Evaluating the Impact of Improvements in Transport: The Case of Transport Corridors in Southern Africa

1
Evaluating the Impact of Improvements in
Transport The Case of Transport Corridors in
Southern Africa

• Sandra Sequeira
• Harvard University
• August 31, 2006

2
Outline of the Discussion

• Motivation Nexus between Transport, Trade and
  Growth
• Policy Intervention Rehabilitating Transport
  Corridors
• Evaluation Project
• Goal measure micro-level impact of changes in
  transport costs
• Primary and auxiliary evaluation questions
• The importance of a micro approach
• Detailed research design
• Empirical Strategy, Robustness of results
• Implementation Details

3
Trade, Transport and Growth

• Freight Costs Firms
• Trade
• Investment
• Congested Ports
• Non-performing Rail
• Poor maintenance of Roads
• GROWTH

4
Transport, Trade and Growth (Cont)

• Hummels (1998), Venables (1999), Yeats (1996)
• increasing importance of transport costs,
  especially for Sub Saharan Africa (SSA)
• SSA commodity based economies, high
  volume, low value goods
• long distances, several border posts
• accumulated neglect of primary transport
  networks
• Goal Reduce costs of exporting raw materials and
  
• importing inputs and final goods
• TRADE-LED GROWTH

5
Policy Goals Guiding the Transport
Infrastructure Shock

• Important to understand
• Extent, Determinants and Impact of transport
  Policies
• Impact of transport costs on firms
• Motivating questions
• Do investments in transport infrastructure
  reduce transport costs for firms?
• To what extent and how do transport costs affect
  firm productivity, profits, employment generation
  and export propensity?

6
Policy Intervention What, Where, Why?

• Project Evaluation of Transport Improvements in
  South Africa and Mozambique
• Why? High impact area due to
• High overland transport costs
• Long distances to ports
• Commodity based economies

7
Geographic Focus

• South Africa
• Important industrial, mining and agricultural
  region far from the coast (600 - 1,200 km)
• Current accelerating growth strategy constrained
  by transport (Eg Import Parity Pricing, Limited
  Response to Asian Commodity Boom)
• Acute sense of importance of transport, 15-20 of
  GDP equivalence

8
Geographic Focus (cont)

• Mozambique
• Third lowest road and railway densities in
  Southern and Eastern Africa
• Acute sense of importance of transport (product
  based studies 60, 80 of product value)
• Historically a transport economy serving a strong
  hinterland (South Africa, Zimbabwe, Malawi,
  Zambia, Botswana, Swaziland)

9
Policy Intervention to be Evaluated

• Promotion of Spatial Development Initiatives
  (SDIs)
• Goal stimulate growth, trade and investment
  along key international transport corridors
• Strategy bilateral agreements that identified
  cluster cross-sectoral investment opportunities
  for PPPs. Transport the promotion of PPPs based
  on matching users to infrastructure providers
• Target Regions key transnational transport
  corridors
• Expected Outcomes virtuous cycle of expanding
  regional markets, increased regional integration,
  a regional approach to FDI, job creation, Small
  and Medium Size Enterprise (SMMEs) development,
  strengthened planning and managing capacity of
  local governments and rapid economic growth.

10

• SDIs in Southern and Eastern Africa

11
(No Transcript)
12
Focus of the Evaluation Project

• Maputo Development Corridor (MDC) consists of
• Gauteng-Maputo toll road (2001)
• Private Port of Maputo (2004)
• Wittbank-Ressano Garcia Rail line (operational by
  the end of 2006)
• Why focus on the Rehabilitation of the
  Gauteng-Ressano Garcia Rail Link
• i. Rail volumes guarantee ship calls at the port
  and justify regular dredging
• ii. Technological developments and
  containerization broadened the range of
  rail-friendly cargo, also rail relative
  cost-effectiveness. Allows for measurement of
  cross-industry and cross-firm effects
• iii. Development Debate Redux role of rail in
  development and growth (Rostow, Fogel, Fishlow,
  etc) measuring the magnitude of the reduction in
  transportation costs associated with rail, who
  benefits, what are the effects on trade,
  structure of regional output, rise in per capita
  income, etc.

13
(No Transcript)
14
Goal of Evaluation Project

• Empirical investigation aims to
• i. Measure the impact of changes in transport on
  firm behavior, investment and growth.
• ii. Measure the interaction between transport
• modes and transport corridors
• iii. Contribute to the broader cost-benefit
  analysis
• on the cost-effectiveness of investments in
• transport

15
Evaluation Questions

• Primary evaluation questions
• i. Does introducing competition between
  transport corridors increase efficiency in the
  transport system and reduce transport costs for
  firms?
• ii. Do reductions in transport costs promote
  regional growth through improved trade and
  investment opportunities?

16
Evaluation Questions (cont.)

• Auxiliary questions
• a. Transport Analysis
• i. Are investments in rail and road
  complements or substitutes?
• ii. What is the nature and magnitude of
  spillover or network effects
• across transport modes and transport
  corridors?
• b. Firm-level Analysis
• i. Do investments in infrastructure
  translate into uniform reductions
• in unit transport costs for all
  commodities and firms?
• ii. Do changes in transport costs affect firm
  dynamics regarding
• investment patterns, export behavior,
  factor productivity and
• patterns of entry and exit from
  the market?
• iii. Is there variation on firms' responses
  to changes in transport costs?

17
Evaluation Questions (cont.)

• c. Growth Analysis
• i. Through which mechanisms do investments
  in
• transport infrastructure affect firms'
  productivity,
• investment and growth?
• ii. Are investments in transport
  infrastructure a cost-
• effective and sustainable
  strategy to promote
• growth?
• iii. Do SDIs cause convergence or divergence
  with
• other non-targeted regions?
• iv. Are the benefits commensurate with the
  costs? What are
• the social and economic
  rates of return to investments in
• transport?

18
Innovative Research Design

• Policy Relevance
• a. Learning potential for future SDIs Policy
  Cascade through
• NEPAD
• b. The importance of micro-data
• 1. Understanding distributional effects of
  transport investments leads to a more targeted
  demand-driven transport policy
• 2. Matching effects to different elements of
  transport corridor (rail, road, ports)
• 3. Survey methodology allows us to get at micro
  effects that escape macro evaluations concerning
  types of investment in transport eg. trade off
  between transport time and cost for different
  firms and industries

19
Innovative Research Design (cont.)

• c. New approach to impact evaluation of transport
  infrastructure
• Existing studies on transport infrastructure
  are based on
• 1) ex ante cost benefit analyses
• 2) aggregate inferences based on crude country
  level
• measurements of port efficiency,
  road density and
• length of railways and their impact on
  trade,
• investment and growth.
• Before and After analysis of distributional
  effects allows us to get at Who benefited? By
  how much? Can we accurately measure the impact on
  trade, on employment, investment and growth?

20
Detailed Research Design Indicators

• Estimating the Effect of Competition in the
  Transport System
• Secondary Data Collection along targeted and
  alternative corridors

21
II. Estimating the effect of transport corridors
on growth and development (Secondary
Data Collection on treated and control
regions) Indicators
22
II. Estimating the effect of transport corridors
on growth and development (cont.)
(Secondary Data Collection on treated
and control regions) Indicators
23
III. Estimating the Effect of Transport Costs
on Industries and Firms (Primary
Data Collection)

• Empirical Strategy
• i. Population of Interest
• ii. Sampling Criteria
• iii. Survey Output and Indicators
• iv. Robustness Tests
• v. Implementation

24
Empirical Strategy

• i. Cross Sectoral Firm Level Survey
• Geographic Location
• i. Treatment regions of Mpumalanga, Limpopo
• and Northeastern Gauteng in South Africa
• Maputo, Southern Gaza and Southern
  Inhambane in
• Mozambique
• ii. Control regions of Southern Gauteng,
  Kwazulu
• Natal, Western and Eastern Cape in South
  Africa and
• the Mozambican provinces of Northern
  Inhambane,
• Sofala, Manica, Nampula, Niassa and Zambézia

25
(No Transcript)
26
(No Transcript)
27
Empirical Strategy (cont.)

• ii. Sampling Criteria
• 800 firms in South Africa and Mozambique
• Cross section of industries Mining,
  Manufacturing, Retail, Tourism and Agribusiness.
• Stratified Random Sampling determined by
  industry, location and firm size

28
Empirical Strategy (cont.)

• Treatment and control groups will be matched
  by propensity scores based on a vector of firm
  and industry level characteristics.
• iii. Survey output
• Firm-level organizational, financial, transport
  and performance indicators
• Chief indicators Profits, Factor Productivity,
  Investment and Risk, Export propensity, Inputs,
  Sales, Output mix.

29
iv. Robustness Tests

• I. Durban and Maputo
• de facto alternatives or complements?
• Is Maputo a real choice?
• Comparability of port technologies, cost and
  productivity
• Depth
• Productivity
• Storage Capacity
• Distances and Cost
• Safety

30
Durban Quay Lengths and Depths

31
Durban vs Maputo (cont)

• Productivity and Growth Potential
• Berth Occupancy 30 Maputo
• 100 Durban
• Crane Moves per hour 15 TEU in both
• Safety Maputo was first African certified ISPS
  port
• Storage Capacity
• more in Durban but limited expansion, 3 days free
  storage
• less in Maputo but more potential for expansion,
  offering 21 days free storage

32
Road and Rail Distances (km)
33
Road and Rail Costs

• Road costs Durban cheaper, more backhauling
• Rail costs Maputo cheaper
• Port Charges Maputo cheaper
• Unit Rands per 20 ft container

34

• Historical Legacies-
• Path-Dependence of Route Choice? No
• In the past, ships would flip from Durban to
  Maputo depending on how many days delay. 45 of
  SA cargo in 1975 was going through Maputo.
• Most firms willing to undergo short-term
  relocation costs even if small cost savings per
  ton km
• Common Practice of Hedging between Ports
• The Curse of the Hub and Spoke Model?
• No clear consensus on the direction of the
  shipping industry
• Economies of flow from consolidation in hubs
  versus higher timing and inventory costs
• A system of more loops with smaller vessels bears
  less risk and could eventually turn out to be a
  cheaper option than running very large vessels on
  few loops. Most likely that both business models
  will prevail.

35
Nature and Expected Impact of the Intervention?

• Prior to the Intervention.expect change

36
Expected Impact of the Intervention
37
Potential Growth of Port Throughput with Rail
38
Comparability of Treated and Control Areas?I-
Variation on Geographic Location of Production

• Sugar growing areas in South Africa are located
  on the coastal strip and hinterland around Durban
  as well as on the area east of Nelspruit, along
  the Maputo corridor. In Mozambique sugar grows in
  central provinces. Each of these areas is
  naturally placed in the catchment area of
  different corridors.
• Fruit Western and Eastern Cape, Nelspruit,
  Mpumalanga, Limpopo, Nampula
• Retail Gauteng, Western Cape, Durban, Maputo,
  Nampula
• Cashews Central, Southern and Northern
  Mozambique
• Car Manufactures Eastern Cape, Pretoria, Durban

39
Mining
40
Mining
41
II- Variation on Firm Size
42
Robustness Tests (cont.)

• Spurious Effect Asymmetric Information?
• What drives the choice to ship through Maputo
  Cost or information? Encouragement Design through
  MCLI (Maputo Corridor Logistics Initiative)
• Time Horizon short enough to avoid confounding
  effects, long enough to see a difference.
  Preliminary qualitative
• research 8 months
• Search for natural discontinuities based on
• a. Geographic location
• b. Product characteristics

43
v. Implementation

• Research Partners
• Witts University, SA- Centre for Applied African
  Micro-Economic Research (CAAMER), Economics
  Department
• School of Development Studies, University of
  Kwazulu Natal, University of Cape Town
• Universidade Eduardo Mondlane
• Universidade Católica da Beira
• Stats SA, CSIR, INE

44
v. Implementation (cont.)

• Advantages of University partner
• - more cost effective
• - better use of local knowledge
• better acceptance of the surveying exercise
• due to higher credibility (problem of Survey
  Fatigue and Mistrust)
• - higher returns to knowledge generation
• promoting applied policy research at Universities
  (capacity building)
• incentives for sustainable ongoing data collection

45
Tentative Implementation Schedule
46
Annex I- Trucking Industry

• Proposed Project Expansion of the Trucking
  Survey currently underway in 6 African countries
  to South Africa and Mozambique
• Goal Measure the Impact of Investments in Rail
  Infrastructure on the Trucking Industry
• Methodology before and after survey to a sample
  of 20 formal firms (PPS) and 30 owner truckers in
  each country
• Primary Evaluation Question
• Do investments in rail infrastructure affect
• i. The structure of the trucking industry?
• ii. The cost of trucking services?
• iii. Employment in the trucking industry?

47
Survey Data

• Ownership characteristics
• Indicators of firm productivity
• Percentage of backhaul trips
• Load factor (ratio of actual load to capacity on
• loaded trips)
• Level of competition in the trucking industry
• Exposure to corruption
• Exposure to overloading fines and choice of
  transport routes
• Average number of hours driven
• Major constraints for growth and formalization
• Nature of interactions with the formal sector

48
ANNEX-II

• Beira Corridor
• Key Transport Infrastructure
• Rail a) Sena line connecting to Tete, including
  spur lines to the Moatize
• coal mines and to the Malawian border (600 km)
  
• b) Machipanda line connecting to the
  Zimbabwean border (300 km).
• Repairs and maintenance concessioned to the
  Beira Railroad Company (CCFB), a consortium
  dominated by the Indian Rites and Ircon
  International who also own the 25 year lease over
  the Machipanda line since 2004.
• Port of Beira concessioned in 1998 to Cornelder
  Mozambique. Undergoing significant change
• Main Problems Dredging of port, delays with
  rail, Zimbabwes economic collapse
• Potential growth vast Moatize hard coking coal
  deposits in Tete Province as well as mineral
  agri/fisheries, tourism, cotton and sugar
  industries along the Zambezi valley.

49
Beira Corridor
50
Nacala Corridor

• Key Transport Infrastructure
• Port of Nacala concessioned in 2005 to Corredor
  Desenvolvimento de Nacala (CDN)- consortium
  between Central East African Railways CEAR, Edlow
  Resources and Railroad Development Corporation
• Rail Nacala-Lichinga, Nacala-Malawi
• Major Problems unstable concession of port and
  rail, significant delays in the
• rehabilitation of the port and the major
  rail lines
• Potential for Growth Capture significant cargo
  from Malawi and
• Zambia, CDN hold enormous potential
  in
• agriculture, forestry and
  fisheries, mining and tourism
  development.

51
Nacala Corridor
52
Nacala Line Forecast (thousand tons)
53
Machipanda Line Forecast (Thousand Tons)
54
Sena Line Forecast (Thousand Tons)
55
Mozambican Exports (US Millions)
56
Mozambican Imports (US Millions)