UTA leadership required quantitative projections of demand, revenue, and ridership impact under three fare-structure scenarios (5%, 10%, 15%). Previous fare changes had been implemented without forecasting, leading to unexpected revenue fluctuations. This report documents the data pipeline, modeling approach, and scenario outcomes that guided the phased 8% increase decision.
The pipeline extracts 18 months of ridership and fare data, computes price elasticity by route and rider segment via log-log regression, layers an ARIMAX forecast with exogenous demand drivers (gas prices, employment, weather), and simulates three fare scenarios with 95% confidence intervals. Actual post-implementation results fell within forecast confidence bands, establishing this as the template for future fare-change planning.
Note. Elasticities estimated via log-log OLS over 18-month panel; segment-specific coefficients statistically significant at p < 0.05. Confidence intervals and route-level variance documented in the full forecasting workbook.
Scenario B at a moderated +8% level captures ~$2.3M incremental annual revenue while holding volume loss below the 10% elasticity threshold. Confidence bands (±5%) gave leadership comfort to proceed. Post-implementation outcomes fell within forecast CIs, validating the methodology and establishing it as the template for all future fare-change planning.