysis within the United Kingdom. The primary contributions include a critical comparison of GVAR models against established macroeconometric frameworks, an extension of GVAR models to facilitate structural shock identification paralleling Dynamic Stochastic General Equilibrium (DSGE) models, and the novel development of a GVAR-based regional model for the UK.

This thesis provides a comprehensive examination of the theory and application of Global
Vector Autoregressive (GVAR) models within the broader context of macroeconometric modelling. It begins by delving into various macroeconometric approaches including, unrestricted and structural VAR models, FAVAR models, and notably, the GVAR model itself. The study thoroughly explores the technicalities, empirical applications, and dynamic analysis inherent to these models, laying a particular emphasis on the GVAR approach and its comparative advantage in capturing global economic dynamics. Initiating with an analytical comparison, the study highlights the GVAR model's unique ability to capture global economic dynamics, demonstrating its enhanced adaptability and robustness in scenario analysis and forecasting across varied datasets. The research advances by extending GVAR models to estimate structural shocks, adopting a structural approach that integrates economic theory and methodological innovations, thus improving shock identification and analysis. This extension is exemplified through the analysis of economic data from the COVID-19 pandemic, illustrating the model's effectiveness in assessing the economic impacts of global shocks.

The thesis progresses by dissecting the economic theories guiding these models, focusing on structural cointegrating approaches, production technology, arbitrage, solvency, and liquidity conditions. It further investigates the GVAR model's forecasting abilities, comparing it with alternative macro models for scenario analysis and forecasting, highlighting its adaptability and robustness in handling diverse datasets.

A significant part of the research is dedicated to extending the GVAR model to estimate structural shocks, aiming to align its capabilities with Dynamic Stochastic General Equilibrium (DSGE) models. This includes a detailed examination of pre- and post-pandemic scenarios, offering insights into the model's versatility and effectiveness in capturing economic dynamics under varying conditions.

Furthermore, the thesis presents an empirical analysis of UK regions using the GVAR approach, marking a novel contribution to regional economic modelling. This model assesses the impact of various shocks across UK regions, providing a nuanced understanding of regional economic interdependencies and responses.

A significant innovation of this thesis is the creation of a GVAR-based model for the UK's regional economies, a pioneering effort that utilises the GVAR framework to explore economic dynamics and shock responses across the UK's diverse regions. This model provides a sophisticated analytical tool for policymakers and economists, enabling a more granular understanding of regional economic interdependencies and variations in shock impacts.

The findings reveal that the extended GVAR model offers a refined framework for understanding global and regional economic dynamics, outperforming comparable models in terms of forecasting accuracy and shock analysis. The regional UK model uncovers pronounced disparities in shock impacts across regions, highlighting the necessity for tailored regional policy measures.