Critical Analysis of TRACECA Study on the International Logistics Centre at Mărculești Dornier Consulting / NTU / Inros Lackner, Final Report – Annex 7, February 2011

General conclusion

The document is a concept study prepared in the promoter’s interest and presented as a feasibility study. Its own figures show that the logistics centre is not financially viable without grants and concessional loans. The demand estimate rests on top-down proxy indicators and on unsigned expressions of interest. Capacity is sized far above projected demand. Significant costs are missing, and the site was chosen at the Government’s request rather than through a comparison of options. The authors themselves acknowledge many of these problems, but the executive summary and the conclusions play them down.

1. The site was chosen politically, not by comparing options

• Section 9.3 contains the key admission. The consultants initially shortlisted Chișinău and Giurgiulești. The Government then asked that Mărculești be studied instead. Nowhere in the study is there a multi-criteria comparison between Mărculești and the alternatives.
• Bălți is never assessed as a site, even though the study itself describes it as:
  • the main demographic and industrial centre of the catchment area;
  • a railway node with existing warehousing (Aquila);
  • the seat of a new FEZ.

Bălți International Airport appears only once, as a possible “synergy” partner (§4.1.1). For a centre intended to serve northern Moldova, excluding the region’s main economic hub is a serious methodological gap.

• The catchment-area reasoning (§4.1.3) concedes that Bălți already dampens Chișinău’s influence in the north. That argument supports Bălți as the hub, not a greenfield site 30 km to its east.

2. The demand estimate is weak and partly circular

• Stacked top-down proxy indicators. National containerisable trade (1.0 Mt, 2007) is multiplied by the catchment area’s share of retail turnover (22%), then by an assumed market share of one third. The result is approximately 73,000 t/year. Every factor is an assumption. Retail turnover is used here as a proxy for freight volume: the authors assume that, because the catchment area accounts for 22% of the country’s retail sales, it would account for roughly the same share of containerisable trade. The choice is methodologically fragile, because retail sales mainly reflect household consumption and imports of consumer goods, whereas a large part of the ILC’s justification concerns agricultural exports. A region may have relatively low retail sales and, at the same time, substantial agricultural production and exports, or vice versa. Absent a demonstrated correlation between the two variables, the 22% share is not a robust basis for estimating freight flows.
• The base year is 2007, before the 2009 crisis (a 9% fall in GDP). The authors assume a full return to 2007 levels by 2011, followed by growth of 5% per year.
• The agricultural estimate lacks an adequate empirical basis. The formula (5% of total agricultural output plus 20% of non-containerised exports, divided by 3) has no empirical foundation. The authors call it “reasonable”.
• The stakeholder evidence is weak:
  • five firms with “varying degrees of interest”;
  • no letter of intent, no pre-lease, no committed volume;
  • the only fruit processor scored 2/5;
  • logistics service providers scored 3/5.
• The only concrete business line is tax-driven. The vehicle import and re-export scheme (2,000–3,000 vehicles per year) relies on holding cars free of customs duties and VAT under the FEZ regime. This is regulatory arbitrage, not logistics demand, and it disappears if the FEZ rules change.
• Irrelevant “evidence”. The Leon shopping centre in Bălți is cited as evidence of regional logistics demand. It would be supplied via Bălți, not via Mărculești.
• The authors acknowledge the underlying problem (§8.5.4):
  • the market is smaller than an ILC normally requires;
  • the region’s trade could be handled entirely from Romania and Ukraine;
  • the need for intermodal transit is low.

3. Capacity is out of proportion to demand

Item	Value
Projected demand captured by PAILM in the base year (containerisable)	~73,000 t (~30 TEU/day)
Design capacity for logistics and industry	357,500 t/year
Design capacity of the container terminal	425,000 t/year (43,500 TEU)
Total capacity vs. containerisable demand	~10×

• The container terminal calculation is inflated:
  • each wagon is counted twice (arrival and departure);
  • a track occupancy factor of 1.3 is applied, implying more than one full train rotation per day on a line with no current container traffic;
  • it states that “only one track has been considered”, yet still arrives at 174 TEU/day;
  • loading units (LU) and TEU are treated as interchangeable, at 10 t each.
• The industrial volume is unjustified. 225,000 t/year from two 3-ha food-processing plots is based on “European empirical data”, with no tenant identified.

4. The investment cost (€16.7 m) omits major items

• Paving of the container terminal is missing.
  • §5.6 provides for heavy-duty paving for 4-high stacking (point loads of about 8,000 kN/m²) over a terminal area of 80,000 m²;
  • Table 22 assigns the “container storage platform” to IAM;
  • yet capex item C covers only the 14,000 m² rail-road area;
  • §8.4.2 then states that rents are low because “no paving will be provided”.

These positions cannot all be true. At the study’s own rate of €45/m², paving the terminal alone costs about €3.6 m, roughly 20% of total capex.

• Terminal and service-centre items are omitted. The terminal tractors and the 8 chassis in Table 15 are not costed. Nor are the Container Service Centre buildings planned for phase 3.
• Land is included in capex although it is not being acquired. “Land acquisition” (€400 k at €1/m², a figure supplied by IAM itself) appears in capex, while §5.1 states that the question of acquisition “does not arise”.
• The allocation of funding responsibilities is inconsistent. Table 22 provides that the terminal operator buys the mobile equipment and superstructure, yet €1.05 m of stackers appear in the project capex.
• The estimate is not based on site data. There are no soil or groundwater data (§7.2), yet foundations and earthworks are priced. On a former military airfield, contamination (fuel, fire-fighting agents, possible munitions) is a real cost risk and is not mentioned at all.
• A 5% contingency is low for a pre-design estimate on an uninvestigated brownfield site.

5. The financial and economic analysis is weak

• The FIRR is 8% nominal, about 5% real at 3% inflation, against the study’s own discount rate of 15%. The project destroys value in financial terms.
• The 18% return for IAM depends on grants plus IFI loans of 67%.
• Sensitivity is severe even on the study’s own terms. Capex +20% reduces the FIRR to 4%. Revenues −20% reduce it to 3%. In light of sections 2–4 above, both scenarios are likely outcomes, not stress tests.
• Table 44 is mislabelled. Headed “Financial”, it includes external benefits (load factor, modal shift) and reports an “EIRR” of 15%. A financial analysis must not include externalities.
• The modal-shift benefits ignore network problems. It is assumed that 10 percentage points of freight shift from road to rail, over rail distances of about 780 km. The study itself notes:
  • the gauge change at Ungheni (1,520 mm);
  • the risk of transit through Transnistria;
  • that the only local rail competitor is losing freight because of poor handling and delays.
• Benefits start too early. Revenues appear in 2012, the first year of construction.
• The risk analysis is narrow. The Monte Carlo simulation uses triangular ranges chosen by the consultant. It does not model the main risk that the study itself names: that the market simply fails to materialise.
• The underlying financial data cannot be verified. The business plan’s “six tables” and the Annex 3 methodology are referred to but not included, so the FIRR cannot be reproduced.

6. The legal and institutional assumptions rely on the promoter’s assurances

• Ownership is described inconsistently. The site is said to be “wholly owned” by the state enterprise IAM. Yet the Ministry of Defence and the Ministry of Economy are described as directly involved in operating the FEZ, and the land is a former military airfield. In Moldovan practice, a state enterprise normally holds state assets under economic management (gestiune economică), not in full ownership, and public-domain land is subject to its own regime. The study contains no analysis of title. It relies on IAM and on the Florești District Council.
• Essential conditions were still unmet at the date of the study. Extension of the FEZ to the land, zoning, and clearing the land of farmers were all “assured” for the end of 2010.
• The land value (€1/m²) comes from a meeting with IAM, not from an independent valuation.
• Tenant protection is weak. The authors point out that leases are protected against termination for only 10 years, which discourages tenants from investing in buildings. This is a real bankability problem, yet it appears only in the recommendations.

7. The “trimodal” character and the “airport” status are unsubstantiated

• There is no air cargo. Air freight is explicitly excluded from the projections. The study states that there are no scheduled flights and that almost all of Moldova’s air cargo passes through Chișinău.
• IAM’s main business was leasing helicopters to the UN in Afghanistan.
• The marketing claims are unsubstantiated. Mărculești’s “reputation for efficiency, reliability, safety and security in the provision of aviation services” is presented as a marketing asset, without any evidence.
• The airfield itself is not assessed. There is nothing on runway condition, certification or operational status.
• The irregular Chinese cargo aircraft (about twice a month, clothing for Odesa, “to avoid additional costs and delays”) is mentioned without any analysis of the costs avoided.

8. The environmental and social assessment is contradictory

• The noise argument contradicts the flight data. The EIA states that airport noise will “mask” ILC noise, yet the same study shows there are no scheduled flights.
• Pollution is rated “No” without data. “Baseline pollution level” is marked “No”, even though there are no soil or groundwater data.
• Wetland and river risks are played down. 3.9% of the study area is very-high-risk wetland and the Răut River is nearby, yet no mitigation measure is costed.
• Farmers’ livelihoods are ignored. “No compensation” is proposed for the farmers cultivating the land. This runs counter to EBRD Performance Requirement 5 on the economic displacement of informal users, which matters if the EBRD is the intended lender.
• The labour-market analysis contradicts itself. It cites unemployment of 3.1% and a “limited workforce”, then presents the weak local economy as a recruitment opportunity.

9. Factual and internal inconsistencies

Issue	Discrepancy
IAM land area	270 ha (§1, §8.1) vs. 266 ha (§2) vs. 286 ha (§3.3.4)
Access road to the R13	4 km (§2, §5.2.1) vs. 0.4 km new + 1.2 km rehabilitated (§5.2.3, §8.1, capex)
Existing rail spur	300 m (§2, §5.2.1) vs. 0.24 km (§8.1)
Rail geography	The site is placed on a “Ukraine (Klimentovo) – Romania (Ungheni)” line, but §3.4.2 routes Klimentovo via Rîbnița. The line through Mărculești appears to be the Bălți–Rîbnița line, towards Transnistria; this needs to be verified.
Train length	Trains limited to 850 m including locomotives, yet 42 wagons × 19.6 m take up the entire 850 m
Government funding	National funding “is not realistic” (§10), yet Table 46 shows a €0.2 m government grant
Paving	Provided for in §5.6, assigned to IAM in Table 22, not to be provided (§8.4.2), not costed

10. Relevance to the Mărculești files

• It is an EU-funded technical assistance study, not an EU funding decision. It recommends applying for NIF grants linked to an EBRD loan, conditional on the project’s bankability, which the study’s own figures call into question. If anyone cites it as evidence of European funding or support for Mărculești, the distinction is clear from the document itself.
• It concerns a logistics centre on the land north of the runway, not the airfield. It says nothing about the runway, PCN or certification.
• The site selection process (§9.3) is useful in itself. The consultants’ shortlist did not include Mărculești, which was added at the Government’s request, and Bălți was never assessed.
• The 2011–2022 implementation period has expired. The strongest test of this plan would be to compare it with what was actually built.

11. Available transport statistics and ex post testing of the demand assumptions

The official statistics available today allow a far more direct check of the market than the retail-trade proxy used in the 2011 study. These data do not automatically demonstrate traffic capturable by an ILC and are not equivalent to an origin-destination matrix, but they show that any update of the study must start from actual transport flows, not from shop sales.

In 2010, the National Bureau of Statistics reported 8.381 million tonnes of freight transported in the Republic of Moldova: 4.394 Mt by road, 3.858 Mt by rail, 127.2 thousand t by inland waterway and only 1.3 thousand t by air. Enterprises in Bălți municipality accounted for 14.4% of reported road volume, i.e. about 633 thousand t if the percentage is applied to national road volume. This is a measure closer to actual transport activity than the retail-trade share, although it too identifies neither the origin/destination of the freight nor whether it is containerisable. Source: NBS, “Transportul de mărfuri şi pasageri în Republica Moldova în anul 2010” (Freight and passenger transport in the Republic of Moldova in 2010), 28.01.2011.

Recent data confirm that air freight remains marginal nationally: 1.4 thousand t in 2021, 1.2 thousand t in 2022, 1.3 thousand t in 2023 and 1.6 thousand t in 2024, compared with 17.616 Mt carried by road and 2.448 Mt by rail in 2024. In 2024, the North Development Region accounted for 13.3% of road volume, about 2.34 Mt. Source: NBS, “Moldova în cifre 2025” (Moldova in Figures 2025) and the press release on transport in 2024.

In January–September 2025, the North Region accounted for 15.5% of the 12.307 Mt carried by road, i.e. about 1.91 Mt; in January–June 2026, the North’s share was 15.1% of 8.862 Mt, i.e. about 1.34 Mt. In the same half of 2026, national air freight amounted to only 0.2 thousand t. These series reinforce the need to separate the North’s road/rail logistics market from the hypothesis of an air market at Mărculești.

Methodological consequence: a new study should rebuild demand from customs and trade data by origin-destination, commodity type, seasonality, current transport mode, door-to-door cost, transit time and shippers’ willingness to change route/mode. Only after that stage can the market share capturable by Mărculești, or by an alternative site, be estimated.

12. Governance, conflicts of interest and investigations subsequent to the study

These facts post-date the TRACECA report and cannot be attributed to the 2011 consultants. They are, however, relevant to an ex post assessment of the governance, bankability and internal-control assumptions on which the project rested. The wording must respect the presumption of innocence and distinguish administrative findings from criminal charges.

2009–2015: an analysis by the National Anticorruption Centre of the activity of the state enterprise Mărculești Free International Airport (Î.S. Aeroportul Internațional Liber Mărculești) flagged poor management, financial losses and links between businesses operating on the airport’s premises and the enterprise’s management. These findings became public in 2016.

2015: a transaction involving four of the airport’s trucks and a semi-trailer later became the subject of a criminal case; in 2022, the former director and a former valuer were sent to trial, with prosecutors alleging damage of almost MDL 2 million. This is an allegation subject to judicial review, not a final finding of guilt.

2016: the National Integrity Commission found breaches of the legal regime on conflicts of interest in connection with the hiring of relatives and contracts with firms associated with relatives of the former director. This episode is distinct from the later criminal cases.

2012–2018 / investigation made public on 1 April 2022: the Information and Security Service (SIS) and the Prosecutor’s Office for Combating Organised Crime and Special Cases (PCCOCS) announced a case concerning alleged embezzlement, tax evasion and money laundering at Mărculești International Airport. According to the official press release, the facts under investigation allegedly occurred in 2012–2018, the preliminary damage exceeded MDL 10 million, and a significant proportion of the service-providing companies were allegedly affiliated with a decision-maker. The SIS also reported the suspicion that important documents had been removed in order to prevent their archiving and examination. Source: SIS, press release of 01.04.2022, updated 07.08.2026. The persons concerned enjoy the presumption of innocence.

Relevance to TRACECA: the study assumed that IAM would be the owner/developer, negotiate long-term contracts, attract operators and financiers, and manage a complex contractual structure. The governance problems documented later do not, on their own, disprove the technical feasibility of a logistics centre, but they should have become an explicit institutional risk in any reassessment of the project and in any financier’s due diligence.

13. The distinction between the ILC and the development of a civil passenger airport

The TRACECA report is not a feasibility study for opening or upgrading Mărculești as a civil airport with scheduled passenger services. Its subject is an intermodal logistics centre on the land associated with the airport. The aviation component is presented as a locational advantage, but air freight is explicitly excluded from the traffic projections as speculative, and passenger traffic does not form the basis of the financial model.

This sets an important limit on how the document can be used: the figure of about €16.7 m is the capex estimated in 2010 for the ILC infrastructure, not the cost of certifying and upgrading the airfield for scheduled traffic. It has not been shown to cover works such as rehabilitation/strengthening of the runway and aprons, airfield lighting, radio navigation, a passenger terminal, aviation security, rescue and fire-fighting, or remediation of possible contamination.

Likewise, the EBRD interest described in the report must be read as interest in principle and as a financing hypothesis conditional on bankability; it is not equivalent to a loan approval, a financing commitment or an EU decision to fund Mărculești. Recently, the EBRD replied that EBRD is not currently involved in this project (the Mărculești airport project announced in 2025), nor is there any intention for the Bank to participate in it.

There is also an internal tension between the “trimodality” marketing argument and the business case: aviation is used to differentiate the site, but generates no demonstrated air traffic in the model. The quantified economic engine remains road-rail, warehousing, containers and agro-logistics.Moreover, the report describes Bălți as the main demographic, industrial and commercial centre of the catchment area, yet Bălți-Leadoveni is not tested as an alternative site. The report therefore cannot support the conclusion that Mărculești is superior to Bălți-Leadoveni for serving the north; such a conclusion would requ