Geographical simulation analysis for logistics enhancement ...

Institute of Developing Economies, JETRO

Geographical simulation analysis for

logistics enhancement in Asia

Satoru KUMAGAI

KEOLA Souknilanh

Ikumo ISONO

October 2018

Contents

• Background of IDE-GSM

• Model and Data

• Example of analyses

• Conclusion

2

What is IDE-GSM? (IDE-Geographical Simulation Model)

• Simulate the long-term evolution of population and the location of industries, Based on Spatial Economics

• Estimate the economic impacts of infrastructure development and customs facilitation at the sub-national level.

• Traditional CBA on infrastructure needs various detailed information thus it takes time and money, and the scope of analyses is limited within nearby region or lacks ground view to consider multiple projects .

• General equilibrium frameworks enables total evaluation of economic impacts beyond directly involved regions.

• IDE-GSM provides estimation of broader impacts on transport volumes and modal shifts

Geo-economic data in IDE-GSM

2018/11/8 4

Sub-national level: 98 economies (3,017 regions) National level:+ 71 countries

GDP by sector -Agriculture/Mining -Automotive -E&E -Textile & Germent -Food Processing -Other Mfg. -Services Population

GDP per capita by region(2010)

Estimating regional GDP by Satellite Imagery

5

Nighttime Lights Imagery of Myanmar Estimated GRDP per capita

+ Land cover

Transport network data in IDE-GSM

2018/11/8 6

Over 14,000 routes, including -Road -Sea -Air -Rail

Multi-modal transport model

Road network data

Manual/Automatic Compilation of Route data

• We consider major routes connecting big cities and provinces.

• We build the dataset manually based on the maps, existing database and field surveys.

• We are trying to construct the data from the third source data (such as OSM) and script, but still adjustment is needed.

7

Trial of building route data by script

In the field trip, we could drive the fastest in this section

8

1990

ASEAN

China

JapanIndia

EU USA

16.1

8.111.3

14.0

25.8 28.0

22.629.3

36.82.71.2

24.9

23.4

11.9

8.7

4.6

3.0

92.5

52.1

1999

ASEAN

China

JapanIndia

EU USA

118.5

19.325.8

52.7

65.5 88.2

45.745.0

54.34.44.9

35.2

60.6

42.6

33.8

12.9

14.7

158.0

66.0

2009

ASEAN

China

JapanIndia

EU USA

283.9

75.9125.5

296.8

143.8 84.5

62.372.5

72.724.421.8

70.0

97.6

114.3

126.4

75.2

101.3

89.5

57.8

31.0%

2.8%

17.6%

2.3%1.2%

24.7%

22.1%15.6%

5.9%3.3%2.8%

24.3%

1 4.5 6 3.8 3.2 9 1 5.2 6.5 4 3.2 9

1995

2015

1990

1999

2009

GDP share in the world

Trade (billion USD)

Background:

Big shift in the world

economy

EU28 Japan

China

ASEAN India

9

Source: ASEAN Secretariat, “ASEAN Community in Figures: Special Edition 2014”.

Progress of economic integration in East Asia

10

The evolution of production networks in East Asia

Source: Compiled from Hiratsuka (2006)

Thailand

•Spindle Motor

•Base

•Carriage

•Flex Cable

•Pivot

•Seal

•VCM

•Top Cover

•PCBA

•HGA

•HAS

Japan

•Cover

•Disk

•Screw

•Seal

•Ramp

•Top Cover

•Latch

•Plate Case

•Label

•Filter

•PCBA

•Suspension

USA

•Disk

•Head

•Suspension

Mexico

•Head

Taiwan

•Top Clamp

Hong Kong

•Filter

Philippines

•D. Plate

•Coil Support

•PCBA

Indonesia

•W. Suspension

•VCM

•PCBA

Malaysia

•Base

•Pivot

•Spacer

•VCM

•Card

•Top Clamp

•Disk

Singapore

•Cover

•Screw

•Pivot

•PC ADP

•Disk

China

•PCBA

•Carriage

•HGA

•Base

•Head

•Suspension

Global Purchasing by a HDD Manufacturer in 2005

11

Urbanisation

Source: ERIA(2015), originally made by Keola Souknilanh

12

Poor state of road infrastructure

National Road No.1, Bhutan

National Road No.1, Myanmar

Bypath of national road No.3, Indonesia

JAKARTA

Source: taken by Ikumo Isono

Sihanoukville airport upgrade

Highway: Kanchanaburi - Dawei

Expressway: Phnom Penh - Sihanoukville

SKRL: PP – Loc Ninh – HCMC

Mekong bridge in Neak Loung (NR1)

Reconstruction of NR3: Phnom Penh -Kampot

Dawei deep sea port

Cross-border facility at Bavet-Moc Bai

Sihanoukville port expansion

Upgrading Dawei airport

Da Nang airport: Passenger terminal

Yangon port: Quay cranes

Thilawa port improvement

Savannakhet airport improvement

Phnom Penh port rehabilitation

Da Nang port improvement

Koh Kong Industrial Estate

Poipet Industrial Estate

SKRL: Poipet - Sisophon (48km)

Pharma & biotech city in Ayutthaya

IT & ITES part in Phthum Thani CDZ

North-South High Speed Railway (to HCMC)

SEZ/FTZ in Dawei

Multimodal logistic park in Dawei

Special Border Zone at Myawadi

Route No.8: Kawkareik-Mawlamyine - Thaton

SEZ in Savannakhet

Border Trading Zone in Dansavanh

Nam Theun 2 and Nam Ngum 2 Hydropower Plant

Ports: Cai Lan, Lach Huyen

Noi Bai Airport terminal 2

Expansion of Cat Bi Airport

Rail link : Hanoi – Noibai AP, Lang Hoa Lac

Hoa Lac high tech park, Vietnam space center

ICT infrastructure enhancement

Hanoi

Yangon

Mawlamyaing

Dawei

Myawadi Khon Kaen

Vientiane

Savannakhet

Hanoi

Bangkok

Phnom Penh

Sihanoukville

MIEC

EWEC

Ho Chi Minh

Da Nang

Chennai

Bangalore

MRT network

Suvarnabhumi airport: Phase 2 development

Laem Chabang port: Phase 2 development

Highway management improvement

Bangkok

： Industrial Agglomeration

Project Progress Status

：Conceptual Stage

：Feasibility Study Stage

：Construction Stage

：Operation Stage

Legend

Selected Prospective Projects in Mekong Sub-region

Chennnai Port- Ennore Port accesssway

Outer ring road in Chennai: Phase 2

Ennore Industrial Park and SEZ

Chennai

High speed railway: Chennai - Bangalore

Expressway: Chennai - Bangalore

Bypass and express ways around HCMC

Cai Mep–Thi Vai port: development, improvement

Rail link: HCMC – Vung Tau, HCMC –My Tho

Petro-chemical complex

Transmission line: Can Tho – HCMC

Power plants in O Mon

Software technology park

Ho Chi Minh City (HCMC)

Selected Prospective Projects in Mekong Sub-region (2013)

ASEAN and East Asia still need infrastructure.

13

High demand of IDE-GSM in Asia

• High demand for transportation infrastructure

Heavy congestion in the big cities

Very poor road in rural areas

• Severe financial constraints of each country

• Poor socio-economic data

• Infrastructure improvement may change economic structure

• To prioritize infrastructure development, economic analysis using regional-wide dataset must be conducted and cannot wait for accurate data to be provided.

14

IDE-GSM: the model

We don’t have:

• Government expenditure

• Taxes

• Savings and investments

• Foreign exchange fluctuation

• Inflation/Deflation caused by change in the money supply

We have

• Firms

• Consumers

• Natural population growth

• Migration within a country

• Price fluctuation by domestic and international competition

15 15

Specification of Industries

Agricultural Sector

Mobile Labor Transport costs Manufacturing Sector

-Input-Output structure in M-sector

Service Sector

Transport costs

-Dixit-Stiglitz monopolistic competition

Transport costs All regions

-Dixit-Stiglitz monopolistic competition

-Constant-returns technology & monopolistic competition

16

Utility function

15

1

5

1

S

jjMA

S

j

MASjM

j

A QQQU

1

0

1

1

0

1

)(

)(

S

S

SS

S

jM

jM

jMjM

jM

jj

diiqQ

diiqQ

n

SS

n

MM

•Agriculture • 5 manufacturing • Service

17

1

0

1

) (

A

A

A

A

di i q Q R

A A

Wage and Price Index

• Agricultural sector

Nominal wage

Price Index

18

)(

)(

)()()(

1

rprL

rFrArw A

A

AA

)1(

1

1

)1()1()()(

AAA

R

s

A

srAA TrprG

• Manufacturing sector

Price index

Nominal wage

19

)1(

1

1

)1()1(11)()()()()(

MMMMM

R

s

M

srMMMMM TsGswsAsLrG

wM (r)

AM (r)

1

M E(s)s1

R

TrsM 1 M

GM (s)(1 M )

1

M

GM (r)1

1

Wage and Price Index(cont.)

• Service sector

Price index

Nominal wage

)1(

1

1

)1()1(1)()()()(

SSSS

R

s

S

srSSSS TswsAsLrG

S

SS

sGTsYrArw S

R

s

S

rsSS

1

)1(

1

1)()()()(

20

Wage and Price Index(cont.)

Labor migration

• Labor migration among sectors within a region

• Labor migration in a country

21

Agriculture 3.8 0.61 0.040

Automotive 4 0.57 0.020

Electronics 6 0.56 0.026

Textile 8.4 0.64 0.018

Food 5.1 0.61 0.033

Others 5.3 0.59 0.172

Service 3 1.00 0.687

Mining 5.6 0.25 0.004

Parameters specifying each industry

• Elasticity of substitution:

• Share of labor input:

• Share in consumption:

22

Transport Costs

First, we estimate firms’ behaviors on modal choice by MNL.

• Firms tend to use trucks for domestic partners.

• Larger the distance between trading partners, the more likely the firms are to choose air or sea.

• Machineries industry tend to use air transports for foreign partners without smooth border transaction at the borders.

Truck as a basis

Coef. S.D. Coef. S.D.

Abroad 3.573 *** 0.736 2.915 *** 0.428

ln Distance (Food as a basis) 0.444 *** 0.170 1.268 *** 0.167

*Textiles 0.104 0.126 -0.151 0.094

*Machineries 0.300 ** 0.135 0.112 0.086

*Automobile 0.201 0.174 -0.104 0.154

*Others 0.148 0.106 -0.068 0.066

Constant -5.711 *** 0.760 -9.621 *** 0.993

Country dummy: Indonesia as a basis

Philippines -0.336 0.470 0.364 0.446

Thailand -2.239 ** 0.904 -0.794 0.624

Vietnam -2.483 *** 0.683 -0.437 0.419

Statistics

Observations

Pseudo R-squared

Log likelihood

Air

1,312

0.3407

-321.5

Sea

23

Second, we develop a simple linear transport cost function with the parameters on time sensitiveness.

• A representative firm in machineries industry will make a choice between truck and air transport and choose the mode with the higher probability of the MNL result.

• A representative firm in the other industries will make a choice between truck and sea transport and choose the mode with the higher probability of the MNL result.

Air Sea Air Sea

Food 60,300,000 3,699 19,254 371

Textiles 2,022,900 11,218 2,968 825

Machineries 44,009 1,899 361 229

Automobile 225,394 7,693 886 628

Others 684,540 5,909 1,634 520

Domestic International

Probability Equivalent Distance with Truck (Kilometer): Domestic and International Transportation from Bangkok

24

Modal Mix of E&E Industry (from Ayutthaya, Thailand)

To Lamphun, Thailand To Phnom Penh To KL

Food Textile Machineries Automobile Others

ctime s 15.7 17.2 1803.3 16.9 16.5

Truck Sea Air Unit Source

cdist M 1 0.24 45.2 US$/km Map

Speed M 38.5 14.7 800 km/hour Estimation

ttrans MDom

0 11.671 9.01 hours Estimation

ttrans MIntl

13.224 14.972 12.813 hours Estimation & Map

ctrans MDom

0 190 690 US$ Map

ctrans MIntl

500 N.A. N.A. US$ Map

25

Making Scenarios: Trade Costs in GSM 26

26

Development of economic corridors

27

Scenario Speed of

trailers

Time required at

national borders

Money costs at

national borders

Without Baseline 38.5 km/h 13.2 hours 500 USD

With Corridor +

Facilitation 60 km/h 6 hours 250 USD

Upgrading roads

≈

Customs Facilitations

How to Estimate the Economic Impacts?

28

2010 2015

with infrastructure projects

already completed by 2015

Baseline

with no or limited development

projects by 2020

Alternative Scenarios

with corridor development by 2020

2030

Compare the differences

(10 years after)

GRP

Economic Impact

(% compared with

the baseline in 2030)

2020

Example 1:

Kuala Lumpur-Singapore high-speed rail (HSR)

• Only passengers can use HSR, no freight

• HSR’s cost per km:

8x of conventional train

2x of road

1 /2 of air

• Waiting time at stations

Longer waiting time => Lower frequency

• Over 200km/h (average speed)

The speed of conventional train is 40km/h

29

- 0

0 - 10

10 - 20

20 - 30

30 - 50

50 - 100

100 - 200

200 - 300

300 -

(Million USD)

Malaysia 1,005

Labuan -3 Johor 611

Kedah 17

Kelantan -8 Melaka -16 N. Sembilan -8 Pahang -26 P. Pinang 51

Perak 39

Perlis 3

Selangor 136

Terengganu -16 Sabah -24 Sarawak -32 Kuala Lumpur 278

Singapore 88

TOTAL 1,093

KL

Singap

ore

Iskand

ar

B. Pah

at

Mu

ar

Ayer Kero

h

Seremb

an

Pu

trajaya

Example 1: SG-KL HSR, result

30

31

https://www.nst.com.my/news/nation/2018/07/395111/japanese-study-finds-malaysia-could-make-us1589bil-annually-hsr

http://beltandroad.zaobao.com/beltandroad/news/story20180729-879026

Reported by Local

Newspapers

Example 2:

Infrastructure development in ASEAN

• Proposing a set of infrastructure development projects to achieve both high economic impact in one country and positive impacts in many provinces.

• Assuming additional NTB reductions.

32

Economic Impact in 2040

Source: ERIA (2010)

Japan's ministerial meeting for economic cooperation and infrastructure development (2013)

Analysis (previous version) is adopted as regional/national policy

Master Plan on ASEAN Connectivity (2010)

Source: ASEAN (2010) Source: Prime minister's office website

33

Comprehensive Asia Development Plan, independent research by ERIA

Example 3: Highways in Mongolia

• Some positive impacts are observed mainly Eastern Mongolia and Northeast China.

Source: Calculated by IDE-GSM

(Compared with Baseline in 2030)

Example 3: Changes in Traffic

Source: Calculated by IDE-GSM

• Traffic between Northeast China and Northwest China, through Mongolia East-West increased.

(Compared with Baseline in 2030)

Example 4:

Suez-Koper Link

• We incorporated NUTS 2 level data for EU while EU is still considered as a rest of world in the model.

• A new shipping link between Suez, Egypt and Koper, Slovenia will benefit some regions in Europe and many regions in East Asia.

Economic Impact in 2030

36

Dilemma in making a CGE model in Asia

2018/11/8 37

More realistic model

Lack of data

Parameters and data required

Need to estimate/interpolate data/parameters

Model should be realistic but simple

Dilemma in Making a CGE model in Asia

2018/11/8 38

Huge diversity in economic structure

Drastically changing in economic structure

Parameters and data required

Parameters/data should be changing dynamically

Model should be realistic but simple

Dilemma in Making a CGE model in Asia

2018/11/8 39

Rapidly developing infrastructure project

No integrated data set/ standard model

High demand in CGE analysis on the project

Model/data should be developed at a time

Model should be realistic but simple

Conclusion

• A CGM model for Asia should be realistic, but simple.

• There is no data and no standard model in Asia, but cannot wait for perfect data/model. Agile system development is required.

• There might be a ‘free lunch’ in Asia …? Often, the source of finance for infrastructure is outside country and term and conditions are not determined. So, a tool like IDE-GSM is required, before conducting detailed CBA.

40