1. Stefan Flügel, Nils Fearnley, Marit Killi
TØI, Norway
Cost Benefit Analyses:
Passengers' Valuations
of Universal Design
for Public Transport
Transportforum, Linkjøping, 14. January 2010
© Transportøkonomisk institutt
15.01.2010 Side 1

2. Outline
 Introduction: Universal Design
 The Valuation Study
 About the project
 Choice Experiments
 New recommended unit values
 Briefly: Use of values in cost benefit analyses
 Summary
© Transportøkonomisk institutt
15.01.2010 Side 2

3. Definition of “Universal Design” (UD)
for public transportation
Adopted by the Norwegian Road Authorities:
“The design of infrastructure, transportation systems or their
surroundings to accommodate the widest range of potential
users regardless of their impairments or special needs”
Focus: UD has potential value for all passengers
-> Applicable for cost benefit analyses (CBA)
© Transportøkonomisk institutt
15.01.2010 Side 3

4. UD includes an
extensive set of provision
Access to station
Facilities at station
Information
(at station and on-board)
Access to vehicle
Facilities on-board
© Transportøkonomisk institutt
1/15/2010 Side 4

5. Project
 Main goal: Obtain monetary unit values for different UD
provisions
 On behalf of the Norwegian Public Roads Administration
 Conducted in three Norwegian cities in 2009
 Two parts of the study
 Qualitatively: focus interviews, on-board study
 Quantitatively: valuation study
© Transportøkonomisk institutt
15.01.2010 Side 5

6. Existing provisions in the three cities
City Drammen (3 Kristiansand Oslo (One bus
bus lines) (2 bus lines) line and 2
tram lines)
Provisions at the stop:
Easy access to the stop is without physical barriers X X X
Tactile paving (rough tracks on the ground for blind and
X X X
vision impaired)
Shelter at the stop X X X
Sitting places at the stop X X X
Real time information systems X
Elevated kerbs to reduce vertical gap X X X
Provisions on the bus/tram:
Bus/tram clearly marked with destination and route id on
X X X
the outside
Bus/tram has space for pram, bicycle, wheelchair X X X
Announcement of next stop on board the bus/tram X X
Announcement of next stop on screen on board the
X X
bus/tram
© Transportøkonomisk institutt
15.01.2010 Side 6

7. Qualitative Part
 Focus interviews:
 Terms and notions like “universal design”, “real-time information”
mostly unknown -> use pictures for illustration in the study
 Real-time information by screen very popular
 On-board study:
 About 1/3 of the passengers noticed upgraded UD
 A majority view UD provisions as beneficial for all passengers and
not only for people with special needs
 UD would lead to more public transport trips
 However, price, reliability and frequency are ranked higher than UD
on an average score
© Transportøkonomisk institutt
15.01.2010 Side 7

8. Valuation study
 Stated Preferences (SP) study to find monetary value of
different UD provisions
 Recruitment: invitation cards on-board (bus and tram lines)
with log-in information to self-administrated internet survey
 Pilot: May 2009
 Main study: July 2009 in the same 3 cities as on-board study
 Sample size and response rates
 Pilot: 103 (11%)
 Main study: 350 ( 5,3%) , big difference between cities
 Merging pilot and main study
© Transportøkonomisk institutt
15.01.2010 Side 8

9. Structure of questionaire Reporting the reference trip Exclusion of
and main attributes in Respondents
N=453
Choice
Experiments (CE)
CE1: Cost, In‐Vehicle Time,
Information at station
N=417
In each CE:
6 choices per respondent
CE2: Cost, Information on
between board, Accessibility to vehicle
2 alternatives
N= 411
Random
allocation
CE3a: Cost, Shelter, Cleanness CE3b: Cost, Shelter, Ice/Snow
N= 225 removal
N=183
CV Questions and Final
questionnaire N=406
© Transportøkonomisk institutt
15.01.2010 Side 9

10. Presentation of alternatives in CE1
1. Introduction of attributes and levels
3. Presentation of alternatives
2. Explaining the choice decision
Choices:
Definitely A, Probably A, Probably B, Definitely B
© Transportøkonomisk institutt
15.01.2010 Side 10

11. Presentation of alternatives in CE2
Explaining the accessibility levels Enhancing the gap between floor and
Pilot station ground in the main study
© Transportøkonomisk institutt
15.01.2010 Side 11

12. Presentation of choices in CE2
© Transportøkonomisk institutt
15.01.2010 Side 12

13. Presentation of alternatives in CE3
© Transportøkonomisk institutt
15.01.2010 Side 13

14. Estimation approaches
 Paradigm: Random utility maximisation (RUM)
 Multi-nominal logit (MNL)
 Fixed coefficients in utility function ; Monetary valuation for UD as the
marginal rate of substitution between the cost and changes in UD
 MNL for unit value determination most comprehensible and robust
 Mixed logit models (ML)
 Random coefficient model; unobserved heterogeneity; panel structure
 Estimating mean and standard deviation of the predefined parameter
distribution function
 Results depended on distribution assumption and on further assumptions
about truncating and censoring
-> General estimation results
 Expected sign and order (MNL) for all parameters ; significant different fro
zero with just a few exceptions
 Taste heterogeneity in the light of mixed logit results high
© Transportøkonomisk institutt
15.01.2010 Side 14

15. Valuation of information at station
WTP for information at the station (from a situation with just a NOK USD
timetable)
Map over local area 0,43 0,08
Speaker about changes in departure 0,69 0,12
Screen with real-time information 4,05 0,72
All three information devices 4,62 0,81
 Real-time information system is clearly the most valuable information source at the
station; consistent with on-board study and focus interviews
 Relatively low monetary valuations for map over the local area and a speaker about
deviating departure times
 Package price of all three lower than the sum of the single provisions
 Loudspeaker seems almost unessential when screen with real time information
is available as the valuation of map and screen almost equals the valuation of
all three information devices
 However, special needs (e.g. reduced sight) not accounted for
© Transportøkonomisk institutt
15.01.2010 Side 15

16. Valuation of information on-board (CE2)
WTP for information on board (from a situation without any NOK USD
information)
Next station via speaker 3,63 0,64
Next station via screen 3,68 0,65
Next station via speaker and screen 4,20 0,74
 If the next station is announced via screen or speaker seems equally beneficial,
around 3,65 NOK is the estimated monetary valuation
 The additional information adds only 55 øre in value
 For the average user (with good hearing and sight) one information seems to be
sufficient
© Transportøkonomisk institutt
15.01.2010 Side 16

17. Valuation of improved accessibility on-board
WTP for accessibility improvements (from a situation without any NOK USD
adjustments)
Lowered vehicle floor 1,67 0,30
Lowered vehicle floor and adjusted ground at the station 2,07 0,37
 In comparison with other UD provisions the values seem relatively low
 Lowered floor has an average value of 1,87 NOK, an additional adjustment
(elevation) of the station ground is valued 0,4 NOK
 Valuation of persons with special needs:
WTP in USD ALL Respondents with physical
problems (*) or heavy baggage (**)
N=2466 N=594
Lowered vehicle floor 1,67 2,88
Lowered vehicle floor and adjusted 2,07 4,01
ground at station
*) limited moveability, walking stick or crutched, pregnant **) big/heavy luggage, a lot of shopping bags, trolley, small kids
The WTP for “steppless” access on the bus or tram is valued up to an average value
of 4,01 for people with “physicals problems” or heavy baggage.
© Transportøkonomisk institutt
15.01.2010 Side 17

18. Valuation of attributes in CE3
WTP for a shelter (average value of CE3a and CE3b) NOK USD
Shelter without sitting place 3,12 0,55
Shelter with sitting place 5,10 0,90
WTP for satisfactory cleanness and ice/snow removal
Cleanness 3,62 0,64
Ice/snow removal 4,97 0,88
 Valuation of shelter seems relatively high (3,12 NOK), if the shelter has a sitting
place the value increases to 5,1 NOK
 Satisfactory cleanness is valued at 3,62 NOK
 Ice/snow removal has an estimated value of 4,97 NOK
 The values seem high but can be (partly) explained with the high standard of
cleanness and ice removal in Norway so that dirt and missing removal can course
high disutility and thereby a relatively high monetary valuation
 Additional analysis: Men value cleanness and snow/ice removal less than women
© Transportøkonomisk institutt
15.01.2010 Side 18

19. Comparison between CE and CV
CV CE CE CE
Total valuation for Valuation on a Valuation on a Valuation on a transport
the package price transport line transport line with line with relatively high
(list of provisions) without UD (just relatively low standard of UD
timetable) standard of UD
Information devices at the 4,62 4,19 (*) (assuming0,14 (**) (assuming that
station: map, speaker and that there is a map
there is a map and a real-
real-time information via at the station”) time information at the
screen station”)
Information devices on 4,2 0,57 (assuming no 0,57 (assuming no
board : Announcement of speaker, but a speaker, but a screen)
next station via screen)
Low-floor bus and 2,07 2,07 (assuming no 0,40 (assuming low-floor
elevation of bus station low-floor bus) bus but no elevation of
station
Shelter with sitting place 5,10 1,98 (Assuming 1,98 (Assuming shelter
shelter but no sitting but no sitting place)
place)
Cleanness 3,62 Assuming cleanness Assuming cleanness
Ice/snow removal 4,97 Assuming ice Assuming ice removal
removal
Sum 4,35 NOK 24,56 NOK 8,81 NOK 3,09 NOK
© Transportøkonomisk institutt
15.01.2010 Side 19

20. Comparison with former recommended values
Provisions Old values (2005 From this study (2009
values) values)
Real time information at bus stop 2.10 4.05
Local map at bus stop 0.70 0.43
Speaker with info of changes, disruptions 2.10 0.69
All three information devices: map, speaker and 4.62
RTI
Sign on board in bus indicating next stop 2.43 3.67
Next stop information announced by driver 1.22 3.62
Both: next stop via speaker and screen 4.20
Bus shelter 1.05
- without seating 3.12
- with seating 5.10
Clean bus stop 2.56 3.62
Snow and ice removal at stop 2.56 4.97
Lights at bus stop 0.67 2.82
Lowfloor bus 0.61 1.67
Elevated curb for level-free boarding 0.31 0,40
Old values based on a review of international studies conducted 1996-2002 (Nossum/Killi (2006))
UD publically discussed in Norway in recent years; Increased purchase power
High standard of UD on investigated bus/tram lines -> we might actually measure compensation prices, that are often found to be
higher than purchasing prices (reference dependency)
Problem of self selection due to low response rate (?)
15/01/2010 Page 20

21. Cost Benefit Analyses
 Systematic evaluation of a projects’ benefits and costs measured both in
monetary units
 With this study we provide monetary values for passengers benefits of UD
investments
 In addition: possible benefits for other passengers and the operators
E.g. Low-floor buses lead to time-savings for all passengers and obtain
efficiency effects for bus companies
 Subtracting the present value of costs (mainly investment and maintenance
costs ) from present value of benefits gives the net present value (NPV)
 Projects with a positive NPV are socioeconomically profitable
 As different profitable project might be competing for government
money, the NPV is divided by the share of finance through government
budget to obtain a measure (“Benefit-cost-ratio”) that can rank different
projects
© Transportøkonomisk institutt
15.01.2010 Side 21

22. Improved welfare case for UD with new values
UD investments with standardised cost assumption (Fearnley 2007)
Benefit cost ratio over 25 years by passengers per year
Old values
New values
15/01/2010 Page 22

23. Summary
 Universal Design is beneficial for all passengers
 Possible to derive monetary values for single provisions with
straightforward choice experiments
 Estimation results indicate that the former recommended
values might have been too low assessed for Norway
 Using the new values in CBA increases the calculated NPV
of UD provisions. Relatively low numbers of passengers are
required to make investments in UD socioeconomically
profitable
© Transportøkonomisk institutt
15.01.2010 Side 23