(Institute for Transport Studies, University of Leeds, 2003)
This note focuses on projects which may cause significant economic restructuring. Two specific examples are used for illustrative purposes, these are: new urban rail lines and major new barrier crossings (see Sections 2 and 3). However, these serve simply as examples of a much wider issue - restructuring effects are equally likely to be stimulated by a major new road in a resource rich region, or other types of major project.
The key issue is that whenever projects bring about a large step change in transport costs, there is a stimulus for a reorganisation of economic activity outside the transport sector. For example, a manufacturer may change the source of their raw materials or relocate their production, a retailer may centralise their operations to serve a larger market area, or a farmer may change their crops to a more marketable combination and hire labour.
These effects are as likely to be stimulated by a major new road in a resource rich region as by either of the examples above. The key point is that these projects raise issues for modelling (Section 4) and for appraisal (Section 5). This note seeks to explore the issues, but also to give firm guidance to project evaluators on the practical methods which can be used within what is a complicated and difficult area The recommendations are summarised in Section 6.
It is perhaps worth noting that these projects are – in effect – extreme cases of generated traffic (see Note Treatment of Induced Traffic [Link]); anything which applies there applies in these cases with greater emphasis. Finally, there is also a link with the advice provided in the Note Low Volume Rural Roads [Link]; both that case and the cases discussed in this note involve restructuring outside the transport sector – the difference is that the low volume rural roads case is simple enough to deal with by focusing on producers in only one sector, or a small number of sectors. This note addresses projects where the economic context is more complicated.
Transport networks can play a key role in the economic development of countries and regions (SACTRA, 1999) []. Communications were crucial to ancient civilisations throughout the world, were crucial in the industrial revolution, and remain crucial in the knowledge-driven economies today. Whilst telecommunications offer an alternative to transportation in some situations, research suggests that so far the development of telecommunications has tended to complement, rather than substitute for, the demand for transportation.
Transport plays a special role in economic development since it provides the connection (across distance) between consumers and producers in a wide range of markets:
- Transport brings together employers and employees in the labour market;
- It facilitates trade by bringing together buyers and sellers, and their products;
- It has the potential to radically change the pattern of activity in the land market; and
- To some extent it facilitates other economic sectors such as the capital market, education, government, culture and tourism.
Figure 1 gives a overview of some of these connections through the transport sector, emphasising the point that transport facilitates a wide range of economic activity, and affects a wide range of economic decisions. It is not surprising that transport receives attention in economic development policies throughout the world (eg. CEC, 2002 []; World Bank, 2002 []). But what evidence is there that individual transport projects can contribute significantly to economic change?
A common theme in the literature on transport and the economy, is that the presence of adequate transport is a necessary condition for economic development, but is not sufficient to guarantee it. Other key local conditions include available human capital/skills in the workforce; availability of suitable land and property; and a stable political and legal context. It is also very much easier to make the case for ‘restructuring effects’ where there is a well worked-out economic strategy for the country or local area concerned. One of the roles of the analyst in economic appraisal is to investigate and report on this wider context, because it can be highly relevant in the process of decision.
When the conditions are right, transport projects have the potential to contribute to significant local, regional or even national economic restructuring. In this note we consider two examples (Sections 2, 3), and consider the implications for modelling (Section 4) and appraisal (Section 5). A summary of key recommendations is made in Section 6.
Urban rail projects, whether metros, trams or heavy rail, have the potential to alter the structure of the urban economy. Whether the impact is significant or not will depend upon the scope of the network, and the quality and price of the service offered.
Among the key effects identified by monitoring studies are:
- Expansion of labour supply at the centre, and to a lesser extent throughout the network,
- Bringing workers from peripheral districts dominated by agriculture into the manufacturing and services sectors in the city;
- Opening up access to education;
- Intensification of land use around rail stations, especially in the urban centre, with corresponding changes in property prices; and
- Relocation of business towards rail served corridors and expansion of existing businesses.
These often take the form of bridges (eg. Jamuna bridge; Tagus river crossing) or tunnels (Japan) across bodies of water, although there are other possibilities (eg. a causeway across a flood plain, mountain crossings). In some cases a barrier can be political, eg. a recently opened national border, rather than physical.
Figure 2: Estuary or river crossing - sketch
In transport network terms, these projects serve to provide a link between disjoint networks. In other words, they link up the networks on either side of the barrier, creating a wider range of destination and route options for people and goods.
In economic geography terms, these projects create new opportunities for spatial competition and economies of scale. Suppose these are significant. Firms located on each side of the barrier gain the opportunity to compete in each other’s markets, which may initiate a process of consolidation and concentration. Residents located on each side of the barrier gain the opportunity to relocate, taking advantage of property market inequalities across the barrier. Labour becomes mobile across the former barrier, expanding the pool of workers and skills available on both sides. Taking advantage of the improved accessibility of the areas on both sides of the crossing, firms and households are attracted to the region.
As proposed by the ‘new economic geography’, clusters of skills (human capital) and knowledge may develop, leading to endogenous economic growth (Fujita, Krugman and Venables, 1999) [].
The modelling task in cases of economic restructuring can be onerous, if it is decided to attempt a full economic model of the problem. Therefore, it is worth reviewing the following three broad options at the start.
(I) Fixed Trip Matrix. The reality is that most World Bank transport project appraisals are conducted using a fixed trip matrix model. This is the most simplistic form of transportation network modelling, since it assumes that there are no changes in trip ends as a result of the project – in other words, there is no reorganisation of freight flows, no shifts in population or changes in production in the local/regional economy, and no changes in the places where people do business, learn or shop. Bearing in mind the restructuring effects discussed in Sections 2 and 3, it may seem perverse to model these projects using such a method, however there are clear reasons why this is sometimes done.
The overriding reasons are lack of data and limited availability of transport modelling expertise. With a suitably-experienced professional in place, and a modest input of resources appropriate to the size of the project, it should be possible to produce not only a ‘Do-Minimum’ but also a ‘Do-Something’ (with project) trip matrix, which will enable a Variable Trip Matrix appraisal (II) to be conducted. Data availability is, of course, a limitation, however within the cost of a major project it is expected that money will be spent on gathering the data needed to demonstrate the economic case for the project.
A secondary reason for the frequent use of Fixed Trip Matrix models, even for projects with restructuring effects, is (we suggest) that there is an ingrained tradition of doing so which tends to perpetuate itself. If a project is appraised using a Fixed Trip Matrix, everyone ‘knows’ what has been done. The FTM is widely seen as a reasonable, conservative estimate of the out-turn trip matrix, whereas forecasts of generated or redistributed trips are seen as open to a wide range of methodologies and – frankly – to exaggeration where this adds to the case for the project.
The difficulty with this viewpoint is that generated and redistributed traffic, if not forseen, can seriously undermine the benefit estimates for the project – in either direction. On one hand, unforseen traffic can rapidly lead to congestion on new infrastructure (even in well-developed networks – see the M25 around London). On the other hand, some projects provide substantial new trip opportunities which lead to economic restructuring and generated/redistributed traffic. Williams and Moore (1990) [] have shown that under plausible assumptions about demand and supply conditions, the extra benefits due to induced traffic can be large.
In summary, if it can possibly be avoided, we do not recommend using a Fixed Trip Matrix model when there are likely to be economic restructuring effects.
(II) Variable Trip Matrix – transport model only As discussed in the Note Treatment of Induced Traffic [Link] transport demand that varies between the Do Minimum and Do Something can be modelled through elasticity methods or through the modelling of specific behavioural responses (e.g. mode shift).
(III) Variable Trip Matrix – full transport and economic model. Sometimes, the transport model only approach (II) may face problems, because the restructuring caused by the project is too deep-rooted, large or complicated for a pure transport model to predict. A good test is: has there been a similar project whose effects can reasonably be transferred across (in the form of transport demand elasticities) into the local/regional economy of this project? If not, a model of the restructuring effects themselves may be required.
In this case, the analyst faces two important decisions:
1. To model production OR NOT.
Freight demand, and the demand for business travel, are driven by firms and other production sector organisations. The key decisions are: what to produce, in what locations, using what inputs? The outcomes of these decisions directly affect the loads on the transport network.
Advantage: The key advantages of attempting a spatial production model are that the modelling process can yield insights into the linkages between transport and the local economy; and can help to focus the policy process on users’ needs.
Disadvantage: On the other hand, such models are data hungry - detailed spatial Input-Output matrices are required; they are more suited to policies or whole networks than to individual projects; and the methodology is still subject to research questions.
2. To model residential location choice OR NOT.
Land-use and transport interaction (LUTI) models offer the chance to forecast changes in land use in response to changes in the transport network, and the feedback effects into transport demand. In particular, these models have an advantage in predicting shifts in patterns of residential location (Mackie et al, 2001) .
For urban public transport projects, and for inter-urban projects which open up new commuting possibilities, the residential location response can be an important source of demand. These models are also very demanding in terms of data requirements.
If it is decided NOT to attempt a production sector model or a residential location model, then a great deal of importance is placed on certain key parameters at the trip generation stage and trip distribution stage in the transport model. These are:
- The planning assumptions: concerning population and employment in each zone in the base year, and population and employment in each zone in future years both with the project and without;
- Demand and distribution: the procedures for deriving trip rates from a zone and the pattern of demand between different zones (eg. the gravity model or other aggregate demand modelling techniques);
- Freight demand matrices (are these realistic in the do-something scenario?);
- Elasticities of demand for travel in response to reductions in generalised cost.
In constructing defensible models, it is always advisable to validate the model against the existing situation (see Note Demand Forecasting Errors [Link]). Sometimes it is also useful to seek out evidence from comparable situations elsewhere, though caution should be exercised when considering the transferability of travel behaviour particularly between countries. Peer review of the demand forecasts by appropriately experienced individuals will add further to their credibility. Of the above, the base planning assumptions, and the impact of the project upon them are probably the most critical for the appraisal.
The economic appraisal issues are closely related to the modelling issues raised in the last section.
In transport cost-benefit analysis, user benefits are usually measured in the transport market itself. This means that the change in consumer surplus is calculated from the flows and costs on the transport network using the Rule-of-a-half (see Framework [Link] and the Note Treatment of Induced Traffic [Link]).
When there are significant restructuring effects, it is possible - although not necessarily true - that benefits might arise in other markets (especially land, labour, product markets) which are not fully reflected by the Transport CBA result (see SACTRA, 1999 ; Mackie et al, 2001 ). The main causes of these additional benefits (or disbenefits) are: monopoly power; non-constant returns to scale; and externalities in the wider set of markets affected by the restructuring (see Framework - Section on Wider Economic Impacts) [Link]. The example given in Section 3 of economies of scale in production leading to endogenous economic growth is one example of this.
The problem we face currently is that there is lack of experience in using the alternative model forms for transport project appraisal, and the initial results from research versions of models are not yet validated. For the moment, there is insufficient evidence to be able to say with confidence whether these models offer a better method than the conventional Variable Trip Matrix method.
The advice, therefore, is to measure the user benefits in the transport market, using the rule-of-a-half as set out in the Framework [Link], and to base this analysis on a Variable Trip Matrix model of demand.
This means that particular attention should be paid to the realism of the key parameters listed in Section 4:
- spatial population and employment assumptions;
- trip generation and distribution functions; and
- elasticities of demand.
The core recommendation is to undertake a good quality transport cost-benefit analysis. Where transport schemes are incremental in nature and the network is already well-developed the assumption should be that there are no additional wider economic impacts. But in the cases considered in this note, the transport benefits will not capture the sum total of the benefits to the economic system as a whole. Nor will it allow properly for the final incidence of benefits as they percolate through the economy into changes in prices, wages and land rents.
In such cases, we recommend the use of a qualitative approach. A commentary should be provided, focusing on the following issues:
- the linkages between transport and the regional economy, focusing on the specific linkages impacted by this project (which markets are expected to be affected – housing? labour? goods and services? - through improvements in which types of transport – commuter transport? inter-city business travel? freight and logistics? - is there any relevant evidence about the direction and magnitude of these effects?);
- the competitive advantage of the regions connected by the improved transport link, in traded sectors (for example, competitive advantage may flow from natural resources and their role in agriculture, fishing, tourism or manufacturing, or it may flow from a regional pool of skilled labour with specific skills) – this will influence the changing pattern of employment and output as a result of the project.
Market research methods including interviews, surveys and focus groups can be used to gain an insight into these effects. The research would be conducted primarily among employers (and potential employers) in the regions concerned, aiming to understand how the project will impact on their decisions about production, employment, location, and transport. Similar work could be conducted with property market professionals if commuting patterns and residential location are vulnerable to change. For further information on these techniques, see Barrett (1999) []. This work would provide qualitative evidence in support of the core transport cost-benefit analysis.
The overall recommendations of this Note are as follows:
- The first priority in any transport project appraisal should be to get a realistic model of the existing trip matrix – the steps needed to achieve this are discussed in the Framework [Link] and the Note Demand Forecasting Errors [Link];
- It is not recommended to evaluate projects where there are likely to be economic restructuring effects using a Fixed Trip Matrix model – instead, a modest expenditure on data and modelling expertise is required, in order to produce a defensible (if not perfect) ‘Do-Something’ matrix – the appraisal may then be conducted in Variable Trip Matrix terms, as set out in the Framework [Link] and the Note Treatment of Induced Traffic [Link]; and
- Except where there is clear evidence to the contrary ‘wider economic impacts’ should not be claimed as a source of additional benefits which help to justify schemes which fail in transport terms. This is because many transport benefits are in fact converted through the economic process into final impacts on prices, wages, rents, etc. Thus the final impacts of transport projects may represent double counting, as well as being very difficult to measure (see Framework [Link] and Note Distribution of Benefits and Impacts on Poor People [Link]). Evidence of wider economic impacts maybe derived from a full economic and transport model, in the case of a large urban project, or market research methods in the case of a low volume rural road (see Note Low Volume Rural Roads [Link]).
- A good qualitative assessment of the wider economic impacts, using market research methods, has been found to be useful and practical, and will help to underpin the changes in trip matrices predicted for the ‘Do-Something’ scenario.
[Available online at:
[] Mackie PJ, Nellthorp J, Kiel J, Schade W and Nokkala M (2001), IASON Project Assessment Baseline. Deliverable 1. IASON Project (Integrated Appraisal of Spatial economic and Network effects of transport investments and policies). Delft: TNO Inro. [Available online at: http://www.wt.tno.nl/iason/]