Investigating the Relationship between Built Environment and Subjective Wellbeing

Abstract

Identifying causal impacts of the built environment (BE) and the associated locational and lifestyle choices on subjective wellbeing (SWB) has been a burgeoning research field since the COVID-19 pandemic. However, key gaps remain in the literature. Specifically, i) quantifying SWB impacts based on disparate environmental attributes ignores the human experience of BE as an integrated system, and risks significant confounding effects; ii) categorising people based solely on the location of their employment, without considering the differences in lifestyle, overlooks the growing flexibility and heterogeneity of modern working; iii) SWB is multi- dimensional and the measurement requires clear conceptualisation and robust instrument targeted at a specified research question or policy goal. This thesis aims to bridge these gaps by introducing a new analytical framework that i) quantifies the causal effects of relocation and how changing holistic BE type impacts long-term SWB; and ii) identifies latent lifestyles based on intra-day time use patterns and how flexible working impacts moment-to-moment SWB. Combining the two perspectives allows this thesis to delineate the complex pathways linking SWB to where people choose to live and how they choose to work within the BE. From a causal inference perspective, this thesis applies econometric and machine learning methods to model the relationship between residential relocation, BE change, and SWB. A decade of UK national-level panel data on residential relocations emulate largescale quasi- experiments of individuals moving within and across holistic BE types, as measured by Census Area Classifications. Ensemble results from staggered difference-in-differences and generalised synthetic control methods show that residential relocation has immediate and enduring positive causal effects on SWB, with changing BE as a key causal factor. Relocating increases SWB by 8-15%; without a change in BE type the effects become statistically insignificant. Further analysis uses causal random forest to quantify heterogeneous effects across worker subgroups, such as long commuters who tend to benefit more from changing BE type. This thesis offers fresh and robust insights for planning policies pertaining to residential mobility and the post-pandemic future of work. From a time use perspective, this thesis identifies ten latent lifestyles across three distinct working arrangements including homeworking, commuting, and hybrid working. Leveraging population-representative time use survey data from the UK (2015, 2016 & 2020), the rapid adaption of flexible working during the COVID-19 pandemic represents a paradigm shift in where and how people work. The changing relationship between time use and SWB is interpreted qualitatively using tempogram visualisations and analysed quantitatively using logistical regression models. While SWB generally decreased during the pandemic, significant heterogeneity is identified between working arrangements within the 6am-6pm usual working hour window, during which hybrid lifestyles managed to improve SWB. A spatio-temporal framework of flexible working is proposed to formalise the observed heterogeneity. The conceptualisation of flexible working as lifestyle transitions provides a new perspective for examining workplace management practice. Together, the causal inference and time use perspectives reconcile the slow process of BE changes with the fast process of lifestyle changes by measuring complementary dimensions of SWB – as evaluated over time and as experienced in the moment. Important policy implications can be drawn from the empirical studies. Future planning policies should i) facilitate residential mobility, reducing frictional costs of relocation and develop diverse housing options across BE types; ii) support lifestyle transitions by understanding the nuanced spatio-temporal flexibility demands of workers on inter- and intra-day bases; and iii) incorporate both evaluative and experiential SWB as key outcome measures for human-centric planning policies.