Is the Economy Shaping the Project, or Justifying It?
Choosing infrastructure around what producers need to compete
I have spent most of my career preparing infrastructure projects, producing studies and reviewing work prepared by others. In some assignments, the client was looking for demand forecasts that would justify a project it had already chosen. A port needs cargo, a railway needs freight and an industrial zone needs tenants. The temptation is to treat the volumes required to make the project viable as the demand it is likely to attract.
World Bank guidance on toll-road PPPs identifies political and commercial pressures involving both awarding authorities and bidders. Where procurement rewards high traffic forecasts without adequately testing their realism, it can place “substantial pressure on the traffic forecaster to produce sufficiently optimistic forecasts”. These incentives help explain how the commercial objective can start to shape the demand assumptions.
The consequences may emerge years later, when the infrastructure is built but the expected volumes fail to materialise. Depending on the contractual arrangements, users may face higher tariffs, the government may have to cover a shortfall, or investors and lenders may incur losses. Reaching financial close does not make the demand appear.
Calculating the volume a project needs to be viable is useful work, but it is not a demand forecast. A forecast needs evidence about the firms or users expected to generate demand, and a client’s determination to reach financial close cannot supply it. Even a sound forecast, however, does not settle the investment choice. Firms may expect to use a new port because it offers a better service, while their main obstacle to competing lies elsewhere. The appraisal must therefore establish how much the port would improve their competitiveness and whether another intervention would achieve more.
In the first article of this series, I proposed a test: describe what a project is for without naming the asset. “Build a port” does not say what economic result we are trying to produce. “Make processed cashew competitive in a defined export market” begins to. Once the asset has been named, can the economic analysis still change it?
This article focuses on infrastructure investments justified by the production, trade or industrial development they are expected to support, whether publicly or privately financed. Access to essential services and territorial cohesion are also legitimate objectives, and investments serving them must be assessed against those objectives.
Four propositions carry the rest of this article:
- Before choosing infrastructure to support production, establish what firms intend to produce, for which markets and with what service requirements.
- Compare how much each investment would improve producers’ ability to compete, and at what cost.
- Different processing stages require different infrastructure, whose cost and reliability can affect which activities are viable.
- Under uncertainty, distinguish what can proceed now, what depends on specific conditions being met, and which options are worth preserving for later.
Identify the constraints holding production back
A damaged road or an unreliable electricity supply may justify attention, but deciding where to invest requires understanding how these problems affect production. Sector policy can identify the activities a country wants to develop. Infrastructure planning should test which services those activities need and where investment would make the greatest difference. The findings may also lead the country to reconsider its original production ambitions.
Public authorities and their advisers should work with producers and prospective buyers to understand what prevents firms from winning or fulfilling orders. Transport costs may be one obstacle; unreliable power, product standards or access to working capital may be others. Discussions with businesses, supported by available data and comparisons with competing suppliers, can help identify which constraints infrastructure could address. Where production is still only an ambition, the analysis must examine what would make firms invest.
The proposed project should then be assessed against the improvement it would bring. If unreliable transport prevents producers from meeting delivery dates, would the investment provide a dependable service at a price they can afford?
An independent evaluation of Millennium Challenge Corporation-funded road rehabilitation in Senegal, summarised in an MCC evaluation brief published in December 2021, re-estimated the economic rate of return at 5.3 per cent for the RN2, against 10.9 per cent at appraisal, and at 2.0 per cent for the RN6, against 11.3 per cent at appraisal. MCC treats 10 per cent as the threshold to proceed with an investment, so on the revised numbers neither road clears the bar its own funder sets. The revisions reflected lower assumptions for baseline traffic and traffic growth on both roads, a better assumed initial road condition on the RN2, and higher actual project costs and assumed operating costs on the RN6. Separately, the evaluation found that shorter travel times had not translated into significantly lower transport prices. The appraisal has to test both the benefits attributed to an asset and how those benefits are expected to reach the producers whose competitiveness it is meant to improve.
A terminal can reduce logistics costs while producers still lack the inputs, working capital, skills or customers needed to expand. Where forecast traffic depends on resolving these constraints, the appraisal should identify the actions required and test what happens if they are delayed.
The investment decision then turns on how much improvement each option offers, at what cost and over what period. The comparison should include operational or regulatory changes, any complementary investments required, and evidence that firms are likely to respond.
What a country wants to produce should shape its infrastructure
Take an iron ore deposit. If the plan is to export ore, the system has to move bulk material reliably from mine to ship: rail capacity, stockyards, bulk handling, loading, and the utilities the mine itself needs.
Now plan to make steel instead. A gas-based direct reduction plant supplying electric furnaces requires dependable gas, a substantial and reliable electricity supply, industrial water and a site suited to the process. The transport system still has to bring ore to the plant, and now has to move steel products to customers. Energy and markets may decide where the plant belongs, which connections matter, and what has to be built first.
Same mine. Same geography. Two different infrastructure systems.
Neither strategy is automatically right. Ore quality, energy costs, scale and markets decide it. But a railway and an ore terminal can serve the mine exactly as intended and still leave the steel industry the country said it wanted without the conditions it needs to compete. A forecast of tonnes through the port cannot make that choice for a country.
Kamoa-Kakula illustrates the same logic in copper. In December 2025 its new smelter produced its first copper anodes, moving the export product from concentrate at around 45 per cent contained copper to anode at 99.7 per cent. Ivanhoe Mines expects the project’s logistics costs to fall sharply as a result. The same decision required dedicated protection against voltage fluctuations on the domestic grid and created a sulphuric acid stream for local sale.
The same applies to timber, cashew and cotton: each processing stage changes the transport services, utilities and standards required. An ambition to add value locally therefore needs to specify the processing stage envisaged and the services it would require. Sector policy and infrastructure planning should develop and test those choices together.
Specifying the stage does not settle the commercial case. If the required services cost too much or cannot be delivered within a credible period, a different processing stage, another location, a regional arrangement or a slower sequence may be more viable. Both the production objective and the infrastructure design must remain open to revision.
Looking beyond the scope of the assignment
The scope of the assignment can limit this comparison. A port authority may commission a detailed expansion study without asking whether an intervention elsewhere along the corridor could deliver greater benefits. That alternative may fall under another institution’s mandate or budget. Each institution can carry out its own responsibilities while the comparison between investments remains unaddressed.
Public planning needs to bring these alternatives into the same decision process, involving the institutions, investors and users concerned. Before major commitments are made, a clearly designated body must have the authority to compare the options and resolve questions of scope, sequence and funding. This can be done for a specific investment without waiting for a complete national industrial strategy.
Boundaries can be legitimate. Existing commitments, service obligations and political priorities are real constraints, and they should be stated as boundaries. A decision already taken should not later be presented as the conclusion of an analysis that was never allowed to question it.
If the case for public infrastructure depends on firms expanding output, their requirements and intentions have to be tested. A sector target is a reason to open the conversation, not evidence that the investment will follow.
Commit without pretending to know the future
In North Kivu, a study of mini-grids operated by Virunga Energies found electricity demand hard to predict even with detailed survey data on the households and firms being connected. In response to low and uncertain demand, the operator and its partners built industrial zones to host processing units and lent to small businesses through a bank partnership, with repayments collected through electricity purchases. Organising the conditions for production is work with a budget and an owner, not an assumption in a spreadsheet.
No government can know precisely how production will evolve over the life of a railway, a port or a grid. Infrastructure can enable activities that were previously impossible, so requiring firm commitments from every future user would rule out worthwhile investments. Some investments can proceed despite that uncertainty; others need firmer commitments or can be prepared for later.
Some investments hold up across several credible production scenarios. That has to be shown for the specific investment rather than the category: would this connection, at this location and this capacity, still be worth its cost if the preferred scenario arrived late, or arrived in a different form? A universal need does not validate a particular asset.
For investments that depend on identifiable users, construction can be tied to milestones such as a commercial commitment, financial close or demonstrated demand. Those conditions must reflect what each party can control. A private investor may need confidence that the public connection will be delivered before committing to a factory, while the public authority may need evidence that the factory will proceed. The investment sequence must make both commitments possible.
Some early decisions preserve room for later expansion: reserving land, protecting an alignment or designing a site to accommodate additional equipment. These choices have costs, including the alternative uses forgone, and should be justified by the value of retaining that flexibility. The case is particularly important where decisions are difficult to reverse. Equipment can often be added later; changing a location or recovering land already developed for another use is much harder. The appraisal should identify these choices before they constrain future development.
One page, before the asset becomes the assumption
Before the preferred asset organises every subsequent study, I would ask for a short decision brief answering six questions.
- Which productive activity and market are we trying to serve, and under what credible demand scenarios?
- What prevents producers from competing today, and what evidence identifies the main constraint?
- What would this intervention change, what else has to happen, and who owns those complementary actions?
- How does it compare with alternative ways of achieving the same outcome in terms of cost, timing and risk?
- What can we commit to now, what must wait for specific conditions to be met, and what should remain open?
- What evidence would lead us to revise the preferred option, its scale or its timing?
The brief does not replace the analysis. It should record what the analysis supports, what remains uncertain, and what has to be resolved before the next commitment. Retaining the original project is a valid conclusion if the alternatives have genuinely been tested.
What are we trying to make competitive?
Naming an asset gives a programme something concrete to announce. It can also narrow the economic choices before they have been examined. Once the assignment is to make a particular port or industrial zone viable, subsequent studies may strengthen the investment case without establishing whether it is the best way to support the intended production.
Designing backwards starts with the productive activity and its market. Establish what would allow firms to compete, identify the services they need, and compare the investments that could provide them. Where the production is still prospective, test what would make it viable and put those enabling conditions in place alongside the infrastructure.
At CPCS, being evidence-based is one of our core beliefs: we do not bend our findings to suit a client’s agenda. For me, that means helping clients act on what the analysis shows, including when it points towards a smaller investment, a different sequence or a decision to wait.
I would encourage anyone considering an infrastructure investment to ask one question: have we established what producers need in order to compete, or only what the project needs them to produce?
→ Subscribe here to receive the next Decided Upstream article in English or French.
Note on sources.
The description of procurement pressure on traffic forecasters is from Matt Bull with Anita Mauchan and Lauren Wilson, Toll-Road PPPs: Identifying, Mitigating and Managing Traffic Risk, PPIAF and the World Bank Group, chapter 7. The Senegal figures, the 10 per cent threshold and the finding on transport prices are all from the Millennium Challenge Corporation’s Unlocking Market Access via Road in Senegal, December 2021.
The North Kivu findings are from Lunanga, Stoop, Verpoorten and Desbureaux, From Demand Deficit to Development Strategy: Navigating Mini-Grid Viability in a Fragile Context, IOB Working Paper 2025.09, June 2025. The Kamoa-Kakula example draws on Ivanhoe Mines’ announcement of first anode production, 2 January 2026; the expected reduction in logistics costs is a company expectation at that date.
The gas-based direct reduction route is described in the IEA’s Iron and Steel Technology Roadmap (2020). Comparative cost and constraint analysis is illustrated in the World Bank’s Light Manufacturing in Africa (2012). CPCS’s approach to evidence is set out in its core beliefs.