Advances in ICT and its Implications for

Travel Behavior: A Case Study of Korea

I- TED 2014 April 9-11 2014, Dallas U.S.

Sungwon LEE, Gyeng Chul KIM, Seung Kook WU and Jieun OH

Contents

Ⅰ

Ⅱ

Ⅲ

Ⅳ

Introduction (ICT Connectivity and Transport)

Stated Preference Survey and Analysis

Policy Implications from SP

Discussions (Q & A)

Proliferation of ICT devices in Korea and in the world The world is becoming more and more connected

Advances in ICT and its Implications

Introduction

3

Intelligent multi-purpose devices with ICT convergence Provide various transport services integrated with ICT

Changes in our lifestyle and its implications for transport

4

Travel costs and travel time are the main attributes in modal choice models

A number of studies found that Information technology development and penetration of ICT devices (like smartphone) change the life, challenging the notion about travel (Lyons and Urry, 2005) and the disutility of transit modes.

Sustainable Development Commission (2010) reported that “the ability to stay connected and turn travel time into productive work time can be a significant attraction for business travelers.

Advances in ICT and travel demand models

Literature Review

5

Travel time can be productive with ICT Lee-Gosselin and Miranda-Moreno (2009)

A survey of Chicago Transit Authority train riders suggests that the riders use time and money in better way than drivers (Frei and Mahmassani, 2011)

Connolly, Caulfield, and O’Mahony (2009) found that multitasking is extremely common while traveling by rail and smartphone users would benefit from internet connection in trains

Gamberini et al. (2012) found that even if the travel length is relatively short, train riders engage in several activities especially for those using mobile ICT devices

Advances in ICT and its implications for travel

Literature Review

6

7

• City bus patronage is increasing

• Passenger kilometer traveled is also increasing

Intracity Bus Transportation Statistics

Year Passenger Passenger-km

2002 4,523,037 24,295,022

2003 4,408,604 22,730,884

2004 4,451,840 25,756,855

2005 4,536,634 25,902,141

2006 4,615,930 27,587,636

2007 4,768,437 28,730,674

2008 5,068,671 30,792,310

2009 5,168,492 32,037,348

2010 5,246,965 31,684,079

2011 5,535,686 32,557,560

Table 1 Intracity Bus Transportation Statistics

*Source: National Bus Transport Association of Korea (http://www.bus.or.kr), 2011 Bus Statistics.

Mass Transport in Korea: Intracity Bus

(Units: Thousand people, passenger-km)

8

Intracity Bus

• Total 348 bus companies operate 32,575 intracity buses in Korea

Intracity Bus Company and Vehicles

Table 2 Intracity Bus Company and Vehicles

City/Province Company Vehicles Vehicles per company

Total 348 32,575 94 Seoul 67 7,525 112 Busan 33 2,511 76 Daegu 26 1,561 60

Incheon 41 2,323 57 Gwangju 10 972 97 Daejeon 13 965 74 Ulsan 8 674 84

Gyeonggi-do 52 9,883 190 Gangwon-do 9 564 63

Chungcheongbuk-do 10 540 54 Chungcheongnam-do 9 731 81

Jeollabuk-do 13 822 63 Jeollanam-do 11 677 62

Gyeongsangbuk-do 15 1,118 75 Gyeongsangnam-do 27 1,637 57

Jeju 4 172 43

*Source: National Bus Transport Association of Korea (http://www.bus.or.kr), 2011 Bus Statistics.

9

Intercity and Express Buses

• Intercity bus patronage is decreasing while passenger-km is increasing

• Express bus is in decreasing trend

Intercity Bus and Express Bus Transportation Statistics

Table 3 Intercity Bus and Express Transportation Statistics

Year Intercity Bus Express Bus

Passenger Passenger-km Passenger Passenger-km

2002 329,192 10,985,630 42,132 9,413,822

2003 282,782 9,821,406 40,140 8,790,072

2004 262,051 10,232,042 38,877 8,432,356

2005 245,678 9,861,530 37,687 8,125,011

2006 246,359 11,955,403 38,873 8,460,067

2007 239,668 13,667,261 38,569 8,470,892

2008 242,121 16,726,963 40,451 8,799,997

2009 235,761 16,881,312 38,095 8,235,563

2010 226,111 16,590,512 38,204 8,247,025

2011 222,121 16,364,201 37,005 8,111,058

*Source: National Bus Transport Association of Korea (http://www.bus.or.kr), 2011 Bus Statistics.

10

Intercity and Express Buses

<Table> Intercity Bus and Express Bus Company and Vehicles Province

Intercity Bus Express Bus

Company Vehicles Vehicles per company

Company Vehicles Vehicles per company

Total 83 7,636 92 8 1,891 236

Gyeonggi-do 16 1,815 113 4 878 220

Gangwon-do 10 719 72 1 113 113

Chungcheongbuk-do 5 478 96 1 85 85

Chungcheongnam-do 5 865 173 - - -

Jeollabuk-do 5 493 99 - - -

Jeollanam-do 8 592 74 1 637 637

Gyeongsangbuk-do 8 1,018 127 - - -

Gyeongsangnam-do 21 1,399 67 1 178 178

Jeju 5 257 51 - - -

Intercity Bus and Express Bus Company and Vehicles

*Source: National Bus Transport Association of Korea (http://www.bus.or.kr), 2011 Bus Statistics.

11

ICT amenities in public transport in Korea

Information Utility in Intracity Bus

City/Province Wi-Fi availability in Bus

Seoul ◦ 97×× Bus ◦ Metro Bus, Gyunggi Province Bus

Busan ◦ 63 bus lines

Daegu ◦ 11 bus lines

Incheon ◦ Circular 7 Lines ◦ Metro 18 Lines

Gwangju ◦ 17 lines

Daejeon ◦ Kyungick Bus KT Wi-Fi available ◦ Other lines SK Wi-Fi available

Ulsan -

Gyeonggi-do ◦ Gyunggi buses ◦ Metro Bus

Gangwon-do -

Chungcheongbuk-do ◦ Woojin and Dongil Bus KT Wi-Fi available

Chungcheongnam-do -

Jeollabuk-do -

Jeollanam-do ◦ Gyangyang City Bus KT Wi-Fi available ◦ Mokpo City KT Wi-Fi available

Gyeongsangbuk-do ◦ Gumi City KT Wi-Fi available

Gyeongsangnam-do ◦ Machang and Dongyang Bus KT Wi-Fi partial

Jeju ◦ 12 Bus lines

*Source: http://gall.dcinside.com/board/view/?id=bus&no=266582

Figure 1 Seongnam City Bus Wi-Fi

Figure 2 Wi-Fi users in bus

12

Rail Transport in Korea

• Rail passenger transport at 1,149 million and distance traveled is 42,492

million passenger-km in 2012

• Both regional rail and metro rail patronage have been increasing since 2005

Korail Passenger Transportation Statistics

Table 4 Korail Passenger Transportation Statistics

*Source: KORAIL (http://www.korail.com), 2012 Rail Statistics.

Year

Total Regional rail Subway in

Metropolitan Region Airport rail

Passenger Passenger-

km Passenger

Passenger-km

Passenger Passenger-

km Passenger

Passenger-km

2005 950,995 31,004,212 115,002 19,076,105 835,993 11,928,106 - -

2006 969,145 31,415,976 114,331 19,078,533 854,814 12,337,443 - -

2007 989,294 31,595,987 110,631 18,680,339 878,664 12,915,648 3,752 91,812

2008 1,018,977 32,026,528 113,098 18,671,356 905,879 13,355,172 6,078 145,159

2009 1,020,319 31,299,106 107,733 17,816,751 912,586 13,482,355 7,340 163,964

2010 1,060,941 33,012,479 112,093 19,018,617 948,848 13,993,862 10,044 211,489

2011 1,118,621 36,784,264 121,769 21,603,203 996,852 15,181,061 32,521 656,660

2012 1,149,340 42,492,561 125,817 22,111,981 1,023,523 20,380,580 49,138 977,392

(Units: thousand people, thousand passenger-km)

13

Railroad

• KTX (Express Rail)

- KTX provides free wifi service

- KTX car number 5 and 13 have PC’s for internet

• Subway

- Metro rail stations and vehicles provide Wi-Fi services (SKT and KT)

ICT Amenities in Rail

Figure 3 KTX Wi-Fi (Left) and KTX-Sancheon business room (Right)

Figure 4 Wi-Fi use case in subway(Left) and Wi-Fi equipment in passenger car (Right)

14

Stated Preference Survey: Base Statistics

Survey Statistics

• Gender • Occupation

• Generation • Residential

15

Base Statistics Car User Characteristics

• Vehicle characteristics • Types of fuels

• Frequency of using private vehicle for commuting

• Reasons for using private vehicles

Average 4.3times

16

Commuting Travel Behavior

• Average time required for commuting • Average travel cost per month

(Unit: minutes) (Units: thousand won)

Average 38.5 min

Average 200,804 won

• Modes of transportation seperate from commuting

17

• Recognition of free Wi-Fi service in public transportation

• Use of ICT devices on public transportation

• Network services when using ICT devices on public transportation

Based on 219 ICT

device users

• Wi-Fi condition of ICT devices on public transportation

Based on 52

Wi-Fi users

Not well connected

but always available

Well connected and fast

Well connected, but often slow

ICT Device Usage in Public Transportation

18

Unstable network connection

Provided with unlimited data

Not sensitive to data usage

Inconvenience of connecting to Wi-Fi

Not familiar with the connection procedures

Too slow

• Reasons for not using Wi-Fi on public transportation

Using SNS (Facebook, Twitter etc.)

Music, Videos, etc.

Playing games

Business purposes (checking -mail etc.)

Instant messaging (Kakaotalk, etc.)

Internet searching

• Purposes of using ICT devices on public transportation (Multiple responses)

* Based on 219 ICT device users

* Based on 3G, 4G (LTE) users

ICT Usage Behavior in Public Transportation

19

ICT Service Usage in Public Transportation

• Monthly average mobile phone bill (Unit: Won)

• Monthly data allowance

• Average time of using mobile phones per day

Average 3.25 GB

Units : hour

Very Sufficient

Sufficient

Just OK

Somewhat Insufficient

Not-sufficient

Quantitative Policy Impact Analysis

Stated preference methodology for impact analysis of hypothetical transport policy measures

- Bases for scientific transport policy intervention Econometric testing of transport policy related

hypotheses

- Perceived vs. real cost of transport

Stated Preference Analysis for ICT’s Impact on Travel Behavior

20

SP Methodology and Estimation Results

If variables are too numerous and too widely varied impossible to create all possible sets of SP

questionnaires Use fractional factorial plan which analyzes only main

effects and guarantee the orthogonality of variables following Kocur et al. (1982) and Hensher (1994)

SP design of mode choice between passenger cars and alternative modes of bus and subway

Explanatory variables travel expense, travel time, and service levels (ICT

amenity levels: free Wi-Fi and etc.)

21

where mass = bus or subway, Surveyed on 240 car users binary choice with

multiple levels of attributes 3,840 effective data points

Ct i meCfuelUcar ⋅⋅ ++= 21 ββα

Mt i meMi ctU ⋅⋅ += 43 ββmass

Utility functions

22

Coefficient Estimation for Respondents Who Declared Bus as Alternative

Coefficient Standard Error Prob. |z|>Z*

CFUEL -.14258D-05*** .4319D-06 -3.30

CTIME -.04316*** .00761 -5.67

MICT .33463*** .06910 4.84

MTIME -.04156*** .00666 -6.24

A_CAR .25930 .17834 1.45

Note: nnnnn.D-xx or D+xx ⇒ multiply by 10 to -xx or +xx. Note: ***, **, * ⇒ Significant at 1%, 5%, 10% level

23

Elasticity Estimates for Respondents Who Declared Bus as Alternative

CFUEL CAR MASS

Elasticity -.2036 .1069

CTIME CAR MASS

Elasticity -1.0322 .5769

MICT CAR MASS

Elasticity -.4390 .2302

MTIME CAR MASS

Elasticity 1.0965 -.6464

24

Most attribute variables are statistically significant, except car constant

Positive car constant But not statistically significant: No intrinsic car preference

Time related demand elasticity is much higher than cost related ones

ICT amenity related elasticity is positive Implying fairly positive role of attracting car users into public transport

Coefficient Estimation for Respondents Who Declared Subway as Alternative

Note: nnnnn.D-xx or D+xx ⇒ multiply by 10 to -xx or +xx. Note: ***, **, * ⇒ Significant at 1%, 5%, 10% level

Coefficient Standard Error Prob. |z|>Z*

CFUEL -.94589D-06*** .3209D-06 -2.95

CTIME -.02434*** .00530 -4.60

MICT .23738*** .06988 3.40

MTIME -.02399*** .00463 -5.18

A_CAR -.11309 .18785 -.60

26

Elasticity Estimates for Respondents Who Declared Subway as Alternative

CFUEL CAR MASS

Elasticity -.1976 .0902

CTIME CAR MASS

Elasticity -.9036 .4363

MICT CAR MASS

Elasticity -.3245 .1502

MTIME CAR MASS

Elasticity .9875 -.4979

27

Most attribute variables are statistically significant, except car constant

Negative car constant But not statistically significant: No intrinsic car preference

Time related demand elasticity is much higher than cost related ones

ICT amenity related elasticity is positive Implying fairly positive role of attracting car users into public transport

Similar result to Bus User case

Demand elasticity of ICT amenity lies just between cost and time variable

Implying higher response than price related policy options Providing higher ICT connectivity in mass

transit could attract car users to MT by making travel time more enjoyable or productive

Reducing travel time is still most powerful policy measure for modal shift towards MT

Policy Implications and Conclusion

No intrinsic car preference in our survey unlike previous ones

ICT revolution will change the way people

travel Riding MT could be productive

Policy Implications and Conclusion

Thank You. (swlee@koti.re.kr)