Insight, Infrastructure & Regional Development
A corridor is measured in asphalt laid and hours saved. Neither number tells you the thing that matters, which is that lowering transport costs for some places raises the relative cost of everywhere else. Corridors redistribute market access before they create it.
The Joint Research Centre and DG INTPA have been running a programme called Corridors and Urban Systems in Africa, which underpins the Global Gateway corridor investments. Its most recent output is a pilot analysis of Abidjan and Ouagadougou, the two largest cities on West African Global Gateway Transport Corridor No. 2, covering demography, socio-economic development, access to services, transport and connectivity and environmental vulnerability.
The programme's ambition is unusually broad for transport work. The first phase employed more than 140 quantitative indicators to identify 55 strategic corridors, narrowed to a shortlist of 11 using 32 indicators covering economic welfare, equity, social inclusion, environmental impact and impact on vulnerable groups. That is a considerably better instrument than kilometres and travel times, and it deserves saying, because the criticism that follows is not that nobody is measuring.
It is that the unit of analysis is the corridor, and the interesting effects happen to places that are not on it.
Paul Nugent's chapter on the Abidjan to Lagos route, in the collection Transport Corridors in Africa, makes a point that ought to be uncomfortable for any corridor programme. Almost all corridors are labels placed upon routes that already existed, often for a very long time, and most analysis infers what a corridor is from its effects rather than from any intrinsic property.
That matters for evaluation in a specific way. If the route already carried the traffic, then the intervention is an upgrade to an existing flow rather than the creation of a new one, and the counterfactual is not "no road". It is "the same road, slower". Those produce very different estimates of impact, and programme documents routinely use the first.
The JRC's own consolidated assessment contains a related and easily missed observation. Along the corridors, lower-performing areas are found on the core infrastructure lines themselves, including the Abidjan to Bouaké stretch, while the high performance concentrates in the large agglomerations. Being on the corridor is not the same as benefiting from it. Much of the value pools at the ends.
Here is the argument, and it is not complicated.
Transport infrastructure lowers the cost of reaching a market from the places it serves. Market access is a relative quantity. A trader in a town on the upgraded road can now deliver to the capital more cheaply, more predictably and in better condition than before. A trader twenty kilometres off it cannot.
Before the upgrade, both faced a bad road and competed on roughly equal terms. After it, one of them has a structural cost advantage on the same product in the same market. Nothing was taken from the second trader. Their costs are unchanged. Their competitive position is worse.
A corridor redistributes market access before it creates it, and the losers experience no measurable event. No road was closed. No price was raised. Their margin simply erodes as somebody else's falls, and there is no line in any monitoring framework where that appears.
This is the standard result in economic geography rather than a novel claim, and it has a second-order effect that matters more. Lower transport costs also make it cheaper for goods to flow into the region from the capital or the port. A local producer of a tradeable good who was previously protected by the cost of reaching them now faces competition from a larger, cheaper producer at the other end of the improved road. Corridors expand markets in both directions, and the direction that gets modelled is usually the export one.
Whether the net effect on a given place is positive is an empirical question with a genuinely uncertain answer. It depends on what that place produces, whether it is a net seller or buyer of tradeables, whether its producers can scale, and whether anybody there has the working capital to take advantage of a market that has just become reachable. None of that is knowable from a gravity model.
There is a specific fact about Corridor 2 worth putting alongside the programme documents.
A World Bank assessment of Burkina Faso's regional connectivity found that the Ouagadougou to Lomé corridor serves as the artery for about 40 per cent of all cargo entering the country, while the Ouagadougou to Abidjan road and rail corridors play their crucial role in allowing Burkina Faso's exports to reach global markets.
Those are different functions, and they suggest an obvious question that a corridor investment should answer explicitly rather than implicitly. Is the objective to lower the cost of what the country sells, or of what it buys? They are not the same intervention and they do not benefit the same people. Export corridor improvements benefit producers of tradeables and the traders who aggregate them. Import corridor improvements benefit consumers and the firms that use imported inputs, and they compete with domestic producers of substitutes.
A programme optimising for trade efficiency in aggregate can improve both totals while making a specific group of domestic producers worse off, and report success accurately.
The evaluation design follows directly from the mechanism, and it is a field design rather than a modelling one.
Sample by distance from the corridor, not along it. Settlements at zero, ten, twenty-five and fifty kilometres from the improved road, matched on what they produce and their baseline market access. The comparison that matters is between the corridor town and its neighbour, and it is the comparison nobody constructs because the programme boundary is the corridor.
Ask traders and transport operators, not planners. The people who know whether a road changed anything are the ones deciding daily what to load, where to sell it, whom to sell it to and what it now costs. Freight rates, waiting times at the border, load factors, and which markets became reachable are all held in the heads of operators and in almost no dataset.
Measure prices in both directions. Farm-gate prices for what the region sells, and retail prices for what it buys. A corridor that raised the first and lowered the second is working. One that lowered both is displacing local production, which is not necessarily wrong and is certainly worth knowing.
Establish who could not participate, and why. Working capital, storage, aggregation, quality standards and information. A market that becomes reachable is only reachable by someone who can finance a consignment and meet a buyer's specification. This is the same structural point about who can adopt that runs through adaptation research, arriving in logistics.
Take the baseline before construction. Almost none of this is recoverable afterwards. A corridor programme with a ten-year horizon that starts measuring at completion has forfeited the only comparison worth having, and the cost of a baseline is a rounding error against the asphalt.
Global Gateway is being evaluated, fairly or not, against a comparison with other infrastructure financiers, and the claim distinguishing it is that European investment produces better development outcomes rather than merely assets. That claim is currently supported by the design of the programme rather than by evidence about its effects, which is the gap between commitment and outcome we have described before in the European relationship with Africa.
The evidence that would support it does not come from more indicators. The JRC has 140 of them and they are good ones. It comes from a different unit of analysis, sampled off the corridor as well as on it, and from asking people who move goods for a living what changed. That is a modest addition to a large investment, and it is the difference between reporting that a corridor was built and knowing what it did.
The Lab works on this through entering a new context and measuring change, and the distributional questions here are the ones we ask in the places where an infrastructure decision lands on a household rather than a map.
If you are financing or evaluating a corridor and need the off-corridor comparison built before construction starts, tell us what you need to know.
This is an independent insight piece by Transitions Lab. For the Lab's applied work, see Measuring Change. See also The EU and Africa on the distance between European commitment and measured outcome, Behind the Border on why variance rather than distance is what stops firms trading, and Europe Has Enough Demonstrations on sampling the people a programme did not reach. To discuss a study, see Contact.