# The Road That Doesn't Count (plain version)

**Verdict:** Transparency rankings of foreign-funded infrastructure often measure the shape of the money and the publishing culture of financiers, not the openness of governments. Three boundary rules, settled before scoring, fix the comparison.

## The story, plainly

Two expressways, financed from the same country.

Road one is Uganda's Entebbe-Kampala Expressway. The Export-Import Bank of China lent the Government of Uganda USD 350 million at 2 percent fixed, over a 20-year maturity with a 7-year grace period; Uganda added USD 126 million; a Chinese state contractor built it; it opened in 2018. The money passed through the treasury and left a paper trail: loan agreement, debt record, budget lines, audits. Every monitoring definition counts it as foreign-funded, because the definitions were written for exactly this shape: follow the loan, find the project.

Road two is an unnamed expressway of roughly two billion dollars, a real case used as an archetype, built under a build-operate-transfer concession. A foreign company financed it, built it, and operates it for decades, repaying itself from tolls before handing it over. The state granted the concession, but no loan touched the treasury. Ask definitions built on sovereign loans and aid flows whether it is foreign-funded and the answer is no. It drops out of the study before scoring begins.

That is the core defect: definitions keyed on financing instruments (sovereign loans, official development assistance) see the finance that still flows the old way and go blind to concessions, equity stakes, joint ventures, supplier's credit, and guarantees. The blindness is not neutral, and it is not about any one country's money: both roads here have the same financier country. The problem is the shape, not the flag.

Two distortions follow.

First, scoring that credits any published information, whoever published it, flatters projects financed by institutions with strong publishing cultures. A funder's rich project page can earn a government a "transparent" score while that government publishes nothing. You graded the bank's website.

Second, scoring that demands records a structure never generated condemns the newer shapes. A build-operate-transfer project whose concession was negotiated directly with the company that financed and built it never held a tender. Writing a zero for the missing tender record measures the deal's structure and calls it secrecy. A zero can mean three things: the record was hidden (the only real failure), the record never existed, or the record does not exist yet. And where a structure skips the tender, the public's questions do not vanish; the disclosure expectation shifts to the concession award and its terms.

## The three boundary rules

1. **Define eligibility by purpose, material stake, and control, not instrument.** Public asset or public service under government-granted concession? A material project interest (financing, ownership, revenue, guarantee, concession) held by a foreign-controlled party? Yes to both: it counts.
2. **Ask why a record is missing before scoring it.** Hidden scores zero; never-existed is excluded with the expectation shifted to the records that do exist; not-yet-due waits. The denominator adjusts so honest reporting never looks worse than silence.
3. **Credit the publisher, and give every figure a passport.** Classify information by who controls its publication. A funder's page proves existence; only government publication is government disclosure. Every figure carries source, retrieval date, currency, and exchange rate with its date.

## Key figures

| Item | Value | Source |
|---|---|---|
| Road one loan | USD 350 million, 2 percent fixed, 20-year maturity, 7-year grace, signed 18 May 2011, Exim Bank of China | AidData project 14235 (retrieved 25 July 2026) |
| Road one, Uganda contribution | USD 126 million (total USD 476 million) | AidData project 14235; Uganda Ministry of Works |
| Road one contractor / opening | China Communications Construction Company / June 2018 | Uganda Ministry of Works |
| Road two | Kampala-Jinja Expressway, indicative USD 1.5 billion, 30-year build-operate-transfer concession, contracting authority UNRA | Uganda PPP disclosure portal entry, Internet Archive capture 12 August 2022 |
| Documents listed, loan-financed road | 50 (World Bank NERAMP, Uganda, P125590) | World Bank document API, retrieved 25 July 2026 |
| Documents listed, concession road | 3 (feasibility study, RFQ, request for qualification) | Uganda PPP disclosure portal, archived capture |
| Kampala-Jinja co-financing | AfDB approved USD 229.5 million for phase one, so the concession is a hybrid, not purely private | African Development Bank appraisal report, 15 March 2021 |

## Provenance

Both roads are public record, both in Uganda, both contracted by the Uganda National Roads Authority, and both part of the same Kampala ring-road strategy. One arrived as a sovereign loan and one as a toll concession, and that is the only difference the argument needs.

- AidData, "China Eximbank provides $350 million preferential buyer's credit for Entebbe-Kampala Toll Road Construction Project": https://china.aiddata.org/projects/14235 (retrieved 25 July 2026)
- Uganda Ministry of Works and Transport, Kampala-Entebbe Expressway: https://www.works.go.ug/component/k2/item/25-kampala-entebbe-expressway
- World Bank, Uganda North-Eastern Road Corridor Asset Management Project (NERAMP), P125590: https://projects.worldbank.org/en/projects-operations/project-detail/P125590 (document list from the World Bank document API, retrieved 25 July 2026)
- Uganda PPP disclosure portal, Kampala-Jinja Expressway entry, last updated 25 October 2021. The portal is offline as of 25 July 2026, so the record is the Internet Archive capture of 12 August 2022: http://web.archive.org/web/20220812111959/http://disclosure.pppunit.go.ug/project/11/kampala-jinja-expressway-kje-ppp

Three honest qualifications. The Kampala-Jinja entry was captured while the project was still in procurement, so some of its thinness is stage-appropriate rather than a disclosure failure. The concession is a hybrid, not purely private: the African Development Bank approved USD 229.5 million for phase one, which cuts the same way as the argument, since the disclosure that does exist comes from a funder rather than the government. And a Wayback index check found seven archived project pages on the PPP portal, none of them the Entebbe-Kampala Expressway: the loan-financed road does not appear in the PPP register, and the concession road does not appear in the aid-flow datasets. That absence was verified, not assumed.

Figures are as reported by these sources. Nothing here scores, ranks, or characterises any country.

**The closing line:** Before you compare governments, make sure you are not comparing their bankers.