Journal of Medicine Surgery and Public Health

Abstract Criminal networks arise from the attempt to balance a need of establishing frequent ties among affiliates to facilitate coordination of illegal activities, with the necessity to sparsify the overall connectivity architecture to hide from law enforcement. This efficiency-security trade-off is also combined with the creation of groups of redundant criminals that exhibit similar connectivity patterns, thus guaranteeing resilient network architectures. State-of-the-art models for such data are not designed to infer these unique structures. In contrast to such solutions, we develop a tractable Bayesian zero-inflated Poisson stochastic block model (ZIP–SBM), which identifies groups of redundant criminals having similar connectivity patterns, and infers both overt and covert block interactions within and across these groups. This is accomplished by modelling the weighted ties (corresponding to counts of interactions among pairs of criminals) via zero-inflated Poisson distributions with block-specific parameters that quantify complex patterns in the excess of zero ties in each block (security) relative to the distribution of the observed weighted ties within that block (efficiency). The performance of ZIP–SBM is illustrated in simulations and in a study of summit co-attendances in a complex Mafia organization, where we unveil efficiency-security structures adopted by the criminal organization that were hidden to previous analyses.

Abstract The interplay between maritime commerce and macroeconomic aggregates such as gross domestic product (GDP) and international trade statistics is key for the understanding of global economic dynamics. In this paper, we construct a new indicator, the Port Saturation Index, based on near-to-real-time port satellite images, which are processed with deep learning techniques to extract relevant information on the size of ongoing maritime commerce. We investigate the efficacy of the index with an empirical application to nowcast economic aggregates and forecast expansions and contractions of European GDP, import, and export. Results show that the index is able to closely track real economic activity, particularly GDP, when the figure is still far from its official release, yielding a novel tool for policy makers to support decision-making.

Abstract When banks fail amidst financial crises, the public criticizes regulators for bailing out or liquidating specific banks, especially the ones that gain attention due to their size or dominance. A comprehensive assessment of regulators, however, requires examining all their decisions, and not just specific ones, against the regulator’s dual objective of preserving financial stability while discouraging moral hazard. In this article, we develop a Bayesian latent class estimation framework to assess regulators on these competing objectives and evaluate their decisions against resolution rules recommended by theoretical studies of bank behaviour designed to contain moral hazard incentives. The proposed estimation framework addresses the unobserved heterogeneity underlying regulator’s decisions in resolving failed banks and provides a disciplined statistical approach for inferring if they acted in the public interest. Our results reveal that during the crises of 1980s, the US banking regulator’s resolution decisions were consistent with recommended decision rules, while the US savings and loans (S&L) regulator, which ultimately faced insolvency in 1989 at a cost of $132 billion to the taxpayer, had deviated from such recommendations. Timely interventions based on this evaluation could have redressed the S&L regulator’s decision structure and prevented losses to taxpayers.

Abstract Identifying distinct subgroups within correlated data is essential for tailoring policies to specific needs, ensuring optimal outcomes for each group. In the context of model-based clustering, we introduce an innovative approach for clustering grouped data using linear mixed models with discrete random effects and exponential family responses (e.g. Poisson or Bernoulli). Our method uncovers the latent clustering structure, net of fixed effects, by assuming that random effects follow a discrete distribution with an a priori unknown number of support points. We refine this process within a modified Expectation–Maximization algorithm, collapsing support points of the discrete distribution with overlapping estimated confidence intervals or regions, derived from the asymptotic properties of maximum likelihood estimators. This approach offers a transparent interpretation of the latent structure, distinct from existing tools for discrete random effects, which often rely on discretionary tuning parameters or predetermined cluster counts. Through simulation studies, we compare our approach with traditional parametric methods and state-of-the-art techniques, demonstrating its effectiveness. We apply our model on real-world data from the Programme for International Student Assessment, aiming to classify countries based on their impact on low-achieving student rates in schools. Our methodology provides valuable insights for effective policy formulation.

Abstract Using a novel, flexible, and easily interpretable dynamic Bayesian state space model, we analyse historic and future longevity trends across 18 high income countries over the last 100 years and 16 large population emerging markets from 1950. Our results show the key driver of global life expectancy is now late-life mortality whose importance is projected to increase further. We find no sign of any impending limit to average life expectancy but project a slowdown in future life expectancy gains despite continuing improvement in later-life mortality. Gains to later-life mortality are increasingly driven by the modal age of death with a slowdown in improvements in the speed of ageing and compressions of mortality. The consequence is a projection increase in the upper bound of age at death and a slowdown in lifespan equality improvements. Whereas the 20th century saw widespread cross-country convergence in longevity indicators the projections are for divergence both within high-income countries as well as large population emerging markets. A particular outlier is the U.S. where our model predicts substantial increases in the modal and upper bound for observed age at death but only small improvements in life expectancy and so an increase in lifespan inequality.

Abstract To understand patterns of social inequality, social science research has typically relied on statistical models linking the conditional mean of an outcome variable to a set of explanatory factors. A prime example of this approach is the Kitagawa-Oaxaca-Blinder (KOB) method. By fitting two linear models separately for an advantaged group and a disadvantaged group, the KOB method decomposes the between-group outcome disparity into two parts: a part explained by group differences in background characteristics, and an unexplained part often dubbed ‘residual inequality’. In this article, we explicate, contrast, and extend two distinct approaches to studying group disparities, which we term the descriptive approach, as epitomized by the KOB method and its variants, and the prescriptive approach, which focuses on how a disparity of interest would change under a hypothetical intervention to one or more manipulable treatments. For the descriptive approach, we propose a generalized nonparametric KOB decomposition that considers multiple explanatory variables sequentially. For the prescriptive approach, we introduce a variety of stylized interventions, such as lottery-type and affirmative-action-type interventions that close between-group gaps in treatment. We illustrate the two approaches to disparity analysis through an application to the Black-White gap in college completion rates in the U.S.

Abstract R. A. Fisher was a devoted Darwinian, and, like Darwin, created science out of nothing. The list is long, but one thinks of likelihood-based estimation, analysis of variance, principles of experimental design, and randomization as standing the tests of time. Such accomplishments ‘from scratch’ (or nearly so) can amaze the fine statisticians who made meaningful incremental contributions to work begun by others, the few ‘greats’ among us who invented something important, and the unusually perceptive introductory statistics student, alike. Fisher thought of randomization in the context of agricultural experiments, but it has impacted most profoundly the science of medicine. Bradford Hill brought randomization to clinical trials. The concept of randomization-based inference, now resurrected in causal inference, was largely forgotten as design and analysis became segregated, perhaps due to analysis software packages. This talk will attempt to give the historical context of randomization and randomization-based inference from Fisher to the present day, including newer concepts such as response-adaptive, covariate-adaptive, and covariate-adjusted response-adaptive randomization. It will be challenging to condense a year of material into one hour, but a devoted Fisherian should be able to be efficient and sufficient.

Abstract We propose an adaption of the multiple imputation random lasso procedure tailored to longitudinal data with unobserved fixed effects which provides robust variable selection in the presence of complex missingness, high-dimensionality, and multicollinearity. We apply it to identify social and financial success factors of microfinance institutions (MFIs) in a data-driven way from a comprehensive, balanced, and global panel with 136 characteristics for 213 MFIs over a 6-year period. We discover the importance of staff structure for MFI success and find that profitability is the most important determinant of financial success. Our results indicate that financial sustainability and breadth of outreach can be increased simultaneously while the relationship with depth of outreach is more mixed.