Khanh Duong

The apple does not fall far from the tree. Social scientists call this the intergenerational correlation of education, one of the most durable findings in modern mobility research. In many economies, parents’ education remains a strong predictor of their children’s outcomes. This suggests life chances are still tied to the circumstances of birth. This persistence is not merely a family-level phenomenon but is structurally amplified by economic inequality — a relationship captured by the Great Gatsby Curve. At the same time, liberal democracies are philosophically committed to the meritocratic principle that an individual’s prospects should depend on their own effort and talent, not on their parents’ social position. A high intergenerational correlation, therefore, challenges equality of opportunity, signalling a rigid opportunity structure and a possible limit on the education system as an engine of mobility.

This dissertation argues that navigating this tension requires moving beyond the raw correlation, which acts as a “black box” hiding the mechanisms inside. The central research question is: How can a society committed to equality of opportunity accommodate the advantages that arise from family life? To answer this question, the dissertation does not simply re-measure the correlation but disentangles and scrutinises its constituent channels, focusing on the family’s role. It posits that the family is more than a channel for financial resources or genetic material; it is the primary site for generating Familial Relationship Goods — the intimate, constitutive, and partial practices like shared reading, shared meals, and authoritative guidance that are essential for human flourishing. The dissertation, therefore, seeks to distinguish which components of educational persistence reflect remediable social disadvantage and which arise from inevitable (even valuable) familial processes.

Pursuing this aim, the dissertation has three core objectives:
1. Empirical refinement: To estimate the transmission of Familial Relationship Goods to intergenerational educational persistence, correcting for methodological biases in traditional behavioural-genetic models.
2. Normative benchmarking: To develop and apply the “Fair Progress” benchmark — a realistic standard that treats a portion of intergenerational correlation as permissible.
3. Empirical quantification: To quantify the unavoidable inequality, or “cost,” imposed on social mobility by preserving the transmission of Familial Relationship Goods.

The argument unfolds across six chapters, each building on the last. Chapter 1 establishes the problem, situating the intergenerational correlation within the literature on the Great Gatsby Curve and introducing the concept of Familial Relationship Goods. Chapter 2 provides the indispensable theoretical foundation. It integrates behavioural-genetic variance-components models with political philosophy to decompose a child’s outcome into genetic endowment, shared family environment, and a residual proposed to capture Familial Relationship Goods. This chapter establishes a taxonomy: financial resources are widely considered problematic; genetic transmission is contested (permissible under meritocracy, arbitrary under luck egalitarianism); and non-financial investment splits into problematic social privilege and permissible Familial Relationship Goods. The chapter operationalises the transmission effect of familial relationship goods as the environmental spillover of parents’ cognitive exposure to randomness on the child’s gene–environment system, transmitted through the direct, co-resident family relationship.

Guided by this framework, Chapter 3 fulfils the first objective. It implements a novel correction method, combining twin and adoption study estimators to account for gene-environment interplay, and applies this method to the Swedish context. The key empirical finding is this: a significant and identifiable portion of the intergenerational correlation — approximately 0.15 — can be attributed to the transmission of Familial Relationship Goods. This finding quantifies the “relational work” of the family, isolating the permissible transmission channel from the mixture of natural and social advantages.

This empirical parameter becomes the foundation for the normative reconstruction in Chapter 4, achieving the second objective. The chapter shows that the conventional benchmark of “perfect mobility” (a statistical correlation of zero) is a flawed and misleading false benchmark. This benchmark is blind to the difficulty of educational transitions, indifferent to the harms of downward mobility, and — most critically — incompatible with the family, because achieving it would require abolishing Familial Relationship Goods. In its place, the dissertation offers the “Fair Progress” framework. This model accepts the approximately 0.15 parameter as a realistic baseline for a society, a baseline representing the permissible level of persistence stemming from family life. From this baseline, the framework derives an “Adjusted Mobility” metric, which reframes the central question from “Has all familial influence been eliminated?” to “Who benefits from societal growth and educational expansion?” This reframing shifts the goal from engineering statistical independence to assessing whether new opportunities are broadly shared or captured solely by the already-advantaged.

The synthesis of the empirical discoveries leads directly to the strategic and empirical analysis in Chapter 5, fulfilling the third objective. The chapter operationalises the “decoupling” strategy: a society seeking to balance mobility and family life must accept the private transmission of relational goods while using public institutions to mitigate the conversion of all forms of advantage (especially financial ones) into hardened inequalities. Applying Double/Debiased Machine Learning within this framework yields the thesis’s central quantification: preserving the transmission of Familial Relationship Goods imposes a measurable, irreducible constraint on aggregate mobility. The “price” of protecting the family sphere is a reduction in educational mobility of 0.24 to 0.27 standard deviations. This estimate provides an empirical measure of the trade-off at the heart of the equality of opportunity trilemma.

The concluding synthesis, therefore, presents three findings. First, the empirical refinement isolates and quantifies the transmission of Familial Relationship Goods at approximately 0.15. Second, the normative benchmarking uses this finding to move beyond “perfect mobility” and establish the “Fair Progress” benchmark. Third, the empirical quantification identifies decoupling as the imperative and quantifies its necessary but limited ambition, given the 0.24 to 0.27 standard deviation cost. Together, these discoveries reframe the debate: the goal is not to eliminate the gravitational pull of the family tree but to flatten the slope of inequality — the Great Gatsby Curve — so that the apple’s roll is determined more by a child’s own agency and less by the elevation of the ground on which the tree stands.

The implications of this synthesis are significant for measurement, theory, and policy. For measurement, the dissertation argues for moving beyond the false benchmark of perfect mobility to the Adjusted Mobility metric, which asks a different question about the distribution of progress. For social theory, it provides a workable compromise between Liberal Equality of Opportunity (which tends to ignore the family’s role) and Radical Egalitarianism (which would abolish it), a compromise grounded in the empirically defined baseline of permissible intergenerational correlation. For policy, it suggests shifting from pointless attempts to engineer mobility toward managing trade-offs via the decoupling strategy. This strategy involves, first, mitigating material stakes through progressive taxation, wage compression, and universal high-quality early childhood education (addressing the capacity gap); and second, broadening access to developmental goods through public mentoring and community programmes (addressing the incentive effect) — all while preserving family autonomy.

The dissertation candidly acknowledges its limitations — its domain specificity to education, its contextual roots in the Nordic context, and the persistent measurement frontiers concerning gene-environment interplay. These limitations, however, chart a clear course for future research: an expansion agenda to apply the Fair Progress benchmark to health, wealth, and well-being; a recoupling agenda to evaluate how state action can supplement relational goods without undermining the family; and an institutional agenda to compare how different welfare state models succeed at the decoupling strategy.

In its final reckoning, this dissertation argues that preserving the family as a sphere of particularistic love requires accepting a foundational limit to equality of opportunity. The goal is not a world without familial influence — an impossible and arguably undesirable goal — but a world where that influence is not destiny. The aim is not to cut down the tree, but to equalise the landscape. A society that balances these concerns is found in the perpetual, careful balancing of commitment to intimate bonds with dedication to ensuring that no child’s future is foreclosed by the circumstances of their birth. The price of intimacy is real, quantified, and unavoidable; the promise of fair progress lies not in severing the roots that nourish us, but in tending to the ground on which all children grow.