Short-Term Transit Ridership and Revenue Forecasting

Transcription

1 Short-Term Transit Ridership and Revenue Forecasting presented to The 5 th Transportation Research Board Conference on Innovations in Travel Modeling presented by Cambridge Systematics, Inc. Feng Liu, Ph.D., Jay Evans, Xuemei Liu, Ph.D. Michael Eichler, Washington Metropolitan Area Transit Authority April 29, 2014

2 Background WMATA operates Metrorail, Metrobus, and MetroAccess 2

3 Objectives Short-term ridership forecasts bridge the gap between current ridership trends and regional travel-demand modeling Current Ridership Trends Past to now, end of FY Travel Demand Forecasting 5 to 25 years in the future Useful for multi-year budgeting and capacity planning 3

4 Previous Work Thousands 25,000 In-house forecasts developed for budgeting and fare change impacts 20,000 15,000 10,000 5,000 0 Jul 04 Jul 05 Jul 06 Jul 07 Jul 08 Jul 09 Jul 10 Jul 11 Jul 12 Jul 13 Jul 14 Observed Estimated Forecast 4

5 State of Practice WMATA s use of the combined distance-based and peak pricing for Metrorail is unique Common agency practice does not address WMATA s needs» An elasticity-based approach is limited by its ability to evaluate the impacts of the other drivers on ridership and revenue» Direct demand approach addresses the impacts of station area developments (e.g., BART)» Others have flat fare Many agencies reported they were working to develop models as sophisticated as WMATA s Opportunities for further market segmentation 5

6 Modeling Approach Econometric modeling approach» Monthly ridership» Average fare» Log-log functional form Tested many market segmentations» Metrorail (peak weekday Metrorail, off-peak weekday Metrorail, and weekend Metrorail)» Metrobus (total Metrobus, weekday Metrobus, weekend Metrobus)» Peak, off-peak, peak-of-the peak 6

7 Variables Ridership, revenue, and fare variables» By market segments Thousands $50,000 $40,000 $30,000 Socioeconomic variables» Employment, population» Gas prices» Transit subsidy benefit Tourism/business travel variables» Hotel rooms sold» Smithsonian visitations $20,000 $10,000 $0 Jan 04 May 05 Oct 06 Feb 08 Jul 09 Nov 10 Apr 12 $5.00 $4.50 $4.00 $3.50 $3.00 $2.50 $2.00 $1.50 $1.00 $0.50 $0.00 Jan 04 May 05 Oct 06 Feb 08 Jul 09 Nov 10 Apr 12 7

8 Variables Metrorail/Metrobus-related variables» Percentage of bus pass trips using SmartTrip» Metrorail or Metrobus fare collection system change indicator Weather variables» Days of extreme cold, hot, rain, snow (weekdays and weekends)» Days the Federal Government was closed due to snow Special events» Inauguration, mall event, peak bloom month, peak bloom days Seasonal and month variables 8

9 Data Model estimation periods» From FY 2005 to FY 2011 Model validation period» From July 2011 to March , , , , , , , ,000 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

10 Short-Term Ridership and Revenue Forecasting Ridership Metrorail weekday peak Metrorail weekday off-peak Metrorail weekend Metrobus Average Fare Metrorail weekday peak Metrorail weekday off-peak Metrorail weekend Metrobus Revenue Forecasts 10

11 Significant Variables Ridership Models D.C. Employment is a strong explanatory variable in all models Cost (average fare by mode and market segments) was significant» Metrorail weekday peak and off-peak, not weekend» Metrobus fare» Metrobus most elastic, Metrorail weekday peak least elastic $2.50 $2.00 $1.50 $1.00 District/regional $0.50 population NOT significant in models $0.00 Sep 02 Jan 04 May 05 Oct 06 Feb 08 Jul 09 Nov 10 Apr 12 11

12 Significant Variables Ridership Models Cost of Alternatives (gas price, lagged) is significant for the Metrobus ridership model, with a small elasticity» Not significant in explaining rail ridership changes Indicators of Tourism and Business Travel» Smithsonian Visitation explains off-peak/weekend rail ridership» Hotel rooms sold explains bus ridership as well as off-peak/weekend rail Extreme weather drives down transit ridership» Number of days of extreme weather (cold, rain, snow) had negative impacts on ridership of some sub-models 12

13 Average Fares Average rail fares are a function of distance, impacted by factors that change the balance of where people ride» Tourism/business travel» Special events» Seasonal factors Average bus fare not impacted by external factors $3.50 $3.00 $2.50 $2.00 $1.50 $1.00 $0.50 $0.00 Sep 02 Jan 04 May 05 Oct 06 Feb 08 Jul 09 Nov 10 Apr 12 13

14 Conclusion Fare elasticity values are within the range reported in the literature Gas prices (lagged) elasticity values are in the low end of the range reported in the literature Estimated ridership and revenue values were compared well with the observed values for the model validation period (from July 2011 to March 2012) Econometric modeling is valuable to address both internal and external factors driving short-term ridership and revenue, including the WMATA s unique distance-based and peak pricing structure 14