PREDICTIVE LABOR MARKET ANALYTICS FOR IDENTIFYING AND RESKILLING WORKERS DISPLACED BY AI: A SIMULATION-BASED METHODOLOGICAL FRAMEWORK AND POLICY STRESS TEST

Paper ID: EIJTEM_2023_10_4_200-218

Author's Name: Yousuf Md Shahan, Sanjida Akter, Farzana Parvin Popy, Shila Das, Sayem Sarwar, Md Fahim Uddin

Volume: 10

Issue: 4

Year: 2023

Page No: 200-218

Abstract:

The rapid diffusion of artificial intelligence (AI) across enterprise workflows has accelerated structural transformations in labor demand, prompting widespread calls for proactive, predictive workforce analytics (Acemoglu & Restrepo, 2022). However, evaluating individual-level displacement risk and longitudinal wage recovery is severely impeded by federal statutory constraints (Titles 13 and 26 of the U.S. Code), which legally prohibit merging public-use survey microdata (such as the CPS Displaced Worker Supplement) with state administrative Unemployment Insurance wage records using personal identifiers. To address this empirical impasse, this study presents a Monte Carlo computational simulation and multi-scenario policy stress test (N = 5,000 workers across B = 1,000 replications) evaluating the diagnostic viability, parameter recovery, and construct validity of predictive workforce targeting. Data/Methods: Rather than asserting an unfeasible empirical merge of public data, we construct a mathematically rigorous simulation environment calibrated to empirical distributions from the Bureau of Labor Statistics (BLS) Current Population Survey (CPS) Displaced Worker Supplement (2018–2022 waves), the Occupational Information Network (O*NET 26.0), the Felten et al. (2021) AI Occupational Exposure (AIOE) index, and state longitudinal training benchmarks (Washington State Workforce Training and Education Coordinating Board, 2022). We evaluate: (1) Monte Carlo estimator recovery and 95% confidence interval coverage; (2) multivariable logistic classification discrimination across baseline and AI-enhanced models; (3) operational precision (Positive Predictive Value [PPV]) under formal Bayesian base-rate recalibration to national displacement incidence (~3.5%); (4) construct-validity sensitivity under cognitive re-labeling confounding (Rai, ; Autor & Dorn, 2013); and (5) reskilling wage recovery across institutional pathways. Key Findings: Across 1,000 Monte Carlo replications, logistic regression recovered pre-specified AIOE parameters with minimal bias (Mean beta = 0.8420 vs. True = 0.8373, Bias = +0.0047, 94.2% empirical coverage). AI exposure elevated classification discrimination from a baseline AUC of 0.692 [95% MC interval: 0.681, 0.703] to 0.821 [0.812, 0.830]. However, our policy stress test reveals that because national annual displacement incidence is low (~3.5%), operational Positive Predictive Value (PPV) drops from 74.5% in balanced research samples to just 9.6% in the general population, generating 9.4 false alarms per true displacement. Furthermore, when AIOE is simulated under Rai's cognitive re-labeling scenario, model AUC collapses to 0.584, disproportionately misclassifying highly educated professional workers. Reskilling simulations mirror longitudinal state benchmarks, where employer-sponsored models (+$2.40/hr, +9.6%) substantially outperform community colleges (+$1.25/hr, +5.0%) and public workforce programs (+$0.30/hr, +1.2%). Implications: This simulation provides a methodological blueprint for state workforce agencies operating inside secure data enclaves (e.g., FSRDCs or state longitudinal data partnerships). Crucially, the severe base-rate penalty mathematically proves that predictive algorithms must be restricted to voluntary upskilling outreach and strictly barred from guiding adverse personnel actions. We discuss legal liabilities under Title VII and the ADEA, illustrated by the Meta AI layoff class-action litigation and EEOC regulatory guidance.

Keywords: Artificial Intelligence (AI), Worker Displacement, Predictive Labor Analytics, Monte Carlo Simulation, Policy Stress Test, Construct Validity, Base-Rate Recalibration, Reskilling Pathways.

References:

1. Abraham, K. G., & Houseman, S. N. (2019). Making work pay: Policies to improve living standards for less educated workers. RSF: The Russell Sage Foundation Journal of the Social Sciences, 5(5), 1–21. https://doi.org/10.7758/RSF.2019.5.5.01
2. Acemoglu, D., & Restrepo, P. (2020). Robots and jobs: Evidence from US labor markets. Journal of Political Economy, 128(6), 2188–2244. https://doi.org/10.1086/705716
3. Acemoglu, D., & Restrepo, P. (2022). Tasks, automation, and the rise in US wage inequality. Econometrica, 90(5), 1973–2016. https://doi.org/10.3982/ECTA19815
4. Autor, D. H., & Dorn, D. (2013). The growth of low-skill service jobs and the polarization of the US labor market. American Economic Review, 103(5), 1553–1597. https://doi.org/10.1257/aer.103.5.1553
5. Autor, D. H., Levy, F., & Murnane, R. J. (2003). The skill content of recent technological change: An empirical exploration. The Quarterly Journal of Economics, 118(4), 1279–1333. https://doi.org/10.1162/003355303322552801
6. Autor, D., Mindell, D. A., & Reynolds, E. B. (2022). The work of the future: Building better jobs in an age of intelligent machines. MIT Press. https://doi.org/10.7551/mitpress/14107.001.0001
7. Brynjolfsson, E., Mitchell, T., & Rock, D. (2018). What can machine learning do? Workforce implications. Science, 358(6370), 1530–1534. https://doi.org/10.1126/science.aap8062
8. Card, D., Kluve, J., & Weber, A. (2018). What works? A meta-analysis of recent active labor market program evaluations. Journal of the European Economic Association, 16(3), 894–931. https://doi.org/10.1093/jeea/jvx028
9. Chalfin, A., Danieli, O., Hillis, A., Jelveh, Z., Luca, M., Ludwig, J., & Mullainathan, S. (2016). Productivity and selection of human capital with machine learning. American Economic Review, 106(5), 124–127. https://doi.org/10.1257/aer.p20161029
10. Eloundou, T., Manning, S., Mishkin, P., & Rock, D. (2023). GPTs are GPTs: An early look at the labor market impact potential of large language models. arXiv preprint arXiv:2303.10130. https://doi.org/10.48550/arXiv.2303.10130
11. Equal Employment Opportunity Commission [EEOC]. (2023). Select issues: Assessing adverse impact in software, algorithms, and artificial intelligence used in employment selection procedures under Title VII. EEOC Guidance Technical Report. U.S. Equal Employment Opportunity Commission.
12. Felten, E., Raj, M., & Seamans, R. (2021). Occupational, industry, and geographic exposure to artificial intelligence: A novel dataset and its potential uses. Strategic Management Journal, 42(12), 2195–2217. https://doi.org/10.1002/smj.3286
13. Jacobson, L. S., LaLonde, R. J., & Sullivan, D. G. (2005). Estimating the returns to community college schooling for displaced workers. Journal of Econometrics, 125(1–2), 271–304. https://doi.org/10.1016/j.jeconom.2004.04.009
14. Varian, H. R. (2014). Big data: New tricks for econometrics. Journal of Economic Perspectives, 28(2), 3–28. https://doi.org/10.1257/jep.28.2.3
15. Washington State Workforce Training and Education Coordinating Board. (2022). Workforce program participant outcomes: Longitudinal net impact and cost-benefit evaluations. State of Washington Research Report.

View PDF