Ideas & Debate

Why the Sh134bn Nairobi to Naivasha SGR line may not give Kenyans value for money

As the debate about the Standard Gauge Railway (SGR) in Kenya continues, a review of the facts and figures raises more questions than answers.

The economic justifications used by the government to promote the necessity of this massive project depend on highly questionable assumptions and sources of information.

For example, the recently released Environmental and Social Impact Assessment (ESIA) for Phase 2A of the project initially predicts a carrying capacity of 13 million tonnes for 2030, but on the very next page states this to be 21.8 million tonnes! (Based on latest data this would require a consistent annual gross domestic product (GDP) growth of 10 per cent in all transit destination countries including DR Congo, Burundi and South Sudan and the Kenya SGR railway capturing all transit, nothing by road, nothing through Tanzania).

It is not clear what these numbers are based on. The figures contradict those in the East African Railways Masterplan, conducted by a reputable and independent consulting company in 2009.

These professional projections were done in three scenarios: Base (most likely), High and Low. This masterplan projects a Base of 7.5 million tonnes, with a high of 9.2 million tonnes and a low of six million tonnes.


This masterplan was prepared for the East African Community (EAC) and thus projection is for the whole of the Kenyan railway system.

If we assume that half of the freight traffic is for Nairobi, and half for Kisumu and export, then the High projection for the SGR Phase 2A comes to 4.6 million tonnes.

Therefore, a comparison of the figures between those in the ESIA and the well argued, evidence-based Masterplan projections, even if we take the high scenario, shows us that the ESIA figures are exaggerated between 283 per cent and 480 per cent.

Forecasts revenue

Global rail freight charges vary between $0.02 per ton per kilometre (tkm) in India to $0.08/tkm in Italy. The ESIA predicts a freight charge of $0.13/tkm.

For the 120 km of SGR Phase2A, this would create gross revenue of Sh21 billion ($202.8 million) using the ESIA’s lowest mentioned growth projection of 13 million tonnes. With the higher projection of 21.8 million tonnes, it would be Sh34 billion ($325 million).

The High projection of the EAC masterplan (9.2 million tonnes), using the same $0.13/tkm rate, yields gross revenue of only Sh7.4 billion ($71.8 million).

Forecasts expenditure

The capital costs of the SGR Phase2A are very high. Kenya took out a commercial loan of Sh134 billion ($1.3 billion). The grace period for repayment is likely to be 10 years.

This means that until 2026 only interest is payable at today’s rate of about one per cent at 4.6 per cent, or about Sh622 million ($6 million) per year. This calculation uses LIBOR, an international benchmark used to calculate interest rates on loans.

By 2026, however, we will have to start repaying the principal over 10 years, or Sh1.3 billion ($130 million) per year, plus interest. LIBOR2 is low now, but tends to go in roughly 10-year cycles with peaks at around five per cent.

If that were the case in 2026, interest would be Sh11.6 billion ( $112 million), or a total payment of Sh25.1 billion ($242 million). Only with the highest projections of tonnage, and at a very high projection of revenue per tkm, would the railway be able to pay capital cost.

In addition, there will also be operational costs. Track and rolling stock will have to be maintained, and locomotives fuelled, staff will need salaries, and the whole management system will need financing.

According to the World Bank most railways in developing countries have roughly 30 per cent operational cost, 70 per cent capital cost.

In our example above, that would be Sh10.7 billion ($104 million), making total annual expenditure Sh35.9 billion ($346 million).

As demonstrated previously, the most optimistic scenario described in the ESIA would create revenues of Sh34 billion. Even in this scenario, for which sound basis is lacking, the SGR will need huge subsidies until 2036.

Construction costs

Many comparisons have been made in the Press about the cost of building railways in Africa, by comparing gross cost per kilometre of railway line.

In Ethiopia a 781 km railway was constructed at a cost of Sh290 billion ($2.8 billion) for 756 km, or Sh384 million ($3.7 million) per km.

Tanzania recently signed a contract for Sh934 billion ($9 billion) for 2,560 km, or Sh363 million ($3.5 million) per km. Here, the gross cost of SGR 1 is Sh3.8 billion for 609 km, or Sh643 million ($6.2 million) per km.

All three railways are constructed to Chinese standards and are standard gauge. We have no details about the Tanzanian line yet, but the Ethiopian line is up and running while Phase 1 in Kenya is 90 per cent complete.

The Ethiopian line is electric, which is more expensive per kilometre than the diesel line in Kenya. Moreover, in Ethiopia about 115 km is double-track, against all single-track in Kenya.

The terrain in Ethiopia is more rugged, as the line goes through the volcanic Rift Valley, and it has many tunnels. In Kenya there are more bridges, to allow the passing of wildlife.

However, it seems inescapable that Kenya did not get a very good deal at Sh643 million per km against Ethiopia at Sh384 million per km, moreover for a better product.

Ethiopia financed approximately 50 per cent of the railway out of its own resources, about 10 per cent from a concessionary Chinese loan, and 40 per cent out of a commercial loan at LIBOR + 3.75 per cent.

This allowed a competitive international tendering resulting in the best global price.

Kenya had to borrow 90 per cent, part at LIBOR + 3.6 per cent, part as a concessionary loan from the Chinese State-owned EXIM Bank.

As a consequence, our mandarins had to accept that one, State-controlled Chinese company did the design and the construction, and as it later turned out, the supervision and quality assurance.

This goes against the public procurement laws of Kenya, which according to the then managing director of Kenya Railways and later Transport Permanent Secretary, do not apply in State-to-State deals.

The contract for the SGR Phase 2A, from Nairobi to Naivasha, is also such a State-to-State arrangement. This time it seems to be a 100 per cent commercial loan, again with EXIM Bank.

Because of this Kenya was not able to negotiate a competitive tender, and a contract was signed with a single State-owned company for the design and the construction.

The contract does not include rolling stock or other extras, but the cost has gone up to Sh134.9 billion ($1.3 billion) for 120.8km, equivalent to Sh1.12 billion ($10.8 million) per km or almost three times as much as Ethiopia paid.

The reason for this high cost was said to be the terrain that required many bridges and tunnels, so that comparison was said to be impossible.

We indeed do not know how many bridges and tunnels there are in Ethiopia, but thanks to the ESIA we have a good idea how much of these there are in the proposed construction of Phase 2A in Kenya.

A paper by the World Bank in June 2014 (Ollivier, Sondhi, & Zhou, 2014), provides an interesting analysis of the recent construction costs of railways in China.

It isolates the civil works (bridges, tunnels, and embankments including stations, each separately), track, signalling, electrification, and land acquisition and resettlement. All of this is costed per km, and a range given.

The ESIA states the exact mix of tunnels, bridges and embankments, so this allows calculation of what the SGR Phase2A would cost if built in China under a competitive tender.


Building in Kenya can be more expensive than building in China, for example because equipment has to be mobilised. For the SGR Phase2A that is not an issue, as the equipment is already in place from SGR Phase 1.

Labour cost is an issue, as the senior staff has to be flown in and paid more than in China. A lot of low-skilled labour is required for construction, and in China the minimum monthly wage is Sh31,143 ($300), while in Kenya low skilled labour is around Sh20,762 ($200).

We assume that this evens out. Cement and sand prices are roughly the same in both countries, but steel is more expensive in Kenya at Sh41,524 ($400)/tonnes against China’s Sh31,143 ($300).

Track is only five to seven per cent of total cost, and includes labour, so this is of limited impact.

In the table (shown left) we show costing for an SGR equivalent in China, on basis of the highest cost to the client in very remote places with extremely difficult terrain.

This cost typically includes a profit margin of between five and 10 per cent for the construction company which is not unusual in large-scale projects. The actual cost to the company is probably not higher in Kenya, than in very remote regions of China.

Kenya, however, pays $1,300 million, giving the company an additional profit of Sh49.7 billion($479 million) in this model.

Interestingly, this China costing is not for a single-track cargo line such as our SGR. It is for a double-track mixed cargo and high speed line, with speeds of 200 km/h for passengers and 120 km/h for cargo!

Double-track is usually 150 per cent of the price of single-track with high-speed specifications adding 30 per cent.

That would mean that the cost to the company for this low speed single-track line would be Sh43.7 billion ($421 million) or Sh363 million ($3.5 million) per km, but sold to Kenya at a price of Sh134 billion ($1,300 million) or Sh1.1 billion ($10.76 million) per km.

It is significant that the costing model used above comes to a km cost of Sh363 million. This is almost the same as Tanzania and Ethiopia are paying under competitive international tendering with some rolling stock added in.

This strongly suggests that Kenya is paying Sh134 billion for a product with a value of Sh43.7 billion an over-payment of $879 million (Sh90.5 billion). Put differently, we are paying a price that is three times the value.

While it is not the mandate of any Environmental and Social Impact Assessment to analyse value for money in detail, it must look at socio-economic impact.

A grossly exaggerated cost, as suggested to be the case by this analysis, would have a negative impact on the economy and thus socially.

The final version of the ESIA should therefore include some analysis of value for money, the more so as in this case, it also threatens the viability of the whole railway, and thus the wellbeing of all concerned.

These issues gain more prominence due to the current cuts in development spending by Treasury and the current projections estimating that 40 per cent of our tax revenue is to be spent on servicing loans.

The Kenyan population however deserves to have insight in the source of the data presented in this ESIA. Kenya deserves a realistic assessment of the socio-economic impact to be expected.

Nema should only issue a construction licence once these economic issues are clarified in a robust manner, in a much-improved ESIA.

As it stands now, the ESIA approves a considerable environmental sacrifice, but there is no robust analysis whether any benefits are to be gained.

Vishwanath is a Kenyan environmentalist with international experience in the fields of water resource management, landscape and wildlife conservation, community participatory approaches, advocacy and lobbying, strategic planning, and communications. He currently works at the International Union of Conservation of Nature in the People and Landscapes Programme, and is also the Acting Chair of the Friends of Nairobi National Park. He has a background in Environmental Studies.