Technical Discussion on Guarantees for Private Infrastructure

By SDS/IFM (08/99, En) See also Infrastructure and Financial Markets

MFI Working Group on Support for Private Infrastructure

May 25, 1999

Washington, DC

 

Agenda

List of Participants

Table: Guarantees Issued by Multilateral Development Banks

 

 

Under the auspices of the MFI Working Group on Support for Private Infrastructure, the World Bank and the Inter-American Development Bank (IDB) cosponsored a seminar on May 25, 1999 to discuss the technical issues related to the provision of guarantees for private infrastructure by the multilateral financial institutions (MFIs). The purpose of the meeting was to provide an opportunity for guarantee experts from the various MFIs to exchange information, discuss technical topics (e.g. pricing and scoring of guarantees), and share ideas regarding best practice (see attached agenda). In addition to representatives from the World Bank and the IDB, participants included representatives from the Asian Development Bank (ADB), the Andean Development Corporation (CAF), the Caribbean Development Bank (CDB), the European Bank for Reconstruction and Development (EBRD), and the Multilateral Investment Guarantee Agency (MIGA) (see attached list of participants).

Role and Scope of Guarantees

The first session of the meeting focused on the role and scope of the various guarantee products offered by the participating institutions. To begin the session, each institution provided an overview of its guarantee program:

The discussion following the introductory presentations began by focusing on the relationship between guarantees and policy reform, specifically on the question of whether there should be some requirement that sector reform in the country concerned be sufficiently advanced so that the guarantee will not serve as a substitute for policy reform. The point was made that guarantees are meant to allow for a test of a new policy and regulatory environment, therefore MFIs need to ensure that guarantees are testing and not substituting for reform. Guarantees should also be limited to covering political rather than commercial risk.

While loans and guarantees have similar characteristics once disbursed (acceleration is a legally permitted remedy for loans, though it has not been exercised), guarantees can be difficult to revoke once they have been issued (i.e., disbursements cannot be suspended should a government abandon its reform efforts). To provide some incentive for governments to persist on the reform path, the World Bank is considering including covenants in its counter-guarantee agreements to limit the Bank's liability under a guarantee to the disbursed loan amount, should a government default or breach its obligations. Private lenders also typically include these covenants in their loan agreements such that they can suspend disbursements if the covenants are breached (although these do not generally transfer political risk to the project). While some viewed it as unfair to project sponsors to suspend a guarantee based upon government violation of a covenant unrelated to the specific project (e.g., a default on other debt owed to the World Bank), it was argued that, empirically, a government is unlikely to default to a multilateral institution and continue respecting its obligations related to the guaranteed project.

The question was raised as to whether the eligibility criteria for government access to guarantees could be tightened. One consequence would be that the multilaterals would issue fewer guarantees. The point was also made that, with reforms progressing so quickly in the infrastructure sectors, MFIs need to avoid constraining reform possibilities by pre-establishing overly conservative criteria for sector reform.

Other issues raised briefly but not discussed included: (i) guarantees for cross-border projects; (ii) how to determine when a guarantee is triggered (at time of breach or only after arbitration), and how debt servicing should be handled between the time of the breach and the conclusion of arbitration; (iii) subrogation rights; (iv) preferred creditor status; and (v) indemnity agreements.

Guarantees without Sovereign Counter-Guarantees

The second session focused on the provision of guarantees to private sector entities without the use of sovereign counter-guarantees. Participants discussed the difference in exposure for the multilaterals when there is no counter-guarantee. Many felt that the difference was not material, given that MFIs typically have much implicit influence with governments. While most institutions said that they would not suspend all operations with a government whose actions resulted in the call of a guarantee, they would certainly pressure that government to conform with its obligations. MIGA can require a sovereign counter-guarantee, but typically does not. In a breach of contract situation, it would proceed against the government, pursuant to the decision of an arbitrator (MIGA breach of contract coverage insures payment of arbitration award).

Guarantee Capacity. Both MIGA and IDB discussed initiatives to cooperate with private sector co-insurers and re-insurers. IDB explained that the number of its guarantees had been limited in the past by project caps, however new, higher limits would allow it to work more with private sector re-insurers and guarantors. MIGA has a "facultative reinsurance agreement" under which it works with private re-insurers to increase its insurance capacity. The maximum exposure per project MIGA can carry alone is US$110m, but coinsurance allows this cap to increase to US$200m. The reinsurance agreement allows this exposure to increase even further. Even with re-insurer participation, the insurance is provided according to MIGA decisions on tenor, country policies, and pricing, such that the buyer is typically unaware that private re-insurers are involved.

Environment and Governance. Participants also discussed the issues associated with enforcing MFI standards on the environment and corruption in the context of guarantees. The IDB can place policy conditionality on the provision of a guarantee, but, once the guarantee is provided, there is little leverage for enforcing this conditionality. The World Bank conducts due diligence to ensure that the sponsor has sufficient funding to comply with environmental requirements, but a call on a guarantee could not be rejected if the sponsor ending up failing to comply with these requirements. The World Bank's only recourse in this situation would be to cap its liability at the amount of the disbursed loan. Of course, the World Bank could refuse to work with the concerned sponsor in the future and could pressure private lenders to apply their powers of moral suasion with the sponsor. If the government violates environmental standards, the liability could again be capped and future loans could be suspended. In the case of MIGA, a guarantee can be canceled for violation of the World Bank Group's environmental and child labor standards.

Scoring of Guarantees

IDB scores guarantees on a one-to-one basis with loans. According to IDB's statute, it must be absolutely certain that it has sufficient "headroom" to cover the guarantee should it be called. Thus, the IDB is forced to treat guarantees the same as loans for internal scoring purposes.

While the IDB focuses on "headroom" as the binding constraint with respect to scoring, the World Bank bases its scoring on capital adequacy provisions. At the World Bank, there is little debate regarding scoring of partial credit guarantees: they are scored at their net present value. Partial risk guarantees for a given country are also scored the same as loans to that country, because empirically the probability of a call on the guarantee and the probability of sovereign default on a World Bank loan are thought to be closely linked. In other words, it is unlikely that a country would default on a World Bank loan and still respect its guaranteed obligations under a given project.

Both IDB and World Bank agreed that, even if a guarantee is not callable until sometime in the future, provisioning must begin at the time the guarantee is issued. Since the banks do not know what their income will be in the year the guarantee is callable and the commitment is irrevocable, they need to smooth provisioning over time.

Pricing of Guarantees

During the last session, each institution offered an explanation of its pricing policies.

Last updated: 05/08/07

Return to Top