Intergenerational Mobility and Lifecycle Profile - Methodological and Empirical Insights

This thesis examines intergenerational mobility with a particular focus on the role of lifecycle bias in estimating the transmission of socioeconomic outcomes across generations. It consists of three self-contained empirical studies unified by a common methodological concern: the bias introduced when shot-run outcome proxies are used instead of lifetime measures. The first study investigates the intergenerational transmission of criminal behaviour and applies a method for lifecycle bias based on the generalized errors-in-variables (GEiV) model. Because short-run proxies vary systematically over the life course, estimates of intergenerational associations are often biased when lifetime data are unavailable. The GEiV model addresses this issue and performs well across both the extensive and intensive margins of criminal offending. The analysis reveals that traditional estimates substantially understate intergenerational associations in criminal behaviour, even when both generations are observed during their teenage years. The corrected elasticities are larger, statistically similar across ages, cohorts, and offence types, and indicate stronger persistence of criminal behaviour across generations. Associations between mothers and children tend to be weaker than those between fathers and children. The second study extends the literature on income mobility by applying the GEiV correction strategy to earnings data from New Zealand. The analysis shows that intergenerational earnings rank correlations and elasticities are significantly understated when relying on short-run proxies. After applying the GEiV correction, estimates become more stable and consistent across cohorts and age groups, revealing stronger associations—particularly among Māori and daughters. The final paper applies the GEiV model correction strategy to the United States using data from the Panel Study of Income Dynamics (PSID). By comparing intergenerational mobility estimates based on lifetime income and short-run proxies, the study shows that lifecycle bias leads to downward-biased estimates of both rank correlations and elasticities. The GEiV-based correction produces estimates that closely approximate “true” intergenerational mobility, even in relatively short income panels. Collectively, these studies underscore the importance of accounting for lifecycle bias and demonstrate the effectiveness of the GEiV model in improving the accuracy and comparability of intergenerational mobility estimates across different contexts and outcomes.