Worker Heterogeneity and Labour Market Dynamics: Insights from Machine Learning and Macroeconomic Modelling

Abstract

The thesis studies the labour market experience of different groups and the role of heterogeneity in shaping labour market outcomes during economic shocks, including the COVID-19pandemic. It consists of four chapters, and combines machine learning techniques and macroeconomic modelling to analyse heterogeneity in the UK labour market and simulate the impacts of shocks and policy interventions. Chapter 1 presents an introduction to the subsequent chapters.

Chapter 2 employs machine learning techniques, specifically a cluster analysis, to identify key dimensions of labour market heterogeneity, investigating differences in individual and job characteristics across workers. Using data from the Annual Population Survey, the cluster analysis highlights productivity, proxied by education and occupational skill level, as the primary factor.

Chapter 3 develops a search and matching model with frictions and heterogeneity across workers and jobs. The model is designed based on the findings of the previous chapter to distinguish workers by education (graduates and non-graduates) and jobs by skill level (low-, medium-, and high-skilled occupations). I analyse the different labour market experiences of graduates and non-graduate using a labour productivity shock to reflect fluctuations in business cycles.

Chapter 4 integrates the search and matching model into a New Keynesian DSGE framework to simulate the impacts of the COVID-19 pandemic on the labour market and evaluate the effectiveness of policy responses. The results suggest that economic shocks disproportionately affect non graduates, exacerbating existing disparities, while policy interventions such as wage subsidies mitigate some adverse effects on the labour market.

By incorporating labour market heterogeneity into macroeconomic models, this research provides insights into labour market segmentation, the propagation of economic shocks, and the role of policy in limiting job loss and supporting wage growth. The findings contribute to the broader understanding of labour market resilience and the design of targeted interventions during periods of economic disruption.

Date of Award	10 Sept 2025
Original language	English
Awarding Institution	University of Bath
Sponsors	Economic and Social Research Council
Supervisor	Christopher Martin (Supervisor) & Matt Dickson (Supervisor)