Transcript PPT - Bureau of Infrastructure, Transport and Regional Economics

An inland rail route between
Melbourne and Brisbane?
The North-South rail corridor study
BTRE Transport Colloquium
14 June 2007
Structure of presentation

Introduction

Total freight market

Mode choice modelling

Access prices

Passenger market modelling

Project outcomes

Areas of future research
Slide 2
The task

Background: freight growth, Auslink upgrades,
lobbying for inland rail route

To consider options for the Melbourne Brisbane rail corridor over the next 25 years

Consortium:
• Ernst and Young (project leader, financial anaylsis)
• Hyder (route options, other infrastructure,
environmental)
• ACIL Tasman (demand, access prices)
Slide 3
Analytical approach
• Identify current total freight market
• Forecast growth in total freight market for 25 years
• Determine current mode shares
• Estimate mode shares over 25 years and their
sensitivity to changing service quality
• Estimate rail freight over 25 years
• Integrate with other models in consortium
Slide 4
Path of far western route
Slide 5
Slide 6
Freight movements (excl. coal)
5.4m
5.1m
0.8m
19.3m
9.2m
7.0m
4.5m
1.2m
9.0m
2.9m
5.3m
To/from WA & SA
into NS Corridor
4m
10.3m
Slide 7
Analysis of base (2004)
• Origin - destination tonnages by commodity
• Data from rail operators, BTRE, ABS, FDF
• FreightSim for forecasting model. Structure:
production, imports, consumption
• ACIL Tasman model of freight inducement effects
Slide 8
Drivers of future demand

GDP growth (production, imports,
consumption)

Transport freight to GDP growth ratio
• Growth of industrial concentration
• Growth of agricultural production
• Growth of imports, service sector
• Scenarios
– High/medium/low GDP
– Growth transport/GDP ratio
Slide 9
New East Coast demand with an
inland route

Coal in southern QLD, northern NSW
• But questions about which route, which port

Little else

Mainly a through route with a stop at Parkes
Slide 10
Freight diversion – northern NSW

27% of grain from Northern Plains to Brisbane
from Newcastle

50% of cotton from Northern Plains to Brisbane
from Port Botany

Brisbane-Perth freight via Parkes

No other material freight diversion
Slide 11
Future demand modelling results
30,000
25,000
Ktonnes
20,000
15,000
10,000
5,000
2004
2009
Melbourne-Brisbane (case A)
Melbourne-Sydney (case A)
Sydney-Brisbane (case A)
2014
2019
Melbourne-Brisbane (case B)
Melbourne-Sydney (case B)
Sydney-Brisbane (case B)
2024
Melbourne-Brisbane (case C)
Melbourne-Sydney (case C)
Sydney-Brisbane (case C)
2029
Slide 12
Road, rail, sea or air ?

Current market shares by mode

Drivers of mode choice

Sydney

Convenient departures

Logit model

Results
Slide 13
2004 snapshot

Current Rail mode shares
• Melbourne-Brisbane ~30%
• Melbourne-Sydney ~7%
• Sydney-Brisbane ~11%

Rail more price competitive on longer routes,
less on shorter routes because of PUD time and
costs

Rail outperformed by road in service quality
Slide 14
Mode performance 2004 – M to B
Road
Relative
price
Sea
0-20%
20-40%
above d-d below d-d
rail
rail
Rail
linehaul
Rail door
to door
30-45%
below
road
0-20%
below
road
Reliability
95%
90%
35-45%
35-45%
Availability
99%
10%
40-45%
40-45%
3-3.5 days
36 hrs
42 hrs
Transit time 21-27 hrs
Slide 15
Explanatory variables

Price
• $/tonne (but complexities)

Reliability
• within 15 minutes of scheduled arrival time

Availability
• % of times the freight carrier is available within an hour of
customers’ preferred time

Survey:
• how does demand for rail (at the expense of road) vary with
changes in price, reliability, availability?
Slide 16
Reliability - problems in Sydney
Brisbane
ARTC
Network
(exception of
north of the
Queensland
boarder)
Trains traverse
the ARTC &
Railcorp
Network from
MelbourneBrisbane
(arrow does
not follow
actual train
path)
40% on-time reliability
( average for both
directions)
80% on-time
reliability
Loss of 50% reliability
in Northern Sydney
30% on-time
reliability
40% on-time
reliability
Sydney
RailCorp Network
60% on-time
reliability
Loss of 20% reliability in
South Western Sydney
ARTC
Network
Melbourne
Slide 17
Transit time & availability
% of Total departures per hour
8%
7%
6%
Current
departure
time from
customer
5%
4%
2009
departure
time from
customer
3%
2%
1%
0%
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time of day
24 hour distribution of truck departures
Slide 18
Survey results – all customers
6%
7%
10%
41%
15%
21%
Price
Reliability
Availability
Transit time
Flexibility
Loss/damage
Slide 19
Survey results - manufactured
23%
51%
26%
Price
Reliability
Availability
Slide 20
The manufactured market
MelbourneBrisbane
MelbourneSydney
SydneyBrisbane
Express freight
5%
5%
5%
Freight sensitive
to reliability and
availability
60%
70%
70%
Price-sensitive
freight
35%
25%
25%
Slide 21
Mode choice

Logit model to predict modal shares
• Calibrated to explain current shares
• Forecast changes based on expected route
characteristics
– Price (incl fuel price, driver shortage)
– Reliability following AusLink upgrade
– Availability following AusLink upgrade

Key parameters estimated from
• Surveys
• Econometric analysis of past data
Slide 22
Logit model

Gives probability (%) of freight forwarder choosing
mode n

Simple logit (two modes) formula

Where U is utility of using rail or road
• linear utility functions: constant + variable1 x coefficient1 + etc
Slide 23
Logit model – hierarchical structure
• Logit model can be used at each level of the freight
decision making process
• Nested logit useful for inland rail analysis
Slide 24
Values/inputs to use in the model

Current estimates of road and rail performance
and mode share were derived from
• Rail operators, BTRE, ARTC, surveys

Future estimates of road and rail performance
were derived from
• ARTC, BTRE, ACIL Tasman, Hyder, freight operators

Standalone road and rail freight pricing model
developed
• Access prices, fuel costs, labour
Slide 25
Scenarios

Case A (reference), Case B (high rail), Case C
(low rail)

Fuel price assumptions

Road and rail labour cost assumptions
Slide 26
Elasticities and coefficients
•
•
Choice elasticities were estimated
Price
Reliability
Availability
Melbourne- Brisbane
-0.5-1.2
0.4-0.7
0.5-0.8
Melbourne-Sydney
-0.7-1.15
0.3-0.7
0.08-0.36
Sydney-Brisbane
-0.3-0.9
0.3-0.7
0.3-0.8
All routes
-0.3-1.2
0.3-0.7
0.08-0.8
From these – coefficients were calculated
Slide 27
Testing of the logit model
• Observed performance (Melb-Perth 2000)
Price (ave
charge $/tonne)
Reliability
Service
availability
Observed
mode share
Road
186
95%
99%
30%
Rail
113
66%
80%
70%
• Logit modelling of same input variables
Price (ave
charge $/tonne)
Reliability
Service
availability
Observed
mode share
Road
186
95%
99%
27%
Rail
113
66%
80%
73%
Slide 28
Results – rail market share
Slide 29
Revenue maximising access charges
Revenue ($ million)
60
50
40
30
20
10
0
$0.00
$1.26
$2.53
$3.79
$5.05
$6.31
$7.58
$8.84
$10.10
$11.36
$12.63
Access charge ('000gtk)
0% Efficiency gains
15% Efficiency gains
5% Efficiency gains
20% Efficiency gains
10% Efficiency gains
Slide 30
Financial and economic results

Summary of results

Interpretation of results

Coastal route: problems and solutions

Inland route: problems and solutions
Full report including Ernst & Young and Hyder
chapters not covered here, on
www.aciltasman.com.au
Slide 31