Who Runs the Gateway: Chattogram Port's Foreign-Operator Concessions

1. Verdict

Confirmed on the terms, uncompleted on the transfer — with the mechanism now documented in unusual detail. The single strongest piece of evidence is CPA's own financial evaluation of DP World's proposal, not outside criticism of it.

The thesis holds on both legs. First, on pricing power: CPA's own internal financial-adviser analysis — obtained and published by Bonikbarta from shipping-ministry documents — found that under DP World's proposed revenue-sharing structure at a representative $120/TEU gross revenue level, CPA's net margin would fall to $17.18/TEU, versus $65.25/TEU if CPA continued operating NCT itself [E/C]. That is a ~74% reduction in CPA's own per-container economics under the operator's own submitted terms, calculated by CPA's own evaluation committee, not by protesters or opposition politicians. Second, on opacity as mechanism: three separate award processes at Chattogram (NCT, Laldia, Patenga) have each proceeded on a government-to-government (G2G) or single-negotiation basis rather than open competitive tender, a pattern a High Court judge explicitly found exceeded the interim government's legal authority under the PPP Act 2015 (in a split verdict later left standing on procedural grounds) [C].

Strongest evidence for the thesis: CPA's internal financial analysis (above) — a primary-tier, self-authored government document showing the counterfactual is dramatically more profitable for the state than the negotiated deal on the table.

Strongest evidence against/complicating the thesis: The deal, as of the most recent evidence (September 2026), has still not been signed — 30+ months after the original 2019 MoU, through two governments, multiple court challenges, and at least four rounds of worker strikes. The negotiation and legal-review process, however flawed procedurally, has repeatedly stopped or delayed the most operator-favorable terms (the $42/TEU rate workers alleged BIDA was pushing; the 30-year term DP World sought against a 15-year financial model) from being finalized. This is evidence that Bangladesh's institutions — courts, CPA's own technical staff, organized labor — have functioned as a genuine, if messy, check, which cuts against a clean "opacity as mechanism" reading and toward "contested, adversarial, and still-open process."

Worst clause/process failure to target first, if the thesis is right: The revenue-sharing (percentage-of-realized-tariff) royalty model DP World proposed, as opposed to the fixed-revenue-per-TEU model CPA's own approved RFP required and that Laldia/Patenga both use. A tiered percentage-of-revenue model both caps CPA's upside if NCT's performance improves and makes CPA dependent on auditing DP World's own reported revenue to verify what it is owed — the single mechanism CPA's own evaluation committee flagged as creating "substantial risks in evaluating actual revenue figures" [E/C]. A renegotiation fixing this one term (mandating fixed revenue/TEU, as in the RFP) would resolve much of the fiscal leakage risk without touching the sovereignty/security questions, which are a separate and harder problem.

2. Findings

Finding 1 — Deal terms, exactly: NCT (DP World) — most detailed on record; Laldia (APM Terminals) — largely settled; Bay Terminal (PSA/DP World/Abu Dhabi Ports) — framework only; Patenga (RSGT) — signed, operating

NCT [E/C]: Operator: DP World (UAE, state-owned). Terminal: 950m quay, 5 berths (14 of the port's 18 quayside gantry cranes), 292,287m² yard, 1.19 million TEU annual handling capacity, built 2007 for Tk2,000 crore [C]. Original MoU: February 17, 2019, under the Awami League government. Structure debated: PPP, G2G basis. Term: RFP/financial model built on a 15-year operate-and-maintain period (with up to 6-month transition); DP World's active bid counter-proposal (per its June 2026 submission reviewed by Bonikbarta) instead sought a 30-year term while still relying on the 15-year financial model's assumptions [E/C] — a term-mismatch CPA's negotiators had not resolved as of the most recent reporting. Transaction adviser: IFC (World Bank Group) [C]. Financials: DP World proposed a $50 million upfront fee plus a tiered royalty (percentage-of-realized-tariff) structure ranging from $42.50/TEU (avg. revenue <$105/TEU) to $141.50/TEU (avg. revenue ≥$210/TEU), with a $52 million minimum revenue guarantee against CPA's current NCT revenue of roughly $145.74 million/year [E]. CPA's own approved RFP and draft concession agreement instead specified a fixed-revenue-per-TEU model. Current operator: Chattogram Dry Dock Limited (a Bangladesh Navy entity), which took over July 7, 2025 after Saif Powertec's 18-year private-operator contract expired [C]. Status as of Sept 19, 2026: unsigned; fresh worker protests announced for Sept 22 and 27, 2026 [C].

Laldia (LCT) [C]: Operator: APM Terminals (Denmark/Maersk). Structure: PPP, G2G basis, Design-Build-Finance-Operate-Transfer (DBFOT). Term: 33-year initial concession, extendable by 15 years, up to 48 years total. Scale: 3 jetties, 800,000 TEU/year capacity. Model: fixed-revenue-per-TEU (unlike DP World's NCT proposal). Transaction adviser: IFC. Status: agreement concluded/signed in the Nov 2025 period per Prothom Alo reporting, ahead of and without the extended public contestation NCT faced, though also criticized by labor/SKOP for lack of disclosed detail [C].

Patenga (PCT) [E/C] — the one fully executed and operating precedent: Operator: Red Sea Gateway Terminal (RSGT, Saudi Arabia, 40%-owned by Saudi PIF). Concession signed December 6, 2023 under the Hasina government, G2G basis with Saudi Arabia. Term: 22 years. Capacity: 450,000-500,000 TEU/year, $240M facility, RSGT capex ~$170M (debt+equity). Model: fixed-revenue-per-TEU (per Bonikbarta's NCT-analysis article, which uses Laldia and Patenga as the RFP-consistent comparators against DP World's outlier proposal) [E/C]. This is Bangladesh's first operating foreign-terminal-concession precedent and the only one with a multi-year operating track record to test tariff and investment behavior against (see Finding 4/7).

Bay Terminal [C]: Not a single concession but four planned terminals with different financing/operating splits still evolving: CT-1 and CT-2 (container) MoU'd to PSA Singapore and DP World respectively (2021-2024 reporting, unconfirmed as final signed concessions); a Multi-Purpose Terminal 3 (Abu Dhabi Ports Group proposed, ~$1bn) and Terminal 4 (gas/oil, unallocated). Marine infrastructure (breakwater, access channel, dredging) is separately financed: World Bank/IDA ~Tk10,272 crore, CPA's own funds ~Tk4,636 crore [C], via a phased 2024-2031 project — i.e., the state builds and finances the shared marine infrastructure while foreign operators build and run the individual container terminals on top of it, a capital-risk split the original brief specifically asked about. This split is confirmed: Bangladesh/World Bank capital is exposed on the enabling infrastructure; operator capital is exposed only on their own terminal superstructure and equipment — asymmetric risk-sharing that favors the operators' downside protection.

Finding 2 — The award process: G2G/sole-negotiation dominates; competitive tender is the exception, not the rule — and independent domestic bidders exist but were sidelined at NCT

Evidence tier: [C]. Of the four terminals reviewed, none was awarded through an open, multi-bidder international competitive tender with published scoring. Patenga and Laldia were G2G-negotiated (Saudi Arabia and Denmark/UAE-adjacent government channels respectively) with IFC as transaction adviser but no evidence of competing international bids being solicited or scored. NCT's process is the most contested and most documented: two Bangladeshi business groups — Saif Powertec (the incumbent, 18-year operator, alleging capacity to handle up to 1.8 million TEU/year domestically) and MGH Group (which claimed its own 15-year concession model could generate ~$1.68 billion in total payments to CPA, a higher-revenue offer than competitors according to its CEO; separately, TBS reporting on reviewed documents put DP World's per-container offer at $93.50–97.50 against MGH's $98.50 — second-sourced corroboration that the domestic bid was competitive on price) — were reportedly in the running before "lobbying by two advisers of the Muhammad Yunus-led interim government" narrowed the field toward DP World, per Northeast News' account of the internal process, which also alleges MGH had positioned itself as a DP World local partner and secured an informal understanding with the son of the then-Shipping Adviser before the then-National Security Adviser (later Foreign Minister) intervened to delay it [single-source allegation — Northeast News only; treat as allegation pending corroboration]. Separately, a five-way race emerged for the adjacent CCT/GCB terminals (DP World, RSGT, MGH, plus local entities), with Prothom Alo (En) and TBS — citing port sources and port data (NCT 44%, GCB 36%, CCT 16%, Patenga ~4%) — reporting that a DP World NCT+CCT combination would control about 60% of container handling, versus ~55% for an RSGT CCT+GCB combination [C — two independent outlets, same port-data basis]. A Saif-Cosmos-Everest consortium (linked to two sitting MPs) also submitted an NCT proposal on April 28 [TBS] — further evidence of domestic interest, though local proposals are reportedly not under active consideration.

• Evaluation authority: CPA board recommends; Shipping Ministry and the PPP Authority (headed by BIDA Executive Chairman Ashik Chowdhury, who "prioritised bringing a foreign operator into Chattogram Port management" according to port officials cited by Bonikbarta) hold effective decision authority; the Chief Adviser's office/Cabinet Committee on Economic Affairs approves. A 12-member CPA "special support team" was formed in mid-2026 specifically to feed the negotiating committee technical/financial data — itself a sign the process was still being institutionally reconstructed as of June 2026, more than seven years after the original MoU [E/C].
• Confidence band: 70-80% that the NCT process, specifically, involved material off-process influence favoring DP World at least at some stage (multiple independent outlets converge on this, though the most detailed account is single-sourced); 90%+ confidence that no NCT, Laldia, or Patenga award met a standard "3+ qualified bidders, published scoring" competitive-tender bar, based on the consistent G2G/sole-negotiation framing across all primary reporting reviewed.
• Falsification condition status: "A genuinely competitive bid with 3+ qualified bidders and published scoring produced the award" — not met for any of the three finalized/near-finalized terminals reviewed.

Finding 3 — Fiscal forensics: CPA's own numbers show self-operation beats the DP World proposal by a wide margin

Evidence tier: [E] (CPA's internal financial-adviser analysis, as reported by Bonikbarta from shipping-ministry documents — the single best primary-tier artifact located in this research). See §4 table for the full breakdown. Headline: at $120/TEU gross revenue, self-operation nets CPA $65.25/TEU; the DP World proposal nets CPA $17.18/TEU — DP World's offered revenue share is "almost equivalent to the port's own net margin and is insufficient even to cover the CPA's fixed administrative costs" per the evaluation committee's own conclusion [E]. Separately, port worker leaders (a less independent source, but consistent in direction) alleged the negotiating range under discussion in early 2026 ran from CPA's own preferred $105/TEU figure down to a BIDA-pushed $42/TEU figure [C] — both below the $65.25/TEU self-operation benchmark. Fiscal year 2022-23 CPA revenue from NCT alone: Tk1,216 crore, with Tk574 crore net income after costs [C] — a ~47% net margin under self-operation, broadly consistent with the per-TEU analysis above.

• CPA also flagged a structural risk specific to any revenue-share (as opposed to fixed-fee) model: verifying the concessionaire's actual revenue would be "extremely difficult and complex," requiring specialised expertise and additional oversight CPA does not currently have [E] — i.e., even the nominal royalty rate understates the real risk, because enforcement/audit capacity is itself a gap.
• On the $50M upfront fee: CPA's financial adviser revised the assumed return-on-investment on that fee from 3% (achievable via BDT fixed deposits) to 12% in a Feb-2025-dated recalculation, which lowered the implied minimum concession fee from $99.54/TEU to $94.96/TEU [E] — a modeling choice that itself materially changes the headline economics and was not obviously justified by market fixed-income rates at the time (3% appears to be the realistic reinvestment rate for a USD-denominated upfront payment; using 12% is a favorable assumption for the deal's optics).
• Regional benchmarks: not independently sourced in this research pass at comparable per-TEU granularity for Colombo, Mundra, Port Klang, or Djibouti; this remains a gap. Patenga's own $170M capex against a 500,000 TEU facility and 22-year term is the only regional-adjacent, same-country data point with confirmed figures, but its per-TEU revenue-share terms were not located in the sources reviewed (unlike NCT's, which were obtained via leaked/published evaluation documents).
• Confidence band: 85-90% that CPA's self-operation counterfactual, as CPA's own committee modeled it, is materially more profitable per TEU than DP World's proposed terms — this is now a documented internal finding, not an outside estimate.
• Falsification condition status: "CPA's projected net revenue per TEU rises under the concession vs the self-operation counterfactual" — not met; the opposite is CPA's own documented finding. This is the single most direct falsification-condition result in either of the two research directives completed to date.

Finding 4 — Tariff and pricing power: contested and structurally ambiguous

Evidence tier: [C]/gap. No source located specifies with contractual precision who holds final tariff-setting authority under the draft NCT concession — whether CPA, a BERC-equivalent regulator (none clearly exists for ports as it does for power), or the operator itself within a band. The royalty-vs-fixed-fee distinction (Finding 1/3) is itself a proxy fight over this question: a percentage-of-realized-tariff royalty structure gives the operator a direct financial interest in raising realized tariffs (since CPA's cut rises with the operator's own pricing), whereas a fixed-fee-per-TEU model is tariff-neutral to CPA's take. This is a structurally important, underappreciated point (analysis, not a sourced finding): DP World's proposed model does not just risk lower CPA revenue at current tariffs — it creates an incentive alignment where CPA's own financial interest would tilt toward tolerating operator tariff increases, which is the opposite of consumer/exporter protection. No source located models a specific 20-30% handling-charge increase scenario or its RMG-export pass-through; this is a genuine gap requiring either the actual signed concession's tariff-revision clause (not yet public, as the deal is unsigned) or CPA/BERC-equivalent rate-setting rules not identified in this research.

• Confidence band: low/informational — this finding should be revisited once (if) a signed concession document becomes available.

Finding 5 — Sovereignty and security provisions: the strongest, most concrete evidence in the entire research set

Evidence tier: [E/C]. In the December 4, 2025 split High Court verdict, Justice Fatema Najib — the senior judge, whose opinion found the process unlawful — explicitly grounded her security finding in specific, named facts: the port's proximity to Myanmar, its location near Bangladesh's largest naval base, and its adjacency to the country's submarine cable [E/C]. This is a judicial, on-the-record finding, not press characterization or political talking point — a rare instance in this research where a security objection is backed by a named judge citing named infrastructure rather than being asserted rhetorically. The same judgment separately found the interim government lacked constitutional jurisdiction to make a major, multi-decade policy commitment of this kind, reasoning that an elected parliament should have the opportunity to weigh in, especially given a February 2026 election was imminent at the time of the ruling [E/C]. The junior judge on the same bench dismissed the petition entirely on standing grounds, producing the split verdict that sent the matter to a single-judge bench (Justice Zafar Ahmed) for final disposal. Justice Najib further found the 2019 MoU itself was 'a non-binding MoU, not a G2G deal,' noting the Dubai government never signed it — a direct challenge to the G2G framing both governments used [New Age]. The clearance was then upheld by the Appellate Division on March 12, 2026, which dismissed the leave-to-appeal and removed any legal barrier to CPA contracting a foreign operator [TBS]; the underlying sovereignty/security objections were not thereby resolved on the merits, only procedurally cleared.

• Data-access and step-in-rights contractual language was not located in any source (deal remains unsigned/undisclosed in full text) — a genuine, currently unfillable gap.
• Currently, operational control sits with Chattogram Dry Dock Limited, a Bangladesh Navy entity — meaning as of this research, the "worst case" sovereignty scenario has not yet occurred; the Navy interim-operation arrangement is itself informative as a revealed preference for keeping strategic control in state/military hands pending resolution.
• Confidence band: 85%+ that the security argument (naval base/submarine cable/Myanmar proximity) is a substantive, not merely rhetorical, concern given it was adopted by a sitting High Court judge with named specifics, though it is also true that a second judge on the same panel found no legal merit at all — genuine judicial disagreement exists, not manufactured controversy on one side.

Finding 6 — Labor: workforce protections are not contractually confirmed; friction is real and escalating, not organized-political theater alone

Evidence tier: [C]. At NCT specifically: Saif Powertec (the prior private operator) reported employing ~3,800 workers and stated, upon its 2025 contract expiry, that all employees would remain and cooperate with the Navy-led interim operation [C] — no evidence of an immediate mass layoff at the Navy-handover stage. Separately, earlier (2024-era) reporting on the original DP World plan cited port-worker concern that over 1,000 workers at NCT could lose jobs if a foreign operator took over [C] — a materially different, smaller headcount figure than Saif Powertec's own 3,800, suggesting confusion or inconsistency in public figures about what "NCT workforce" actually means (direct Saif Powertec employees vs. broader port-linked labor). No draft concession text with explicit employment-guarantee or headcount-protection language was located — this remains contractually unconfirmed. What is confirmed: multiple, escalating, dated strike actions — a 6-day work stoppage in late January/early February 2026, an indefinite strike called for February 8, 2026 (later reportedly paused after government assurances), and fresh protests announced September 19, 2026 for September 22 and 27 — indicating labor unrest is a recurring, not one-off, feature of the process and has materially disrupted port operations on at least three separate occasions across 2026 alone [C].

• Confidence band: high confidence (routine, multiply-corroborated wire reporting) that labor disruption is real, sustained, and operationally significant; low confidence on any specific promised-vs-contractual headcount protection figure, because no contract text is public.

Finding 7 — Precedent: DP World and Djibouti (Doraleh)

See §5 table.

Finding 8 — The interim-government legitimacy question: directly and substantively litigated, not merely argued in the press

Evidence tier: [E/C]. This is unusual among the research questions in that it received a direct judicial ruling, not just political commentary (see Finding 5). Justice Najib's finding that "this government has no jurisdiction to make such major policy decisions" and that the matter should await an elected parliament is as close to a definitive answer to Research Question 8 as exists in the public record [E]. The government's own Additional Attorney General countered that the Chief Adviser and Advisers of the interim government can lawfully exercise the powers the PPP Act vests in the Prime Minister and Cabinet — a live, contested legal question that the split verdict left formally unresolved on the merits (resolved only procedurally, by sending it to a single bench, which appears to have allowed the process to continue). Risk of a future elected government repudiating or renegotiating the deal: not separately assessed by any source located beyond the general observation, embedded in Justice Najib's reasoning, that a multi-decade strategic-asset commitment made by an unelected government facing imminent elections carries this risk inherently. Given Bangladesh held elections in February 2026 (per the "days before national elections" framing in the sourcing) and a new elected government has since taken office (consistent with FY27 budget references to a new Finance Minister and a BNP-linked government in the companion Bangladesh banking-sector research), the deal's fate now sits with a government that did not negotiate it — which is itself a highly relevant, confirmed structural fact: it was NOT signed before the election despite the interim government's evident intent to do so (DP World itself sought more time and no signing occurred during the interim government's tenure, per Bonikbarta's February 2026 reporting) [C]. This is arguably good news for legitimacy (an elected government will decide) and bad news for deal certainty (operators face real re-negotiation/cancellation risk, which likely explains DP World's aggressive push for a 30-year term to lock in value before any political change).

• Confidence band: 90%+ that the interim-government-authority question was a real, judicially significant obstacle (not spin) — directly evidenced by a High Court judge's written finding; 70%+ confidence that the still-elected/new government retains genuine optionality to renegotiate or cancel, based on the deal's unsigned status carrying through the political transition.

Finding 9 — Bay Terminal financing: capital-at-risk split confirmed, asymmetric

See Finding 1 (Bay Terminal) above — World Bank/IDA + CPA's own funds finance the shared marine infrastructure (breakwater, channel, dredging); operators finance their own terminal superstructure. This is a materially different (and more state-capital-exposed) model than NCT/Laldia/Patenga, where the operator/concessionaire bears the brownfield or greenfield terminal capex directly. Bay Terminal is therefore the more fiscally exposed of the two models for the state, because Bangladesh/World Bank capital is committed regardless of whether the container terminals built on top of it (CT-1, CT-2, MPT-3) ever reach the throughput needed to justify the enabling infrastructure spend — a sequencing risk not raised by any source reviewed but implied directly by the phased financing structure itself.

Finding 10 — Early-warning dashboard

See §6.

3. Deal Structure — Dated Decision Timeline (NCT, the most contested case)

4. Fiscal Comparison Table

Read: On CPA's own numbers, the self-operation counterfactual outperforms the on-the-table concession terms by roughly 3.8x on a per-TEU net-margin basis ($65.25 vs $17.18). This is the single sharpest, most quantified finding across both Bangladesh research directives completed to date, precisely because it comes from the government's own internal evaluation rather than a triangulated outside estimate.

5. Precedent Table

Synthesis: All three precedents point the same direction: DP World's international arbitration and enforcement record is exceptionally strong across very different host-state contexts (an authoritarian small state, an EU member state). The dispute-resolution forum clause in any Chattogram concession is therefore one of the highest-leverage, highest-stakes provisions in the entire deal — more consequential than the headline royalty rate, because it determines what happens if Bangladesh later wants out. This was not explicitly flagged as the top priority by any source reviewed but follows directly from the comparative evidence assembled here.

6. Early-Warning Dashboard

7. Sources (accessed September 2026)

• Prothom Alo (En) — "Chattogram port operations stall over NCT deal"; "Chittagong Port: Foreign companies to operation two terminals"; "Maersk's APM to build and run Laldia terminal" (Laldia terms)
• The Business Standard (tbsnews.net) — "Legal clouds, labour fury trail interim govt's push on Ctg port deal"; "New Mooring Container Terminal lease process explained"; "Ctg Port workers announce fresh protests..."; "Application filed with HC seeking status quo..."; "Fresh petition filed in SC..."; "Port workers allege pressure to hand over NCT to DP World at excessive low rate"; "Ctg Port resumes operations after 7-day disruption..."
• The Daily Star — "New mooring container terminal: Govt moves for UAE lease amid protests"; "Apart from NCT, race on for two other Ctg port terminals"; "HC to rule on writ petition challenging NCT deal..."
• Bonikbarta (En) — "DP World seeks 30-year operating rights on NCT based on a 15-year financial model" (primary-tier: CPA's internal financial evaluation, upfront fee, royalty tiers, per-TEU cost/margin analysis) — single most important source in this research
• Northeast News — "Renewed push to lease Chittagong Port's New Mooring Container Terminal to DP World" (internal lobbying/advisor account — single-source, flagged)
• Maritime Gateway — "Chattogram Port executes four projects under PPP mode"; "Opposition rises over handover NCT operation..." (NCT history, 2007-2024 operator data, worker headcount figures)
• New Age (newagebd.net) — "HC bench delivers split verdict on interim government's NCT deal move" (judicial reasoning, security findings)
• Financial Express (thefinancialexpress.com.bd) — "High Court asks for appeal again on NCT deal"
• BSS News — "HC verdict on New Mooring Container Terminal on Dec 4"
• PortNews (en.portnews.ru) — "DP World, RSGT and MGH fight for CCT rights at Chattogram port" (throughput share by terminal)
• The Sun (Malaysia); Hellenic Shipping News; Container News; Saudi Exchange filing; Enterprise (enterpriseam.com) — RSGT/Patenga concession terms and capex
• World Bank project document (P176812) — Bay Terminal Marine Infrastructure Development Project, financing structure
• Global Arbitration Review (globalarbitrationreview.com) — DP World v. Djibouti (Doraleh) LCIA award/enforcement chronology; DP World v. Belgium (Antwerp) ICSID ruling
• Observer BD (observerbd.com) — Saif Powertec contract expiry, Navy handover, workforce figures
• UNB, Bangladesh Monitor — strike/labor-action wire reporting
• tob.news — "NCT deal under scrutiny despite court clearance" (throughput share, domestic-capacity counter-argument)

Flagged as unresolved / not publicly obtainable within this research pass:

• The full text of any draft or signed NCT concession agreement — none located; all financial terms here are drawn from leaked/reported evaluation documents, not the contract itself.
• Regional per-TEU benchmark data for Colombo, Mundra, Port Klang specifically — not sourced; only Djibouti (via the arbitration record) and Patenga (partial capex data) are confirmed regional/domestic comparators.
• Exact, contractually specified tariff-setting authority and cap/floor/indexation provisions — not public, because no deal is signed.
• Precise, single-sourced-vs-corroborated status of the "advisor lobbying" allegation (Northeast News account) — treated as [S]/single-source and flagged accordingly; the general finding that DP World was informally favored over Saif Powertec/MGH is more broadly, though not identically, corroborated by the CCT-share reporting and worker-alleged $42/TEU figure.
• Whether the December 2019 or 2019-generally MoU date is accurately Feb 17, 2019 per multiple outlets, given the 2023 Awami League "approval" date reported elsewhere — a minor date inconsistency across sources that a primary CPA/shipping-ministry timeline document would resolve.