inequality measurement

This study addresses the lack of cross-national comparability in global Gini coefficient estimates, which arises from discrepancies in data sources, welfare metrics, and methodological choices. By harmonizing 12 international databases, the authors construct a unified dataset encompassing 222 countries and over 122,000 observations. They provide the first systematic quantification of pairwise inconsistencies among alternative Gini estimates for the same country-year and rigorously assess the influence of key factors—including welfare indicators, reference units, equivalence scales, and survey design. The analysis reveals that Gini coefficients for identical country-years can differ by as much as 50 percentage points, with the choice of welfare metric identified as the primary driver of cross-country incomparability. The paper proposes a methodological framework to enhance temporal and spatial comparability, establishing a standardized foundation for measuring economic inequality.

This study investigates how socioeconomic factors influence internet performance across regions with varying population densities, thereby elucidating localized drivers of the digital divide. Integrating 170 million crowdsourced speed tests with U.S. Census Block Group–level demographic data, the authors employ hierarchical modeling based on random forest regression, permutation importance analysis, and sampling bias correction. Findings reveal that population density significantly affects network performance only at macro scales; once density is controlled for, household income and racial composition emerge as dominant predictors, with race exhibiting greater explanatory power for download speeds than either income or education. These results underscore the highly localized nature of internet inequality and highlight the urgent need for place-specific policy interventions rather than reliance on monolithic national narratives.

This study investigates whether classical inequality measures satisfy the decomposability axiom in the context of three-person income distributions and reveals the geometric manifestations of their violations. By modeling such distributions on a two-dimensional income-share simplex, the decomposition of overall inequality into within- and between-group components is recast as a geometric constraint. The paper provides the first visual characterization—within the simplest nontrivial setting—of the decomposition behavior of prominent indices, including the mean log deviation, Gini coefficient, coefficient of variation, and Theil index. It clearly identifies the distinct geometric patterns through which each measure deviates from strict decomposability, thereby deepening the understanding of their structural properties and offering an intuitive basis for selecting and comparing inequality measures.

This study addresses the inconsistency in causal effect estimates between observational studies and randomized controlled trials (RCTs) by proposing the first unified framework for decomposing causal effect heterogeneity. The framework systematically identifies and quantifies three sources of heterogeneity: differences in covariate distributions, variation in mediating pathways, and shifts in outcome-generating mechanisms. Methodologically, it formally defines effect decomposition across data types (observational vs. experimental), integrating causal inference, sensitivity analysis, and decomposition modeling, while enabling robust parameter estimation under multiple hypotheses. Evaluated through simulation studies and an empirical analysis of the “Moving to Opportunity” experiment, the framework demonstrates improved interpretability, robustness, and policy generalizability in synthesizing evidence from heterogeneous data sources.

This study addresses the challenge of comparing multidimensional socioeconomic outcomes across groups when data exhibit clustered structures and within-cluster correlations. The authors propose a distribution-free, robust cross-group comparison method that treats clusters as independent units and integrates longitudinal rank-sum tests (LRST) with order statistics to construct a rank-based multivariate aggregation framework. This approach effectively synthesizes high-dimensional, correlated indicators into an interpretable composite ranking while avoiding reliance on parametric modeling assumptions. Innovatively combining rank fusion with the inherent clustering structure of the data, the method is applied to evaluate the refundable Earned Income Tax Credit (EITC) policy, revealing systematic differences in county-level multidimensional outcomes between states that implemented the policy and those that did not. These findings remain robust across varying cluster sizes and resampling schemes.

This study addresses a critical flaw in the existing Beta Lorenz curve, whose parameter space fails to satisfy the theoretical constraints inherent to Lorenz curves, leading to systematic bias in estimating poverty and inequality from grouped income data. The authors explicitly identify this deficiency for the first time and propose a novel four-parameter family of Lorenz curves that rigorously adheres to all formal properties of genuine Lorenz curves while retaining practical usability. Through parametric modeling, derivation of necessary constraints, and extensive empirical validation across more than 2,000 datasets, the new model demonstrates superior performance in estimating poverty and inequality metrics. Specifically, it significantly reduces the systematic underestimation of poverty levels observed in over 80% of cases compared to the widely used General Quadratic (GQ) Lorenz curve.

Existing training monitoring tools struggle to simultaneously track dynamic changes across multiple metrics and diagnose fairness disparities among subgroups. This work proposes a TensorBoard plugin that, for the first time, integrates multi-metric linked visualizations with slice-level fairness analysis within a unified interactive interface, enabling real-time monitoring of both performance and fairness without modifying the training pipeline. By combining multi-view charts, user-defined subgroup slicing, standard fairness metrics, and correlation analysis across heterogeneous indicators, the approach successfully uncovers hidden demographic and environmental biases in high-performing models on the YOLOX architecture and BDD100k dataset, facilitating early detection and diagnosis of fairness issues during model training.

This study addresses how infra-marginality—differences in data distributions across groups—complicates judgments of AI fairness, as conventional statistical parity metrics often fail to align with human perceptions of fairness. Through a controlled user study involving 85 participants in a hypothetical medical decision-making scenario, the authors systematically investigate how group-specific model performance and training data availability shape fairness judgments. They find that when group-wise performance is equal or unknown, participants favor outcome equality; however, when performance disparities are attributable to data imbalance, models preserving these differences are perceived as more fair. These results demonstrate that human fairness judgments are not solely based on outcome equality but are significantly influenced by beliefs about the underlying causes of disparities, thereby challenging the prevailing assumption that statistical parity should serve as the default standard for algorithmic fairness.

This study addresses how algorithmic prioritization in public-sector resource allocation under scarcity can exacerbate relative inequalities among intersectional identity groups and erode public perceptions of institutional fairness. By integrating fairness analysis, intersectionality theory, and policy implementation modeling within real-world constraints, the research systematically uncovers the adverse effects of algorithmic prioritization mechanisms driven by an “efficiency-first” logic. Findings reveal that greater resource scarcity intensifies the inequality-amplifying effects of such algorithms and strengthens individuals’ perceptions of systemic injustice. These results challenge the prevailing narrative that equates technical efficiency with optimal allocation of public resources, highlighting the need to critically reassess algorithmic design in policy contexts where equity and legitimacy are paramount.

This study investigates whether the pronounced inequality in user interactions on online social platforms stems from structural mechanisms inherent to digital environments. By constructing cross-platform user–post bipartite networks that integrate both posting and interaction behaviors, the authors quantify interaction inequality using Kullback–Leibler divergence, the inverse coefficient of variation, and a log-transformed Gini coefficient. The analysis provides the first systematic evidence that such inequality remains stable over time across diverse platforms, scales, and governance models, indicating it is not a stochastic artifact but rather driven by deep-seated structural constraints. These findings reveal the systemic origins of how visibility and participation are allocated in online spaces, underscoring the role of platform architecture in shaping user engagement disparities.