nfrastructure finance in Latin America, particularly since the end of the last century, has traditionally meant Chile. The Chilean roads programme has been a lucrative source of fees for monolines, and has proved popular with sponsors. But Chile's population ? at 17 million ? is roughly that of metropolitan New York, and dealflow will be ramping down in the future.

In theory, therefore, this would be a good opportunity for newly investment grade Mexico to come back to the markets. Indeed the Secretariat of Communications and Transport (SET) held presentations in New York (for the banks) and Madrid (for the likely sponsors) in conjunction with its legal advisors, Coudert Brothers. Indeed, it plans to launch nine new concessions over the coming years, and expects to name a preferred bidder on the first shortly.

While Mexico, rated BBB by Standard & Poor's, is a hot destination for capital, as witnessed by the popularity of its power projects, doubters will be more vocal in this sector. The programme of the 1990s was not a great success, as earlier concessions struggled with uneconomic structures, overleverage and the collapse of the domestic construction sector. The currency crisis of 1994-5 hit the dollar-denominated debt profile of many roads hard.

Many of these road projects were nationalised or brought back on sponsor balance sheets, and the experience left a sour taste in many mouths. These roads are now undergoing restructurings and refinancings ? of which two have been public bond deals. The first, the Maxitunel concession, was a MxP180 million ($19 million) issue led by BBVA. This refinanced a tunnel concession between Acapulco and inland Mexico that was in the hands of local group ICA. It gained a AA.Mx local scale rating from S&P, had a 15-year maturity and a 9.6% interest rate when it ran in May 2001.

A second issue, for the Cancun-Kantunil concession (Autopistas del Mayab) went out in June 2002. Again led by BBVA, the MxP1.8 billion series gained a rating of Aa3.Mx (local) and Ba1 (global) from Moody's. Moody's is examining the sector for further refinancings in the near future.

This time, however, roads will be operating under a new framework (this can be found at www.sct.gob.mx). The commitment towards the programme is laid out by Oscar de Buen Richkarday, head of the toll roads unit for Mexico. He says ?we have a roads system that requires a lot of new investment, and the thrust behind the programme is the scarcity of available public funds. It has been developed internally by ourselves and Banobras, the national development bank.?

The programme consists of nine concessions of varying length and size, and uses existing legislation laid down for the earlier programme. These are listed below. It will benefit from a road using population that has been exposed to tolling for almost 15 years. Indeed, the sector is probably less politically sensitive than energy, an often emotional issue.

Mexico's roads programme

Size ($

Con- Length million

cession Name Lanes (km) equiv.)

1 Matehuala Bypass 4 14 37

2 Perote-Amozoc 2 103 150

3 San Blas Junction-Escuinapa 2 151 230

4 México City Northern bypass 4 220 700

5 Morelia-Salamanca 2 81 150

6 Laguna Verde-Gutiérrez Zamora 2 127 250

7 Saltillo-Monterrey 4 40 223

8 Libramiento de La Piedad 2 17 34

9 Allende-Juárez 2 42 60

Source: SCT

The government has borrowed from the Chilean experience ? itself a reaction to Mexico's earlier troubles. The earlier model, says de Buen, had both positive and negative features, ?but we wanted to create a model that drew on the positive elements from 10 years ago, and attract investors. We did look at the Chilean model, but have adapted it to come up with a framework that is suitable for Mexican conditions. The feature we found most useful is the idea of the minimum revenue guarantee, to ensure that the project company is paid back and allow debt to be serviced.?

Where Chile's experience diverges is in the type of concession offered. Most of Chile's roads come with an existing history of tolling, mainly require upgrades, and are concentrated around the capital, Santiago, or the Ruta 5 (Pan-American highway). Mexico's roads head in several directions and compete far more heavily with rail traffic.

The other key feature of the new model is that the Mexican government will put into the project a subordinated contribution commitment (SCC) that would be designed to make up the distance between anticipated total revenues and principal and interest. This will be disbursed through the Fund for Investment in Infrastructure (FINFRA), administered by Banobras, as will minimum revenue payments, according to de Buen.

Repayment of the concession's costs, according to guidance from Banobras, will be shared above a fixed and expected rate of return to the sponsor. This may be split 80:20 in favour of the concessionaire, after the repayment of the SCC, debt and the expected return. Bids will be judged, subject to prequalification technical abilities, on requiring the lowest possible SCC and revenue guarantee.

Nevertheless, the new framework looks viable, according to Mario Juarez, visiting attorney at Chadbourne & Parke. ?From the standpoint of lenders, the new model's terms look favourable. The sponsor is constrained in certain ways, as with the Chilean experience. They will need to determine if their return on investment is sufficient.?

The strength of the concession structure is matched by its rigidity ? there will be little extension of the works to be approved by the government. Such overruns were among the main reasons why earlier concessions became uneconomic. Here the government is encouraging bidders to do as much of their own due diligence as possible, and to avoid relying overtly on traffic studies provided by the government. As such, bidders will need to work out their cost of financing in advance ? and there is not going to be a shortlist phase in which bids can be refined, since this is not allowed under Mexican law.

This is not the only area where Mexican law will clash with the ability of sponsors to come up with robust structures. The aspect of the previous programme most frequently cited by bankers as a fault was the requirement that the concession feature a free alternative. This would violate the principle that a truly economic road be in some way essential, or at least that its alternatives be sufficiently unattractive.

The first concession to be offered, for the Matehualpa bypass, has brought back memories for nervous lenders. Matehualpa is a 14km length of 4-lane road that will divert long-distance traffic round the City of Matehulapa, in the north-east (see Map 1, below). It has an estimated cost of MxP370 million, and this size has not made it attractive to some of the larger players.

More significant, however, is the requirement that the winning bidder construct a new road that skirts the city far more closely. The only tinkering with this requirement is that the municipality has offered to prevent heavy trucks from entering the city, a promise that is not irrevocable. As a result, according to reports in the Mexican press, there have only been about three bidders for the road.

Project bankers are usually keener to start off such new programmes with a bang, if only to bolster their own pitch books. But de Buen says that the choice of Matehualpa was a deliberate one. ?It is a small project,? he says, ?and we went ahead because of the size, so that it will be a useful learning process. We do not expect much interest from international players.?

The need for an alternative route has its basis in the Mexican constitution, which has until now interpreted the right to freedom of movement around Mexico as providing free alternatives. The unwillingness to budge on this issue is no doubt, however, a factor of current political conditions. Memories of the protests on the site of Mexico City's proposed new airport are fresh, and acquiring rights of way will be a tortuous process, as some oil pipeline developers have discovered.

De Buen stresses, however, that upcoming concessions will be much more attractive, saying ?we have a $150 million equivalent road coming up in the shape of the Amozoc-Perote concession (see Map 2, below), where we expect to see a lot of interest from international, specifically Spanish, players.? This road, at 103km, connects the major urban centres of Puebla and Xalapa, has an existing road of variable quality ? not, at present, part of the package. This would be much more attractive to the Europeans.

Whichever corporates take up the challenge, they will likely have to work within the framework of the Mexican banking sector. The SCT has explicitly said that it is unlikely to offer any hedge to overseas lenders, unlike Chile, which was prepared, on the Autopistas del Maipo deal, to offer a hedge when the cross-border bond market became the only one likely to provide the concession's debt. While there is a slim possibility, although scant evidence, that an exchange rate devaluation insurance product might be of interest, peso debt would be likely to predominate.

In this context, interest is likely to shift towards the condition of Mexico's banks. In this sphere, de Buen offers guarded optimism, pointing out that ?we can now go beyond ten years, which was a difficulty with the previous programme. Fifteen-year debt should be possible, when cushioned by the equity and public contributions.? Nevertheless, he adds that the unit will see how the first two concessions go before assessing the quality of the domestic bank market.

This would be in a more robust condition than previously, and Citigroup, with its interest in Banamex, and Santander, through Serfin, have both expressed confidence in the sector. The two will be among the main candidates from international institutions for financing mandates. The bond market may also play a role, although in-depth involvement on the part of the monolines has been slight so far. MBIA is one of the players thought to have the most active interest.

The SCT's claim to be prepared to study what works and to be pragmatic is emphasised by De Buen. He will not completely rule out offering a hedge in response to market conditions, and may find the Chilean negotiating experience more directly useful to the programme than formal structures. Nevertheless he believes that the minimum revenue guarantee and SCC will be sufficient to make concessions robust.

The final area where sponsors will need to examine their options is in the potential for the securitization of toll revenues. At present this will be encouraged insofar as it makes debt cheaper and is directed towards amortization of bank and SCC debt. This is designed to ensure a commitment by sponsors to concessions, but may potentially hinder the redeployment of capital. De Buen suggests that limited recovery of equity through securitization may be possible, subject to authorization, although current Banobras guidelines for financings to not envisage this.

A preferred bidder on Matehualpa could be announced in April 2003, with financial close possible by July, while proposals are expected for the Amozoc-Perote by July 2. At present the second concession, in an encouraging sign, has received about 16 expressions of interest, although some of these are from financial institutions. After the initial two, de Buen hopes to get as many as two out per year. Banobras is considering extending the template to other sectors.