An Explorative Econometric Model of Sydney Metropolitan Rail Patronage. Safe, clean and reliable passenger services

Transcription

1 An Explorative Econometric Model of Sydney Metropolitan Rail Patronage Dr Neil Douglas Managing Director Douglas Economics George Karpouzis Chief Economist Rail Corporation (RailCorp) NSW Safe, clean and reliable passenger services

2 With a little help from our colleagues: The authors would like to thank: CHARLIE LIN, ATIQUR RAHMAN and KERRIE TANDY of RailCorp And also the anonymous ATRF Reviewers title Page 2

3 Overview Introduction Literature review Model Patronage trip rate Office employment Gross state product (GSP) per head Rail fare Train kilometres operated title Page 3

4 Overview Other explanatory variables Omitted variables Correlation between variables Estimated model Impact of the global recession Conclusions title Page 4

5 Introduction The aim of the research paper: - to provide a short to medium term forecasting model of rail patronage. The model: - an explorative time-series econometric model; - demand for Sydney metropolitan rail; - annual data 1969 to 2008; - regression analysis year on year change in rail demand; - dependent variable trips per head using factored population; - explanatory variables variation in rail trip rate; and - office employment and gross product per head important determinants of rail patronage. title Page 5

6 Literature Review econometric models of rail demand Time series models: - intercity models based on commuter rail services in the UK; - cross sectional or gravity type models; and - hybrid time series - cross section approach. Elasticities: - price; - service level; and - service quality. Mode share models: - Sydney travel model Revealed and stated preference title Page 6

7 Comparison with other studies Sydney metropolitan rail patronage: - used revenue accounts data; - a time series model was fitted to the data; and - data periodicity annual from 1969 to TfL underground and bus travel in London: - a time series model was fitted to the data; - data periodicity four weekly data; and - larger number of observations. title Page 7

8 Comparison with other studies Charles River Associates (CRA): - a time series model was fitted to the data; and - shorter time series; and - absolute level of rail patronage. title Page 8

9 Model Variation in the year on year change in Sydney metropolitan rail patronage Variables: - socio-economic; - fare; - service level; and - event. 38 years of accounting data from title Page 9

10 Model Dependent variable: - rail patronage divided by factored Sydney metropolitan population; - year on year change in rail trip rate. specified in ratios to reduce auto correlation. Then Logged title Page 10

11 Model title Page 11

12 Model = Qt = number of rail trips (millions) Pop t = population of Metropolitan Sydney in millions RF = RF t RF t 1 TKMS = ln TKMS TKMS t t 1 title Page 12

13 Model MOE = ln MOE MOE t t 1 RGSPP = ln RGSP RGSP t t 1 MI t = major incident on CityRail taking a value of 1 if an incident occurred else zero. OLY t = Sydney Olympics taking a value of 1 if financial year = else zero. title Page 13

14 C i Model AFC t = Automatic Fare Collection taking a value of 1 if financial year is or later. C = constant indicating the average year on year ratio change in patronage level. i = parameters to be estimated with tdenoting year. title Page 14

15 Patronage trip rate Trend in CityRail Patronage title Page 15

16 Patronage trip rate Trend in Sydney Population title Page 16

17 Patronage trip rate Trend in CityRail Trip Rate title Page 17

18 Office employment Trend in Office Employment title Page 18

19 Gross state product per head Ratio Change in Real Gross State Product per Head title Page 19

20 Average Revenue $ Per Trip Rail fare Trend in Average Revenue 2.50 Average Revenue per Trip Nominal Av. Revenue Real Av. Revenue 0.00 title Page 20

21 Train kilometres operated Trend in Train Kilometres title Page 21

22 Other Explanatory variables Sydney Olympics Major Incidents (4) Automatic Fare Collection title Page 22

23 Omitted variables Petrol Price Currie Car Parking $$$ in CBD Bus Fares Rail Service Level Tourism Retail Activity Accessibility to the Rail network title Page 23

24 Correlation between variables Correlation Matrix title Page 24

25 Estimated model title Page 25

26 Patronage and fare elasticity title Page 26

27 Jul-91 Jul-92 Jul-93 Jul-94 Jul-95 Jul-96 Jul-97 Jul-98 Jul-99 Jul-00 Jul-01 Jul-02 Jul-03 Jul-04 Jul-05 Jul-06 Jul-07 Jul-08 Jul Months Passenger Journeys Millions Thousands Estimated CBD White Collar Employment Impact of the Global Recession Estimated CBD White Collar Employment vs Total CityRail Rolling 12 Month Passenger Journeys Month Journeys Estimated CBD White Collar Employment title Page 27

28 Conclusions Annual rail patronage for 38 years modelled; Data: socio-economic, fare, train operating kilometres and other factors; Two variables describing economic activity Metropolitan office employment (MOE) and Real Gross State Product per Head (RGSPHH). Strong positive relationship with rail patronage. 10% increase in MOE increases patronage 6.1% 10% in RGSP raises rail patronage 7.4%. title Page 28

29 Conclusions Fare and train kms reasonable Others reasonable but with estimation error The model is exploratory some variables not yet included include when annual data replaced by four weekly or quarterly data. title Page 29

30 Questions title Page 30