Worker well-being: antecedents, measurement, and consequences

This thesis consists of three essays in applied microeconomics and behavioural economics. While each essay is stand-alone, they share a common underlying theme, worker well-being (WWB). This thesis was motivated by the growing numbers of organisations that are investing in worker wellness programmes (WWPs) to tackle the rising tide of worker ‘un-wellness’. While the personal and economic costs associated with a dissatisfied workforce are indisputable, there is scant evidence that WWPs are effective, either at enhancing WWB or at improving economic outcomes. There is a danger that organisations may introduce ill-advised initiatives which their workers do not want, or need, and which have no documented links to the behaviour they are targeting. While there are many factors which may explain why WWPs often fail, I hypothesise that one key factor relates to organisations' lack of understanding of WWB. This thesis addresses this issue by examining the conceptualisation, measurement, stability, and potential enhancement of WWB, through an economics lens. Chapter 1 provides an overview of the thesis and its contributions. Chapter 2 details the theoretical framework. Much of the existing literature relies on pre-existing datasets which contain only a few, narrowly defined, WWB measures such as life satisfaction. Chapter 3 details how I overcome this data gap by using two novel purpose-built surveys to compile a rich, longitudinal WWB dataset. Chapter 4 focuses on one of the most pertinent consequences of poor WWB for organisations, voluntary turnover. Given evidence that pecuniary factors (e.g., salary) are poor predictors of quit intentions (QI), research is increasingly focused on non-pecuniary (e.g., meaningful work) QI antecedents. These factors are typically subsumed under WWB - a latent construct which captures the extent to which individuals report enjoying a ‘good working life’, as defined by them. Historically, economists have relied on job satisfaction to proxy for WWB. I investigate whether this approach is justified. I use novel survey data and multivariate regression to examine the relationship between WWB measures and QI. I find that a measure of worker engagement performs as well as job satisfaction. However, a composite model (job satisfaction + affect + engagement) performs best, explaining 29.4% of variation in QI versus 23.8% for job satisfaction alone. This suggests that different WWB measures may tap into distinct aspects of the WWB – QI relationship. While certain dimensions of well-being are relatively stable, there is evidence that well-being is impacted by external shocks. In Chapter 5 I investigate the impact of COVID-19-related restrictions on WWB. Building on Chapter 4, we incorporate a second wave of longitudinal data and use a fixed effects model to examine the impact of the first lockdown in the UK on full-time workers. We find that lockdown, and the large-scale transition to homeworking, were associated with unchanged or improved WWB. This result may reflect adaptation or sample composition. Organisations are increasingly using WWPs to tackle the social isolation and burnout which may be linked to protracted homeworking. The impact and cost-effectiveness of WWPs relies on their ability to attract sufficient employees. In Chapter 6, we employ an RCT design to investigate whether insights from behavioural economics can be used to ‘nudge’ employees to attend EDI themed wellness events. We examine the impact of four classic nudges (simplification, changing the messenger, social proof and defaults) on registration and attendance. We find that using a default is an effective way of increasing attendance. Taken together, my thesis makes a substantial contribution to the labour, happiness, and behavioural economics literatures. My findings also have important policy implications for organisations seeking to adopt a data-driven, evidence-based, approach to WWP design and evaluation.