It has taken almost five years, but Mexico's PPP programme has produced its first closed transaction. PPS, short for Proyectos para Prestación de Servicios, as the concept is known in Mexico, has been developed in the teeth of the country's low investment grade rating and the intricacies of its budgeting process. And one deal down the line, the finance ministry, or Hacienda, is set both to transfer the PPS concept to sectors other than roads, and to increase the complexity of the concessions on offer.

The PPS concept has survived the disputed Mexican presidential election – the centre-right PAN candidate, Felipe Calderon, has been confirmed as winner, beating Andres Manual Lopes Obrador (AMLO) by roughly a quarter of a million votes. The long and fraught process of deciding the winner of the election involved a blockade of the Paseo de la Reforma, the avenue running through the city centre.

Were this to have continued, Project Finance's Mexican PPS roundtable would have been its second to be disrupted by protestors (although, unlike in its Canadian incarnation, Mexico would have been an incidental casualty). But PPS, as opposed to the potential private investment in the Mexican energy industry, is less controversial. Nevertheless, the victory of Calderon, an avowed and enthusiastic free marketer, over AMLO, whose pronouncements on infrastructure had tended to emphasise the role of state finance, will make the private sector rest more easily.

During the interim, ICA closed the Irapuato-La Piedad financing, a Ps730 million ($67 million) financing, the first under the new PPS regime. ICA won the concession in mid-2005, on the back of its experience with financing Latin toll roads and its slate of civil works contracts. The lead arranger of the Ps580 million debt, which features a Ps189 million partial guarantee from the International Finance Corporation, is Banco Santander.

CONIPSA, as the special purpose vehicle for the concession is known, will be familiar in structure to observers of the UK and Canadian PPP road deals. It features no tolling, a shadow traffic volume-related payment equivalent to roughly 15% of total project revenues, and a credit that rests on the performance of the sponsor as contractor and operator coupled to the credit of the transport and communications ministry, or SCT.

Why guarantees are useful

The last element is by far the most important, and casts a shadow over the entire PPS programme. While more developed countries have little problem persuading lenders of their ability to make concession payments, those that are less creditworthy have to resort to more complex structures.

Thus, jurisdictions like Peru and the Dominican Republic have used minimum revenue guarantees backed by multilateral-backed insurance products, or have carved out dedicated revenue streams designed to mimic, or even achieve a higher rating than, sovereign debt (search "nordeste" and "IIRSA" at projectfinancemagazine.com for more details). Mexico falls awkwardly between the two, which explains the IFC's deployment of the guarantee facility.

While Mexico's country rating, which now stands at BBB/A1 (S&P/Moody's), is robust enough to ensure financeable concessions, lenders have still expressed qualms about the budgeting process for PPS. Appropriations to pay concessionaires must be approved by the country's congress each year, and congress is a disparate and factious body. While Calderon's party – the PAN – won the most seats in congress, AMLO's PRD came in second. The PRI – formerly the largest party – is now third largest, but still holds the balance of power in the legislature, and is somewhat divided on the subject of private participation in infrastructure.

The issue of multi-year budgeting for PPS has been a frequent source of criticism from lenders almost from the start of the programme. It has been raised in discussions with the SCT and Hacienda, and the government has largely been receptive to such issues. It has bumped up the priority of payments to PPS concessionaires to just behind the payment of government salaries, and ahead of other government obligations.

Even with this concession, it took Santander several months to get credit approval for the CONIPSA loan. If Santander, a bank with a long history in Mexico, could take so long, the broader market sentiment is likely to be, if anything, more cautious. Santander is likely to hold most of the debt, with the exception of the sale of small participations to Banobras and possibly Banorte. ICA and Santander are now working on a Ps700 million loan for the adjoining section – Querétaro–Irapuato – the PPS concession for which ICA also won.

It should be noted, though, that the 75km CONIPSA concession has an indicative international rating that is solidly investment grade, and that the Mexican government is not likely to renege on its commitments. And Santander was able to get approval for the loan even without the enhancement of the IFC partial guarantee.

This might lead the casual observer to conclude that the IFC facility, while a small source of lender comfort, might be better suited to countries with worse ratings. The IFC facility thus might be viewed as an experiment designed to be perfected elsewhere, and CONIPSA as a slightly over-engineered test-bed.

But Mexico is likely to provide several further opportunities, because the cap on PPS projects at a federal level is set at $15 billion, and any transactions beyond this cap will need to go forward at a subsovereign level. The states and the municipalities, however, are much less creditworthy, and subsist to a greater or lesser degree on transfers of money from the federal government.

While most of these states have national scale ratings that fall into the single-A category, some, such as the state of Mexico at mxBB+, are much lower. At these levels, guarantee products are the only way to make a concession with a state counterparty financeable. When Project Finance asked roundtable participants whether such guarantees were essential or merely useful, the responses largely consisted of boilerplate statements of how unique project finance deals are, and how nothing is a given.

But enhancing some of the less robust subsovereign credits will be difficult, and the IFC, with the Tlalneplantla water project from 2003, has proved that the product is acceptable to Mexican lenders. The alternative for states is to put together financing structures that revolve around the sale of the rights to federal government transfers.

The Oaxaca template

The state of Oaxaca does not have a rating, is located in the south of Mexico, and rivals Chiapas as the least developed of the states. In 2004, it carried out a Ps500 million securitization of federal receipts, which garnered an A+ rating from S&P. Nevertheless, the state has an ambitious slate of infrastructure projects, advised by Estrategia Financiera and Cal Y Mayor, and has demonstrated a sustained interest.

However, the state's first structured infrastructure financing (this description is used advisedly) does not, as many participants were eager to point out, really qualify as a PPS transaction. The reason for this is fairly simple – a lender would not be willing to take on uncovered Oaxaca payment risk on a concession.

The financing for the state's administrative headquarters best resembles a leasing contract, of the sort familiar to lenders active in the German, French and Japanese markets, whereby very little performance risk resides in the concessionaire, and the borrower is simply relying on lease payments from government. While this arrangement does not produce huge levels of risk transfer, it does produce financeable concessions.

Oaxaca's project uses the assignment of the right to payments from the federal government to make lenders comfortable with the project. The arrangement was sufficient to bring in Banorte as a lender, and progress to date on the project has been strong. Indeed, the experience may have much more relevance to other states than the PPS roads programme to date.

Probably the biggest weakness in the use of Oaxaca's template is not so much that it does not conform closely enough to the PPS structure, but that states will need to be wary of the ultimate budgeting and ratings impact of dedicating large proportions of their federal government receipts to financing infrastructure projects.

More for later

The federal programme, for now, still has most of the momentum. Indeed, the Secretaría de Educación Pública (SEP), Mexico's education ministry, has just awarded to Acciona the Universidad Politécnica de San Luis Potosí project. This would be the first of four polytechnics to go forward using the polytechnic route. Moreover, Acciona won over a consortium led by local construction firm Marhnos and Currie and Brown, proving that outside investors are able to compete on the more complex PPS projects.

In the roads field, the next project – Nuevo Necaxa-Tihuatlan – will be tendered as a mixture of tolled and free sections, and combine work to be performed by the concessionaire and work that the SCT will complete and then transfer to the concession. The concession is an 85km section with a capital cost of $700 million.
Nuevo Necaxa-Tihuatlan is the final section of a 293km route between Mexico City and the port of Tuxpam on the country's east coast. Much of the remainder of the road is tolled under a concession held by a Banamex trust, the Fidecomiso de Autopistas y Puentes del Golfo Centro. The road goes through hilly and difficult terrain, and presents considerable technical and construction challenges. But the revenue potential for the section, which is designed to spur economic development rather than benefit from it, is not high.

So the SCT has divided the road into two sections – Nuevo Necaxa-Avila Camacho and Avila Camacho-Tihuatlan – of which the first will be built as a PPS and the concessionaire paid from the SCT's budget, and the second section will be tolled, and built using funding from the SCT. Toll revenues from the second section will go towards decreasing PPS payments on the first. Since the 37km PPS section has a cost of $600 million, and the 48km tolled section will cost $100 million, combining the two has a strong logic.

Meanwhile, the SCT has a further three real toll concessions out to bid. The first is the 92km Monterrey to Saltillo section, incorporating a bypass to the west of Saltillo, which has a cost of Ps2.5 billion, and two bridges, Ps80 million San Luis Rio Colorado II and the Ps650 million Reynosa Anzalduas. The bids, for which awards are due shortly, follow seven successful awards with a total value of Ps14 billion.

Initial criticisms of the real toll concession centred on the robustness of the traffic studies, as well as the ability of foreign bidders to cope with the aggressive nature of the concessions. So far Mexican construction companies and the occasional Spanish operator have been the main winners under the programme. But the SCT has not lacked for bidders, even if these have largely steered clear of non-recourse financing.

The biggest individual winner has been the Impulsora del Desarrollo Económico de America Latina (IDEAL) fund, which spun off late last year from Carlos Slim's Inbursa Group. Slim's presence in the market has not gone unnoticed in the business press, but until recently the likely lender to IDEAL's projects was assumed to be a combination of Inbursa institutions and Mexico's development bank Banobras.

However, one lender independent of Inbursa and active in the market says that IDEAL has recently begun approaching project lenders to work on financing concessions. The lender would not provide any further details of these discussions, but one piece of speculation in the Mexican market is that IDEAL may look to a capital markets refinancing of completed projects.

Tolls and Toluca

While banks can comfortably provide maturities of beyond 20 years to newbuild PPS projects, the capital markets, particularly with enhancement, can provide 30 with ease. The Toluca Toll Road refinancing, a 22-year UDI1.13 billion (UDI is an inflation-linked version of the peso, the equivalent total is $397.4 million) applies the terms on offer to a restructuring of the issuer's tariff structure.

The issue was wrapped by MBIA, and underwritten by HSBC and BBVA, and is designed to pass on lower debt service costs to motorists by reducing tolls and luring drivers away from a competing free route to the 21km road. The reductions in tolls average between 25% and 45%, and essentially result from debt providers' increased comfort with the Mexican macroeconomic environment.

Nevertheless, the restructuring did require that an existing subordinated debt facility, held by the Hacienda, was not fundamentally altered, and that the issuer decrease in tolls in such a way that the monoline insurer and bondholders would still have access to sufficient cashflow. The SCT also had to provide an extension to the concession.

The deal closed on 7 April, comfortably ahead of the elections, and priced for a 5% coupon, slightly above Toluca's 2003 refinancing, which had a 4.5% coupon but only a 15-year maturity. The bonds were issued under MBIA's Programa AAA de Infraestructura Mexico (PADEIM), which allows a borrower to benefit from an expedited registration process.

Metro technical

Outside of the established PPS and toll roads processes there are a small number of other more bespoke transport concessions looking for financing. They can look to the Eu207 million ($248 million) financing for a 15km suburban project in Mexico City. CAF, a Spanish train manufacturer, closed this financing in July.

The project involves the rehabilitation of a line running from Buenavista to the northern suburb of Cuautitlán, and the provision of new rolling stock. The project will be reliant on passenger revenues for debt repayment, although the government has provided the project with a Ps100 million liquidity facility. The debt facility has a debt service coverage ratio of 1.86x.

The commercial logic behind the project is sound. Several bankers were forced to take to public transport and abandon their cars during the blockade of La Reforma. They noticed how crowded the city's public transportation system is, since the city's small number of metro lines is stretched to capacity. The suburban train project will have two terminals and five intermediate stations, and will carry between 420,000 and 490,000 passengers per day, paying roughly Ps8.5 ($0.78) per trip.

The main challenge in structuring this deal was matching peso revenues to project costs denominated in Euros. CAF was able to get hold of 14-year debt and persuade Spanish export credit agency CESCE to insure it. The two tranches of Eu190 million and Eu27 million were then swapped into Pesos to isolate the project from devaluation risk.

The deal is very much a one-off – the SCT has stressed that it is unlikely to provide any hedging products to road deals, and there are few other light rail schemes at anything beyond a conceptual stage. The SCT has announced plans for a $12 billion high-speed rail link between Mexico City and Guadalajara, and plans to bid this out shortly. But that deal, which officials have billed as the first use of a bullet train in the western hemisphere, faces even more unique technical and financial challenges. n