Thursday, Aug 6: 8:30 AM - 10:20 AM
6042
Contributed Papers
Thomas M. Menino Convention & Exhibition Center
Room: CC-212
Main Sponsor
Business and Economic Statistics Section
Presentations
This document introduces a novel business-cycle turning-point analysis method that leverages the nonparametric coincident profile tool to construct confidence intervals for turning-point dates. Through a numerical study and two empirical applications: one using economic data from the United States and the other from Colombia, we demonstrate the accuracy of the method in identifying turning points, closely aligning with the reference cycle in each case. In addition, in our analysis of United States economic data, we conduct a pseudo-out-of-sample analysis that further validates the method's superior performance in predicting turning-point dates.
Keywords
Business cycles
Turning points
Non-parametric test
Coincident Profile
Confidence intervals
Longitudinal studies on occupational outcomes are often disadvantaged by a lack of reliable data. Examination of failure‑time distributions can help formulate hypotheses on underlying mechanisms. Data from administrative records of United Kingdom (UK) registered physicians who qualified between 1968 and 2019 was linked with data on judicial decisions (Medical Practitioners Tribunal Service, MPTS) on charges of malpractice. An MPTS case was treated as a 'failure', facilitating the use of the methods of reliability theory. The merged dataset had 454 men (81.65%) and 102 women 18.35% (n= 556). Men were about 4.5 times as likely as women to face MPTS. The time to MPTS showed a strong right skew and lognormal and Weibull fits showed the latter fits to be superior. Males had a higher Weibull shape parameter (p=5.41) than females (p=1.75), suggesting differences in offending between the sexes. The superior fit of the Weibull, particularly with a high shape parameter for males, suggests a weakest‑link or early‑vulnerability model of malpractice risk. This is the first study to empirically examine malpractice risk in physicians using UK administrative data. These findings are of importance in planning for risk mitigation.
Keywords
System reliability
Computer graphics
Economic statistics
Speaker
Arun Chind, Proshen Health & Risk Consulting Ltd.
Income inequality has long challenged economists, gaining renewed attention after the Great Recession of 2008. Empirical evidence shows that market income inequality, measured by the Gini coefficient, has risen in industrial economies since the mid‑1970s. Prior research offers mixed conclusions on whether inequality promotes, hinders, or weakly relates to economic growth. Using data adapted from Brueckner and Lederman (2018), this study examines the relationship between inequality and GDP per capita growth in ECOWAS countries. Key variables include the Gini coefficient from the SWIID, along with GDP per capita, investment, government consumption, and educational attainment from established international datasets. We estimate models in which growth effects depend on initial income levels. To capture the dynamic evolution of inequality and growth, we integrate Functional Data Analysis (FDA), which models entire trajectories rather than discrete observations. This approach provides deeper insight into how long‑term inequality patterns shape economic performance across countries.
Keywords
ECOWAS
Functional Data Analysis
Panel Data
Inequality
Gini Coefficient
We develop a principal–agent model to analyze the renting-versus-owning decision, where maintenance effort is costly and unobservable under renting but internalized under ownership. Effort reduces depreciation, increasing long run property quality and neighborhood stability. The model yields closed-form solutions for optimal effort and ownership thresholds:
e_r=β/c_e,
e_o=(β+αP)/c_e
The model also provides conditions under which households prefer owning based on property value and financing terms. To complement the theory, we implement simulation based sensitivity analysis using R, incorporating uncertainty in behavioral and economic parameters. Visual analytics (such as ΔU(P) curves, rate-threshold heatmaps, and affordability trajectories) quantify the joint effects of incentives, prices, and product design (e.g., FRM vs ARM).
This work illustrates how statistical modeling and simulation can inform housing policy, mortgage product design, and credit risk analytics by linking behavioral incentives to measurable outcomes. Planned empirical validation integrates loan-level data and predictive modeling. By combining optimization, simulation, and statistical visualization, this
Keywords
Principal–Agent
Housing
Mortgage
Credit Risk
Finance
Simulated Sensitivity Analysis
A widely used index of market concentration is the Herfindahl-Hirschman index (HHI) for firms in an industry. Defined as the sum of the proportion of scale (i.e. outputs, inputs) in the industry for firm. As the value of HHI increases, the level of concentration increases.
This index has spawned a large and growing number of investigations in the literature. In addition to comparisons of industrial concentration, it has been used to measure the competitive balance of sporting leagues, the performance of university departments, the variations in language translations, among other applications.
To make comparisons of the value of HHI between groups (industries, sporting leagues, locations), it is necessary to construct HHIs that are based on a comparable number of individuals. Unfortunately, the direct comparisons of HHI for one group will be contaminated by differences in the sample size of each group. In this paper we examine a number of alternative methods for weighting the HHI to account for differences in sample size. Then we survey the recent applications where these transformations may influence the inferences drawn in these contributions.
Keywords
Herfindahl Hirschman Index
Sample moments
Measure of sporting league competition
Measure of industrial concentration
Neural language model-based text embeddings are increasingly used in labor economics to measure occupational similarity and skill distance, yet standard practice treats them as fixed-point estimates. We show that downstream analyses, such as wage regressions or automation risk scores, can be sensitive to embedding uncertainty, leading to overconfident and biased inference.
We present an uncertainty-aware framework. First, we construct embedding-based similarity measures from BERUFENET, a database of expert-written German occupational descriptions, enabling finer comparisons than standard taxonomies such as ISCO or the German KldB 2010. Second, we quantify embedding uncertainty via Monte Carlo dropout and ensemble methods, and propagate it through downstream estimators via simulation to obtain valid confidence intervals.
We demonstrate our approach to occupational mobility analysis, showing that ignoring embedding uncertainty can substantially underestimate standard errors. Our methods apply wherever embeddings serve as inputs to statistical models.
Keywords
Uncertainty Quantification
LLM
Econometrics
Labor Economics
Text as Data