Paper 96

Time Series Analysis and Forecasting of U.S. Teenage Female Unemployment

September 24, 2025: 10:30 AM - 11:00 AM

Careers, Training, and Education (CTE), Breakout Room #7

Authors	Abstract
Sekyiwaah Nuamah	This study analyzes monthly unemployment patterns for U.S. females aged 16-19 from January 1950 to April 2025 using SAS time series procedures. The research addresses critical socioeconomic needs to understand youth unemployment dynamics, particularly among teenage females who face unique labor market challenges including limited work experience, seasonal employment patterns, and systemic inequalities. This demographic historically experiences unemployment rates exceeding 15%, with heightened vulnerability during economic downturns like the 2008 financial crisis and COVID-19 pandemic. Using PROC ARIMA, we implemented a comprehensive time series analysis workflow including stationarity testing via Augmented Dickey-Fuller tests, model identification through ACF/PACF analysis, and SARIMA model fitting. After evaluating multiple candidate models, SARIMA(0,1,1)(0,1,1)[12] emerged as the optimal specification based on AIC criteria and diagnostic checks. The model successfully captured both seasonal patterns and short-term dynamics, with residual analysis confirming model adequacy. The study generated 12-month forecasts with 95% confidence intervals, revealing expected seasonal fluctuations consistent with school-year employment patterns. These findings provide valuable insights for policymakers developing youth employment programs and workforce development strategies targeting vulnerable demographic groups.

Sekyiwaah Nuamah

This study analyzes monthly unemployment patterns for U.S. females aged 16-19 from January 1950 to April 2025 using SAS time series procedures. The research addresses critical socioeconomic needs to understand youth unemployment dynamics, particularly among teenage females who face unique labor market challenges including limited work experience, seasonal employment patterns, and systemic inequalities. This demographic historically experiences unemployment rates exceeding 15%, with heightened vulnerability during economic downturns like the 2008 financial crisis and COVID-19 pandemic. Using PROC ARIMA, we implemented a comprehensive time series analysis workflow including stationarity testing via Augmented Dickey-Fuller tests, model identification through ACF/PACF analysis, and SARIMA model fitting. After evaluating multiple candidate models, SARIMA(0,1,1)(0,1,1)[12] emerged as the optimal specification based on AIC criteria and diagnostic checks. The model successfully captured both seasonal patterns and short-term dynamics, with residual analysis confirming model adequacy. The study generated 12-month forecasts with 95% confidence intervals, revealing expected seasonal fluctuations consistent with school-year employment patterns. These findings provide valuable insights for policymakers developing youth employment programs and workforce development strategies targeting vulnerable demographic groups.