Racing with AI: How complementarity shapes regional labour market outcomes and the future of work

Abstract

This thesis explores the role of complementarity in shaping the impact of AI on labour market outcomes, analyzing it from three levels: macro (regions), meso (industries), and micro (individuals). It is organized in five chapters: Chapter 1 presents the motivation for the study, identifies the gaps in the existing literature, outlines the research questions, and provides an overview of the thesis structure. Chapter 2 explores the complementarity between AI and regional labour markets. It finds that the entry of AI technologies is associated with an increase in employment for non-routine cognitive analytical or interpersonal occupations. However, employment in routine-intensive occupations tends to decline. At the regional level, we find that AI-intensive regions are increasingly specialized in non-routine occupations that require creativity and interpersonal skills. This might imply that AI-intensive regions are likely to be more resilient and prosper from the AI transition, while regions that specialize in routine occupations are at risk of losing out. Chapter 3 shifts to an industry-level analysis. It presents an explorative study on the relationship between the rise and fall of industries and intra-regional wage inequality. We adopt an empirical-driven approach using linked employer-employee data in NUTS-3 regions in the Netherlands during the period 2010-2019. Our study shows that related diversification in less complex industries tends to reduce wage inequality in a region. This implies that it remains a policy challenge to combine smart and inclusive growth in different regions. We find no significant relationship between exits of industries and regional inequality, with one exception: unrelated low-complex exits tend to increase intra-regional wage inequality. Overall, these findings suggest that related diversification in less complex industries tends to bring benefits in terms of inclusive growth, while unrelated exits in less complex industries tend to do the opposite. Chapter 4 moves to a more granular level of analysis, focusing on AI-skill complementarity at the individual level. It explores which skills are complementary to AI technologies and how this complementarity will shape workers' return to wages and employment when adopting AI technologies. I introduced a novel concept called "complexity intelligence," which refers to the diverse and complementary skill sets that a worker possesses. I argue that complexity intelligence is a complementary skill to AI as it allows workers to develop new skills faster, strategically allocate tasks with AI and excel in complex tasks where AI capabilities fall short. I assess the hypothesis using O*NET and American Community Survey data in 2005-2021. This empirical setting leverages a novel data set measuring AI adoption at the occupational level with BERT (Bidirectional Encoder Representations from Transformers) language models. Most notably, the findings show that complex occupations are more likely to adopt AI technologies. In addition, there is a positive correlation between AI adoption and employment growth. AI adoption is associated with an increase in wage growth on average, with a larger increase for complex occupations. Chapter 5 summarizes the main conclusions and discusses policy implications. It also reflects on the key limitations and proposed directions for future research.